)
A
• ■ ■■
■ r
) -r-
h£-f.* ■ \ ■
Hi
6
WISCONSIN ACADEMY
OF
SCIENCES, ARTS AND LETTERS
VOL. XXXI
NATURAE SPECIES RATIOQUE
MADISON, WISCONSIN
1938
TRANSACTIONS
OF THE
WISCONSIN ACADEMY
OF
SCIENCES, ARTS AND LETTERS
VOL. XXXI
NATURAE SPECIES RATIOQU E
MADISON, WISCONSIN
1938
OFFICERS OF THE WISCONSIN ACADEMY OF SCIENCES,
ARTS AND LETTERS
President
Chancey Juday, University of Wisconsin
V ICE-PRESIDENTS
In Science: Paul W. Boutwell, Beloit College
In the Arts: S. C. Wadmond, Delavan
In Letters: Julia Grace Wales, University of Wisconsin
Secretary-Treasurer
Loyal Durand, Jr., University of Wisconsin
Librarian
Gilbert H. Doane, University of Wisconsin
Curator
Charles E. Brown, State Historical Museum
Council
The President, ex officio
The Vice-Presidents, ex officio
The Secretary-Treasurer, ex officio
The Librarian, ex officio
E. A. Birge, past president
Charles S. Slichter, past president
Louis Kahlenberg, past president
Henry L. Ward, past president
M. A. Brannon, past president
L. J. Cole, past president
S. A. Barrett, past president
Charles E. Allen, past president
Rufus M. Bagg, past president
Committee on Publication
The President, ex officio
The Secretary, ex officio
H. A. Schuette, University of Wisconsin
Committee on Library
The Librarian, ex officio
A. L. Barker, Ripon college
Ira A. Edwards, Milwaukee Public Museum
W. S. Marshall, University of Wisconsin
L. E. Noland, University of Wisconsin
Committee on Membership
The Secretary, ex officio
E. F. Bean, Geological and Natural History Survey
P. W. Boutwell, Beloit College
W. E. Rogers, Lawr Vcge.
0. L. Kowalke, Uni isconsin
Correspondence relating to publica. cz actions or to other Academy business should
be directed to the Secretary, Loyal Duv . r , IS ence Hall, Madison, Wisconsin. Publications
Intended for the Library of the Acaden . .. t directly to the Librarian, Gilbert H. Doane,
University of Wisconsin Library, Madi
CONTENTS
SOU 73
f (/(/ ~7 00 03
Page
A Laboratory Study of the Absorption of Light by Lake Waters (54
text figures) Harry R. James1 and E. A. Birge ................. 1
Microstratification of Inland Lakes (10 text figures) Lester V. Whit¬
ney . . . . . . . . 155
Continuous Solar Radiation Measurements in Wisconsin Lakes (11
text figures) Lester V. Whitney . . . . . 175
Transmission of Solar Energy and the Scattering Produced by Sus-
pensoids in Lake Waters (4 text figures) Lester V. Whitney ... 201
Mineral Content of the Lake Waters of Northeastern Wisconsin (7
text figures) C. Juday, E. A. Birge and V. W. Meloche . . . 223
The Estimation of Magnesium in Lake Water Residues V. W. Meloche
and Katherine Pingrey . . . . . ... _ 277
Sodium and Potassium Content of Wisconsin Lake Waters and their
Residues (18 text figures) D. Lohuis, V. W. Meloche and C.
Juday .......... . . . . 285
Geology and Ground Water of the Trout Lake Region, Vilas County,
Wisconsin (6 text figures) Carl Fries, Jr . . 305
The Distribution of Heterotrophic Bacteria in the Bottom Deposits
of Some Lakes (6 Text figures) Arthur T. Henrici and Eliza¬
beth McCoy . . 323
The Silica and Diatom Content of Lake Mendota Water (8 text figures)
V. W. Meloche, G. Leader, L. Safranski and C. Juday _ ..... 363
Photosynthesis of Aquatic Plants at Different Depths in Trout Lake,
Wisconsin (13 text figures) Winston M. Manning, C. Juday and
Michael Wolf . . . . .......... 377
Amount and Distribution of the Chlorophyll in Some Lakes of North¬
eastern Wisconsin (3 text figures) Zygmunt Kozminski ........ 411
A Study of the Fish Parasite Relationships in the Trout Lake Region
of Wisconsin Samuel X. Cross . . . . 439
Age and Growth of the Sucker, Catostomus commersonnii (Lacepede),
in Muskellunge Lake, Vilas County, Wisconsin (Plate I and 5
text figures) William A. Spoor . . . 457
A Second Report on the Growth of the Muskellunge, Esox masquinongy
immaculatus (Garrard) in Wisconsin Waters (1 text figure)
Clarence L. Schloemer ......... . . . . . ....... 507
0GT95 1938
Growth of the Buffalo in Wisconsin Lakes and Streams (7 text figures)
David G. Frey and Hubert Pedracine . 513
A Census of the Fish Caught by Anglers in Lake Kegonsa (1 text
figure) Chancey Juday and Lawrence E. Vike . 527
Fish Records for Lake Wingra Chancey Juday . 533
Professor C. Dwight Marsh and His Investigations of Lakes Mrs.
Florence W. Marsh . 535
A New Species of Receptaculites ( R . pedunculatus ) from the Silurian
Strata of Eastern Wisconsin (2 figures) W. H. Twenhofel . 545
Topography of Abandoned Beach Ridges at Ellison Bay Door County,
Wisconsin (3 text figures) 0. L. Kowalke and E. F. Kowalke . . . 547
A North American Record for J uncus capitatus Weig. S. C. Wadmond 555
Landlocked Salmon in Wisconsin (Plate II) T. E. B. Pope . 559
Proceedings of the Academy . 567
A LABORATORY STUDY OF THE ABSORPTION OF LIGHT
BY LAKE WATERS
Harry Raymond James
Professor of Physics, Hastings College
With the Cooperation of
Edward A. Birge
Biologist , Wisconsin Geological and Natural History Survey
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 72.
TABLE OF CONTENTS
Page
Introduction . 7
Chapter I. Apparatus and methods . 15
Monochromator . 15
Optical system . 17
Water tubes . 18
Thermopile . 19
Galvanometer . 20
Filters . 20
Room . 21
Reading and calculation . 22
Percentile absorption curves . 23
Chapter II. Absorption of light by distilled water . . . 25
Apparatus and methods . 25
Results, Table I . 30
Comparison with other observers, Table II . 31
Purity of water . 37
Molecular scattering . . . 38
Effect of temperature of water . 39
Chapter III. Absorption of light by lake waters . 45
Section 1. General conditions . 45
Suspensoids, turbidity . 46
Color, standard of . 48
Computation of . 52
Amount and sources . 54
General action of lake waters . 56
Section 2. Observations on lake waters . 59
Treatment of waters . 61
Settling . 61
Filtration . 62
Dilution . 64
Control of bacteria . 64
Spread of single observations . 66
Section 3. Mean results . 67
Settled waters . 68
Filtered waters . 69
Dilutions . 71
Three groups of lakes . 71
Variation of observations . 73
2
Comparison with results of Pietenpol . 74
Results of von Aufsess and Erikson . . 76
Section 4. Absorption curves from typical lakes . 78
Settled waters. Devils and Little Papoose lakes . 79
Filtration. Waters that filter to color 0 . 81
Filtration. Waters of color 6 . 83
Filtration. Groups I and II compared . 85
Filtration. Waters of Group I . 88
Filtration. Waters of Group II . 90
Dilutions . 93
Absorption at 3650 A . . 95,
General tables of percentile absorption . . . . . . 99
Settled waters. Table X . . . . 100
Filtered waters. Table XI . 103
Dilutions. Table XII . 107
Chapter IV. Analysis of absorption of light by lake waters . 110
Types of curves . Ill
Section 1. Remainder curves . 112
Section 2. General factor curves . 114
2- component analysis . , . 114
Single waters . 116
Marl Lake . . 116
Horsehead Lake . 118
Dilutions . . 119
Little Long Lake . 119
Helmet Lake . 120
Factor curves, 7,000-8,000 A . 121
Section 3. Suspensoid curves . . . 123
3- component analysis . 123
Methods . 123
Nagawicka Lake. Factor curves . 125
Nagawicka Lake. Additive curves . 128
Single waters. 3-component . 130
Crystal Lake. Color 0 . 130
Mean colors 4, 6, 10, 20 . 131
Muskellunge and Marl lakes . 133
Lake George. Color 24 . . . . . 136
Adelaide Lake. Color 28 . . . . . 137
Rudolf Lake. Color 50 . 138
Helmet Lake. Color 236 . 140
Use of mean color curve . 142
Two waters of same color . 143
Dissolved and colloid colors . 144
Mean curves for color absorption; 0-180 . 145
Mean color transmission; 7 wave-lengths . 146
Summaries . 148
3
LIST OF ILLUSTRATIONS, PART I.
Page
Figs. 1 and 2. Diagrams of monochromator . 16
Percentile absorption by distilled water.
Fig. 3. General diagram, 4000-8000 A. 4 curves . 26
Fig. 4. Percentile curves, 6500-8000 A. 6 curves . 27
Fig. 5. Percentile curves, 4000-6500 A. 8 curves . 28
Fig. 6. Same as Fig. 5, plotted on semi-logarithmic scale . 34
Fig. 7. Effect of temperature on absorption . 39
General diagrams for lake waters.
Fig. 8. Solar energy curve, with visibility curve . 50
Fig. 9. Absorption by dilutions and by filtrates of same color . 51
Fig. 10. Absorption by observed and computed colors . 52
Fig. 11. Relation of organic carbon and color in lake waters . 55
Fig. 12. Departure of single observations from mean; color 6 . 65
Fig. 13. Departure of filtered waters; colors 3, 4, 8, 16-20 . 66
Mean total absorption curves.
Fig. 14. Settled waters . 68
Fig. 15. Filtered waters . 70
Fig. 16. Dilutions . 71
Fig. 17. Comparison of results, James and Pietenpol . 75
Fig. 18. Von Aufsess and Erikson. Percentile curves . 77
Effect of settling and filtration.
Fig. 19. Settling; Devils Lake and Little Papoose Lake . 79
Fig. 20. Filtration; waters which filter to color 0 . 82
Fig. 21. Filtration; waters of color 6 . 84
Fig. 22. Filtration; waters of Groups I and II . 86
Fig. 23. Filtration; waters of Group I . 89
Fig. 24. Filtration; waters of Group II . 91
Fig. 25. Dilution; filtered water of Helmet Lake . . 93
Fig. 26. Dilution; unfiltered water of Mary Lake . 94
Fig. 27. Absorption curves, 3650-7600 A. 6 settled waters . 95
Fig. 28. Same for 10 filtered waters . 96
Analysis of absorption; remainder and factor curves.
Fig. 29. Total and remainder curves; settled waters . . 113
Fig. 30. Total and remainder curves; filtered waters . 114
Fig. 31. Remainder and factor curves; settled waters . . 115
Fig. 32. Remainder and factor curves; filtered waters . 116
Analysis of absorption; 2-component analysis.
Fig. 33. Marl Lake . 117
Fig. 34. Horsehead Lake . 118
4
Fig. 35. Little Long Lake; dilutions . 119
Fig. 36. Helmet Lake; dilutions . . . 120
Analysis of absorption; 3-component analysis.
Fig. 37. Nagawicka Lake; color 12 . 124
Fig. 38. Nagawicka Lake. Additive Curves . 128
Fig. 39. Crystal Lake; color 0 . 130
Fig. 40. Mean; waters of color 4 . 132
Fig. 41. Muskellunge Lake; color 6 . 133
Fig. 42. Mean; waters of color 6 . 134
Fig. 43. Marl Lake; color 6 . 134
Fig. 44. Mean; waters of color 10 . 135
Fig. 45. Mean; waters of color 20 . 135
Fig. 46. George Lake ; color 24 . 137
Fig. 47. Adelaide Lake; color 28 . 138
Fig. 48. Rudolf Lake; color 48 . 139
Fig. 49. Helmet Lake; color 236 . 140
Fig. 50. Use of mean color; Columbian Lake; color 6 . 142
Fig. 51. Comparison of two waters; Helmet and Mary; color 14 .... 143
Fig. 52. Dissolved and colloid colors . 144
Fig. 53. Mean curves, color absorption; colors 0-180 . 145
Fig. 54. Mean color transmission; 7 wave-length, 4000-7000 A . 146
5
LIST OF TABLES. PART I.
Table I. Absorption of radiation by one meter of distilled water.
3650-8000 A. Stated as coefficients and as percentages ... 30
Table II. Percentile absorption by 10 observers. Stated for each
100 A . . . 31
Table III. List of observers . 36
Table IV. Data for theory of molecular absorption . . 38
Table V. Effect of temperature on absorption by water . 41
Table VI. Observations of Collins and Baldock; stated as coeffi¬
cients of extinction; 5500-8230 A . . 42
Table VII. Observations of Ganz, 12°, 87°; stated as coefficients at
each 100 A, 6950-8200 A . 43
Table VIII. List of Lakes examined; with data . 60
Table IX. Effect of filtration on color and on percentile absorption . . 63
Table X. Settled waters. General table of percentile absorptions . . .100
Table XI. Filtered waters. General table of absorptions . 103
Table XII. Dilutions. General table of absorptions . 107
Table XIII. Mean percentile absorptions at 3650 A . 97
Table XIV. Nagawicka Lake. Computation of factor curves . 126
6
INTRODUCTORY NOTE
Edward A. Birge
This Survey acknowledges with gratitude the assistance
which has been generously granted by the Brittingham Trust
Fund. The entire cost of the present investigation of the rela¬
tions of light and lake waters has been met from this Fund. It
has borne all expenses from the designing and building of the
monochromator to the completion of publication. This aid has
enabled the Survey to undertake and carry through a research
so extensive that it was quite beyond the other resources of the
Survey, and yet was one needed as a foundation for other inves¬
tigations, such as those on photosynthesis in lakes and on the
reaction of their organisms to light.
Grateful acknowledgement is also made of the active cooper¬
ation in the study, of the Department of Physics, University of
Wisconsin. The work of Dr. James was done in the Physics
Laboratory, under the immediate direction of the late Professor
C. E. Mendenhall. His untimely death in 1935 has deprived us
of his aid and counsel in the final revision of the Report.
The present Report was planned as a contribution to our
knowledge of the ecology of lakes and it must be judged from
that point of view. The earlier study of Dr. Pietenpol (1918)
had fully informed us of the variety of the conditions to be in¬
vestigated and of the complexity of the agents that affect light
as it penetrates the water of lakes. There was no expectation
that we should be able to quantify these conditions in any such
sense as the term is rightly used in pure physics ; but the study
has told us quite as much about the ecology of light in lakes as
could be expected. Much has been learned about the effect on
light of color in lake waters, and in the second part of the Report
the percentile results, set forth in Part I, are applied to the solar
energy spectrum. Thus their ecological bearing becomes more
manifest.
In 1933 Dr. James was called to the chair of Physics in Hast¬
ings College, and its duties have claimed most of his time. As a
result it has been necessary for the Survey to edit the Report for
QfT95 MSS
7
8 Wisconsin Academy of Sciences , Arts, and Letters
the press, and much of this work has been done by myself. This
statement applies with especial force to Chapter IV, which at¬
tempts to give a limnological application to the physical data
found by Dr. James for lake waters. The fullness of the data
has permitted suggestions much more significant than those
which were drawn, in like manner, from the pioneer study of
Dr. Pietenpol.
Two matters in the Report call for special attention. The
first relates to the chapter on the action of distilled water on
light. The tables and diagrams which accompany Chapter II
make it clear that much is still to be learned about this part of
the story. The wide differences between the percentile curves,
reported by observers from Hufner and Albrecht to Collins and
Ganz, are so great that they cannot be the result of accident or
of instrument. The percentile absorption found by Dr. James
is by far the lowest among those recorded in the short-wave
spectrum; it agrees with that of the majority about 5500 A;
it is among the highest at greater wave-lengths. Dr. James
tested many samples of distilled water; his readings were care¬
fully made; his results as stated are the mean of several con¬
cordant series. It was not in the plan of the study to make an
exhaustive investigation of distilled water, in its relations to
light, and of the variations which these present. But the ob¬
servations on filtered lake waters are in good general accord
with those on distilled water; they do not look toward the re¬
sults reached by other students where these depart most widely
from those of Dr. James. It would not be easy to give a reason¬
able interpretation to the observations on filtered lake waters, if
the action of distilled water were to be based, for instance, on
the curve of Sawyer, 4000-5000 A; or on those of Aschkinass
or Sawyer, 6000-7000 A. Lake waters and distilled water, as
found by Dr. James, seem to tell the same general story. But it
is plain that new and much more complete studies are needed on
the relations of supposedly pure water and light.
The second subject for special mention is that found in Chap¬
ter IV ; the analysis of the action of lake waters on light. The
observations of Dr. James were not made with the purpose of
using them in such an analysis. The possibility of this use was
seen only after study had been directed to the great mass of
readings which Dr. James had made and had interpreted from
the physicist’s point of view. A study specifically directed to-
Introduction
9
ward such analysis would have been carried on in a different
way. But it seems to the Survey that the results set forth in the
diagrams of this chapter are worth presenting. It is especially
hoped that they may suggest the use of new and better methods
to obtain more complete analysis and quantitative results.
This is only one of the many aspects of the ecology of lakes
which are calling for quantification. Perhaps another one may
be named, on which the study of Dr. James has given us much
information that would enlarge a more detailed and accurate in¬
vestigation. This is the presence, extent, origin, and variation
of spectral bands of local selective absorption of radiation by the
waters of lakes. This subject is not discussed in the present Re¬
port; but certain especially plain cases are pointed out in the
explanations of the diagrams.
The data of this report state the total percentile effect of a
one-meter column of lake water (settled, filtered, or diluted) on
radiant energy in the form of light. The quantity of energy in
the light entering the water is compared with that emerging
from it. No observations have been made, within the water
itself, on the processes affecting light which go on there.
Madison, Wisconsin. June 10, 1937.
A LABORATORY STUDY OF THE ABSORPTION OF LIGHT
BY LAKE WATERS
Part I. Percentile Absorption of Radiation in the
Visible Spectrum
Harry R. James
Introduction
The transmission of white light through a material medium
is always accompanied by the loss of some of the radiant energy
by absorption; that is, some of the energy is transformed into
heat, chemical energy, or to some other frequency of radiation.
When the light passes through a medium which transmits all
wave-lengths uniformly the emergent light will still be white;
but if some part, or parts, of the spectrum are absorbed more
than others the phenomenon of selective absorption is observed
by the absence, in the transmitted light, of the wave-lengths so
absorbed and the medium will show a color that is determined
by the wave-lengths transmitted. All transparent media exert a
selective action in some part of the spectrum, though the action
may not be in the visible region.
Pure water exerts a selective action which varies greatly
throughout the visible spectrum. It is very transparent to blue
and violet but absorbs red very strongly. If a material which
affects the transmission of light in any way is mixed with the
water, the effects of the water and of the material will be added
and a new absorption spectrum will be given.
The present study was undertaken at the suggestion of Dr.
E. A. Birge, biologist of the Wisconsin Geological and Natural
History Survey, with the purpose of investigating the selective
action on light of waters from the inland lakes of Wisconsin,
with particular reference to the effect of coloring matters con¬
tained in the waters. The work is part of the study of the trans¬
mission of radiation in the waters of inland lakes, which has for
many years been conducted by Dr. Birge and Dr. C. Juday. They
have been attempting to determine the quantity and the com¬
position of the light found at various depths in these lakes ; with
11
OCT 25 1938
12 Wisconsin Academy of Sciences, Arts, and Letters
the ultimate purpose of ascertaining the effect on plant and ani¬
mal life. In this way the transmission of solar energy through
lake water and the varying forms of the solar energy curve can
be given definite values as ecological factors in the life of these
waters.
The relations of solar radiation and lake water offer much
more complex problems in a small inland lake than in the ocean
or in the great lakes. The colors or stains present in such waters,
and derived from plankton or from marshes adjacent to the
lake, exercise an important selective action on the light, which
is practically absent from larger bodies of water. The study,
therefore, of the effect of color of water on absorption of light
is a matter of prime importance in the investigation of the ecol¬
ogy of small lakes.
During the years of 1914-16, early in the work of the Survey
on lakes, Dr. W. B. Pietenpol (1918) made an investigation
somewhat similar to the present one. He examined about 59
specimens of water from Wisconsin lakes; and this was the
first extensive study of the kind made on waters of small inland
lakes. He reported his results in terms of the absorption coeffi¬
cients at 12 wave-lengths from 4700 A to 6630 A; these were
the limits set by his instrument. The waters examined were
filtered through a coarse Berkefeld filter and therefore one with
large pores. To this report Dr. Birge added a brief discussion
of the relation between absorption of light and color of water,
based on Pietenpol’s work.
As the Survey continued its work on lakes, it became evident
that there must be a new and more extensive laboratory study of
the subject. Its necessity was emphasized by the extension of
the Survey’s work to the lakes of the Northeastern District of
Wisconsin. In these waters ecological conditions are extraordi¬
narily various, and, in particular, their color ranges from zero
through many types of yellow and red until it becomes very
dark. Much enlargement of knowledge was, therefore, immedi¬
ately necessary in order that light might receive proper con¬
sideration as a factor in the ecology of lakes. The action of lake
waters on light must be determined at all wave-lengths of the
visible spectrum; and careful attention must be given to color
in lake waters as a factor modifying the composition of the
spectrum. The present report is part of the result ; Part I, the
present paper, is devoted to a discussion of the percentile ab-
James & Birge — Lake Waters and Light
13
sorption of radiation by water and its content as found in lakes
and with especial attention to the effect of color on radiation;
in Part II, which is nearly ready for the press, these percentile
results are applied to a standard solar energy curve, which
shows the amount and the composition of the solar energy de¬
livered to the surface of the lake. Thus are determined the
changes in quantity and in composition which the energy spec¬
trum suffers as it penetrates the waters of lakes. In all of the
study the unit of lake water is taken as a stratum one meter in
thickness; absorptions are stated as percentages, not as coeffi¬
cients of extinction.
In Part I of the report the main subjects discussed are as
follows :
1. A new study of the selective ' action of distilled water on the
visible spectrum. With this are presented results from other
observers and a brief account of the effect of temperature on
the absorption of light by distilled water. The results are
given in tables and in Figs. 3-7.
2. A study of the relations between lake ivaters and radiation in
the visible spectrum. It comprises more than 180 series of
readings on waters, in different conditions, from 50 lakes;
in which were determined their percentile absorption of
radiation at 21 (or 22) wave-lengths of the spectrum. These
were placed from 4078 A (more rarely 3650 A) to 8000 A;
thus covering the visible spectrum. The determinations are
reported in tables showing percentile absorption in settled
and filtered lake waters, and in dilutions made from waters
of high color. From these data curves of total percentile
absorption have been prepared and many of them are shown
in Figs. 14-26. Two diagrams are also given to show ab¬
sorption from 3650 A to 8000 A in settled and filtered waters.
3. Total percentile absorption is determined by the combined
action of several factors . For certain lake waters a partial
quantitative analysis of these factors has been made and
illustrated by diagrams. In a 2-component analysis total ab¬
sorption is assigned to factors of two types, viz., pure water,
whose action is the same in all lakes ; and the united effect of
color, suspensoids, and all other minor factors operating in
the lakes. This division of total absorption can be made in
14 Wisconsin Academy of Sciences , Arts , and Letters
all waters, and some cases are shown in Figs. 29-36. In some
waters a 3-component analysis is possible, in which quantita¬
tive values are assigned to the action of water, of color, and
of suspensoids. Figs. 37-54 contain factor curves which
show the differing action of these factors in different parts
of the spectrum, both in waters of single lakes of different
colors, and in mean results derived from groups of lakes
having the same color rating.
CHAPTER I
APPARATUS AND METHODS
In designing the apparatus provision was made for handling
large quantities of radiant energy and for a sensitive and
steady means of measuring energies.
The optical system consists of a monochromator of special
size with attendant external lens system for forming a parallel
beam of light, to be passed through the water to be tested, and
focused on the slit of a monochromator. After dispersion in the
monochromator the light energies are measured by means of a
thermopile and galvanometer. Two different sources of light
are used; a tungsten strip filament gas-filled lamp, designed
especially for spectroscopic work, supplies the energy for the
spectral region from 8000 A to about 4500 A; and a quartz
mercury arc lamp is used in the region from 3650 A to 5460 A.
The two ranges thus overlap enough to give a close check on the
relative behavior of the two sources. Fig. 1 shows a perspec¬
tive drawing of the arrangement of the optical system, together
with the tubes for holding the water to be tested and the com¬
parison cell ; all as used for the work on distilled water. Fig. 2
shows the plan of the apparatus.
Fig. 1 shows the monochromator, lens mountings, and speci¬
men tube supports, all carried upon a 6-inch steel channel beam
4.2 meters long, to hold all parts in correct relation to each other
when adjustments are made. In this figure the galvanometer
and scale, thermopile, light sources, monochromator cover, shut¬
ters and screens for guarding against stray radiation, are all
removed for clearness in showing the basic system. The tube
mounts are attached to an iron pipe 5 cm. in diameter ; the pipe
is mounted upon adjustable pivot bearings at each end and the
center is supported by a straight bearing to increase the stiffness
of the movable system. Two of the tube supports are fitted with
curved iron arms which extend around and under the edge of the
table, and hold weights to counterbalance the weight of the tubes
and contents so that they are in neutral equilibrium in all posi¬
tions of their swing in and out of the beam of light. This ar-
15
E5~i
16
Wisconsin Academy of Sciences, Arts, and Letters
d °
S a fl
S & §
£
O 0)
XS m fn
W>'& o
3 §^
T g cd
g ©^ i
t> - p3 <
a*s g,~
m 1 CJ —
© <U
> JT<. .
^ — H.;
(3
d &b
fl
g s
> o
^ H\p
,C So
O OT 0)
2 '£ 'S
w a g
fl O *S
fl d
*• fi
|w
d w
. H
s^sl§
■SpqSU^
W <dP-i m
® g
& H
fl 13
S ° •
S “ S
£ &0-P
u a tn
si!
§ c “I
-2 K -B
^•S-g
is a!
© fl ©
» o+j ;
a u :
m S :
43 ojO .
_ fl &0
2*C .S
fl ts 5
flog
5s s
Srfl
g ^
© ©
o ^ I
© ©
© fl &p
§U S
tH tH tH pq gQ C<J
hlHO
S -2 ^ rcs
O fl fl ©
OHO £
,fl ?-,
60 fl
£50
James & Birge — Lake Waters and Light
17
rangement avoids possible torsional effects upon the optical sys¬
tem due to shifting the weights of the tubes. The lenses are
mounted in tubes with rack-and-pinion motions for focusing and
these tubes are carried on adjustable mountings set upon cast-
iron carriages; these carriages have been machined so as to
slide along the steel channel beam to approximately correct fo¬
cusing position, where they may be clamped. Screw motion is
provided for controlling the lateral position of the lens mounts
in the light beam. The system as diagramed in Fig. 1 weighs
approximately 120 kg. when completely assembled for taking
observations.
Details of Construction of Individual Apparatus
Monochromator
The monochromator was designed by the author and was
built by Mr. J. P. Foerst, mechanician for the Department of
Physics in the University of Wisconsin. The design was based
on the familiar Wadsworth mounting for right-angle deviation
described by F. L. 0. Wadsworth (1894:37). The prism is car¬
ried upon a table which is turned by a screw upon which is
mounted a drum marked with a uniform scale for use in cali¬
brating for wave-lengths. In this type of mounting a surface
silvered mirror is mounted with the prism at such an angle as
to reflect the dispersed radiation at 90 degrees with the original
path of the beam. (See Figs. 1 and 2.) The prism is mounted
so that the wave-lengths observed at the emergent slit have come
through at minimum deviation.
The prism was cut from light flint glass of refractive index
1.62 at 5500 A. It is 6.3 cm. high and 9.0 cm. length of face.
The refracting angle is 60 degrees.
The lenses are achromatic, requiring 3 mm. motion in focus¬
ing for wave-lengths between 4000 A and 8000 A. They are
6.3 cm. in diameter, matching the prism height, and the focal
ratio is f :4. The lenses are moved by rack-and-pinion controls
operated by graduated knobs so that settings of the lenses can
be repeated accurately. The whole assembly is covered with a
tightly fitting metal cover, blackened inside, and with a series of
baffle plates or screens built in to cut down stray light from re¬
flections from the lenses and face of the prism. The slits of the
monochromator open symmetrically and are operated by screws
18
Wisconsin Academy of Sciences, Arts, and Letters
fitted with graduated heads for setting to known widths. The
slit widths used in this work are 0.125 mm. for the spectral re¬
gion from 5800 A to 8000 A; and 0.50 mm. for wave-lengths
between 4000 A and 5800 A. The 0.125 mm. slit subtends a
spectral range of 35 A in the neighborhood of 6200 and the 0.50
slit subtends a range of 32 A in the region about 4000 A.
The prism was calibrated for wave-lengths by reference to
55 known lines from the following spectra : quartz mercury arc,
spark spectrum of zinc and cadium, flame spectrum of sodium
and potassium, and the spectrum tubes of argon, helium, and
hydrogen. The lines were identified by reference to “Atlas of
Spectra” by Eder and Valenta (1904) and by the usual com¬
parison methods. The longest visible line used in calibration
was 7703 A, and the setting of the drum for 8000 A was deter¬
mined from the dispersion curve furnished with the prism, and
further checked by the Hartmann dispersion formula.
External Optical System
The lenses of this system are of the same grade and focal
ratio (f :4) as those used in the monochromator. The diameter
of these lenses is 8.3 cm. The prism and all the lenses were
especially made for this investigation by the Bausch and Lomb
Optical Co. The focal lengths of all lenses were measured at 6
points in the spectrum, so that in use the lenses could be set
close to the correct position to focus any wave-length desired.
Water Tube and Comparison Cell
The tubes for holding the water for testing are of brass
8.6 cm. in diameter and made in sections of 0.50 meter and 1.0
meter. The sections are fitted with end collars so that they can
be bolted together, making longer tubes when desired. For the
work with distilled water, the inside walls of the tubes were
silver plated (see p. 29), but for the lake water the tube walls
were blackened by a chemical process and a coating of paraffin
or of white ceresin wax was melted over this to avoid contamina¬
tion of the water.
The ends of the tubes are closed with glass discs cut from
plate glass 3 mm. thick, the faces of which are plane parallel
within a few seconds of arc, as tested by a spectroscope and
Gauss eye-piece. The glass discs are mounted upon rubber gas-
James & Birge — Lake Waters and Light
19
kets at the ends of the tubes and they are held tightly in position
by a heavy brass ring, which in turn is held by screws passing
through the ring and into the collar at the end of the tube. The
rubber gaskets are soaked for a time in hot ceresin wax to coat
them, so that water cannot come in contact with the rubber, from
which it might become contaminated. It was found necessary
to place several gaskets under one of the end-plates, making a
spongy mounting, so that the position of the plate could be
shifted slightly with the screws to bring it parallel with the
other end-plate and perpendicular to the axis of the tube. This
arrangement was necessary to prevent the long tube of water
from acting as a prism and deviating the beam. In working
with distilled water, this end mounting was much modified
(p. 29).
The short tube, T-2, shown in Figs. 1 and 2, is empty except
for the comparison cell mounted at one end; the tube is used
only for convenience in holding the cell in position. The com¬
parison cell is made from two plate glass discs cut from the
same stock as that used in making the end plates for the water
tubes; the absorption of light in each of these was found to be
the same as in the others. The discs in this cell are separated
by a film of water about 0.02 mm. thick. Thus, when light passes
through the glass of the comparison cell, it suffers the same
losses from reflection and absorption as in passing through those
of the long water tube, and the only difference is the absorption
due to the longer column of water in the long tube.
It was necessary to boil the distilled water used in the com¬
parison cell so that air bubbles would not form within the cell
after being assembled.
Thermopile
The thermopile is the single junction compensated type de¬
signed according to the theory given by Dr. C. H. Cartwright,
(1930). The elements are of bismuth and tellurium and the
receiver is a piece of silver foil 5 mm. long and 1.50 mm. wide,
coated with platinum black. The junctions are mounted in a
brass case, with a quartz window, arranged to be evacuated. A
high vacuum is maintained by means of cocoanut charcoal cooled
by liquid air. The tellurium used for the junctions was first
annealed by heating it slightly above the transition temperature
between the alpha and beta forms of the crystal (360 degrees)
20 Wisconsin Academy of Sciences , Arts, and Letters
in an atmosphere of nitrogen and then gradually cooling to room
temperature. This reduced the resistance of the tellurium by
64 per cent and reduced the thermo-electric power against plati¬
num by only 15 per cent, so that there was a distinct improve¬
ment in the character of the metal for use in thermopiles. In
working with this tellurium it was noted that the thermoelectric
power varied with the difference of temperature between the
hot and cold junctions, the value being smaller for small differ¬
ences in temperature.
Galvanometer
The galvanometer is the Type ZC, 15-ohm instrument manu¬
factured by Kipp and Sons of Delft, Holland. It requires 45
ohms of external resistance for critical damping, and this re¬
sistance is approximated in the construction of the thermopile.
The sensitivity of the galvanometer is variable by means of a
magnetic shunt, and this also affords a means of bringing it into
harmony with the external resistance for critical damping.
The period of the instrument is 4.8 seconds and the sensi¬
tivity is 7.8 X 10 10 amperes per mm. at 1 meter. In use the
scale distance is 217.3 cm. and under these conditions the over¬
all sensitivity of the pile and galvanometer system is such as to
give 1 mm. deflection for a radiation of 0.044 ergs per second
incident on the receiver. This indicates a rise in temperature of
1.01 X 10’4 degrees centigrade at the junction.
The pile is protected from sudden temperature changes by
wrapping it in wool, while the entire galvanometer circuit from
the terminals on the brass pile case to the metal box within
which the galvanometer is mounted, is enclosed in metal tubing
or wrapped in tin foil; and the entire shielding grounded at
each end of the line to protect the system from stray electrical
effects. One side of the galvanometer circuit is grounded to pre¬
vent the accumulation of static charges on the coil, causing it to
stick to the pole faces. With these precautions the zero posi¬
tion of the galvanometer rarely drifted more than 1 mm. in 10
minutes.
Infra-red Filter
It is necessary to filter out the strong infra-red radiation
from the strip filament lamp, since it interferes with observa-
James & Birge—Lake Waters and Light
21
tions at the short wave-lengths. This is done by use of a water
cell 25 cm. long, placed in the parallel beam of light; or by a
Jena glass filter, BG-7. Either of these filters is effective but
the water filter is more satisfactory because there is less ab¬
sorption in the range of the spectrum in which observations are
being made.
Room Conditions
The room where the apparatus is assembled is built within
a larger room; it has only one outside window, which is kept
closed and covered with heavy cardboard and beaver board. The
steam radiator is shut off so that the room is heated only by con¬
duction and slow convection from the surrounding rooms. In
this way sudden variations of temperature in the room are made
impossible.
Collimating System
It was found in practice that a parallel beam much more
free from stray light can be obtained by use of a small achro¬
matic condensing unit of about f :1,5 aperture, which is set near
the source and forms an image of the source on a screen fitted
with a variable slit. This slit is at the principal focus of the
collimating lens and is really the source of light for the test
system. The condensing lens is set at such distances from the
screen and light source that the collimating lens is filled with
the light. A stop is placed in the collimating lens so that a para¬
llel beam 5.50 cm. in diameter is formed for the test system.
Filters and Filtration Methods
All filtration s of water are done with a Berkefeld filter; one
of fine and one of medium porosity being used. The water is
filtered into a large bottle within which a partial vacuum is
maintained by means of an aspirator. A mercury manometer
tube is attached to the aspirator tubing so as to show the vac¬
uum conditions within the bottle. The smallest practicable dif¬
ference in pressure is used and usually not more than 10 cm. of
vacuum is required.
There is enough dust in the air of the room to make an ap¬
preciable difference in absorption of light by the clear samples
of water ; and accordingly the air, that passes through the stop-
22 Wisconsin Academy of Sciences, Arts, and Letters
cock used to regulate the vacuum within the bottle, is first drawn
through a moist cotton plug to remove the dust.
Method of Taking Readings
In taking readings on a sample of water, the light source is
first heated to equilibrium temperature and the zero position of
the galvanometer is recorded. A deflection is then observed
through one tube and another zero reading made, after which a
deflection is taken through the other tube and the zero again
read. This procedure gives three readings of the zero position
for each wave-length setting ; but, since the steadiness of opera¬
tion is usually such that the zero does not shift during a series
of several readings, the second zero reading is usually omitted,
thus saving about 30 per cent in the time required to make a
determination of absorption at any wave-length. However, if
the zero position shifts during a reading the whole is discarded
and a new reading made. There are commonly 21 different wave¬
length settings made in running a test of a water and after a
series has been completed a few scattered points throughout the
spectrum are rechecked to note any change in behavior of the
apparatus, or character of the water during the run. If any of
these recheck values varies more than one per cent from the first
determination a recheck is made of all the neighboring points
to locate the source of the variation.
Calculations of Absorption
Calculations of absorption from observations made with the
one meter tube are based on the following: — Let (do) be the
deflection observed when the light comes through the comparison
cell and (d) be the deflection through the cell containing the
water to be tested where (L) and (I) are the corresponding in¬
tensities of light which produces these deflections. Then
kl = cd and KL> = cdo where (k) is the common per¬
centage of loss by reflection and glass absorption in the two cells.
Then also I = Cd and I> = Cdo where C = — 1C—
k
Now the percentage in terms of intensities is P =-— j — - X 100
and therefore by substitution for the intensities from the above
equations
7 d x ioo
do
P —
James & Birge — Lake Waters and Light
23
If the two-meter tube has been used, the calculation of per cent
absorption per meter is conveniently done by first calculating
the absorption coefficient (A) of the equation I = L eAX where
(Io) is the energy transmitted through the comparison cell, (I)
is the energy transmitted through the water tube, (x) is the
length of the tube in meters, and (e) is the base of the Napier¬
ian system of logarithms. From the above relation
A = — — (log Io — log I)
x
and since the galvanometer deflections (do) and (d) are tested
to be proportional to the energies (Io) and (I), the deflections
may be substituted directly in place of the energies and the
value of (A) is given by
A = — ^ (log do — log d)
Method of Presenting Results
The result of observation at any designated wave-length is
stated in terms of the percentage of radiation which is absorbed
by one meter of the lake water which is under examination.
The percentages thus obtained are used as ordinates, with wave¬
lengths as abscissas, to plot percentile absorption curves , from
which come the numerous diagrams of Part I. These are in¬
tended to present to the eye the effect of these waters on the
spectrum.
The lake waters have yielded some 3650 such percentages,
which are recorded in Tables X, XI, XII, pp. 100-109; these
tables contain the data on which the report is based.
The curves of the diagrams necessarily show both the per¬
centile absorption of radiation by a meter of water, and its
complement, the percentile transmission. Absorption is empha¬
sized in the make-up of the diagrams, since such results have
commonly been expressed in terms of coefficients of extinction.
This Survey has found the use of percentages more convenient
than that of coefficients.
In computing the ordinates for the various types of curves
discussed in Chapter IV, recourse is had to percentile transmis¬
sion ; but it has not been necessary to alter the character of the
diagrams. Fig 53, p. 146, offers a diagram as a specimen of per-
24 Wisconsin Academy of Sciences , Arts , and Letters
centile transmission of radiation, as affected by color in lake
waters.
In Part I little reference is made to mean absorption or trans¬
mission, in various spectral regions, as found in any water.
This aspect of the general subject is discussed in Part II, in
which these percentages are applied to a solar energy curve.
References
Cartwright, C. H. 1930. General theory, design, and construction of sensi¬
tive vacuum thermopiles. Rev. Sci. Instru. 1 :592.
Eder, J. Mi., and E. Valenta. 1904. Beitrage zur Fhotochemie und Spectral-
analyse. Wien.
Wadsworth, F. L. 0. 1894. An improved form of Littrow spectroscope. Phil.
Mag. 38, series 5 :137.
CHAPTER II
ABSORPTION OF LIGHT BY DISTILLED WATER
A study of the absorption spectrum of distilled water was
made, extending over the spectral region 3650-8000 A, and thus
covering the visible spectrum. The results furnish a basis for
the comparison and analysis of the observations on lake waters,
which, so far as possible, cover the same region. The study was
needed, since no series of observations on distilled water had
been made, which went through the whole visible spectrum ; and
no later study has attempted to do so.
Distilled water was examined in 1931 and 1932, and some of
the results were checked in 1936. The work was done with all
reasonable care; but it did not come within the general plan of
the investigation to make an exhaustive study of the relations
between distilled water and radiation; still less to furnish a
standard absorption curve for this relation. Readings were
made at 29 wave-lengths of distilled water, only a few more
points than were used for the lake waters. These were quite
enough for ecological purposes but much too far apart to be used
in discovering bands of selective absorption or transmission ; or
in determining the limits of such bands.
Apparatus and Methods
The apparatus employed was the monochromator already de¬
scribed (p. 15). A two-meter tube was employed for readings
in the short-wave region of the spectrum; in the region 7000-
8000 A both one-meter and two-meter tubes were used. The
inside of these water tubes was coated with paraffin or ceresin
for the earlier observations ; in later ones the tubes were heavily
plated with silver.
The samples of water were prepared with great care; the
best samples were obtained by double distillation in a tin still,
whose condensing column was flushed with a continuous stream
of air that had been freed of dust and carbon dioxide by passing
through a suitable train of solutions. The water was run di¬
rectly into the tubes of the monochromator through pyrex glass
25
26
Wisconsin Academy of Sciences , Arts, and Letters
Fig. 3. Curves showing the percentage of radiation absorbed by one
meter of distilled water, as determined by various observers. Only 4 curves
are plotted, in order to avoid confusion. For details and for other curves
see Figs. 4, 5, 6. The main curve (J) shows absorption as found by the
present report; it is lower than other recent work in the region 8000-7500
A, higher at 7000^-6000 A, and lower again at 5000-4000 A. These results
agree with those of Aschkinass (As), 7000-8000 A, so closely that much of
the latter curve cannot be plotted; the curves diverge widely, 6500-7000 A.
Sawyer’s curve (S) begins at 6500 A, where it has absorption lower than
any other observer; in the violet Sawyer’s results are higher than others.
For high absorption in the long wave region the curve of Collins is plotted,
since it shows the greatest absorption to 7000 A.
James & Birge — Lake Waters and Light
27
Fig. 4. Principal percentile absorption curves, 6500-8000 A. The
heavy line (J) shows the results of the present report. Aschkinass’ curve
(As) is plotted so far as it does not run too close to that of James. Note
their wide divergence, 7000-6500 A. The work of Collins (C) and Baldock
(B) at 25° and 2;6° is treated as a single curve, interrupted at 7000-7030
A. The curve of Ganz (G) is that for 12°; it shows absorptions close to
those of Collins, 7400-8000 A, but much lower, 6950-7400 A.
28 Wisconsin Academy of Sciences , Arts , and Letters
Fig. 5’. See opposite page.
4000 4500 5000 5500 6000 6500
James & Birge — Lake Waters and Light
29
tubing that had been well steamed out. The openings where the
pyrex tubing connected with the still and the water tube were
closely covered with moist absorbent cotton, to exclude atmo¬
spheric dust; and all air exchanges incident to the filling, rins¬
ing and emptying of the water tubes took place through filter
plugs of moist cotton. In order to protect the water from con¬
tamination with the brass tubes, their interior was silver plated ;
and to avoid contact of the water with organic matter that might
color it, the gaskets for sealing the joints and end-plates of the
tubes were cut from hard surface filter paper. The filter papers
had been soaked in a solution of cellulose acetate in acetone, and
were thoroughly dried afterward. The cellulose compound bound
the fibers of the paper so that they did not wash into the water
and form scattering particles. The mounting of the end-plates
was made flexible by means of a short section of copper Sylphon
bellows (Sylphon Co., Knoxville, Tenn.), which was also silver
plated inside. Thus the water had no opportunity to take up
organic matter, such as might be found in the water when the
tubes are lined with paraffin or similar organic substance, and
when the gaskets are of rubber or such material.
Water treated as described was very clear and of a delicate
blue color in transmitted light; when viewed at a small angle
with the 5.5 cm. beam of parallel light coming through the two-
meter tube, only an occasional scattering particle could be seen,
and in the best samples these disappeared after settling for 12
hours. In these samples a beam of light 1 cm. in diameter, such
as was used by Pietenpol (p. 74), gave no trace of its presence;
the tube was optically empty. Water samples that did not meet
Fig. 5. Percentile absorption of radiation in the middle and short-wave
spectrum. The results of 8 series of observations are shown: Aschkinass
(As), von Aufsess (v A) Baldock (B), Ewan (E), Hufner and Albrecht
(H & A), James (J), Pietenpol (P), Sawyer (S). Note that 5 of these
curves are in good agreement at 5500 A and that they diverge as they de¬
part from this wave-length. At 6500 A the spread of the curves is about
12 per cent; and two of the most recent, James and Sawyer, have nearly
the maximum difference. The same general statement may be made for
conditions at 4000 A. Aschkinass’ curve shows the two bands of increased
spectral transmission, found by him but not by others. Von Aufsess reports
a great increase of transmission beginning in the same spectral region as
Ashkinass’ second band.
30
Wisconsin Academy of Sciences , Arts, and Letters
Table I
Absorption of radiation by one meter of distilled water; stated as
coefficient of extinction and as percentile absorption.
Note. — Table I shows the absorption of radiation by 1 meter of dis¬
tilled water, as determined from water contained in a tube which was
silver plated inside. From 8000 A to 5970 A these determinations agree
with those made from water in tubes coated inside with paraffin; at shorter
wave-lengths absorption is lower in silver lined tubes. See Table X for
differences.
this standard of clarity invariably showed higher absorption in
the regions of blue and violet.
Table I shows the absorption of radiation by one meter of
distilled water. The results are derived from the four clearest
waters which were treated as described above, and the readings
in the region 7000-8000 A were checked in two other waters,
using a water tube one meter long. In order to show the rela¬
tion of the single waters to the mean, the readings of the four
waters are platted in Fig. 6 at three points nearest the mini¬
mum. They show that the several readings were in good agree¬
ment, and the same concordance was equally marked at all
points of the spectrum.
The data for the general absorption curve are derived from
the observations given in Table I; they are stated in Table II
for each 100 A of the spectrum from 3650 A to 8000 A. They
are better discussed by comparison with results found by other
observers, and this is done in the next section.
In Table II are the data for 8 series of observations on dis¬
tilled water, counting as one series the results of Collins and
Baldock. Each of these series covers 1500 A or more; there is
James & Birge — Lake Waters and Light 81
Table II
Percentage of radiation absorbed by one meter of distilled water as
recorded by several observers .
32
Wisconsin Academy of Sciences, Arts, and Letters
Table II, Continued
Note. — Table II states percentile absorption of radiation by 1 meter of
distilled water, as determined by several observers. The data are derived
from curves which have been plotted from the observations and are given
for each 100 A of the spectral region covered. The results for Aschkinass,
von Aufsess, and Sawyer are derived from the tables in Sawyer’s report
of 1931. The data for Collins, Baldock, and Ganz are stated in Tables VI
and VII. Those of Collins are for water at 26° ; those of Baldock are for
25°; those of Ganz for 12°.
also a ninth series, that of Ganz, for the region 7000-8000 A.
Several of these series extend beyond the visible spectrum, but
in the present paper tables and curves are kept within the limits
of the study of lake waters, 3650-8000 A. The list of observers
follows : ' ■! <
Fig. 3 shows on a small scale the absorption curves, 4000-
8000 A, derived from several of these observers ; other diagrams
are on a larger scale so as to show more curves and greater
detail. Fig. 4 contains the more important curves for the region
6500-8000 A ; Fig. 5 has two diagrams on larger vertical scales,
one giving 4000-6500 A and the other beginning at 5800 A, In
Fig. 6 eight curves are platted, on semi-logarithmic paper, from
3600 A to 6500 A, in order to bring out more clearly their direc¬
tions and their details at minimum absorption.
It might be well to let tables and diagrams speak for them¬
selves, but a few of their more striking features will be pointed
out. Consider Fig. 5 at 5500 A; there 5 curves, including that
of James, are close together, within what may fairly be called
the limits of accidental variation, absorption being 2. 9-3. 6 per
cent. Three curves are much higher; two of them come from
the last century, and their position might be thought a case of
James & Birge — Lake Waters and Light
33
early work, were it not that Baldock*s curve is both the latest
and that with highest absorption, 8.6 per cent.
If the 5 curves in good agreement at 5500 A are followed
either to greater or lesser wave-lengths, they diverge; at 6500
A, three of them, those of Pietenpol, Aschkinass and Sawyer,
are much below the others, while two, those of von Aufsess and
James, have joined the group which had high absorption at
5500 A. James records at 6500 A an absorption of 28.6 per
cent; the older records of Ewan and of Hiifner and Albrecht
are practically the same ; while that of Baldock is a little higher,
31.2 per cent. At the other extreme is Sawyer, with 18.9 per
cent, while Aschkinass has 22.0 per cent. Thus at 5500 A Saw¬
yer and James differ by 0.8 per cent and at 6500 A by 9.7 per
cent; while on the other hand Baldock is 5.1 per cent above
James at 5500 A and only 2.6 per cent at 6500 A.
The table and Figs. 3 and 4 show an interesting relation be¬
tween the observations of James and Aschkinass in the long¬
wave spectrum. For much of the region 8000-7000 A their
curves are so close that both cannot be platted on the same dia¬
gram; at 7000 A they register the same. From this point they
diverge widelv, with a maximum difference of nearly 8 per cent
at 6700 and 6800 A, Aschkinass having the lower value; and
they come together only at 5500 A, and then separate again.
The recent and very careful series from Collins and from
Ganz may be compared in the region 7000-8000 A. Fig. 4 and
Table II show that their curves are close together from about
7400 A to 8000 A; but at smaller wave-lengths Ganz finds a
lower absorption. At 7000 A both Collins and Baldock find ab¬
sorption of about 48.3 per cent (extrapolating Collins* curve
from 7030 A), while Ganz finds only 36.6 per cent, 12 per cent
lower. The readings of Ganz were made at the temperature
of 12° C and those of Collins at 28° ; but both observers agree in
finding little effect of temperatures of 25° or lower at this wave¬
length. It should be noted that the studies of Collins, Ganz, and
Baldock were made with the purpose of determining the differ¬
ential effects of temperature on absorption by distilled water;
and their several series at various temperatures had to show
concordant results. This purpose, therefore, demanded a care
which gives an unusual value to the differences in their results.
34 Wisconsin Academy of Sciences, Arts, and Letters
3600 4000 4500
James c& Birge — Lake Waters and Light 35
In Fig. 6 the course of the absorption curves is especially well
shown near the minimum absorption and in the short-wave
spectrum. The curves for Ewan, Sawyer, and James are rough¬
ly straight lines in the region 4500-3600 A, indicating that the
percentile increase of absorption has a fairly linear relation
with decreasing wave-length. This rate of increase is much
greater in the curves of Ewan and Sawyer than in that of James.
At 3650 A Sawyer registers 21.6 per cent while James has only
3.6 per cent.
The minimum for these curves is found at about the same
point in the spectrum, and it is reached by a regular and natural
approach from both directions. The minimum found by James
is so much below that of others that the readings of the four
waters from which it is derived are platted in Fig. 6 to show
their general concordance.
Figs. 4 and 6 show the bands of increased selective trans¬
mission of radiation found by Aschkinass and von Aufsess. No
other investigators have been able to verify these bands ; in the
present study careful search was made for them, but failed to
find any trace. It is perhaps worth noting that the rapid in¬
crease of transmission found by von Aufsess came in the same
spectral region as the most marked band found by Aschkinass.
Von Aufsess found that this increase of transmission went much
farther and did not recover within the limits of his instrument.
Pietenpol’s curve also has a small increase of transmission in
the same region, so small that the coincidence may be accidental.
Many other matters of this type appear in the diagrams, but
those that have been given serve to emphasize the important
fact that there is a wide divergence in the results reached by the
Fig. 6. Absorption curves of Fig. 5 plotted on semi-logarithmic paper.
This method brings out two main points: A. The curves are widely sepa¬
rated at their minima so that each can be readily traced. B. A regular per¬
centile increase or decrease of absorption is indicated by a straight line
curve. Note that in general, minima are reached by what may be called
a reasonable course; the minimum cannot be much altered, except with
changes in the general course of the curve.
Readings for the single waters used for the mean in the present re¬
port are plotted at the minimum and on each side of it. Readings elsewhere
are in similar agreement. The curve of Aschkinass is much like the others
if the local bands of increased transparency are omitted.
86 Wisconsin Academy of Sciences , Arts, and Letters
several students of the absorption spectrum of distilled water.
Their agreements and differences are so great and so various
that they cannot be attributed to the type of instrument used or
to their care in using it. So far as we can learn, their mean
error of observation was of the same order of magnitude. Among
the older series, that of Ewan (1894, p. 152) contains 10 dupli¬
cate readings. They show variations of the same general mag¬
nitude as those reported by Ganz (1936, p. 337), and more
elaborately worked out by Sawyer (1931, p. 85). The mean
departure of the single readings employed in the present paper
are of the same type, and the same may be said of those of
Pietenpol.
Ganz (p. 334) rightly calls attention to the gain in accuracy
by the use of a narrow beam of light. Quite appreciable errors
may be caused by the use of a wide beam at spectral points where
absorption is rapidly changing, such as 6000 A or 7200-7400 A.
Ewan’s reading at 7300 A seems to be derived from a spectral
band whose limits are 7500-7080 A; and this condition would be
partly responsible for the low absorption there recorded. Hiif-
ner and Albrecht also used wide spectral bands, but this could
not have produced the high absorptions recorded by them for
5500-4500 A. In general, only a small part of the differences in
the run of curves can be thus explained.
Most observers of distilled water have not recorded its tem¬
perature ; no doubt their readings were made at “room tempera¬
ture”, which is far from a constant. Several studies have shown
that in general water at higher temperatures has an increased
capacity for absorption of radiation, and that at certain spectral
Table III
Observers of radiation in distilled water vjhose results are shown and
discussed in the present report.
James & Birge — Lake Waters and Light
37
regions temperature has a variable effect. This matter is dis¬
cussed in the next section.
It seems, therefore, that the differences presented by the
absorption curves recorded by the several students must be re¬
ferred to differences in the character of the water employed, as
their fundamental cause. Distilled water, as used in these *
studies, is very far from a pure or simple substance. Krishnan
(1925, p. 698) rightly emphasizes the extreme difficulty of ob¬
taining pure water and of keeping it when obtained. Hiifner
and Albrecht long ago stated the situation : Die Hauptschwierig-
Jceit liegt am Wasser . The experience of the present study indi¬
cates that a paraffin lining of the water tube may help to give
high absorptions in the region 4500-3650 A; like those there
recorded by Sawyer. In his samples of water the cause may have
lain elsewhere, since the water was in such tubes “only a few
minutes’’ ; but such a difference in absorption as appears at *
3650 A between James and Sawyer is attributable to differences
in the water, and probably in its content of impurities.
At the other end of the spectrum the effect of minute par¬
ticles and of paraffin, etc. is minimal, but there the situation is
complicated by the structure of water itself. Temperature has
a notable and variable effect on the molecular aggregations of the
water; salts in solution and organic ions also modify its “struc¬
ture temperature”, to use the term of Bernal and Fowler. Such
molecular aggregations as are expressed by the terms dihydrol
and trihydrol also affect absorption in a complex manner. But
we have still to learn how these or other conditions bring about
such different absorptions as appear at 7000 A between the re¬
sults of observers so recent, and so competent, as Collins and
Baldock, and Ganz.
For the present, the reasonable conclusion of the matter "
seems to be that each observer has recorded with fair accuracy
the behavior toward radiation of his sample of supposedly pure
water, and that the different results come fundamentally from
differences in the water. We will not discuss whether these dif¬
ferences were unnoticed traces of impurities ; whether they were
in the molecular constitution of the water; or whether there
were also other and less obvious causes. It is sufficient to bring
forward the situation and to emphasize it; since it makes evi-
38 Wisconsin Academy of Sciences, Arts, and Letters
dent that much is still to be learned about the relations of water
and radiation.
Agreement with Theory and Molecular Scattering
The data for absorption of radiation presented in this report
may be compared with values computed from theory of molecular
scattering, as given by Einstein and Smoluchowski (Krishnan,
1925). The results are stated below in Table IV.
The region of agreement of observation with theory is small,
being best at 4940 A and diverging rapidly on the long-wave
side of that region, with a more gradual departure on the side
toward the shorter wave-lengths. This indicates that there is
only a very narrow region in which the water absorbs by scat¬
tering the light from its molecules as discrete particles ; but that
for wave-lengths both longer and shorter than this the absorp-
Table IV
Comparison of absorption data used in this report with those computed
from the theory of molecular scattering.
tion by other effects, such as electron resonance, becomes impor¬
tant.
Temperature of Water as a Factor in
Absorption of Radiation
In the present investigation no observations were made on
temperature as a factor modifying the capacity of distilled water
to absorb or transmit radiation. Such a study is not important
in determining the ecological effects of color and suspensoids in
lake water. But a knowledge of the effect of temperature is
necessary when a quantitative analysis is undertaken, which
shall assign relative values to several groups of factors that
modify the passage of radiation through this water. This ne-
James & Birge — Lake Waters and Light
39
cessity is emphasized by difficulties and irregularities which are
found when such analysis is applied to the spectral region, 7000-
8000 A.
Most of the investigation of temperature as a factor affecting
the relation of water and spectral radiation has been done on the
long- wave spectrum, from 7000 A on, extending into the infra¬
red, and therefore beyond the spectral regions here considered.
The results of two such studies are here presented, so far as they
relate to the visible spectrum. One of these is by Collins (1925),
beginning at 7030 A and extending far into the infra-red; it is
continued by his pupil R. Baldock (1933, still unpublished), cov-
IOO
80
60
40
20
O
%
-20
5500 A 6000
7000
8000
Fig. 7. Effect of temperature of water on its capacity for absorbing
radiation. The curves represent percentile absorptions found by Collins,
7080-8000 A, and by Baldock, 5500-7000 A, at the temperatures indicated.
Note absorption band at 6000 A, slightly indicated at 25° and well marked
at 96°. A similar band is present at about 7400 A but is not well shown by
percentile plotting. See the papers of Collins and Ganz. Note that temper¬
ature effect becomes zero near 7800 A, then reverses, increase of tempera¬
ture diminishing absorption. At about 8100 A the effect becomes positive
again. Curve marked F is a “factor curve” discussed on p. 125. Crosses in¬
dicate factor curve for Collins’ observations at 0.5° C.
40 Wisconsin Academy of Sciences, Arts, and Letters
ering the spectrum from 5500 A to 7000 A. The other is by
Ernst Ganz (1934) which begins at 7000 A and goes into the
infra-red. All of these students have permitted the publication
of their original data, which are not given in their reports. This
Survey is deeply grateful to them for their courtesy and aid.
The reports of Ganz form the more recent and the more de¬
tailed study; they deal not only with the effect of temperature
but also with that of solutes. But they come into the visible
spectrum only as far as 7000 A. The numerical data found by
him are given in Table VII, and the percentile absorptions for
temperature 12° in Table II.
The work of Collins and of Baldock may be treated as a
single study of the spectral region 5500-8000 A and beyond. The
temperatures which they employed are so nearly the same that
their curves meet at 7000 A with sufficient exactness. Fig. 7
contains two percentile absorption curves, one showing absorp¬
tion by water at 25°-26° and the other at 90°-96°. The differ¬
ences in absorption are plainly marked, which the higher tem¬
perature effects. In each curve there is a local band of selective
absorption near 6000 A; clearly indicated for 25° and well
marked for 96°. No other similar band appears until the wave¬
length 7000 A has been passed.
The quantitative effect of this increase of 71° in temperature
may be computed by the methods used in determining the effects
of color (p. 114). The percentile transmission for 96° at any
wave-length is divided by the percentile transmission for 25° at
the same wave-length; the quotient is the percentile factor by
which transmission at 25° must be multiplied in order to reduce
it to the transmission observed for 96°. The results thus ob¬
tained are connected by a broken line (F, Fig. 7) which corre¬
sponds to the “factor curves” shown in the diagrams of Chapter
IV. In the region 6000-7000 A this line shows a percentile effect
of temperature which is roughly uniform; the increase from
25° to 96° reduces transmission to about 80 per cent of its first
value. For example, at 6500 A a meter of water at the tempera¬
ture of 25° transmits 69 per cent of incident radiation; at 96° it
transmits 55 per cent, and transmission at the higher tempera¬
ture is about 80 per cent of that at the lower.
The course of the factor curve between 6000 A and 7000 A
shows that the form of the absorption curve will not be modified
James & Birge — Lake Waters and Light
41
by ordinary variations in the temperature of the water. Per¬
centile absorption as measured by Baldock is roughly parallel to
that found in the present study and is greater by 1.2-4.3 per
cent, with a mean of 2.6 per cent. Some part of this excess is
due to the higher temperature of the water, as observed by
Baldock.
The situation is very different in the spectral region 7000-
8000 A. Percentile absorption rises rapidly, almost reaching 90
per cent at 7400 A and remaining near or above 90 per cent to
8000 A. Throughout the region the relations of water and radi¬
ation are subject to great and rapid variation. There is a strong
local band of selective absorption near 7400 A, so that the trans¬
mission at 90° may be hardly one-third of that at 26° ; at 7790 A
transmission at 26° and 90° is the same; near 8000 A is a se¬
lective band in which transmission increases with rise of tem¬
perature; and near 8200 A such a rise decreases transmission.
These variable and intricate relations are reflected by the
course of the “temperature factor curve” in Fig. 7, which goes
from a transmission of about 82 per cent at 7000 A to about 30
Table V
Percentage of radiation absorbed by one meter of distilled water ; at the
wave-lengths and temperatures indicated.
Note.-— The percentile absorptions stated in Table V are derived from
the coefficients of extinction reported by the several observers and contained
in Tables VI and VII.
42
Wisconsin Academy of Sciences , Arts , and Letters
Table VI
Effect of temperature of water on absorption of radiation.
Observations of Professor J. R. Collins
Observations of Mr. Russell Baldock.
James & Birge — Lake Waters and Light
43
Table VII
Effect of temperature of water on absorption of radiation.
Observations of Dr. Ernst Ganz.
Note. — Tables VI and VII are printed as a sort of supplement to the
brief account of the effect of water temperature on absorption of radiation.
They bring together numerical data, hitherto unprinted, on an important
aspect of the relations of water and radiation. The data have been furnished
by Professor J. R. Collins, Dr. Ernst Ganz, and Mr. Russell Baldock; this
Survey gratefully acknowledges their kindness.
The data from Professor Collins and Dr. Ganz are part of the obser¬
vations used in preparing their papers on the effect of temperature on ab¬
sorption of radiation by water. Their studies extended far into the infra¬
red, beyond the spectral region considered in this report. Mr. Baldock
investigated the spectral region 5500-7000 A in the Laboratory of Pro¬
fessor Collins at Cornell University, as part of his work for the degree of
master of science. The thesis has not yet been published.
Table VI gives coefficients of absorption just as they were received
from Professor Collins and M'r. Baldock, except that the meter is used as
the unit stratum of water. Dr. Ganz observed on an arbitrary scale of
wave-lengths, derived from the instrument which he employed ; he supplied
us with a table, showing the equivalents of this scale for each 100 A
from 6950 A to 9000 A. Table VII gives the readings for these points up
to 8200 A for temperatures 12° and 87°. Dr. Ganz also stated the coeffi¬
cients of absorption in terms of common logarithms; they have been con¬
verted into natural logarithms, the type used elsewhere in the present paper.
It has not been thought necessary to print the original scale or to show the
simple mathematical changes.
In Table V and in the diagrams these coefficients of extinction have
been converted into percentages of absorption; all of these changes have
been carefully made and it is hoped that they are free from errors. But this
Survey is responsible for any mistakes which may be found.
per cent near 7340 A; then passes to 100 per cent or more near
7800 A, returning to 70 per cent near 8200 A. There seem also
to be present in distilled water certain conditions, whose details
are still unknown, but which greatly affect its capacity for ab¬
sorption of radiation. There is no reason to suspect serious error
in the readings of either Ganz or Collins at 7000 A; yet one
44 Wisconsin Academy of Sciences , Arts , and Letters
finds there about 48 per cent absorption and the other 87 per
cent. Probably a study with distilled water from a new source
would yield still other results.
The complexities of the situation are further increased in
practice by the possible large effects of accidental errors. If
series are read in two waters the slightest variation in setting
the wave-length may bring about great errors in their apparent
relations. So also with the unavoidable errors in reading per¬
centages. Ganz places the probable error in absorptions at about
±1 per cent; but v/ith absorptions of 90 per cent or more a
reading which, by error, adds one per cent, changes the factor
curve by 10 per cent or more. This matter is further discussed
from the standpoint of lake waters in Chapter IV, p. 123.
References
Aschkinass, E. 1895. Absorptionsspectrum des fliissigen Wassers. Wied.
Ann. 55 :401.
von Aufsess, O. 1903. Die Farbe der Seen. Doktordissertation. Munich.
Baldock, C. R. 1933. The effect of temperature on the absorption of liquid
water near 6100 A absorption band. Unpublished.
Bernal, J. D., and R. H. Fowler. 1933. A theory of water and ionic solution,
with particular reference to hydrogen and hydroxyl ions. Jour. Chem.
Phys. 1:515.
Collins, J. R. 1925. Change in the infra-red absorption spectrum of water
with temperature. Phys. Rev. 26:771.
Ewan, T. 1894. On the absorption spectra of dilute solutions. Proc. Roy.
Soc. 57:117.
Ganz, E. 1936. Ueber die Absorptionsspektrum von Wasser, wassrigen Los-
ungen u. Alkoholen zwischen 0.70-0.95. Ann. d. Physik. 26: 331.
Hufner, G., und E. Albrecht. 1891. Ueber die Durchlassigkeit des Wassers
fur Licht von verschiedenen Wellenlangen. Wied. Ann. 48:10.
Juday, C. 1914. The Hydrography and Morphometry of the Lakes. Wis.
Geol. and Nat. Hist. Survey. Bull. XXVII.
Krishnan, K. S. 1925. On the molecular scattering of light in liquids. Phil.
Mag. 50: 697.
Pietenpol, W. B. 1918. Selective absorption in the visible spectrum of
Wisconsin lake waters. Trans. Wis. Acad. Sci. 19:562
Sawyer, W. R. 1931. The spectral absorption of light by pure water and
Bay of Fundy water. Canada Biol, and Fish. 1931:75.
CHAPTER III
ABSORPTION OF LIGHT BY LAKE WATERS
Section 1. General
Every natural water affects the solar radiation, which may
enter it, in a much more complex manner than that of distilled
water. This situation is due to the substances, very numerous
and very different in kind and quality, that may be dissolved or
suspended in it. A natural water, however completely isolated,
contains materials, organic and inorganic, dead and alive, which
affect light ; and their action is continually altered by processes
at work within the water.
This statement applies with double significance to small in¬
land lakes, the subjects of the present study, since their volume
of water is small and is much more affected by these processes
than is that of larger waters, like the ocean or even the Great
Lakes. Small lakes also have most intimate relations with their
environment; and they derive from environment, directly and
indirectly, large quantities of material in every possible condi¬
tion.
Many of the substances present in lake waters, need not be
considered in the present study; such are the gases and the
inorganic salts dissolved in the water. These, no doubt, have
their effect on the transmission of light, but this effect is not a
measurable one under the conditions of the present study. No
measurable differences, for instance, have been found between
ordinary “hard water” and “soft water” lakes, which can be
directly assigned to the mineral content of their waters. This
report, therefore does not consider these substances.
But there still remain for consideration innumerable — or, at
least, unnumbered— substances whose presence in lake waters
modifies their relations with light. They may be grouped, pro¬
visionally, under two heads : Suspensoids and Colors.
45
46 Wisconsin Academy of Sciences , Arts, and Letters
SUSPENSOIDS
Under the term suspensoids are included all sorts of small
particles and masses found suspended in lake waters. Some of
them are of larger size, like algae, Crustacea, bits of debris and
grains of silt. These obstruct light and may be so numerous
that their effect is very great ; but they may be easily removed
from the water by settling or by filtering through ordinary filter
paper. All waters examined for this report were allowed to
settle, in cold storage and in darkness, as is indicated for each
water in Table X ; and these larger bodies form no part of the
suspensoids as herein discussed.
After such settling, there remain in suspension a vastly
greater number of particles and masses which are not so easily
removed. These are the substances which give turbidity to lake
waters ; when few they appear as shining particles in a beam of
parallel light passed through the water, when numerous they
create a sort of fog in the water under like conditions.
These particles and masses exist in numerous forms, living
and dead, fresh and decomposing; they are present as organ¬
isms, as masses of colloid from dead organisms, as particles and
colloids extracted from marsh or peat. They are of all sizes,
ranging from those so large as to be easily removed by simple
means to those so small as to pass fine filters, and from these
to molecules distributed through the water. They may be trans¬
parent or opaque, colored or uncolored; they may have a high
or a low refractive index; and the amount and the nature of
their effect on the passage of light is correspondingly various.
Turbidity, in the waters here studied, is chiefly due to or¬
ganic particles; only small amounts of inorganic matter are
present as particles in most of these lakes ; their effect on light
is correspondingly small and is indistinguishable from that of
organic particles. A few lakes contain much marl in the form
of particles; Marl Lake, Fig. 43, as its name implies, has such
a water and the marl has a very considerable effect on the pass¬
age of light. This condition is exceptional ; in most waters the
suspensoids concerned with the absorption of light are mainly of
organic origin and are of various nature. One very important
James & Birge — Lake Waters and Light 47
class is that of the colloids. These constitute the living substance
of plants and animals and they are the chief form which remains
in the water as these organisms die and decompose. They exist
in the water as particles or as minute flocculent masses, too
shapeless to be called particles, and passing from these into col¬
loid solution. Most of these suspensoids are of such size as to
be removed from the water by a fine Berkefeld filter, but some
of them pass even the finest of the filters that have been used in
this study. The aim in filtering was to leave as much color as
possible in the water, while reducing suspensoid absorption to a
small amount. Success was reached in the second purpose, as
judged from the point of view of ecology rather than that of
physics. But the amount of color removed with the suspensoids
varied greatly with the condition of the color material in the
water.
It is certain that some degree of turbidity must remain in all
the waters examined as filtrates ; but so little suspensoid remains
after filtration that its effect on light may be nearly or quite
negligible. This appears from an examination of absorption
curves from filtrates of color zero or with a low number. When
the absorptive effect of water as water has been eliminated,
many of these curves show a minimal percentage of absorption
which is zero or perhaps 1-3 per cent in the middle spectrum.
Waters of higher color have a higher effect in this region; but
there is apparently no good reason for attributing this to sus¬
pensoids. The filter should remove these equally from water of
any color. This matter is more fully discussed on p. 62.
All colors found in lake waters are originally colloid in their
nature or are associated with colloids. This connection of color
and colloid is usually present in any sample of lake water, and
in that case the color is reduced by filtration. Colors so associ¬
ated with colloids are called colloid colors; in their effect on
light are combined the action of the color and the colloid with
which it is combined or to which it is attached. Compare Fig.
52.
The physical form and stability of colloids depends much on
the hydrogen ion concentration of the water in which they are
suspended. The pH of lakes varies and, with this, there are
48 Wisconsin Academy of Sciences, Arts, and Letters
changes in the colloids; under some conditions they tend to col¬
lect into larger masses and under other conditions they tend to
break up into finer particles. Such changes not only vary the
action of colloids on light but also vary the efficacy of the filter
in removing colloids from the water.
When water containing colloids is diluted with distilled water
the above influences may become important, for the normal con¬
ditions within the water are disturbed. Thus the effect of the
colloid as such on the absorption or scattering of light is a vari¬
able factor in any water, due to changes of the pH of the water.
One series of dilutions was made from filtered water, that of
Helmet Lake. In Fig. 9 are compared the absorption curves of
several dilutions and those of filtered lake waters of the same
color. The agreement of filtrate and dilution is so good as to
make it evident that no great or irregular effects have been pro¬
duced by the action of the distilled water on the color materials.
There is no way, at present, to determine this possible effect
upon unfiltered waters used in dilutions.
In all of this discussion it is assumed that changes in the form
of the colloids have no influence on the action of any colors that
may be associated with them. Reference to all of these points is
made later in explaining differences in absorption found in
waters of the same color rating.
Color
Standard of Color
The standard of color used in this report is that of the United
States Geological Survey. It is fully described in Leighton
(1905) and more briefly in Standard Methods of Water Analy¬
sis (6th Ed., 1925, p. 8). It is based on matching the color of
water in question with that of a water solution of platinum-
cobaltous chloride; and the result is stated as a number repre¬
senting the number of liter-milligrams of platinum found in the
solution whose color best matches that of the water. Thus the
scale is known as the U.S.G.S. or the platinum-cobalt standard;
colors are stated as cardinal numbers, such as 0, 3, 8, 16, 54, etc.
This method, which determines color by matching a repro¬
ducible standard and which states its results as a number, has
great merits. It renders its determinations in a quantitative
way which no other method approaches; and these determina-
James & Birge — Lake Waters and Light 49
tions can be used with fair accuracy in computations; such as
those employed in the analysis of color absorption of light, pp,
145-148. It has its limitations and its defects, and these will be
pointed out, since this Survey, and especially the present study,
have probably taxed its capacities quite as greatly as any earlier
use.
The standard relates to colors that range from yellow to red,
and that come from the decomposition of organic substances.
This Survey has not considered inorganic colors, such as that
due to ferrous carbonate in the deeper waters of lakes, which
have consumed all of their oxygen. The standard can not be
applied to green colors, nor can it give a rating to a green com¬
ponent of a mixed color. Thus waters that are examined for
color may appear colorless, or they may have a color which
ranges from yellow to red in all sorts of degrees and mixtures,
and always with a brownish component.
The color of any water is a mixture whose components are
the result of many and complex operations ; matching it against
a single standard obviously involves an averaging-over, whose
results can not be precisely accurate. The color of the water
may have more yellow than the best possible match of the stand¬
ard ; or it may have more red ; more often water and standard
differ in a way that is plain to the eye but can not be stated in
words. The rated color of the water comes from a mixture of
diverse materials; a little of a darker component may give the
same total effect to the eye as that from a larger amount of a
lighter component; but the detailed effect of the two types of
mixtures on the parts of the spectrum might be very different.
One other element in the situation must be mentioned; it
affects the judgment of the observer of the water in a way that
is inescapable, since it depends on the fundamental sensitivity
of the eye to color. In natural waters there is a larger or smaller
amount of dark color material, which strongly absorbs the short¬
wave spectrum. The observer of the mean color of the water
must learn to concentrate his attention on the color and neglect
its opacity ; otherwise his rating of the color will be too high by
an amount dependent on the quantity of suspensoids present in
the water. But the human eye has little or no sensitivity to color
in the spectral region of the darker red, (Fig. 8) and some of
the color in that region is sure to be regarded as opacity; it
50 Wisconsin Academy of Sciences , Arts , and Letters
Fig. 8. Solar energy curve with relative visibility curve. The main
curve is the solar energy curve which has been employed by this Survey as
its standard. Visibility is assumed as 100 per cent at about 5800 A, in the
yellow band. The curve indicates relative visibility for color at smaller and
greater wave-lengths, assuming equal illumination. Sensibility at junction
of blue and green is 23 per cent of the maximum in the yellow. At green-
yellow it is 91 per cent; at yellow-orange, 85 per cent; at orange-red it is
13 per cent. From Eder and Valenta.
would thus be rated as part of the non-selective material of the
water, in its action on light, while in fact it is highly selective.
This situation makes itself especially manifest in waters of little
or no color. In filtered waters rated as of color 0 or trace, as
reported in Table XI, absorption at 4078 A is very variable and
a large part of this irregularity is due to the presence of a little
of this darker color material, which has been rated as opacity.
This irregularity of absorption at 4078 A brings out the weakest
point of this method of rating colors; its dependence on the
human eye affects all results. There should be found a method,
independent of color sensation, by which these mixed colors can
be analyzed and quantitative values assigned to their components.
But in spite of all difficulties and defects, much knowledge
may be derived from the use of this standard ; and two diagrams
are given to illustrate general results of the kind. In Fig. 9 are
given curves for total absorption of the spectrum, 7000-4000 A,
as found in dilutions of the filtered water of Helmet Lake, and
in the filtered waters of four other lakes that had the same color
James & Birge — Lake Waters and Light
51
Fig. 9. Total absorption in dilutions from filtered water of Helmet Lake
(sample 2), compared with absorptions from filtrates from lake waters
having the same color. Continuous lines represent results in dilutions;
absorption in filtrates is shown by signs placed at the wave-lengths where
readings were made in all waters. Data from Tables XI and XII.
A. Helmet, 17.5%, color 45 vs. Tadpole, color 46.
B. Helmet, 10%, color 26 vs. Wisconsin River, color 26.
C. Helmet, 5%, color 16 vs. George, color 16.
D. Helmet, 1%, color 4 vs. Otter, color 4.
rating as the dilutions. These waters were wholly independent
of Helmet Lake, and the correspondence of absorption at the
several points of observation is quite as close as could be looked
for. It shows that the color rating was made objectively and
that similar color ratings come from similar conditions in the
waters.
In Fig. 10 is shown the result of comparing color absorption
curves from observed waters with others which have been com¬
puted for higher colors than those observed. The method of com¬
putation is state below, p. 52. In such computations factor
curves must be employed, the absorptive action of water itself
being eliminated. Such curves show the absorption effected by
all factors except water, as they would act in a medium which
has no effect on radiation.
The diagram offers five factor curves from filtrates, repre¬
sented by continuous lines; their color ratings range from 6 to
46. Along each curve, at the regular points of observation, are
placed signs which show absorption in a water of one-half the
color represented by the line; and for which absorption has
been computed for double the observed color.
Wisconsin Academy of Sciences, Arts , and Letters
52
4000 A 5000 6000 7000 8000
Fig. 10. Color absorption as found in certain filtered waters, compared
with that found in filtered waters of one-half the color rating, but computed
for twice their observed color. Factor curves must be used in such computa¬
tions, since the effect of water as water is much the same whatever the color.
There are five curves for pairs of filtrates; in three of them single waters
are compared; two compare mean curves. One pair of settled waters is
compared. Continuous lines represent observed results; signs show absorp¬
tions computed for raised colors. Data from Table XI; method of com¬
puting stated below. For color curves see p. 12,7.
A. Wisconsin River, settled, color 64 vs. Horsehead, color 33 raised
to 66.
R. Tadpole, filtered, color 46 vs. Rudolf, color 24 raised to 48.
C. Wisconsin River, filtered, color 26 vs. Horsehead, color 12 raised
to 24.
D. Mean 5 waters, filtered, color 16-20 vs. Mean 6 waters, color 8 raised
to 16.
E. Nagawicka, filtered, color 12 vs. Oconomowoc, color 6 raised to 12.
F. Mean 6 waters, filtered, color 6 vs. Mean 6 waters, color 3 raised to 6.
Computations are made from percentile transmission, equal
to 100 per cent minus the observed absorption. Color trans¬
mission for a color twice as great as that observed equals the
square of the observed transmission. Thus in the color curve
for the Wisconsin River (color 26), absorption at 5460 A was
35.1 per cent; transmission was 64,9 per cent. At the same point
Horsehead Lake (color 12) had 20.0 per cent absorption or 80.0
per cent transmission. By computation Horsehead Lake with
color 24 should have 64.0 per cent transmission; a result very
close to the observed transmission of Wisconsin River for color
26. All data indicated by signs in Fig. 10 have been thus com¬
puted. Compare Figs. 46 and 48 and also see Table XIV.
James & Birge — Lake Waters and Light
58
These computations can equally weil pass from higher colors
to lower. If in the above case, the square root had been taken
of the transmission observed for Wisconsin River, the result
would have agreed with that observed for Horsehead Lake at
color 12.
Such good agreements as those shown by the comparisons in
Figs. 9 and 10 are not to be found in every case ; they are pres¬
ent in only a minority of the waters. For their success it is
necessary that the waters agree in the general composition of
the color mixture; only so can they agree in their action on the
different parts of the spectrum. Figs. 12 and 13 show the range
of percentile absorption in waters of the same color rating ; and
this range is caused by different color mixtures, which agree in
their effect on the eye but differ in their action on the spectrum.
Rut the curves of these diagrams illustrate the ways in which
the platinum-cobalt standard may be used to advance the knowl¬
edge of the effect of color in lake waters. 1. Color mixtures in
different waters are often of the same type and have a similar
action on the spectrum; so nearly similar that these compari¬
sons and computations may be applied to them; and with no
more difference in results than might be observed in the same
water on different dates. 2. Suspensoid absorption of light in
these filtered waters is not great or such comparisons would not
be possible. 3. It is safe to raise the color of a filtered water to
that of the same water settled, in order to determine the effect
of suspensoids. If the resulting suspensoid curve comes out as
non-selective, it is fair evidence that the color-mixture of the
settled water was reduced but not greatly altered by filtration;
and it is also evidence that the suspensoid curve is approxi¬
mately correct. A single pair of settled waters is plotted in Fig.
10 as a sort of curiosity. Wisconsin River (color 64) is com¬
pared with Horsehead Lake, color 33, raised to 66. The agree¬
ment is very close; such a case in settled waters is very rare,
since the two waters must agree in the percentile effects of sus¬
pensoid absorption as well as in that of color ; and such similar¬
ity is not often found.
Amount and Sources of Colors
Most lake waters in their natural state exhibit some color.
Only two samples among the 55 which were examined as settled
54 Wisconsin Academy of Sciences , Arts, and Letters
waters, were rated as without color; and both of these came
from Crystal Lake. Colors have been determined in the field
for 546 lakes and lakelets of Northeastern Wisconsin. The re¬
sults for 527 of these waters have been reported in Juday and
Barge (1933: 216-235). Their color ranged from zero to 340 on
the platinum-cobalt scale ; and in small pools, whose waters have
been concentrated by evaporation, colors exceeding 1000 have
been found. The highest color examined in the Laboratory was
264, as reported in Table X.
The report on color by Juday and Birge shows that the mean
color of the entire group of lakes is 43 and that their mean
amount of organic carbon is about 7.1 mg/1; but this mean
comes from two types of lake which differ widely. The mean
color of seepage lakes is 21 and their carbon is about 4.7 mg/1 ;
the figures for drainage lakes are, color 61, carbon 8.8 mg/1.
Many lake waters appear colorless to the eye, and field de¬
terminations reported 56 waters of this type; only 7 of them
were from drainage lakes. Examination with the monochroma¬
tor showed that there are traces of color in almost all filtered
waters which are rated as colorless. Only one filtrate, one of
three samples from Crystal Lake, had an absorption in the short¬
wave spectrum which was close to that of distilled water. Table
XI shows the situation.
Since colors are derived from organic compounds there is
necessarily a quantitative relation between the color and the
organic carbon present in a water. This has not been studied in
filtered waters, but the correlation in natural waters is discussed
in the paper referred to, p. 232. Waters that were rated as zero
in color have carbon amounting to 1.2-4. 5 mg/1; and higher
colors regularly mean greater quantities of carbon; so that
colors of 200 or more are associated with carbons of 20-25
mg/1. In general it appears that there is a rough correlation,
such that the addition of 1 mg/1 of carbon means a rise of 10
color points; but in every mean there is a wide range both of
carbon and color. The quantitative relation as stated holds for
waters of lower colors, but as colors rise the increase effected by
a single mg/1 of carbon becomes smaller. The general situation
is shown in the diagram, Fig. 11.
Thus colors originate in the decomposition of organic sub¬
stances, and the process may take place either within the lake
James & Birge — Lake Waters and Light
55
0«i9 20-39 40-59 60-79 60-99 100-119 120-139 140 - 159 160-179 180-199
Fig. 11. Relation between the color of the water and its organic carbon.
The upper curve, shown by a broken line, represents the maximum amounts
of organic carbon found in the waters of the various color groups ; the lower
broken line shows the minimum amounts of carbon. The solid line curve
between them indicates the mean amount of organic carbon in the several
color groups. From Juday and Birge (1933).
or outside of it. In the first case the plankton is the main source
of organic matter. This consists mostly of particles so large
that they settle rather rapidly after death and soon reach the
bottom, where most of their decomposition takes place. A second
source of organic compounds is the higher plants, which grow
from the bottom of the lake in shallow water ; these also as they
die, sink to the bottom and there decompose slowly. Lake waters
whose sources of color and of organic carbon are internal, usual¬
ly have low colors and these have little red or dark material in
the color mixture. Seepage lakes with sandy margins, like Crys¬
tal and Diamond lakes, offer excellent examples of this type.
Similar conditions may be found in drainage lakes which have
no affluents, like Silver Lake, or in lakes of large volume, like
Green Lake.
Other colors originate outside of the lakes in decomposing
masses of vegetation, like marsh or peat bogs. They are brought
into the lakes by drainage waters, rain or spring, and may be
56 Wisconsin Academy of Sciences, Arts, and Letters
called extractive colors . These may come into the lakes associ¬
ated with colloids or they may be in solution. They are of in¬
numerable types and shades of yellow, orange and red, and are
present in the waters as complex mixtures of such substances.
They almost always include a certain amount of very dark ma¬
terial, both as dissolved color and as very minute particles ; this
gives a brownish tinge to all these colors, even to the lighter
shades of straw-yellow. In general, all of the darker colors of
lake waters are due to these extractive colors, brought into the
lake either from adjacent bogs or from more distant sources by
means of the waters of affluents.
Action of Lake Waters on Light
The action of lake waters on radiation, as shown in the per¬
centile absorption curves of the following diagrams, is the com¬
bined result of many factors, which for the purposes of this
report are grouped under the three heads, water, color and sus-
pensoids. Absorption curves for settled waters, such as Fig. 14
and many others, exhibit the combined effect of all three groups ;
the curve for distilled water, W in Fig. 14 and many others,
shows the effect of water alone ; diagrams for filtrates, like Fig.
15, give primarily the combined effect of color and water ; factor
curves for filtrates, like Fig. 53 and those marked C in many
diagrams, isolate the effect of color so far as this can be isolated
by present methods ; factor curves for suspensoids are given in
Figs. 37-49.
Absorption of radiation at the red end of the spectrum is
mainly the work of water as water. This absorbs an average of
90 per cent or more of radiation at wave-lengths 7400-8000, so
much that other factors add but little to its effect. The mini¬
mum effect of water is one per cent or less, even in water tubes
lined with paraffin ; it is found near 5000 A and rises but little
in the shorter wave spectrum. This action of water is considered
as a constant in all lake waters.
Absorption of radiation at the blue end of the spectrum is
primarily due to color, and therefore varies endlessly with the
degree and the kind of color found in the lake waters. A lake
water may have so little color that its absorptive action at 4078
A may be nearly or quite indistinguishable from that of pure
water. One such water was found, (Fig. 20), the third sample
James & Birge—Lake Waters and Light
57
from Crystal Lake : in the other two samples absorption at this
wave-length was increased by color; which added 3.1 and 24.8
per cent. The settled water of the clearest sample absorbed 28.1
per cent in this region, and was rated as of color zero ; filtration
removed substantially all of the absorbing material present in
the water.
In this region increase of color is associated with rapid in¬
crease of absorption, as is shown in several diagrams. Fig. 54
shows that absorption by color alone at 4000 A reaches 100 per
cent near color 30; mean at color 20 is well above 90 per cent;
and at 3650 A the same percentile absorption may be reached at
color 10 or even lower, Fig. 28.
Absorption by colors decreases very rapidly as its effect is
traced to longer wave-lengths; for low colors, 0-4, there is
usually a minimum where color seems to produce no effect. This
is rarely reached at smaller wave-lengths than 6000 A. Fig. 25
shows that in filtered waters color adds very little to the absorp¬
tive effect of water at wave-lengths of 7500 A and more, even
when the color of the water may exceed 100. But factor curves
in the spectral region 7000-8000 A require discussion, p. 121.
From all this it follows that the percentile absorption curve
of a filtered water of moderate color, as shown in Fig. 15 and
others, has the general form of the letter U, the maximum ab¬
sorption of radiation being at the ends of the spectrum and the
minimum lying somewhere in the middle. This shape is well
marked in Fig. 15 for colors 8-10 and in Fig. 25 by colors 16 and
26. Fig. 28, in which curves are extended to 3650 A, displays
this condition in its most developed form. In lake waters of low
color the color-limb of the U is little developed ; for colors above
20 the color-limb approches 100 per cent ; and at colors of 30-40
the absorption curve loses its upward-facing concavity and be¬
comes nearly a straight line. At still greater colors absorption
becomes 100 per cent at greater wave-lengths than 4078 A;
reaching 5500 A at color 180 ; and in such cases the form of the
absorption curve changes, having a concavity facing downward.
In all such absorption curves for filtrates the minimum is in the
middle spectrum ; in the waters examined it has not been found
at wave-lengths less than 5000 A or greater than 7000 A.
This discussion of the form of absorption curves is of course
based on the percentile method of plotting, which is found the
58 Wisconsin Academy of Sciences , Arts , and Letters
most suitable for the purposes of this report. If coefficients of
absorption were plotted there would be no 100 per cent limit and
the form of curves would not alter at high colors.
Since absorption of radiation by pure water may be regarded
as a constant, it is not difficult to divide the total action on light
of a lake water into two components ; one of these being water
and the other all other factors combined. It is much more diffi¬
cult to divide this second component and to assign their parts to
colors and to suspensoids; but something can be done which
yields results accurate enough for ecological studies.
Berkefeld filters may be found whose pores are such that
they keep back most of the suspensoids and yet permit the pass¬
age of dissolved color material. Such filters may leave so little
suspensoid material that its absorptive action on light is negli¬
gible; this is illustrated by the numerous low colored filtrates,
whose minimum absorption is zero or only one or two per cent.
The color of the filtrate from certain lake waters (e.g. Naga-
wicka Lake, Fig. 37), is the same as that of the settled water.
In such cases it appears that the color material is wholly in solu¬
tion and that only suspensoids are removed by the filter. The
absorption curve of such a filtrate may be regarded as due to the
action of color and water; and from this curve, compared with
that of the settled water, the effect of suspensoids may be com¬
puted, as is done in Chapter IV.
In most cases part of the color material is removed by the
filter, along with the colloids with which it is associated; part
of it is in solution and remains in the filtrate. In such cases the
action of suspensoids in the unfiltered water may often be com¬
puted, if the dissolved color material has the same general com¬
position as that remaining in the suspensoids. If, however, the
color that remains in the filtrate differs from that in suspensoids
the computation can not be made. This subject is fully discussed
and illustrated in Chapter IV.
The action of suspensoids on light shows no such strong se¬
lective effects as does that of water and of color. Suspensoid
curves often have almost or quite uniform ordinates, as in Fig.
39; this holds for the spectral region 4000-7000 A; the long¬
wave region 7000-8000 A requires special discussion. More com¬
monly such curves show a greater absorption in the short-wave
region. This may be due to the greater dispersive effect of parti-
James & Birge — Lake Waters and Light
59
cles on short-wave radiation, but some of it is probably the re¬
sult of color remaining in or with suspensoids.
If then suspensoids reduce the transmission of radiation
about equally in all parts of the spectrum, their effect will be
relatively small in determining the form of the absorption
curve; they add an equal amount to absorption at all wave¬
lengths. Their percentile effect on the form of the total absorp¬
tion curve is greatest in the middle of the spectrum, where that
of both color and of water is least, as is shown in numerous
diagrams. So many variable factors are included in the groups
of colors and suspensoids that their action is best studied in the
several diagrams.
Suspensoids agree in their type of action on radiation, but
they differ widely in their quantitative effect. In the settled
water of Muskellunge Lake (Fig. 41) suspensoid absorption
hardly reaches 10 per cent; in Marl Lake (Fig. 43) its mean is
more than 40 per cent ; in Crystal Lake, whose settled water had
less color than any other, (Fig. 39) suspensoid absorption is
between 15 and 20 per cent. The several diagrams give a fair
notion of the range in the amount of suspensoid absorption.
Water, color, suspensoid — each has its own type of absorp¬
tion curve; and the type is consistently adhered to, at least in
the spectral region 4000-7000 A and also to wave-lengths as
short as 3650 A. Throughout this region there were not ob¬
served any strongly marked local bands of selective absorption.
Rise of temperature in the water produces such a band near
6000 A (p. 39), but any such effect is rarely perceptible in lake
waters, (Fig. 34), or, if present, it is so small that it is hardly
distinguishable from an accidental irregularity. Factor curves
in the spectral region 7000-8000 A have great irregularities in
absorption, whose causes are still only imperfectly known.
Section 2
Observations on Lake Waters
This section presents an account of observations on the
transmission and absorption of light in the waters of Wisconsin
lakes. It states these facts as found in settled and filtered waters
and in dilutions of high colored waters. In all cases the unit of
water is a stratum one meter in thickness and the report deals
Table VIII
List of lakes
Lakes of Northeastern Wisconsin
Lakes of Southern Wisconsin
* Sample taken at Sauk City.
Note. — In Table VIII are given some of the more important data regard¬
ing the lakes whose waters were examined in the course of this study. No
explanation is needed for most of the columns of data. Under the heading
Type, D means a lake with outlet, called in the Report a Drainage Lake ;
S means a lake without affluent or effluent, called in the Report a Seepage-
Lake. The exchange of water in such a lake is mainly underground ; such
lakes usually have exceptionally “soft” water. In the column Color the
number states the color-rating on the U.S.G.S. or platinum-cobalt scale.
These numbers often appear on the curves of the diagrams. The numbers
under Group place the lake in one of the three Groups, as defined in
Chapter III. By Symbol is meant the letter or letters by which the lake
is indicated on the percentile curves of the diagrams.
James & Birge — Lake Waters and Light
61
with the percentage of solar energy removed or transmitted at
certain points in the visible spectrum. These percentages are
stated in the tables; they constitute the ordinates for the ab¬
sorption curves shown in the diagrams, and discussed in the
text.
Observations were made on the waters of 49 Wisconsin lakes
and the Wisconsin River. These were so chosen that their waters
cover a wide range of color, since one main purpose of the study
was to ascertain the effect of color on light absorption. Colors
ranged from zero to 264, as rated by the platinum-cobalt scale.
Complete tests were made on 55 waters, since more than one
water was examined from some lakes. Waters were examined
in the condition of settled in cold storage, as filtrates, and as
dilutions; 181 series of tests were made, each involving deter¬
minations of absorption at 21 or 22 wave-lengths of the spec¬
trum. Some 4000 determinations of absorption were made, re¬
quiring about 18,000 separate scale readings.
Table VIII, gives a list of these lakes with some of the
pertinent facts regarding each; Tables X, XI, and XII, pp.
100-109, give in detail the data for the absorption of light as
found in these waters, at the several wave-lengths where read¬
ings were made. The tables are accompanied with a similar
series for distilled water. This differs somewhat, especially in
the short-wave spectrum, from the data given in Table I, p. 30.
The results in the large tables are from distilled waters observed
in tubes lined with paraffin, like those used for the lake waters.
They show higher absorptions in the short-wave region of the
spectrum, than do those found for distilled water in silver plated
tubes.
Treatment of Lake Waters
1 . Storage and Settling
Table X contains percentile absorptions for 54 settled wa¬
ters ; the color of the water when examined and the time that it
had remained in storage are stated for each water.
Samples of water received at the Laboratory were stored in a
dark room which was kept at a regulated temperature of about
5°. They remained in storage for different lengths of time, as
specified in Table X. During this time the water was not dis¬
turbed, either by moving the container or by convection cur-
62 Wisconsin Academy of Sciences, Arts , and Letters
rents; so that larger particles of plankton, etc., settled to the
bottom. When the water was to be examined it was removed by
a siphon so adjusted as not to disturb the bottom deposit.
Settling had a different effect on the various waters. If the
color material was largely in the colloid form, the color of the
waters was lowered by settling; if color was in solution it was
changed little or not at all. Cases which illustrate both situa¬
tions are later discussed, p. 79.
Waters were kept in jugs of colored glass during the process
of settling. A little mercuric chloride solution was added to each
sample of lake water at the time of collection ; so that there was
no bacterial action on organic matters in the waters. Color may
have passed from colloids into solution during long processes of
settling or it may have altered by processes other than bacterial.
No study was made of such processes if they are present.
2. Filtration
Table XI contains the data for percentile absorptions of 65
waters, which have been filtered by Berkefeld filters. The pur¬
pose of filtration was to remove as much suspensoid matter as
possible and to pass as much color as possible. The first end
was attained to a reasonable degree, as is shown by the small
minimum absorption in filtrates of low color. But in most waters
much of the color is connected with suspensoids and colloids and
is removed with them.
The filters were Berkefeld, of such fine pores as to remove
about 75 per cent of the bacteria; that is, they were of the grade
now known as Berkefeld V. The effect of such filters upon the
color of the several waters is discussed in detail in connection
with the several lakes. In general high colors were greatly re¬
duced by the filters; as for instance the color of Rudolf Lake
was reduced from 50 to 24, that of Lake Mary from 109 to 34.
Fortunately, one very high colored water was found, that of
Helmet Lake, which was altered very little or not at all by filter¬
ing. It retained its color of 200 or more, while no other filtrate
had a greater color than 46.
The filter always reduces absorption in the central part of
the spectrum, (about 5800 A), where the percentile effect of
suspensoids is greatest. Its effect in the violet region (about
4000 A) is variable, depending on the condition of the color ma-
James & Birge — Lake Waters and Light
63
Table IX
Effect of Filtration on Color and on Absorption
A. Lakes of Group I
For Definition of Groups see p. 71
B. Lakes of Group II
C. Lakes of Group III
64 Wisconsin Academy of Sciences , Arts , and Letters
terial. If this is dissolved, filtration has little effect on absorp¬
tion but if color is connected with particles or colloids filtrations
may have a large effect. The various types of cases are illus¬
trated in the following diagrams.
3. Dilutions
The waters of 6 lakes (7 samples) and that of one small Bog
Pool were examined as dilutions in distilled water; the results
are listed in Table XII. The lakes with their original color are
as follows: Helmet, No. 1 (264), Helmet No. 2 (236), Little
Long (74), Mary (109), Tadpole (74), Tamarac (57), Turtle
(43), Bog Pool (1064). Table XII gives the percentile absorp¬
tions for 52 dilutions made from the waters of these lakes. Their
action on light in their 100 per cent condition is stated in the
tables for settled and filtered waters.
Dilutions were made by adding a certain percentage of the
lake water to distilled water ; the percentages, the number from
each water, the resulting color and the effect on the transmission
of light are all shown in Table XII. Altogether, 52 dilutions
were examined; those from Helmet Lake were made from fil¬
tered water, the others were not filtered, since filtration removed
much of the color.
As will be seen in later sections of this report, the examina¬
tion of these dilutions furnished information regarding the effect
of color and of colloids on the passage of light. They were also
used as a basis for the more detailed study of the color materials
present in the lakes, by applications of Beer’s Law, since the
absorption can be studied as a function of concentration in these
waters.
Jf. Control of Bacteria
The waters of the lakes are rich in organic matter and there¬
fore are excellent media for bacterial growths. Such growths
not only effect changes in the organic materials which may alter
colors, but they also bring about turbidity of the waters, which
affects the passage of light. The action of bacteria was effec¬
tively controlled by adding to the waters at the time of collection
a small amount of concentrated solution of mercuric chloride.
The amount of solution thus added was about 10 drops for
each liter of water; experiment showed that this amount did
James & Birge—Lake Waters and Light
65
Fig. 12. Percentile spread of single determinations of absorption in
settled and filtered waters of color 6; 8 settled waters and 6 filtrates are
used. Mean curves are factor curves (p. 115), eliminating effect of water.
Mean absorption in settled waters is 68.7 per cent at 4078 A and 28.1 per
cent at 5970 A; for filtrates the figures are 48.7-3.4 per cent.
The percentile spread of single observations is shown by vertical lines
at each point of observation. It is far greater in settled waters and its
reduction with larger wave-length is much less than in filtrates. Both facts
are correlated with the large effect on radiation due to suspensoids and on
its non-selective nature. In the filtered waters the spread is greatest at
4078 A and decreases irregularly with increasing wave-length. Color 6
presents the maximum difference between settled and filtered waters of the
same color.
not alter the percentile absorption of light by distilled water.
No experiments were made to determine its action, if any, in
modifying the settling of suspensoids while the water was in
cold storage.
The Spread of Absorption
In each of the Tables X, XI, and XII the waters are arranged
by color, in order to bring out the agreements and differences in
absorption among waters rated as of the same or similar color.
The action of water may be taken as a constant, as stated in the
first column of the tables. Several facts appear immediately on
inspection: 1. Variations in absorption in waters of the same
color is greatest in the short-wave spectrum, especially plain at
4078 A. 2. In the spectral region 7000-8000 A other agencies
add little to the absorptive effect of water itself. 3. Settled
waters are more variable than filtered waters of the same color.
66 Wisconsin Academy of Sciences , Arts , and Letters
Fig. 13. Mean factor curves for filtered waters of colors rated as 3, 4,
8, 16-20, with lines at each point of observation indicating the spread of
percentile absorption in the single waters from which the mean is derived.
The number of waters in the several groups is 6, 7, 5, 5, Note that, in
general, the spread is greater for lower colors and for shorter wave-lengths.
4, These differences are not due to irregularities within the
series for each water ; on the other hand, a series which is rela¬
tively high or low at any point in the spectrum is likely to keep
that character through a considerable spectral region.
Two diagrams are given to bring out more clearly the facts
regarding this spread of absorptions in waters of the same color
rating. The first diagram, Fig. 12, compares settled and filtered
waters of color 6. Factor curves are employed, thus eliminating
the effect of water itself. The nature of factor curves and the
method of computing their ordinates are stated on pp. 115-117.
The other diagram, Fig. 13, presents the situation in filtered
waters of several colors. The curves in both diagrams are given
from 4078 A to 6000 A only, since absorption and variation are
both low in filtered waters of these colors at wave-lengths great¬
er than 6000 A.
In Fig. 12 the means are derived from 8 settled waters and
from 6 filtrates. The factor curve for settled waters has the
flatter form characteristic for such waters and due to the effect
of suspensoids ; that for filtrates also has the characteristic form
of a color curve. The spread of single observations is much
greater in the settled waters ; its range at 4078 A is from 89.0
per cent for Oconomowoc Lake to 43.0 per cent for Mendota,
with a mean of about 68.0 per cent. At 5970 A the same lakes
James & Birge — Lake Waters and Light
67
were at the extremes, and the same may be said for the inter¬
mediate readings. The mean for filtered waters was about 43.0
per cent at 4078 A; Oconomowoc Lake had the maximum ab¬
sorption there, about 51.8 per cent; Muskellunge Lake was the
other extreme with 31.3 per cent. In general also the spread
was between these two lakes, the other waters holding inter¬
mediate positions, as may be seen from the tables.
The spread in settled waters remains much the same at all
wave-lengths; this character of the curves as well as others is
due to the presence of large and variable kinds and amounts of
suspensoids in the water ; their effect is little modified by change
of wave-length. It should be noted that color 6 shows the maxi¬
mum spread found for any color and also the maximum differ¬
ence between settled and filtered waters.
Four curves are given in Fig. 13 for filtrates with color 3, 4,
8, 16-20; the numbers of waters in each group are 6, 7, 5, 5.
In general, the spread of individual readings is larger for lower
colors and for shorter wave-lengths ; but this condition is by no
means invariable. Note that for color 3 the mean absorption at
4078 A is slightly greater than for color 4, also the maximum
absorption for color 3 is lower and the minimum is higher than
for color 4. Such irregularities are to be expected, but as the
two mean curves are followed toward 6000 A, that for color 3 is
consistently lower than that for color 4.
Section 3.
Mean Results of Absorption
Figs. 14, 15, 16 present the mean results of the absorption
of light by lake waters of different colors and in different condi¬
tions. The curves are derived from the percentile absorptions
contained in Tables X, XI, and XII, and the number of waters
included in each mean can be seen from these tables. Many of
the conditions represented in these means are discussed in more
detail on later pages, but brief general statements may be made
here. The percentile absorption curves have a fundamentally
similar form in all three diagrams. Absorption for low and
moderate colors is least in the middle spectrum, and the position
of the minimum moves toward 7000 A in the highest colors. All
curves show an increasing divergence as they are followed to¬
ward the short-wave spectrum. This is due to the dominant in-
68 Wisconsin Academy of Sciences, Arts, and Letters
fluence of color as an absorbing agent in that region. In the
long-wave region the curves converge; in the region 7000-8000
A absorption rises rapidly and curves come very close together
about 7400 A; from this point to 8000 A they are nearly para¬
llel and lie close to or above 90 per cent. This situation is due
to the dominant influence of water as the absorbing agent in this
region. The curve marked W shows the effects of water alone.
Settled Waters, Fig . lh, Table X
The absorption curves for settled waters exhibit the effect on
radiation of lake waters which are nearly in their natural condi¬
tion, altered only by the settling out of the larger suspensoids.
The curves for filtrates show the combined action of water and
color, suspensoids having been removed as far as possible by
filtration. Dilutions were mostly made from settled waters ; their
curves present the effects of water, color, and of such suspensoids
as were in the sample of water used in the dilution. The amount
of suspensoids per liter is reduced in the same proportion as the
color materials.
Fig. 14. Settled Waters. Curves for mean percentile absorption, show¬
ing effect of increasing color. Data from Table X; color rating marked on
curves. Color 28 is mean of colors 24-33; 60, of 55-65; 105 from Mary 101,
109; 264, Helmet.
Note that curves are relatively flat, 4078-6000 A; minimum absorption
not below 20 per cent; all absorptions notably above that of water, 4078
A-7000 A. All are characteristics of settled waters.
James & Birge—Lake Waters and Light
69
In the settled waters suspensoids exert their full effect ; the
result is to raise the percentile absorption for every color. Since
suspensoids are non-selective in their action, this effect is rough¬
ly uniform throughout the spectrum, but it is especially mani¬
fest in the middle spectrum, where color and water have their
combined minimum. The result is that curves for settled waters
are flattened in the middle spectrum, as compared with those for
filtrates. The U-form of the curves is far less marked.
This condition is best shown in curves for the lowest colors.
Minimum absorption for color zero is close to 20 per cent at
about 5500 A, far above its position in filtered waters. The same
relation is apparent in other colors, as may be well seen from
the position of the curves for colors up to color 13 in the two dia¬
grams. Absorption in the middle spectrum is greater in settled
waters and the course of the several curves in the middle spec¬
trum is more uniform. The curve for settled water, color 28,
should be compared with that for filtrate, color 45.
The minimum absorption found at 4078 A in a settled water
was 28.1 per cent; this was in a settled water from Crystal
Lake ; its color was rated as zero both as settled and as filtered.
The filtrate gave 3.0 per cent absorption at 4078 A, of which
water was about 2 per cent. This water is one of the best exam¬
ples of a large contribution by suspensoids; as color rises its
relative effect becomes greater and the percentile contribution
of suspensoids is necessarily smaller.
Filtered Waters, Fig. 15, Table XI
The absorption curves of Fig. 15 present the effects on light
of color and water, with only a minimum amount of suspensoids.
The more striking changes made by this removal of suspensoids
are : 1. The lowering of absorption in the middle spectrum, giv¬
ing the curves a more decided U-like form. 2. The small ab¬
sorption by color in waters whose color rating is zero, as shown
by the close correspondence of the zero curve with that for dis¬
tilled water from about 5800 A on. 3. In waters up to color 12,
there is little absorption due to color at wave-lengths greater
than 6000 A. In the diagram the curves for intermediate colors
can not be carried beyond this point. 4. There is only a small
amount of spread at 7000 A between curves for color zero and
for color 46. 5. Curves for all colors agree closely in the region
Wisconsin Academy of Sciences, Arts, and Letters
70
Fig. 15. Filtered Waters. Curves for mean percentile absorption,
showing effect of increasing color. Data from Table XI; color rating marked
on curves. Note U-shaped course of curves, especially from colors 6-30;
curve for colors 28-34 nearly straight line 4078-5800 A; that for color 46
begins to show concave downward near 4500 A. All are characteristics of
filtrates. Note small addition to effect of water by low colors from 6000 A;
shows that waters had very little suspensoids left after filtration. Curves
for higher colors may be found in Fig. 25; derived from dilutions of filtered
water of Helmet Lake.
7400-8000 A; much more closely than do the corresponding
curves in settled waters.
The curve for color 46 represents a single water; that for
Tadpole Lake, whose color as settled was 130 and is shown in
Fig. 14. This water yielded the highest color found in a filtrate,
except that from Helmet Lake, nearly all of whose color passed
the filter. The series of dilutions from the water of this lake is
presented in Fig. 25 and shows the absorption curves for fil¬
trates up to color 180. The filtrate of Tadpole Lake, color 46,
gave an absorption curve very similar to that of the 17.5 per
cent dilution of the water of Helmet Lake, color 45.
The curves of Fig. 15 should be compared with those of Fig.
53, in which is presented the mean effect of color alone. This
Fig. 15 shows how color dominates the short-wave part of the
curves for filtrates ; while the action of water controls that part
from 5500-6000 A to 8000 A.
James & Birge — Lake Waters and Light
71
Fig. 16. Dilutions. Curves for mean percentile absorption, showing
effect of increasing color. Data from Table XII; color rating marked on
curves. General form and position of curves are between those of curves
in Figs. 14 and 15. Most dilutions were made from settled waters, but
suspensoids were reduced in same proportion as color.
Dilutions , Fig. 16, Table XII
Since only one lake yielded a filtrate with a high-colored
water, the curves of mean absorption by dilutions include the
effect of that amount of suspensoid which was carried in the
water used in making the dilution. This would be less than that
of the settled water, but greater than that found in the filtrates.
For dilutions such as 0.1 per cent the quantity of suspensoid
must be very small, but it produced very plain effects. The curve
for dilutions of color zero shows that absorption at 4078 A is
only a little above that for filtrates, but its general form in the
short-wave spectrum resembles that of a settled water. It is
nearly flat, but its percentile absorption is only about one-half
that of the settled waters of zero color.
This general situation is apparent throughout the diagram.
Curves for dilutions lie between those of settled waters and
filtered waters of similar color. They approach the settled waters
as the percentage of lake water in the dilution becomes greater,
and finally are practically indistinguishable from those for set¬
tled waters.
Three Groups of Lakes
Lake waters may be classified on the basis of the relations
between their color material and filtration. Dissolved colors
72 Wisconsin Academy of Sciences, Arts, and Letters
pass the Berkefeld filter much more readily than do those associ¬
ated with colloid or organic particles. Since all colors originate
in colloid mixtures, those that are manufactured in the lake are
likely to be in such condition that a large part is removed by the
filter. Extractive colors, coming from bog or peat, may be in
dissolved form or may be still connected with colloids or other
particles; the nature of their origin and transport to the lake
make it probable that a large part of the material will be in dis¬
solved form. Waters whose color is mainly extractive are likely
to have a relatively large amount of dark color material, since
decomposition has gone further in their sources than it can go
in plankton on its way to the bottom of the lake.
Thus the behavior of waters toward filtration gives a basis
for grouping them ; and since color acts most effectively on radi¬
ation in the violet region of the spectrum, the results of filtra¬
tion are most apparent there.
A very great variety of conditions is brought about by the
filter, but these may be employed to place the lakes in three
groups, whose distinguishing characters are as follows :
Group I. In these waters much or most of the color material
is extractive and is present in the water as dissolved color. In
such waters of low or medium color rating, filtration may make
little or no change in color. In other high colored waters filtra¬
tion may reduce the color but there still remains a large amount
of color material, causing high percentile absorption in the vio¬
let. This situation is emphasized by the presence of much dark
color in the extractive material.
Group II. In these waters color is mainly present in associa¬
tion with colloids or other similar substances, and is therefore
removed in great part by the filter. Still further, in such waters
most or all of the color is produced in the lake itself and there¬
fore the colors are more of the yellow type than of the darker
reds. Filtration causes a considerable reduction both of color
and of absorption of violet and blue radiation.
Group III. In these waters the color material is in much the
same condition as in those of Group II; but it is more com¬
pletely attached to colloids and removed by filtration ; and there
is almost no extractive color in the waters. The absorption re¬
maining after filtration is lower than in waters of the other
types and of the same color ; whether settled or filtered.
James & Birge—Lake Waters and Light
73
The primary differences between these three groups evi¬
dently rests on the quantitative distribution of color materials
between dissolved and colloid. This again comes to depend mainly
on the source of the color, whether internal to the lake or outside
of it. The three types of waters are quite distinct in many lakes
and they provide a convenient method of characterizing the be¬
havior of the waters here studied. Four waters of the 55 exam¬
ined for this study do not fit into the Groups. Certain lakes vary
considerably in the amount and kind of color material present
in their waters and one or two lakes were found whose water
came in Group I in one season and in Group II in another year.
Thus it is not at all impossible that if waters were examined
from the hundreds of lakes in the District the relation of their
color to filtration would place them in diverging series rather
than in separate groups. But for the purposes of the present
report the classification into Groups is that which best facilitates
the handling of the observations. See Table IX, p. 63.
The discussions of single lakes will furnish many illustra¬
tions of these Groups.
Variation of Absorption in Individual Lakes
The former section and Figs. 12 and 13 make it evident that
there is considerable variation in absorption among waters hav¬
ing the same color rating ; and that this variation is greatest in
waters which still contain their suspensoids. The question na¬
turally arises as to the amount of variation in absorption which
may be found in the same lake on different dates. The present
study does not yield much material available for this discussion,
but something may be found.
The three samples from Crystal Lake were all rated as of
color zero; one was taken on May 15, the other two on July 14
and 28. The first sample absorbed 26.9 per cent of the radiation
present at 4078 A; the others gave 5.2 and 3.0 per cent at that
wave-length. In all three absorption practically coincided with
that of distilled water before wave-length 5970 A is reached.
The earlier sample evidently exhibited some effect of the spring
rains and of the early plankton growth; the July specimens
gave normal mid-summer conditions. Variations of this sort are
usually found in all lakes, especially in those of low color.
74 Wisconsin Academy of Sciences, Arts, and Letters
A more interesting comparison may be made between the
results of the present study and those reached some 15 years
earlier by Pietenpol (1918). lie examined spectral absorption
in the filtered waters of 31 lakes, 17 of which are in the list for
the present study. The absorption curves of 7 of these samples
from common lakes are shown in Fig. 17 ; in none of the other
10 lakes were differences between the two sets of curves greater
than those in the diagram.
The curves from Pietenpol cover a smaller spectral range
than do those of the present report ; this limitation is due to the
instrument employed by him. The differences between the curves
are usually more noticeable near the ends of PietenpoFs curves ;
this again is probably associated with the fact that these deter¬
minations were near his instrumental limits. In most cases Piet¬
enpoFs curves show the higher percentile absorption ; he used a
coarse Berkefeld filter which passed larger quantities of sus-
pensoids than the finer filters used in the present study.
But Fig. 17 brings out the significant fact that the general
form of the absorption curves is the same in the waters taken
on dates 15 years apart. This shows that the character of the
color material in lake waters remains much the same over long
periods of time. The complex mixtures of colors do not change
much from year to year. Thus the correspondence between the
two sets of curves not only offers a check on the general accu¬
racy of the work in both investigations ; it also serves to empha¬
size the general stability of conditions in the individual lakes.
In 6 of the 7 cases shown in Fig. 17 percentile absorption is
greater as reported by Pietenpol; in Pine Lake, however, the
earlier study has the lower absorption. This no doubt represents
change in the color conditions of the lake.
Both of these facts show that transmission of the spectral
regions is in harmony with data of “characteristic transmission”
of total light, reported by Birge and Juday (1932: 560). The
table covers observations on 25 lakes extending over a maximum
period of 13 years. In most cases transmission is surprisingly
uniform; in a few cases there is considerable variation. The
range of tranmission of a meter of water in Trout Lake was
from 62 to 71 per cent, as determined in 7 observations, each in
a different year; in a similar way 4 sets of transmissions in
Crystal Lake ranged from 80 to 84 per cent. On the other hand,
James & Birge — Lake Waters and Light
75
Fig. 17. Comparison of absorption curves as found in the present re¬
port and by Pietenpol (1918) 15 years earlier. His curves are shown by
broken lines. Length of his curves depends on limitations of instrument
employed. General good agreement in form of curves indicates the constant
nature of color materials in the lakes. Higher absorptions found by Pieten¬
pol were due to his use of a coarse Berkefeld filter which passed much sus-
pensoid.
76 Wisconsin Academy of Sciences, Arts, and Letters
transmission in White Sand Lake was 50 per cent in 1930 and
70 per cent in 1931 ; this was an extreme case, the table contains
85 series of observations on the 25 lakes and agreement is far
more marked than is variation.
Both the uniformity and the variation that are found depend
on the supply of organic matter. If this comes chiefly from
sources internal to the lake it will be fairly uniform in amount
and quality, since it depends on nutritive conditions within the
lake and these do not alter greatly or rapidly. If the color ma¬
terials are mainly from sources outside of the lake, such as the
higher extractive colors, these also come from areas and de¬
posits which do not alter and which, under average conditions of
rain, etc., yield similar color mixtures. The largest variations
are likely to be found in lakes with relatively large affluents, like
White Sand Lake. The curves of Fig. 14, therefore, report the
ordinary conditions in the Northeastern Lake District, as found
by Birge and Juday (1931: 320).
Spectral absorption as recorded by von Aiifsess and Erikson
The preceding section shows that Pietenpol’s curves for spec¬
tral absorption by lake waters are in good agreement with those
of the present report. There are curves by two other observers
which may also be brought into comparison; they are those of
von Aufsess (1903) and Erikson (1933).
Von Aufsess examined the spectral absorption in 10 lakes of
Germany; he reported his findings in tables and curves giving
coefficients of extinction in the spectral region 6940-4690 A, as
a maximum. He made readings at 10-15 wave-lengths, so that
his curves have ample data. It is noteworthy that all show a
local absorption band near 6000 A.
Fig. 18 shows four of his curves, altered to curves of per¬
centile absorption and including the extreme cases found by him.
Most of the waters that he examined had little color and some
had absorptions at or near 5000 A, which were lower than any
found in Wisconsin waters, even after long settling.
The curve for Walchensee is of the same type as that for the
settled water of Crystal Lake, Fig. 20; but absorption in the
short-wave spectrum is only about 10 per cent in Walchensee,
while that of Crystal Lake is about 20 per cent. Evidently the
Wisconsin lake had more suspensoid absorption. Absorption in
James & Birge—Lake Waters and Light
77
Fig. 18. Von Aufsess and Erickson. Percentile absorption curves
from inland lakes. A — Arbersee; K= Kochelsee; S = Staffelsee; W=Wal-
chensee. All are German lakes reported by von Aufsess. G=Gunflint Lake,
Minnesota; reported by H. A. Erikson.
the Kochelsee exhibited more color effect; its water seems to be
in the same class as that of Little Bass Lake, Table X, No. 6,
Fig. 27. The color of that lake is rated as 4 ; but it is uncertain
whether Kochelsee would continue its absorption along the same
lines into the shorter wave-lengths.
Staff elsee and Arbersee show much more color effects. The
curve for Staffelsee resembles that of Beasley Lake, Table X,
No. 24; this water was rated as of color 9. Absorption in
Arbersee resembles that of the water of Lynx Lake, No. 2, as
shown in Fig. 22 and given in Table X, No. 36. Its color is 27.
Von Aufsess distributes lake waters into 4 Groups, based on
their behavior toward blue radiation. The first group does not
absorb it at all, and the color of the lake is blue. In the other
classes there is more absorption and the color of the lakes passes
through green to yellowish green and finally to yellow and brown
in Group IV. In this group, represented by the Staffelsee, blue
is “vollstandig absorbiert”. This classification evidently refers
to the effect of the lake waters on the eye, not to percentile ab¬
sorption. The water of the Staffelsee would probably have trans¬
mitted some 50 per cent of the blue radiation at 4900 A, and
perhaps some 25 per cent at 4240 A. The Kochelsee represents
Group III (“stark absorbiert”) ; more than 70 per cent of the
78 Wisconsin Academy of Sciences, Arts, and Letters
radiation was transmitted at 4900 A, and there could hardly
have been less than 60 per cent transmitted at 4240 A.
Erikson’s paper gives a single curve from Gunflint Lake,
Minn., obtained in testing a photographic method of recording
spectral absorption. It shows high absorption by color and rela¬
tively little by suspensoids ; that is, absorption at the minimum is
less than would be expected in a natural water which has so
great an effect in the short-wave spectrum. The present list
does not include a Wisconsin water whose natural state is a good
match for this one. A good agreement may be found with the
settled water of Little Papoose Lake, Table X, No. 32, Fig. 19;
this water had the color 18, but its original color, before settling,
was 30. The same may be said of Horsehead Lake, whose color
fell from 33 to 16 during settling, as is shown in Fig. 34. This
water at color 1 6 is also a good match for the water of Gunflint
Lake, but in its original state its absorption was much greater,
as is shown in Table X. The curve for Gunflint Lake, like those
of the other two lakes named, resembles in many ways the curve
for a filtered water; it may be compared with the mean curve
for filtrates, colors 16-20, Fig. 15.
The percentile absorption curves of Birge and Juday (1931)
deal rather with absorption of spectral color bands than with
that at specific wave-lengths ; they will be discussed in Part II
of this report.
Section 4
Percentile Absorption Curves from Typical Lakes
The following pages discuss the effects of settling, filtration
and dilution, as illustrated in the waters of single lakes which
are typical of general conditions. The results are primarily
shown by diagrams, Figs. 19-28 ; these contain 96 percentile ab¬
sorption curves, of which 39 are from settled waters, 38 from
filtrates, while 19 represent dilutions. They come from 30 dif¬
ferent waters whose colors range from 0 to 180. Similar curves
from these and from other waters are employed in the analytical
study of absorption of radiation, as found in Chapter IV, Figs.
29-52.
Settled Waters. Devils Lake, Little Papoose Lake. Fig . 19.
The water of Devils Lake was employed in an experiment on
the effect of long settling. This is a seepage lake, low colored,
James & Birge — Lake Waters and Light
79
Fig. 19. Effect of time of settling on the absorption in Devils Lake.
D I is the first absorption curve and D II is the curve taken 24 hours after,
the water settling in the tube during that time. D III was taken after the
water had stood one month longer in cold storage and D IV was taken 10
weeks later, while D V was the absorption after settling one year in cold
storage. In the course of this time the color of the water changed from 6 to
4. The broken line is the curve of the water from Little Papoose Lake,
settled 10 months. Note that the absorption curve in the long-wave region
is identical with Devils, settled one year. Settling made considerable
change in absorption in the central and long-wave part of the spectrum in
this water, which was color 18. Compare Fig. 33.
and belonging to Group II. Nearly all of its color is associated
with suspensoids and these, in turn, are mostly plankton and its
derivatives. The lake lies near the top of a glacial hill in a gorge
between high quartzite cliffs; its drainage area is small and no
marsh is included in it.
In Fig. 19, D I is the absorption curve of the recently col¬
lected water, of color 6. Absorption at 4078 A was 85 per cent
and was 60 per cent at the minimum, about 5600 A. This ab¬
sorption is much above the mean for settled waters of color 6,
as shown in Fig. 14. The water remained in the tube of the
monochromator and was tested after 24 hours, with the result
shown in curve D II. There was a reduction of absorption,
amounting to about 8 per cent at 5000 A, with a similar quantity
through the middle spectrum. This was brought about by the
rapid settling of the larger particles of plankton, etc. The water
was returned to cold storage and examined again after a month.
The curve D III shows the result; there is a reduction of ab-
80 Wisconsin Academy of Sciences , Arts, and Letters
sorption at 5000 A amounting to nearly 17 per cent, but there is
only 6 per cent at 4078 A, where color is more effective than at
greater wave-lengths. This curve is much like the mean curve
for settled waters of color 6 and given in Figs. 14 and 42; the
main difference between the curves is at 4078 A, where the lake
water has greater effect. A fourth series was read 10 weeks
later, D IV, which showed a small reduction at 4078 A, in the
short-wave spectrum, with a maximum of about 4 per cent;
little or no change was found beyond 6200 A; the color was
rated at 5, and very probably the rating of 6 for the preceding
series was somewhat too high. The water then went back to
storage and remained until a total time of a year had passed;
the readings at that date gave curve D V. Color had fallen to 4 ;
total absorption in the middle spectrum is below that for the
mean color 4, Fig. 14, but is about the same as that at 4078 A.
The results may be summed up in a table, extrapolating
curves to 4000 A and taking 5680 as the minimum :
Thus the percentage removed by settling, 4000-6000 A is much
the same; but the fraction of the total represented by this loss
increases considerably ; loss is less than 50 per cent of the total
at 4000 A and more than 70 per cent at 6000 A. This difference
depends on the relative efficiency of colors like 4 and 6 at differ¬
ent wave-lengths.
The curve L P is placed in Fig. 19 to show the effect of long
settling on a water of a very different type from that of Devils
Lake. Little Papoose Lake is a drainage lake of relatively high
color, mostly of the extractive type. The original color was 30,
and the water was not examined in that state; but the curve
when recently settled would not have differed notably from those
shown in Figs. 14 or 22 for waters of like color. The water re¬
mained in storage for 10 months and was then examined. Color
was reduced to 18 and the curve LP gives the percentile ab¬
sorption. There is evidently a great change in the middle spec¬
trum, but much less at 4078 A. At this wave-length a meter of
James & Birge — Lake Waters and Light
81
water of color 80 would transmit about 1 per cent of incident
illumination, as is shown in Fig. 14; this water at color 18
absorbs about 91 per cent of incident light instead of some 99
per cent at color 30. Absorption at 5700 A would have fallen
more than 40 per cent, from some 70 per cent for color 30, to
27 per cent for color 18. Drainage waters of this type are char¬
acterized by this great effect of settling on absorption at 6000
A, with a small reduction at 4078 A.
The curve for the settled water of Little Papoose Lake closely
resmbles that for the filtrate of Adelaide Lake, also of color 18,
as shown in Fig. 22. Apparently long settling had allowed nearly
all suspensoids of Little Papoose Lake to reach the bottom, and
had produced much the same effect on the absorption curve that
filtration would have done, if it had left the color the same.
Waters Whose Filtered Color is Zero or Trace .
Fig . 20, Tables X, XI.
In Fig. 20 are grouped absorption curves from several waters
which, when filtered, were rated as of color zero. The main
purposes of the diagram are to show the range of color and ab¬
sorption in waters whose filtrates rate as zero ; and to show the
range of absorption in the short-wave spectrum by waters so
rated in color. All waters belong to Group III.
The lakes whose absorption curves are given as settled wa¬
ters are : Crystal, No. 1 and No. 3 ; Edith, Elkhart, and Silver ;
curves for filtrates represent Crystal, No. 3, Edith and Elkhart ;
Weber and Trout are included to fill gaps in the series of waters.
The settled waters of Trout and Weber lakes were not examined
and the filtrate from Silver was clouded with very fine suspen¬
soids and the readings were obviously useless.
Edith and Elkhart lakes show the highest absorptions among
the settled waters, especially in the short-wave spectrum; the
curve for Silver Lake is the lowest in the middle spectrum. Two
factors must be mentioned which may have contributed to these
results. 1. The water of both Edith and Elkhart lakes was col¬
lected on April 9 and probably contained some color material
brought in with spring rains; settling went on for 6 weeks in
each case and the finer suspensoids might still remain in the
water. Silver Lake was visited in August and its water settled
for 9 months before the curve was taken. The low absorption
82 Wisconsin Academy of Sciences , Arts , and Letters
Fig. 2,0. Absorption curves of 5 waters which became 0 in color when
filtered; the original colors 0 to 6. Upper part shows the curves for settled
water and the lower shows the absorptions after filtration. Note the effect
of the filter, particularly at the wave-lengths longer than 6000 A; also
the reduction in the absorption in the central part of the spectrum in all
the waters. The curve marked (W) in the lower diagram represents both
Weber Lake and Crystal Lake, No. 2. The lowest curve is the third sample
from Crystal Lake, the clearest sample found in this investigation. These
are all samples of waters from Group III as described on page 72.
in the middle spectrum is associated with the long time of set¬
tling, as in Devils Lake, Fig. 19 . These are only samples of the
numerous secondary questions which are raised by a study like
the present one, but which cannot be answered. They involve
questions of fact which must be worked out for each lake. The
time of settling has a very different effect on different lake
waters. The series for Little Bass Lake, in Table X, shows that
it settled as long as did Silver Lake, but without any indications
of especially low readings in the middle spectrum.
The absorption curves for filtered waters show the maximum
range for waters of that rating. Crystal No. 3 has absorptions
James & Birge—Lake Waters and Light
83
so close to those of distilled water that departures may fairly be
called accidental, as Table XI shows. It may be that the high
absorption of Crystal, No. 1, at 4078 A, as shown in Fig. 20 and
Table XI, is due to the continuation of spring conditions until
May 15, when the water was collected; but Weber Lake was
collected on the same date and its curve agrees almost exactly
with that from Crystal No. 2, which was taken on July 7. So
also, the two series from Diamond Lake were taken, one in May
and the other in July; but they show no differences of any sig¬
nificance, as is reported in Table XI.
This table shows that the maximum absorption at 4078 A,
for waters of colors 0 and T, is 35 per cent and the minimum is
3 per cent; this difference rapidly diminishes at greater wave¬
lengths, and there is less than 1.5 per cent between maximum
and minimum at 5970 A, in waters of color 0, and about 3 per
cent if waters of color trace are included. This situation is in
marked contrast to that in settled waters; at 4078 A, in the 10
settled waters of lowest color, the range of absorption is 41 per
cent, and it is about 20 per cent at the minimum.
Waters of Color 6. Fig. 21, Tables X, XL
In Fig. 20 are given absorption curves for waters which are
rated as of much the same color after filtration, although differ¬
ing before that process ; Fig. 21 shows the effect of filtration on
6 waters which are rated as of the color 6 in their settled con¬
dition, but which differ in color when filtered. All lakes belong
to Group III, except White Sand Lake, which is of Group I, at
least in the specimen of water here examined.
The lakes are Columbian, Devils, Edith, Marl, Muskellunge,
White Sand; the curve for Edith is not plotted as a settled
water; Table X shows that it agrees very closely with that of
Columbian, so that both cannot appear on the same diagram.
All 6 curves are shown as filtrates, so far as is possible.
Waters of color 6 have the maximum spread among the
observations on single waters of the same color rating, as is
shown in Fig. 12. The present diagram gives the characteristic
appearance of such a set of absorption curves from settled wa¬
ters. There is a wide spread of curves at 4078 A, from White
Sand with 78 per cent to Muskellunge with 35 per cent; at
5040 A this difference is not greatly reduced. Marl having the
84
Wisconsin Academy of Sciences, Arts, and Letters
4000 A 5000 S000 7000 8000
Fig. 21. Absorption curves in settled waters of color 6 in the upper
diagram and the same after filtration in the lower. Note the relatively
small change in the absorption in White Sand at 4078 A when filtered and
that the curve remains above the others to about 7000 A. This is a typical
water of Group I; the others being from Group III. Compare the above
with the effect of filtering Devils Lake, where the absorption throughout
the spectrum was greatly lowered by the filter. Green Lake, color 4, is
included as a settled water, to show overlapping of color groups.
maximum at 55 per cent, and Muskellunge the minimum at 17
per cent; at 5970 A the figures are 52 and 23 per cent, and at
7000 A they are 61 and 49 per cent; at 8000 A Columbian is
lowest with 90 per cent, while White Sand and Marl coincide at
93 per cent. The spread of the curves depends in part on differ-
James & Birge — Lake Waters and Light
85
ences in the color mixtures of the same rating, but in much
larger part on the effect of suspensoids, whose action is reduced
only slowly as wave-lengths increase.
The filter separated these waters into two groups, as rated
by color ; in Marl, Muskellunge, and White Sand lakes color was
altered little or not at all; Columbian and Devils lakes filtered
to color trace and Edith Lake to color 0. Even with this range
of color absorption curves come together at 6000 A, showing
that the effect of these lower colors is nearly limited to the
short-wave spectrum. White Sand Lake is a conspicuous excep¬
tion; very probably fine suspensoids passed the filter in this
water, as the finest particles of Marl certainly did in that water ;
and Table X shows that this water from White Sand Lake has
an exceptionally great absorption throughout the middle spec-
rum, 5000-7000 A; the second sample from this lake filtered to
color trace ; its absorption was the greatest among waters of
that color, so far as to 5840 A, so it is not improbable that the
color materials of the lake are also exceptional.
Two of the three waters which filtered to low colors offer no
points of special interest; Columbian and Devils lakes have ab¬
sorption curves somewhat above the means for waters of color
trace , but not so far from the mean as to call for discussion. On
the other hand, Edith Lake, color 0, has at 4078 A the highest
absorption found in a water of that rating. Its curve is almost
identical with that of Muskellunge, color 6 ; the type of the color
material probably has an influence on the situation, for the water
of Edith Lake was collected on April 9, when early spring con¬
ditions would control color mixtures ; while that of Muskellunge
was collected in August, at midsummer conditions. But, as
already stated, this aspect of the subject has still to be investi¬
gated.
Groups I and II Compared. Fig. 22
Fig. 22 contains curves from three lakes which illustrate the
differing effects of filtration on waters of different types. The
lakes are Adelaide (color 28), Horsehead (color 83) and Midge
(color 21) ; Adelaide belongs to Group I and has much of its
color in solution, the other lakes belong to Group II and most of
their color is associated with colloids. The lakes differ in color
but their absorption curves are very close together, indicating
Wisconsin Academy of Sciences , Arts, and Letters
86
Fig. 22. Comparison of typical waters of Groups I and II in their ab¬
sorption curves before and after filtering. Note that the absorption in
Midge Lake, color 21, is lowest only in the region between 4000 A and 5000
A, while in the central spectrum it is somewhat higher than Adelaide, color
28. Horsehead, with its color of 33, is highest throughout the spectrum.
Note the very great reduction in absorption due to filtering Midge and
Horsehead, in the violet as well as in the central spectrum, characteristic
of waters of Group II. Adelaide illustrates the behavior of waters of
Group I with its small reduction of absorption in the short-wave region.
Note that its color was reduced less than half by the filter. Midge and
Horsehead lakes are typical of the waters of Group II.
that the combined action on light of color and suspensoids in
these waters is almost identical. Only a fraction of one per cent
of the incident solar energy remains at 4078 A and a maximum
of 83-37 per cent at 6700 A.
Absorption curves for the filtrates are shown in the same
diagram; filtration has relatively little effect on the water of
Adelaide and makes great changes both in color and absorption
in the other two waters. Color is reduced about one-third in
Adelaide (28-18), about three-fourths in Midge (21-6), and
more than nine-tenths (33-2) in Horsehead. Evidently most of
the color of Adelaide is dissolved; the water was filtered again
without much effect, indicating that the suspensoids were almost
wholly removed by the first filtration. Its water was optically
good, without much turbidity, and the curve for the filtrate is
practically determined by color (18) and by water. The curve
is typical for a water of Group I with a medium high color;
absorption shows little reduction at 4078 A but a marked change
James & Birge— Lake Waters and Light
87
in the middle spectrum, at 5800 A from 67.4 per cent to 30.5.
This indicates that the settled water had a large suspensoid ab¬
sorption, most clearly evident in the middle spectrum. Filtered
waters of color 26-28 have about 33-36 per cent absorption at
5800 A, somewhat below that of the filtered water of Adelaide
when recomputed for that color.
The color mixtures of these three waters differ in the settled
and filtered states, so that suspensoid absorption can not be
accurately calculated, but its mean value in the middle spectrum
for Adelaide is about 45 per cent, 58 per cent at 5040 A and 36
per cent at 7000 A. (Fig. 47) Adelaide Lake is small and deep;
its water is usually clear in the open lake (Birge and Juday,
1932 : 544) showing that the suspensoids are not of a type that
causes obvious turbidity.
Midge Lake is surrounded with bogs, the lake water is both
colored and turbid, and the bottom of the lake is composed of
soft ooze. (Birge and Juday, 1932: 551). Color is largely of
the colloidal type and filtration effected great changes both in
color and absorption. The filtrate curve shows these changes
throughout the region 4078-6000 A; absorption of light at 4078
A goes down from 99.9 per cent to 37, at 5800 A from 68 to 14
per cent. Evidently most of the color is in colloidal form and is
removed with the suspensoids. In the settled water mean sus¬
pensoid absorption, 5000-7000 A, is computed as about 51 per
cent, somewhat higher than in Adelaide. This sample from Midge
Lake had settled for 10 months, but was much more turbid than
a second sample of the same color after one month settling; at
5800 A, No. 1 absorbed about 68 per cent of the light and No. 2
only about 49, or 28.5 per cent less. In this second sample mean
suspensoid absorption 5000-7000 A, was about 22 per cent; the
filtered water was much like that of No. 1 ; see Table XI. Horse-
head Lake is also surrounded with bog and marsh ; nearly all of
its color was in colloidal form and as the water stood in cold
storage only 3 weeks, it had little chance to pass into solution.
Filtration reduced color from 33 to 2; absorption at 4078 A
from 99.9 to about 30 per cent, at 5800 A from 71 to 7.0 per cent.
It will be noticed that at 5800 A color adds practically nothing
to the absorption of light by water. This water offers an ex¬
treme case of colloidal color when suspensoids are largely or¬
ganic.
88 Wisconsin Academy of Sciences , Arts, and Letters
Settled and Filtered Waters of Group I. Fig. 28.
Fig. 23 contains absorption curves for settled and filtered
waters from lakes belonging to Group I. Such waters have much
dissolved color, which passes the filter ; filtration may not reduce
color at all or it may greatly change it. In any case much of the
color material is extractive and comes from the environment;
and color absorption at 4078 A is high, relatively to the color
rating. All are drainage lakes, with affluents which bring color
to them.
The list of lakes follows, with their color ratings both as
settled and as filtered waters; curves are not given for colors
marked with an asterisk. Lake Anna (62 — 20), George (24 —
16), Helen (74* — 12), Little Papoose (18 — 8*), Lynx (27 — 4),
Mary (101—34, 32), Nagawicka (12—12), North (18*— 12),
Oconomowoc (6—6), Oxbow (63 — 28*).
The water of Lake Mary was filtered twice; the second fil¬
tration made little change in color or in absorption, indicating
that the first had removed nearly all suspensoids and colloids,
and with them had taken about two-thirds of the color. In a
like way, the filter removed about two-thirds of the color from
the water of Anna, five-sixths from Helen and an even larger
part from Lynx Lake. On the other hand, filtering brought no
change to the color of Nagawicka and Oconomowoc lakes and
that of North and George was not greatly reduced.
The Oconomowoc-Waukesha Lake District has three repre¬
sentatives in this diagram, Nagawicka, North, and Oconomowoc
lakes; the others, which appear in the tables, are La Belle,
Okauchee, and Pine. The account of the District, with map, in
Juday (1914: 36), shows that all lakes lie near together; they
have similar environments and resemble each other in area,
depth, etc. (See Table VIII, p. 60). The color material in all
of them seems to be of the same type. All belong to Group I,
except Pine Lake, which has no affluent and is placed in Group
III. Their color acts alike both in settled and filtered waters.
The curves for settled waters in Fig. 23 form a characteristic
series. Absorption of light by color and suspensoids together
adds much to that of water and, for higher colors, shifts the
minimum of the combined effect from near 5800 A to 7000 A;
there the rapidly rising action of water itself takes control of
the curve. The curves for Lynx (color 27) and George (color
James & Birge — Lake Waters and Light
89
Fig. 23. Absorption curves of 9 samples of waters of Group I before
and after filtering. The absorption in all is much reduced in the central
spectrum but relatively little at 4078 A. Note that most of the waters
are in the same order of intensity of color after filterng as before, though
the absorption in Lynx was reduced somewhat out of proportion to the
others; this lake is on the border line between the first and second groups
of waters. A few lakes are shown that are not common to the two diagrams
in order to include more cases, since the general form of the curves is the
same for all in a given group. North, Nagawicka and Helen lakes are
remarkable for their low absorption near 5800 A while still showing
high absorption in the violet.
90 Wisconsin Academy of Sciences , Arts , and Letters
24) differ much more than their color ratings would indicate.
Some part of this difference may be due to the long time which
George Lake had in cold storage, 9 months, as compared with
2 months for Lynx. This suggestion finds some support in the
curve for Little Papoose, which also had 9 months in storage and
agrees with George in relatively small absorption, 6000-7000 A.
But the character of the suspensoids in the several waters prob¬
ably had a part in these diverse effects, since Midge Lake (color
21) was 9 months in storage with no such results, as is shown
by Table X and Fig. 22.
In the curves for filtrates are included three waters of color
12; two of these are from southern Wisconsin and one, Lake
Helen, is from the Northeastern District. They serve to illus¬
trate the relation between color of their grade in their action on
radiation. The percentile absorption for Lynx (color 4) and
Oconomowoc (color 6) at 4078 A is much above the average for
the mean of waters of their colors, as is shown in Fig. 15 ; the
color material of both filtrates adds much to the effect of water
in the spectral region 5000-5600 A.
Many additional details could be pointed out, but the two
sets of curves give a general picture of the effects of filtration
on waters of this Group, and in this fact lies a great part of
their value. In the region 5000-7000 A absorption of light is
much reduced by filtration; color maintains its relatively high
effect, 4000-5000 A; water prevents reduction of absorption in
the filtrates, 7400-8000 A. The percentile spread of the several
curves is widened at 4078 A, little altered at 5000 A, much re¬
duced at 6000, 7000 and 8000 A. The contrasted effects on radi¬
ation of water and of color are responsible for this picture ; and
the percentile method of plotting has much to do at 4078 A.
Waters of Group II. Fig. 21+
In waters of this Group absorption is high in the violet of
settled waters and low in the same region of the filtrates. The
lakes of this Group are listed in Table IX, p. 63, together with
data for their comparison. Fig. 24 shows absorption curves for
five typical lakes of the Group. These are Bear (12*— Tr.),
Beasley (9 — 4), Horsehead (33 — 2), Kawaguesaga (10—4),
Youngs (13 — 3) . Absorption data are given in Tables X and
XI ; the settled water of Bear Lake was not examined.
James & Birge — Lake Waters and Light
91
Fig. 24. Absorption curves of 4 typical waters of Group II before and
after filtering. Note that Horsehead, with the highest original color had the
lowest color after filtration. Note, however, that the filter reduced these
waters to almost identical character. The absorption curve of the filtered
sample of Bear Lake (B), from Group III, is also shown in this figure
for comparison with the second group. Note that it is lower in absorption of
the violet than are any of the other waters.
The water of Horsehead Lake is discussed in connection with
Fig. 22 ; the curves are repeated here in order to show their re¬
lations with those from waters of lower color. The intimate
relation of color and suspensoids in this water is made plain
by the great reduction of color wrought by the filter.
The water of Kawaguesaga Lake was in cold storage one
month and color changed from 14 to 10 during that time. This
indicates that color was associated with suspensoids, and that
many of them were large enough to settle out and carry their
color with them. The filtered water has color 4, and its absorp¬
tion was close to that of Horsehead (color 2) and Beasley (color
4).
Beasley Lake is one of the lakes of the Waupaca Chain O’
Lakes. Table VIII lists six of these lakes among the waters
examined for this report. The map of the Chain is found in
Juday (1914: 104) ; it shows 18 lakelets, of which 12 are in a
connected chain. Three of the waters examined lie in the course
of the chain, Columbian, Long and Rainbow lakes; Marl Lake
is at the head of the chain ; Otter and Youngs lakes are not in
it but are connected through short outlets. The four lakes in
92 Wisconsin Academy of Sciences , Arts , omd Letters
the chain are of Groups II and III; Youngs Lake is placed rather
doubtfully in Group II ; while Otter is definitely of Group I. Thus
the typical lakes of this little chain belong in general to Groups
II and and III, just as those of the Oconomowoc- Waukesha area
belong to Group I.
Beasley and Youngs lakes have absorptions of similar char¬
acter; suspensoids and color are so combined that the two wa¬
ters in their settled state have almost the same curves, although
Beasley has color 9 and Youngs color 13. Their filtrates were
of colors 4 and 3, respectively, and the curves were in close
agreement through most of their course; but that from Youngs’
water departed from that of Beasley in the short-wave spectrum
and had a high absorption at 4078 A. There was therefore pres¬
ent in this water the larger amount of dissolved color.
There is no special diagram for waters of Group III, which
are found in the marly lakes and in other lakes of low color.
Figs. 20 and 21, which give curves for waters that filter to color
0 and for waters of color 6, are made up almost wholly from
waters of this Group. Such waters have lower absorption in
both settled and filtered condition than is usual for lakes of the
same color. Table VIII shows that none of the 15 lakes in the
Group has a higher color than 6 in the settled state or above
trace after filtering. Thus there is little or no extractive color
in the waters.
In the general statement regarding the three Groups, em¬
phasis was placed on the fact that the classification depends on
the varying relation of color materials and suspensoids, chiefly
colloids; also that the presence and extent of this relation is
determined by the action of filters whose effect varies with their
porosity. The Groups therefore do not represent fundamental
facts of nature, but relations which may pass by small degrees
from one type to another. Helmet Lake represents an extreme
on one type, that of Group I, the filters employed had practically
no effect on its great color and a correspondingly great part of
the color material is in solution. Crystal Lake, sample 3, is an
equally extreme case of water without color. No color rating
above zero could be given to it, and the same filter which passed
the color 264 of Helmet, removed suspensoids from Crystal and
left the water to absorb practically the same percentage of radi¬
ation as that taken by distilled water at corresponding wave¬
lengths. Between these extremes there are innumerable possible
James & Birge — Lake Waters and Light
93
combinations of numerous color materials and suspensoids, col¬
loid or other, and a few of these are illustrated by the curves of
the preceding diagrams. The waters of the report have been
grouped into classes in order that the facts and observations may
be handled more conveniently and discussed more intelligently.
The few details that have been given are probably enough to
show that series rather than groups would be a more perfect
classification ; but if series are to be arranged they must be based
on a much larger number of waters, and these must be examined
by methods more accurate than those now available.
Dilutions. Figs. 25, 26, Table XL
Two diagrams are given to percentile absorption curves from
dilutions. Fig. 25 shows 10 curves from dilutions of the filtered
water of Helmet Lake, color 180 ; Fig. 26 has 9 curves from the
dilutions of the unfiltered water of Lake Mary, color 109. With
each curve is shown the percentage of the original water in the
specimen examined, and also the resulting color. There will be
noted a very rough correlation between the percentage and the
color ; this appears both in the single curves of the two diagrams
and in the two sets from lake waters, one of which had an origi¬
nal color nearly twice as great as the other.
Fig. 25. Helmet Lake. Absorption curves for dilutions from filtered
waters. Compare the curves with those for corresponding colors of filtrates
in Fig. 15. These dilutions of filtered water behave much like filtrates; the
addition of distilled water has made no marked change. Note the instance
of local selective action near 6000 A in the curve for color 45. All curves
are of usual form for the indicated colors.
94 Wisconsin Academy of Sciences, Arts, and Letters
The water of Lake Mary was in cold storage for 7 months
before it was examined and most larger particles in suspension
should have settled out during that time. Filtration of other
samples showed that there is in the water a “dark brown slime”,
with which most of the color is associated. The water of Helmet
Lake has a much darker color which passes the filter; in one
sample, settled 9 months, the color rating was 264, and was not
altered by filtration; in another, settled for 6 weeks, filtering
reduced color slightly, from 236 to 212.
The absorption curves for dilutions of Helmet should be com¬
pared with those of filtrates of Group I, as given in Fig. 23.
Curves for medium colors in the two diagrams have the same
pronounced U-form that is characteristic of the Group. Those
for low colors, especially that for the 0.1 per cent dilution, have
high absorption due to presence of very dark material. The color
in the waters of the three greatest dilutions exerts no appreci¬
able effect beyond 6000 A, where their curves coincide with that
of water itself.
The curves for Lake Mary, Fig. 26, indicate different condi¬
tions due to the suspensoids of the unfiltered water. The origi-
Fig. 26. Lake Mary. Absorption curves for the series of dilutions
from unfiltered water. Note flat run of curves for low colors in short¬
wave spectrum; a characteristic of settled waters, and present here in spite
of great dilution. Compare with curves for filtrates of Helmet, Fig. 25. Note
the series of maxima and minima in the curves in the central spectrum,
indicating selective action in relatively narrow bands. This water con¬
tained much suspensoid material and much of the color was associated with
colloids.
James & Birge — Lake Waters and Light
95
nal color was 109, about one-half that of Helmet ; there is there¬
fore little material in the 0.1 per cent dilution, either of color or
suspensoid, which can absorb radiation. At 4078 A Mary ab¬
sorbed 5.2 per cent of the radiation present and Helmet absorbed
9.1 per cent; the effect being much in the same proportion as
the original colors of the two waters. But with all greater
amounts of the waters of Lake Mary, the action of suspensoids,
including the colloids, asserted itself strongly. Compare curves
for dilutions 0.5 per cent and 1.0 per cent in Fig. 26 with those
in Fig. 14 for mean colors 0 and 4 in settled waters. The com¬
parison may be carried farther, and the curve for Mary, color 20,
may be compared with that for mean color 28 in Fig. 14. The
whole assemblage of curves in Mary shows the type for settled
waters ; although suspensoids in the dilutions have been reduced
in the same proportions as the color material, and in the ordinary
settled water there is no necessary correlation of color with col¬
loids or with suspensoids in general.
Absorption of Radiation at 3650 A. Figs. 27, 28
In the present study observations were in general limited to
the spectral region 4078-8000 A; but in a few cases the series
Fig. 27. Percentile absorption curves, 3650-7600 A, in settled waters.
Color is marked on the curves. The lakes are: Columbian; Crystal, two
samples; Little Bass; Marl; Mendota; Muskellunge. The extension of
the curves to 3650 A brings out the increasing influence of color on absorp¬
tion, as curves pass through the violet region of the spectrum. The effect
of suspensoids, including colloids is dominant in the middle spectrum. Note
that the percentage which separates maximum and minimum curves is
about 50 at 3650 A and is still about 37 at 5700 A.
96
Wisconsin Academy of Sciences, Arts, and Letters
Fig. 28. Percentile absorption curves, 8650-7600 A, in filtered waters.
The lakes are given with their colors, in the order of their appearance in
the diagram; Diamond, 0; Crystal, No. 2, 0; Bear, T; Devils, T; Elkhart, T;
Long, 3; Lynx, 4; Oconomowoc, 6; La Belle, 8; Oxbow, 10. The curves
show the rapid rise in the effect of color on radiation in the region 4078
A-3650 A. The U-form of the absorption curve is much more strongly
marked than when curves end at 4078 A. Note that percentile absorption at
3650 A is nearly 100 per cent for color 10. Suspensoid absorption has little
to do with these filtrates. The range of absorption is from about 12 per
cent to 100 per cent at 3650 A, a spread of 88 per cent; at 6000 A the
spread is reduced to about 3 per cent. Compare the diagram in the middle
spectrum with Fig. 27.
of readings was extended by a reading at 3650 A. These are
numerous enough to give some notion of absorption at that
wave-length, though they are too few to permit the extension of
mean extension curves so far into the short-wave spectrum. The
number of waters so examined is 43, of which 24 were filtrates,
10 were dilutions and 9 settled waters. The results are here
presented in a table for the filtrates and in two diagrams, one
for settled waters and one for filtrates. The curves for the dilu¬
tions did not offer any distinctive features.
All waters so examined had low colors. Only 4 rated above
color 6 and only one of these rated so high as color 10. This
was a filtrate and its absorption at 3650 A was recorded as 99
per cent. Probably little more than a trace of radiation was
passed by a meter of this lake water, and filtrates of the color
8-12 absorb 100 per cent of incident radiation at 3650 A, about
as colors 35-40 absorb 100 per cent at 4000 A,
James & Birge — Lake Waters and Light
97
Table XIII
Percentile absorption of radiation at 3650 A, by one meter of filtered
lake water.
Note.— Readings at 3650 A were recorded as whole percentages; in
Table XIII mean results are given to fractions of one per cent, but the
decimal has little value. Minima and maxima are stated at the nearest
whole per cent. Two filtered waters had higher colors than those used in
the table. One water, with color 8, absorbed 96 per cent of the light at
3650 A and 61 per cent at 4078 A; the other water had the color-rating of
10 and its absorptions were 99 and 75 per cent at the respective wave¬
lengths. See Table XI for details.
Fig. 27 contains 7 absorption curves for settled waters; it
shows characteristic curves of this type. The diagram gives three
curves for water of color 6, Marl, Columbian and Mendota lakes.
These nearly cover the range of absorption for that color in the
middle spectrum; the minimum for Marl is about 50 per cent,
that for Mendota is about 14 per cent, while Columbian with
about 28 per cent is not far from the mean minimum. The wide
difference is due mainly to the great amount of suspensoids in
Marl Lake. Only one of the lakes in the list had been in storage
more than a few weeks ; this is Little Bass Lake and its absorp¬
tion had not greatly altered during the time.
In Fig. 28 percentile absorption curves from filtrates are
plotted to 3650 A. The diagram brings out two evident facts:
1. The great increase of absorption between 4078 A and 3650 A;
the percentage for any curve is nearly doubled in this spectral
space of about 500 A. 2. The very pronounced U-shape of the
curves ; more evident in these curves than in those which end at
4078 A, because the minimum is at a lower level. The maximum
absorption at 5685 A for any of these waters is about 14 per
cent, and that of distilled water at the same point is 5 per cent.
The curve for Oxbow Lake, color 10, shows that absorption
for this color at 3650 A is practically 100 per cent; for the curve
98 Wisconsin Academy of Sciences, Arts, and Letters
is beginning to take the form which has its concavity facing
downward, the form that characterizes curves which reach 100
per cent within the spectral region included in the diagram.
References
American Public Health Association. 1926. Standard methods for exami¬
nation of water and sewage. New York.
Birge, E. A., and C. Juday. 1931. A third report on solar radiation and
inland lakes. Trans. Wis. Acad. Sci. 26:383.
Birge, E. A., and C. Juday. 1932. A fourth report on solar radiation and
inland lakes. Trans. Wis. Acad. Sci. 27:523.
Erikson, H. A. 1933. Light intensity at different depths in lake water.
J. 0. S. A. 33:170.
Juday, C., and E. A. Birge. 1933. The transparency, color and specific con¬
ductance of the lake waters of Northeastern Wisconsin. Trans. Wis.
Acad. Sci. 28:205.
Leighton, M. O. 1905. Field assay of water. U.S.G.S., Water Supply and
Irrigation Papers. No. 151. Washington.
For reference to papers by von Aufsess and Pietenpol, see p. 44.
James & Birge — Lake Waters and Light
99
EXPLANATIONS OF TABLES X, XI, XII.
The following Tables, X, XI, XII, contain the record of the observations
which are central to this Report. They give the percentile absorption of
light at various wave-lengths, from 4078 A to 8000A, by a stratum of lake
water one meter thick. There are three tables, one each for waters that
have been settled in cold storage, filtered through a Berkefeld filter or
diluted with distilled water. The tables contain 161 series of readings from
the waters of 48 lakes and 1 river; absorptions were observed at 21 wave¬
lengths, which are stated in the tables. Opacity of the waters reduced to zero
the value of some of these readings in the short-wave spectrum. There
are 54 series from the settled waters of 40 lakes; 65 filtrations from the
waters of 44 lakes; and 52 series of dilutions of 8 waters from 7 lakes.
Readings were also made at 3650 A in 9 settled waters, 24 filtrates, and
10 dilutions; these are also reported in the tables.
In each table the waters are arranged in order of color, so that it may
be easy to compare absorption in waters of the same or closely allied color¬
rating and in like relation to filtration. The color stated is that of the
water at the time of examination. In the case of dilutions it will be noted
that there is a very rough correlation between the color of the original
water and that of a given percentage; e.g., a dilution containing 10 per
cent of Helmet Lake water (color 264 or 180) is recorded with color 25
or 26, one with 30 per cent of water from Lake Mary (color 101) has color
30, etc.
This method of arranging the waters removes from the series of dilu¬
tions the results from the undiluted (100 per cent) water; they will be
found in the table of settled waters, or, in the case of Helmet Lake,
sample 2, among filtrates.
All of these waters when run were contained in a tube lined with
paraffin; the results for distilled water that are given in the tables are
from observations made with such water tubes. For wave-lengths of 5685 A,
or less, they are higher than those reported in Table I for distilled water
run in silver plated tubes; but the differences are so small that it was not
necessary to correct for them.
In Table X is stated the length of time that the several waters re¬
mained in cold storage; it will be noted that long time of settling produced
very different effects in different waters. Table XI states the month and
day of collecting the water; waters collected in spring may contain ex¬
tractive color, and also suspensoids, brought in by spring rains; in waters
collected in mid-summer or early autumn such conditions are likely to be at
a minimum. In Table XII are stated the percentages of the lake waters
which were contained in the several dilutions.
The Report states that 181 series were run; the omitted observatiens
added nothing to the general result. Many of them were partial series,
taken in order to confirm or correct results in the series reported.
Settled waters.
Percentage of radiation absorbed by one meter of lake water.
100 Wisconsin Academy of Sciences, Arts, and Letters
Table X. Settled Waters. (Continued).
James & Birge—Lake Waters and Light 101
Table X. Settled Waters. (Continued).
102 Wisconsin Academy of Sciences , Arts , (md Letters
Filtered waters .
Percentage of radiation absorbed by one meter of lake water.
James & Birge — Lake Waters and Light 103
Table XI. Filtered Waters (Continued).
104 Wisconsin Academy of Sciences, Arts, and Letters
Table XI. Filtered Waters (Continued).
James & Birge — Lake Waters and Light 105
Table XI. Filtered Waters (Continued).
Table XII
Dilutions.
Percentage of radiation absorbed by one meter of lake water.
James & Birge—-Lake Waters and Light 107
Table XII. Dilutions , (Continued)
108 Wisconsin Academy of Sciences, Arts, and Letters
Table XII. Dilutions , (Continued)
James & Birge — Lake Waters and Light 109
CHAPTER IV
Analysis of Total Absorption of Light
The preceding chapters of this report have dealt with the
action on spectral radiation, of lake water and all matters con¬
tained in it. The purpose of the study was to obtain information
along this line, which should extend through the visible spectrum
and which, especially, should give a more accurate knowledge
of the effect on light of colors in lake waters. The filtration of
these waters was intended to be so thorough as to leave little
absorbent in them except color; so that the observations on
the filtrates should result in “color curves”, as the report has
named them.
But after the observations had been ended and tabulated, it
appeared that they might also furnish unexpected conclusions
regarding the relations of lake waters and light. They furnished
data by which the total effect of lake waters on light could be
divided among several factors. This division was necessarily a
very rough one, judged by any quantitative standard. This situa¬
tion is in part due to the conditions present in the waters. Their
action on light is a very variable one, both in general and locally ;
as absorbents vary and as bands of local selective absorption
appear. It follows that to obtain even approximate accuracy
regarding the factors at work in a single water, very numerous
and closely spaced readings of absorptions must be made
throughout the visible spectrum; and if general results are to
be reached, many waters must be thus examined.
Such detail was quite outside of the possibilities of the pres¬
ent study ; but it has seemed wise to present the results of this
analysis, even in their imperfect form. They are offered to the
eye in diagrams, rather than in tables giving percentages. Their
preparation has brought to this Survey suggestions of wider
possibilities in the study of the relations between lake waters and
110
James & Birge — Lake Waters and Light 111
light; and it is hoped that the curves contained in these dia¬
grams may, in their turn, suggest new and more accurate meth¬
ods of quantifying the relations between the lake and the sourcq
of the power that operates its activities.
Chapter III discusses the data which come from observation
and experiment on percentile absorption of light in the visible
spectrum by lake waters. These are given in three large tables,
for settled and filtered waters and for dilutions, and in diagrams
derived from these tables. The present chapter contains results
which are derived from these data by combinations and simple
computation. They are presented in diagrams rather than in
tables ; since the active agents in the absorption of light in lakes
are numerous and are little known in a quantitative way. Under
such conditions diagrams best convey the conclusions that may
be reached.
In the tables and diagrams in Chapter III, percentile values
are given to the united action of water and of all other agents.
In the present Chapter separate values are given to water and
to the combined action of other absorbents; and the attempt is
also made to assign tentative values to the action of color and of
suspensoids in certain waters.
Since a separate value is assigned to the action of water, this
may be used in computing the value of other factors in absorp¬
tion. There thus arise several types of ordinates for as many
types of curves which show, in percentile form, the action on
light of the factors present in lake water.
These types are as follows :
1. Total absorption curves. 4. Factor curves, general.
2. Water absorption curves. 5. Factor curves, color.
3. Remainder curves. 6. Factor curves, suspensoids.
A seventh type might also be named, that of Residue curves,
the effect of the materials removed by filtration; but these are
not considered at present. Numerous curves of Type 1, total
absorption curves, have been presented in Chapter III ; those of
Type 2, water absorption curves, have been similarly presented
112 Wisconsin Academy of Sciences , Arts, and Letters
in Chapters II and III. The ordinates for Type 3, remainder
curves, are derived from those for total absorption and for wa¬
ter. The percentage of total absorption is regarded, at any
point in the spectrum, as the sum of two factors, (A) water and
(B) all other agents combined. At each point of observation in
the spectrum the percentile value of water is subtracted from
that of total absorption ; the remainder is the value of absorp¬
tion by other agents. The same result may also be stated as the
amount of absorption which is added by other agents to that of
water.
Types 4, 5, 6 are Factor curves; that is, the percentage of
total absorption is regarded, not as the sum of the action of cer¬
tain agents, but as their product. One such curve may be com¬
puted, which with that for ivater will yield the observed total
result ; or this general factor curve may itself be assigned to two
components, color and suspensoids. The methods and the results
are given in the following sections of this Chapter.
Section 1 . Remainder Curves
In this section the percentile absorption of light brought
about by a meter of lake water at any wave-length of the spec¬
trum, is regarded as the sum of two factors ; one of these is the
action of water, the other is the combined action of all sub¬
stances contained in the water. The percentile value of water
absorption throughout the spectrum is stated in the first column
of Tables X, XI, XII, and it is easy to determine the amount
added to this by any water at any wave-length, where absorp¬
tions are recorded.
Two diagrams are given to show a series of such curves ; the
first, Fig. 29, is from settled waters, the other, Fig. 30, is from
filtrates. Curves for total percentile absorption are represented
by full lines, light weight ; that for water absorption is a broken
line, light weight; the remainder curves are shown by broken
lines, heavy weight. Total percentile absorption at any point of
the spectrum is equal to that of water, plus that indicated by the
remainder curve. For example, Table X gives the absorption of
Tadpole Lake at 6850 A as 75.4 per cent; that of water at the
James & Bit ge— Lake Waters and Light
113
Fig. 29. Settled waters. Comparison of curves for total absorption and
remainder curves. Three types of curves are given: 1. Total absorption
curve, taken from data in Table X. The symbol of the lake and its color
are given from Table VIII. 2;. Water curve, from Table X; this is marked
W, and its ordinates are taken as constants in all waters. 3. Remainder
curves, shown for each water by broken heavy lines. Their ordinates are
the remainder after subtracting water absorption from total absorption
at any wave-length. They represent the combined effect of all absorbents
other than water, considered as an addition to the effect of water. In other
words, the ordinate for total absorption at any point is equal to the sum
of that for water and that of the remainder curve. The contribution of
water is very small in the short-wave spectrum; that of other absorbents
decreases rapidly in the middle and long-wave spectrum, and becomes
negligible, 7400-8000' A. Note indications of local absorption bands near
6000 A.
same point is 38.0 per cent; the value of the ordinate of the
remainder curve is therefore 37.4 per cent. In fact, the curves
for water and for remainder absorption cross close to 6850 A,
as shown in Fig. 29. All ordinates for the remainder curves are
determined in the same way.
Remainder curves are all of the same type. They are close
to the total curve in the region 4000-5000 A, where water has
little effect on light. They depart more widely, as the action of
water increases, from 5500 A on. In the long-wave spectrum,
7000-8000 A, the effect of other agents than water decreases and
it becomes insignificant from 7400 A to 8000 A.
Irregularities appear in most of these remainder curves. They
represent local bands of selective absorption; their value has
114 Wisconsin Academy of Sciences , Arts, and Letters
Fig. 30. Filtered waters. Comparison of curves for total absorption and
remainder curves. The diagram is quite like Fig. 29 and the same explan¬
ation applies to it. Curve for color 104 is from Table XII, the others are
from Table XI. Figs. 29 and 30 are intended to show the amount of ab¬
sorption of radiation in lake waters which may be added to the effect of
water itself by matters dissolved or suspended in the lake.
been worked out in many cases, especially that of bands which
appear at the same point of many curves, such as those near
6000 A. But the points of observation are spaced some 200 A
apart and this distance is so great that such bands can not be
limited with accuracy. They are therefore left with this men¬
tion.
Section 2. General Factor Curves
2-Component Analysis
In this section total percentile absorption is considered, not
as the sum, but as the product of two factors ; the same as those
of the preceding section. The ordinates for the general factor
curve may be computed from coefficients of extinction; but the
method preferred in this report is to employ percentile trans¬
mission of radiation, the complement of percentile absorption.
For example, in the settled water of Marl Lake, Table X, No. 15,
percentile absorption at 6300 A is 55.5 per cent; that of water
at the same wave-length is 25.0 per cent. Transmission is there¬
fore 44.5 per cent for the lake and 75.0 per cent for pure water;
dividing the first percentage by the second gives a quotient 59.3
per cent. The transmission ordinate, 44.5 per cent, is therefore
James & Birge — Lake Waters and Light
115
4000 A 5000 6000 7000 8000
Fig. 3!l. Settled waters. Comparison of remainder and factor curves.
In Figs. 31 and 32 the remainder curves of Figs. 29 and 30 are compared
with corresponding factor curves. Factor curves are based on percentile
transmission of radiation by a meter of lake water, not on percentile ab¬
sorption. Percentile transmission is the complement of absorption, and is
reckoned downward from the 100 per cent line. Ordinates for total trans¬
mission at any wave-length are regarded as the product of two factors:
1. The percentile transmission of water. 2. The percentile transmission
permitted by all other absorbents, acting independently of water. For
method of computation see p. 115.
Factor curves are very close to remainder curves in the short-wave
spectrum, but depart more widely from it as wave-length or color in¬
creases. The irregularities, 7000-8000 A are discussed p. 121.
the product of two factors, water at 75.0 per cent and other ab¬
sorbents at 59.3 per cent. Other ordinates are determined in the
same way and the general factor curve is plotted from them.
The methods and the results of this computation are presented
in detail in Table XIV for Nagawicka Lake, on p. 127.
Figs. 31 and 32 show the position of the factor curve in the
10 waters whose remainder curves are given in the preceding
diagrams. In all cases the factor curve has a higher absorption
than the remainder curve; the two curves are close together in
the short-wave spectrum; they separate more and more for
higher colors and at greater wave-lengths. Numerous irregu¬
larities appear in these curves ; they are very conspicuous in the
spectral region 7000-8000; this condition is further discussed,
p. 121.
Since the percentile action of water itself on radiation at any
wave-length of the spectrum is taken as a constant in all lake
116 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 32. Filtered waters. Comparison of remainder and factor curves.
The diagram corresponds to Fig. 30 as Fig. 31 does to Fig. 29. The explana¬
tion of Fig. 31 applies to Fig. 32. These factor curves are those employed
in the 2-component analysis of absorption, and may be better understood
from the diagrams of single waters, like that for Marl Lake, Fig. 33.
waters, the combined percentile effect of the other agents can be
readily determined. Percentile ordinates have been computed and
general factor curves have been drawn for each of the 181 wa¬
ters whose total action on light has been examined. This total
action has thus been distributed to two components, one of them
being water and the other the combined effect of substances
contained in the water.
Four diagrams are given, Figs. 33-86, to illustrate 2-compon¬
ent analysis of the action on light of the settled or filtered waters
of four lakes. Altogether 21 curves for total absorption are
given, with the corresponding general factor curve and the curve
for water absorption. There are 3 curves for settled waters,
3 for filtrates, 7 for dilutions of settled waters and 8 for dilutions
of filtered water.
Marl Lake, Fig . S3
Marl Lake represents a simple case of 2-component absorp¬
tion. Curves are given for total absorption of settled water (S),
and for filtrate (Fi), and for that of water (W) ; the two factor
curves are marked F. The percentile transmission at any wave¬
length, of the factor curve, multiplied by that of water, yields
the observed total transmission at that point. Note the occa-
James & Birge — Lake Waters and Light
117
Fig. 33. Marl Lake. 2-component analysis of absorption of radiation.
In Figs. 32 and 33, S= curve for total absorption, settled water; Fi=curve
for filtered water; W — curve for water absorption. These are from Tables
X, XI and are taken as curves of percentile transmission. F= factor curves
computed as stated in text, p. 117. Percentile transmission of water at
any point in the spectrum, multiplied by percentile transmission of factor
curve, equals the observed transmission in total curve.
The curve for water and the factor curve for any lake may also be taken
as percentile absorption curves. They represent the effect on radiation of the
factor or factors, as they would act if operating alone, and not affected by
the presence of other absorbents. See p. 128.
Note that in the region 7400-8000 A the filtrate curve shows higher
absorption than that for the settled water. This not infrequently occurs,
as the result of disturbance of conditions by the filter.
sional irregularities in the curves, and especially that absorption
at 7400 A is greater in the filtered than in the settled water.
Such results not uncommonly follow filtration, in this spectral
region. Factor curves are particularly uneven, 7000-8000 A.
The factor curve of the filtrate is the same as that used as the
color curve in the 3-component analysis of the same water in
Fig. 43.
Horsehead Lake, Fig. 3U
For Horsehead Lake 4 pairs of curves are shown, 2 from set¬
tled water and 2 from filtrates. The first and the last curves are
also given in Fig. 22, to illustrate the effect of filtration. The
water of color 33 had been 3 weeks in storage, that of color 16
had been there for 3 months. Settling had removed much of the
color from this water, as it did from that of Little Papoose Lake,
118 Wisconsin Academy of Sciences , Arts , cmd Letters
Fig. 34. Horsehead Lake. 2-component analysis of absorption. Four
pairs of curves are given, two for settled and two for filtered water. The
curve for color 331 was from water that had settled for 3 weeks; that for
color 2 comes from the same water refiltered. The explanation of Fig. 37
applies to Fig. 34.
The curves for colors 33 and 2 are given in Fig. 22. Note that the
curve for color 2 reaches zero absorption near 5800 A, but reappears near
7000 A. In the curve for color 16 there is a strong band of local absorption
near 6000 A. Indications of such a local effect are common and may be
seen in many diagrams; but so marked a band is rarely found. Compare
Fig. 7.
Fig. 10. Filtering reduced color to 8 and refiltering brought it
to color 2.
In the factor curve for color 2 the percentile value is practi¬
cally zero, 5800-6850 A; other agents adding no measurable
amount to absorption by water. From 7000 A to 8000 A about
5 per cent is registered in the factor curve. Such a situation is
not unusual in low colored waters. A similar condition, but less
regular would be found in the factor curve for the filtrate of
color 8, if that were continued to 8000 A. The factor curve for
the settled water, color 16, has a marked irregularity at 6000 A.
At this wave-length there is a definite band of selective absorp¬
tion by water, as is shown in the temperature curve of Fig. 7.
Small irregularities are often present at this point in factor
curves but rarely one so well marked as is this. At 7400 A the
curve is little above zero, but rises at greater wave-lengths. The
factor curve for color 83 develops characteristic variations,
6850-8000 A. It is practically a straight-line curve, 6000-6850
James & Birge — Lake Waters and Light
119
A, and an extension of the same line would end at the absorption
observed for 8000 A. Compare the factor curves for the higher
colors in the two following diagrams.
Little Long Lake . Fig. 35
Little Long Lake has been chosen as an example of 2-com¬
ponent analysis of dilutions made from settled water. This water
was examined after it had been in cold storage for 8 months;
during that time its color had fallen from 96 to 74. The form of
the curves indicates that much material still remains in sus¬
pension. Curves are given for the undiluted and settled water
of this lake and for 6 dilutions made from it, extending down to
0.1 per cent, or 1 c.c. to 1 liter of distilled water. The color
rating of the dilutions is roughly parallel to their percentage of
dilution; 50 per cent of lake water and of distilled water gives
about half the color of the undiluted water.
The factor curves for the two lowest dilutions, 0.1 and 1.0 per
cent, show very little evidence of color, although they were rated
as having a trace of color. Absorption at 4078 A is little above
Fig. 85. Little Long Lake. 2-component analysis of absorption in 7
dilutions from the unfiltered water of this lake. The color and the per cent
of dilution are stated on each curve. The absorption curves have the char¬
acters of those from settled waters; compare with those of Lake Mary,
Fig. 26. Note especially the course of factor curves, 7000-8000' A; they
show a general similarity with much variation in detail. The factor curve
for 0.1 per cent dilution shows nearly uniform ordinates, 4078-7000 A; such
absorption is mainly due to suspensoids.
120 Wisconsin Academy of Sciences, Arts, and Letters
that at greater wave-lengths. The course of the factor curves is
roughly parallel, 6000-7000 A, indicating a relatively large effect
of suspensoids. This situation is in sharp contrast to that shown
by Helmet Lake, Fig. 36, in which absorption depends almost
wholly on color. Factor curves, 7000-8000 A, are parallel in a
similar sense, and their variations are smaller than in Helmet
Lake. There is probably present some modification of colloids
by the distilled water used in the dilutions.
Helmet Lake. Fig. 36
The diagram for Helmet Lake gives 8 factor curves for dilu¬
tions made from its filtered water. All curves have the form
characteristic for filtrates, in which most of the effect on light
is due to color and little to suspensoids. Factor curves for color
16 in Helmet and for color 14 in Little Long may be compared.
In Little Long Lake the curve has 75 per cent absorption at
4078 A and 21 per cent at the minimum, 6685 A; for color 16 in
Helmet absorption at 4078 A is 75 per cent and only 4.5 per cent
at 6685 A. The factor curves for the two lowest colors fall to
Fig. 36. Helmet Lake. 2-component analysis of absorption in 8 dilu¬
tions from the filtered water of this lake. This is to be compared with Fig.
35. The curves of Helmet are as characteristic for filtrates as those of
Little Long are for settled waters. Note that factor curves for dilutions
0.1 and 1.0 per cent disappear near 6000 A.; that for 1.0 per cent reappears
at 7000 A; and the other also reappears but is not plotted. Note that a
mean factor curve for colors 104 and 180 could readily be drawn through
the oscillations of the curves.
James & Birge — Lake Waters and Light
121
zero near 6000 A; the curve for color 4 is plotted, 7000-8000 A;
that for color zero would be very similar. All factor curves in
this region have sharp and considerable variations in their
course. Note should be made of the possible extension through
this region of the courses of these curves as shown in the region
6000-7000 A.
Factor Curves , 7000-8000 A
There are few or no large irregularities found in the numer¬
ous series of percentile absorptions given in Tables X-XII ; and
the curves plotted from them run quite as smoothly as could be
expected, whether they represent settled waters, filtrates or dilu¬
tions. The same may be said of factor curves in the spectral
region 3650-7000 A; they show obvious minor irregularities,
some accidental, some the result of local selective absorption;
but none which seriously affect the general run of the curves.
In the region 6000-7000 A, for instance, absorption by filtrates
of low color runs close to that of distilled water, Fig. 15, and
few noteworthy departures appear in their relation.
But this picture changes abruptly in the spectral region 7000-
8000 A. Great irregularities appear when the attempt is made
to distribute to factors the small excess absorption which lake
waters add to that of distilled water. These are conspicuous in
Figs. 31 and 32, which compare remainder and factor curves,
and they are well shown in the numerous diagrams which at¬
tempt to analyze absorption in single lakes. There is a sharp
contrast between the smooth relations of the curves for water
absorption and those for total absorption, and the great varia¬
tions in the course of the factor curves derived from them.
The situation is due in part to the large and rapid changes in
the effect of water on radiation, which come in this region. Ab¬
sorption by water rises rapidly to a mean of some 90 per cent,
7400-8000 A; this fact reduces additions by other factors and
adds to the difficulties of distributing these additions to the
proper factors. The largest addition made in this region by a
settled water is that of Helmet Lake, color 264; this raises ab¬
sorption, 7400-8000 A from a mean of 89.8 per cent to one of
95.0 per cent, and the course of the curves is shown in Fig. 49.
Transmission is thus reduced from 10.2 per cent to 5.0 per cent,
and this reduction calls for a factor absorption of nearly 48 per
cent. In other words, an excess absorption of one per cent in
122 Wisconsin Academy of Sciences , Arts, and Letters
this region demands a factor absorption of about 10 per cent;
and a variation of one per cent between adjacent points of ob¬
servation entails a similar change of some 10 per cent in the
factor curve. Thus the high percentile value of water absorp¬
tion in this region gives a large percentile value to relatively
small variations in factor absorption, whether such variations
are accidental, or are due to local variations in color absorption,
etc.
There may also be present a second cause for this situation in
the long-wave spectrum; one that lies in the relations of radia¬
tion and lake waters. The complex relations existing in this
region between absorption and temperature may easily be asso¬
ciated with variations in absorption quite large enough to appear
conspicuous in factor curves. It is not impossible that there is a
similar local sensitivity to variations in the mixture of color
materials present in the water. No studies have been made of
this relation or of other like matters, but the courses of the
factor curves show their necessity.
In the diagram for Muskellunge Lake, Fig. 41, the water
curve and that for total absorption are closely parallel ; in Fig.
48, Rudolf Lake, they are far apart, especially at 7400 A. A
similar situation appears in Fig. 44 at 7200 and 7400 A and in
Fig. 40 at 7400 A. In Marl Lake, Fig. 48, the two curves ap¬
proach closely at 7400 and 7600 A, with corresponding effect on
the factor curves. Many other such cases appear in the dia¬
grams and in the tables ; but in every case the results could be
smoothed out by small changes in the position of either curve,
changes which are well within the normal and, indeed, inevitable
range of variation between two samples of lake water; or of
distilled water, as can be seen from Fig. 6. But no attempt has
been made to do this, since the irregularities emphasize the spe¬
cial need for detailed study on the complex relations of light and
water in this spectral region. That simple distribution of total
absorption of radiation to three classes of absorbents, which
brings satisfactory tentative results in the region 3650-7000 A,
seems to be inadequate in this last part of the spectrum.
This situation offers to the student of lake physics a group
of interesting questions, which as yet are without answer. Are
these irregularities in absorption due to the intricate relations,
in water, of temperature, wave-length and absorption? Do these
James & Birge — Lake Waters and Light
123
affect the relations of other absorbents, which without them
would be simple? Or is there set up an independent and similar
intricacy when color enters as a factor modifying absorption by
water ?
These are examples of problems in the physics of lakes ; but
it should be noted that they have much less significance in lake
ecology. The quantity of radiation, which is subject to these
irregular variations, is only a small part of the entire energy
spectrum; and a part which is rapidly extinguished by absorp¬
tion effected by water.
Section 3. Suspensoid Curves
8-Component Analysis
The preceding section shows that the total action of any lake
water on light, which is represented by the total absorption
curve, may be analyzed into the effect of two components, repre¬
sented by factor curves. These are the water curve and the
general factor curve; the last showing the combined effect of
color and turbidity.
The ordinates of the general factor curves of a settled water
can also be analyzed and assigned to the several effects of two
components, by methods similar to that employed for total ab¬
sorption. The effect on light of color can often be separated
from that of suspensoids to a degree which is not complete but
which is quite accurate enough to show the ecological situation
and to bring out the differences in the action of these two fac¬
tors. This separation may be made by a comparison of factor
curves from settled and filtered waters along five different lines.
1. The action on light of settled and filtered waters may be compared
in waters whose color is not altered by filtration.
2. In waters whose color is reduced by filtration, the ordinates observed
for percentile transmission of the filtered water may be re-com¬
puted for the color of the settled water. Color curve and general
factor curve may then be compared as in Case 1, above.
3. Settled and filtered waters may be compared, which are of the same
color but from different lakes.
4. Mean results may be compared in settled waters and filtered waters
of the same color.
5. None of these methods can be applied to waters which have a
definite color when settled, but whose filtrates have the color 0 or
124 Wisconsin Academy of Sciences , Arts, and Letters
trace, or even 1 or 2. In such cases comparison may be made be¬
tween the settled water and mean results for filtrates of the same
color.
In 2-component analysis the percentile ordinates for total
transmission are divided by those for pure water in order to
obtain the ordinates for the factor curve. In 3-component analy¬
sis the ordinates for percentile transmission of the two factor
curves, for a water settled and filtered, are employed in the same
manner. Percentile ordinates for transmission in the factor
curve for a settled water are divided by the corresponding ordi¬
nates in the factor curve for the filtrate of the same water ; the
Fig. 37. Nagawicka Lake. 3h-component analysis of absorption. Figs.
37-52 are a series of diagrams of 3-component analysis of absorption in
lake waters; the general plan of the diagram and its lettering are the same
in all. T=r curve for total action of lake water; W= similar curve for dis¬
tilled water; the ordinates for both are observed and are found in the tables.
F= general factor curve, the same as that found in preceding diagrams.
This is not one of the 3 factor curves employed in 3-component analysis;
it is used in computing the suspensoid curve, as stated in text p. 125. In
these diagrams it is usually shown in whole or in part. C = color curve. This
is the factor curve for the filtered water, corresponding to curves shown
for filtrates in Figs. 33 and 34. S=f actor curve for suspensoids, com¬
puted as stated in text, p. 125. Color is stated on T and also on C if this
differs from that of T.
At any wave-length the product of the percentile transmission ordinates
of F, and W is equal to that of T. The transmission ordinates of F are
computed from those of W and T; those of S come in a similar way from
C and F. See Table XIV.
James & Birge — Lake Waters and Light 125
quotients are the ordinates for transmission in the suspensoid
curve.
In this process the assumption is necessarily made that the
values of percentile transmissions of the filtrate or color curve
may be taken as accurate, just as those for pure water are so
regarded in the first analysis. But this assumption is never com¬
pletely true; in all cases the curve for absorption by filtrates
represents the effect of some suspensoid absorption in addition
to that by color. In many waters this addition is so small that
it does not vitiate results ; but in other cases it renders compari¬
sons impracticable. A second necessary assumption is that filtra¬
tion does not appreciably alter the composition or mixture of the
color materials; at best this is only approximately correct, and
in some cases filtration so changes color composition as to pre¬
vent comparisons.
The simplest cases for this analysis are those from waters
whose color is not altered by filtration through a Berkefeld filter.
In such waters practically all of the color is found in solution
and passes through the filter, while almost all of the suspensoids
are retained. This situation is present in only a few lakes ; nine
waters are included in the list, and their color ranges from zero
to 12. Besides these, there were two high-colored waters whose
color did not change on filtration; Helmet Lake (color 264) and
Turtle Lake with color 43.
The water of Nagawicka Lake is selected to show in some de¬
tail the method and the results of this analysis. Fig. 37 repre¬
sents the situation ; it contains five curves, all but the first being
factor curves.
1. The curve of total absorption (T) the same as that shown for this
lake in Fig. 23.
2. The curve of water absorption (W).
3. The general factor curve (F) computed from 1 and 2.
4. The curve for color absorption (C). This is the general factor
curve for the filtrate of this water; computed in the same way as
the factor curve for total absorption.
5. The curve for suspensoid absorption (S), computed from 3 and 4.
In computations for this diagram, as in all of this type, the
factor effect of water is taken as a constant. The general per-
126 Wisconsin Academy of Sciences, Arts, and Letters
Table XIV
Nagawicka Lake. Data for factor curves as shown in Fig. 37.
Note. — Table XIV states the method of computing factor curves for
Nagawicka Lake. All data in the table are percentile ordinates of trans¬
mission of radiation by one meter of the lake water. Column A gives such
ordinates for distilled water; B, settled water of the Lake. Table X, No. 28;
C, general factor curve, derived by dividing data of B by those of A; D,
filtered water of the Lake, Table XI, No. 49; E, factor curve of filtrate,
taken as color curve; F, suspensoid curve, derived by dividing data of C
by those of E.
In Fig. 37 are shown five curves plotted from the data in this Table.
These are the curve for total action (T), which comes from the ordinates
of B; the water curve (W) from data in A; the general factor curve (F),
from C; the color curve (C), from E; the suspensoid curve (S), from F.
See Fig. 37 and text for the method of using these factor curves.
centile factor curve (F) is computed in the same way as those
of the preceding section of this Chapter, so that the product of
the transmission ordinate at any wave-length and the transmis¬
sion of water is the ordinate observed in the curve for total
transmission. Thus in Fig. 37 the water curve (W) has a trans¬
mission ordinate at 6000 A of 81 per cent; that of the total
James & Birge — Lake Waters and Light
127
curve (T) is 63 per cent; dividing the second by the first gives
the ordinate for factor curve (F) about 78 per cent.
The curve for color absorption (C) is the factor curve for the
filtrate, derived from water and total for filtrate just as that
for factor of settled water was computed. It is designated as
the color curve because there is evidence that there is very little
absorption by suspensoids in the filtrate. The percentile trans¬
mission ordinates for the suspensoid curve are derived from the
ordinates of this curve and those of the general factor curve.
The transmission ordinates of the factor curve are divided by
those of the color curve and quotients are the ordinates of the
suspensoid curve; such that color multiplied by suspensoid
equals general factor.
There may then be determined for any lake water, three per¬
centile transmission curves such that the product of their ordi¬
nates, at any point in the spectrum equals the ordinate of the
curve of transmission for settled water at the same wave-length.
Thus in the water of Nagawicka Lake, the percentile transmis¬
sion of the settled water at 6000 A is 63 per cent, equivalent to
absorption of 37 per cent, stated in round numbers. At the same
wave-length, transmission of pure water is 81 per cent; that of
color is 94 per cent; and that of suspensoids is 83. The product
of the three factors yields 63 per cent, its value in the settled
water. Similar results will be found at any wave-length of the
spectrum.
Summing up the results of these processes, it may be said
that two series of determinations of spectral absorption were
made on the water of Nagawicka Lake, one on the settled water
and one on the filtrate, each having the color 12. From the per¬
centile transmission ordinates of distilled water and those of
the settled water have been computed those of the general factor
curve . The factor curve of the filtrate has been taken as the
color curve ; and from its transmission ordinates and those of
general factor curve have been computed those of the suspen¬
soid curve. These assign the percentile transmission of radia¬
tion by a meter of the settled water of the lake, to the combined
action of three factors. The quantitative accuracy of this assign¬
ment is sufficient for use in the further investigation of the
effect of light as a factor in the ecology of the lake.
128 Wisconsin Academy of Sciences, Arts, and Letters
Additive Curves, Fig. 38
Factor curves, including among them the absorption curve
for pure water, show percentile absorption of radiation as it
w^ould be effected by the agent or group of agents acting by itself.
But when three groups of agents are acting together in one lake
water, each reduces the results which may be effected by the
others. If therefore, factor curves are treated as if they were
absorption curves their combined effect is greater than that rep¬
resented by the curve of total absorption. This may be easily
seen in Fig. 37 ; if the areas are measured which lie below the
curves marked W, C and S, their sum will be about 12 per cent
greater than the area below the curve marked T, that of total
absorption.
But if the result is desired, it is easy to convert these per¬
centile transmission or factor curves into corresponding curves
of absorption, under the assumption that the factor ordinates for
Fig. 38. Nagawicka Lake. Transformation of factor curves into ad¬
ditive curves. If factor curves are regarded as absorption curves, they
show the effect which each factor would have on radiation if it were oper¬
ating independently of the others. Additive curves show its effect as one of
the group of factors that are operating simultaneously in the lake. For
method of computation see p. 128. The sum of the absorption ordinates of
C, S and W at any wave-length is equal to the ordinate of T. In securing
this result each factor ordinate is reduced by the same percentage. The
additive curve is therefore a percentile absorption curve with the same
general form as the factor curve, which is computed as a percentile trans¬
mission curve.
James & Birge—Lake Waters and Light
129
each factor curve are to be reduced by the same percentage. If
these factor curves are taken as percentile absorption curves, the
sum of the ordinates of the factors at any wave-length is always
greater than the percentage observed in the curve for total ab¬
sorption of the settled water. In Nagawicka Lake, Fig. 37, the
factor curve at 7000 A shows absorption of 15 per cent and that
of water is 45 per cent. The sum of the two ordinates is 60 per
cent, while the curve for the settled water shows absorption of
53 per cent. This is 88.3 per cent of the sum of the two factor
curves. If absorption in factor curves is reduced to 88.3 per
cent of its value, the products are, water 39.7, other factors 13.3
per cent ; the two adding to 53 per cent, the value of absorption
in the settled water.
The same process may be applied to the factor curves for
color and suspensoids. At 4078 A the general factor curve, thus
recomputed, has an absorption of 82 per cent ; that of the factor
curve for color is 76 and for suspensoids 32; their sum is 108;
the percentage is the factor curve 82, or 78 per cent of this sum.
Reducing absorption by color and suspensoids to 78 per cent of
their factor value, gives them 59 and 23 per cent, whose sum is
82 per cent. Adding to this about 2 per cent for water absorp¬
tion at 4078 A, yields a total of 84 per cent, the absorption in the
settled water.
In this way may be derived ordinates which will convert any
factor curve into a corresponding absorption curve. In Fig. 38
are shown the changes which are thus made in the curves for
3-component analysis of Nagawicka Lake. The absorption ordin¬
ates of the factor curves are reduced by the same percentage and
yield new curves which are approximately parallel to the factor
curves, but at a lower level of absorption. The sum of the ab¬
sorption ordinates for W, C and S at any wave-length equals the
corresponding ordinate of T; and the sum of the areas below
the curves W, C and S is equal to that below T. This subject will
be further discussed in Part II of the report.
B. Single Waters, 3-Component Analysis
Figs. 39-52
A series of 14 diagrams is given to illustrate 3-component
analysis of absorption, as found in waters that have a wide
130 Wisconsin Academy of Sciences , Arts, and Letters
range of color and turbidity. After the detailed account of Naga-
wicka Lake, only brief discussions are needed for these waters ;
they call attention to the more striking points in the diagrams.
The factor curves are given, and are taken as absorption
curves, showing the effect of water, color and suspensoids, as
each would operate if its action were not affected by that of the
others; they do not represent additive curves, like those given
for Nagawicka in Fig. 38. The diagrams are left in their pres¬
ent form because each new stage of computation removes results
further from the original data, and in this case without corre¬
sponding increase of knowledge.
Crystal Lake . Fig. 39
Three diagrams show 3-component analysis in waters whose
color was not altered by filtration. These are Muskellunge, Marl
and Crystal lakes ; and the present diagram represents the first
sample of water from Crystal Lake. This was the only lake
Fig. 39. Crystal Lake. 3-component analysis. Curves and lettering
as in Fig. 37. The diagram is from water sample No. 1, as given in Tables
X and XI. Curves C, S and W are factor curves; the product of their per¬
centile transmission ordinates at any wave-length is equal to percentile
transmission ordinate of curve T. Note relatively large effect of suspensoids
and compare with Fig. 41. The general factor curve is not included in the
diagram.
Sample No. 1 from Crystal Lake had a relatively large color absorp¬
tion at 4078 A. If the diagram had been based on sample No. 3, the factor
curve for color would have resembled that in the present diagram at 6000-
6800 A. Curve for total absorption of this filtrate is given in Fig. 20.
James & Birge — Lake Waters and Light 131
whose settled water was rated as zero in color ; the filtrate had
the same rating, but the monochromator readings show that
color was present ; though too little to register on the eye. Table
XI shows that this has always been the case with filtrates whose
rating was zero, except in the third sample from Crystal Lake.
Absorption in the color curves, as given in Fig. 39, is practi¬
cally zero, 6000-6685 A; elsewhere there is a small absorption,
probably indicating the presence of a little colloid in^the filtrate.
The large absorption band at 7200 A is noteworthy ; it was pres¬
ent in all three specimens of water from Crystal Lake and it was
more strongly marked than in many lakes of higher color.
Ordinates for the suspensoid curve show little variation in
the course of the spectrum ; they are greater than in other lakes
of higher color, like Muskellunge Lake, Fig. 41. Part of this
difference was due to the longer time of settling, which Muskel¬
lunge Lake had. Suspensoid absorption in Crystal is about 18
per cent at 4078 A, declining to about 12 per cent at 7000 A;
this course of the curve perhaps indicates traces of color re¬
maining in the suspensoids. The selective effect of suspensoids
in refraction and dispersion, in the short-wave spectrum, has
not been investigated ; but it seems that it must be small.
Mean Results, Colors U, 6, 10, 20
Figs. UO, If 2 , lf.lt> ,lf5
Four diagrams illustrate 3-component absorption as found
in mean results from waters rated as of the same color; their
range covers the colors of most of the larger and deeper lakes
of Wisconsin. The means are derived from Tables X and XI,
and the number of lakes included in each mean is from 5 to 8,
depending on the number rated as of the same color or near to
it. In all cases the absorption of the single lakes is well within
the range for lakes of the assigned color.
Tables X and XI show that very few lakes appear in the
same color group both as a settled and as a filtered water ; only
such lakes can thus appear whose color is not altered by filtra¬
tion; even where the same lake is present in both groups the
water may come from different samples. The diagrams there¬
fore afford a fair test of the objectivity of the color rating.
The results are in general as good as those from single lakes,
if they are judged by the non-selectivity of the suspensoid curve.
132 Wisconsin Academy of Sciences, Arts , and Letters
Fig. 40. Settled waters of mean color 4. ^component analysis. Curve T
is the mean of waters 4-8 in Table X; curve C is computed from mean of
waters 25-31 in Table XI. Curves and lettering as in Fig. 37. Note course
and level of suspensoid curve; also that of color curve, 6000-7000 A. Com¬
pare curves T and W at 7400 A and note effect on S.
That for color 4, Fig. 40, lies close to the 20 per cent absorption
line, 4078-7000 A, and much the same may be said of those for
colors 10 and 20. This would seem to represent the factor effect
of suspensoids in lakes of the type included in the means. The
suspensoid curve for color 6, Fig. 42, evidently carried some
color, which raised absorption, 4000-4500 A, to 40 per cent. The
same is true for the waters of Marl Lake, Fig. 43, which is in¬
cluded in this mean, both as a settled and as a filtered water.
In the diagrams for colors 6, 10, 20, the use of means has
smoothed out many of the irregularities in the region 7000-8000
A ; but this is not the case in color 4, and in color 20 the curves
for color and suspensoids cross three times. The general factor
curve is shown in Figs. 42, 44, 45 ; it is omitted in Fig. 40, since
the curves are more crowded.
Muskellunge and Marl Lakes, Figs. 41, 43
Muskellunge and Marl lakes are selected to illustrate 3-com¬
ponent analysis for waters of color 6. The waters of this color
have been used in Fig. 12 as showing the widest variation found
in waters of the same color rating ; and these lakes represent the
extremes of the group. Color was not altered in either lake by
James & Birge—Lake Waters and Light
133
Fig. 41. Muskellunge Lake. 3-component analysis. This water was
in cold storage for a year and nearly all suspensoids had settled out; hence
small effect shown by curve S. Color curve C shows a definite action on
light, 6000-8000 A, probably due to fine suspensoids. that passed the filter.
This effect is often found. Note the small and regular addition which other
agents add to that of water, 7400-8000 A.
filtration; Muskellunge had been nearly a year in cold storage
before examination, and probably color had been extracted from
suspensoids during that time. The water had an exceptionally
low absorption through the short-wave spectrum ; Table X shows
that very few of the 19 settled waters with color from 0 to 6
inclusive, have at any readings an absorption lower than that of
Muskellunge. Lake Mendota, also of color 6, comes nearest to
it. The ordinates for total absorption are smaller than those of
Crystal Lake, Fig. 39, due to the smaller amount of suspensoids.
The diagram for Muskellunge is placed next to that for Mean
Color 4 in order to show the agreement between the color curves
of the two diagrams, and also the much greater suspensoid ab¬
sorption in the water of the lower color.
Marl Lake, Fig. 43, offers a striking contrast to Muskellunge.
The diagram is placed next to that for Mean Color 6, in order to
bring out the fact that this contrast is primarily due to sus¬
pensoids. The character of Marl Lake has been described, p.
46, and its water had settled one week before examination. The
color curve is almost identical with that of the Mean, 4078-6000
A, it shows more absorption, 6000-7000 A, indicating that some
of the finest particles of marl passed through the filter.
134 Wisconsin Academy of Sciences , Arts, and Letters
Fig. 42. Settled waters of mean color 6. 3-component analysis. Curve T
is the mean from waters 12-19 Table X; curve C is computed from waters
35-39 Table XI. Curves and lettering as in Fig. 37. The course of curve S
shows higher absorption in the short-wave spectrum; probably due to some
modification of color materials by filter. The general factor curve is given.
The transmission ordinates for curve S are derived from those of F and C.
Fig. 43. Marl Lake. 3-component analysis. Curves and lettering as
in Fig. 37. M)arl Lake has a water of lowT color and with maximum amount
of suspensoids; just as Muskellunge Fig. 41, has the same color but with
minimum effect of suspensoids. In Muskellunge suspensoids alone would
permit an average transmission of more than 90 per cent of incident
radiation; in Marl they wouid pass only about 60 per cent. Color in Marl
has almost the average action for color 6, Fig. 42; while that of suspen¬
soids is much greater. Compare diagram also with Fig. 47, showing simi¬
lar effect of suspensoids.
James & Birge — Lake Waters and Light
135
Fig. 44. Settled waters, mean color 10. 3-component analysis. Curve
T is the mean from waters 23-28, Table X ; curve C is computed from waters
42-48, Table XI. Curves and lettering as in Fig. 3i7. Compare curves with
those of Figs. 40, 4-2, 45, and note effect of increasing color.
4000 A 5000 6000 7000 8QOO
Fig. 45'. Settled waters, mean color 20. 3-component analysis of ab¬
sorption. Curve T is mean from waters 31-36, Table X; curve C is com¬
puted from waters 52-56, Table XI. Curves and lettering as in Fig. 37.
Note that position of curve S indicates that suspensoids have less effect on
light than they have in waters of low color. There is no necessary relation
between color-rating and amount of suspensoids in any water. Compare
with Figs. 46, 47.
136 Wisconsin Academy of Sciences, Arts, and Letters
Suspensoid absorption in Marl Lake is the greatest found in
any water of low color ; it is not due to organic particles but to
inorganic particles of marl. Absorption is about 52 per cent at
4078 A, and 32 per cent at 7000 A; indicating by this decline
the presence of some color with the suspensoids. The mean for
suspensoids alone would probably be 40 per cent or something
less.
One effect of this large quantity of suspensoids is to make the
course of the total absorption curve less concave in the short¬
wave spectrum. Comparison with Mean Color 6, Fig. 42, shows
that absorption at 4078 A is some 10 per cent higher in Marl
Lake; at 5000 A the excess is about 17 per cent, and is 15 per
cent at the minimum, 6685 A. The situation is further empha¬
sized by comparison with other waters, both of similar and of
higher color. Compare the total absorption curve for Marl Lake
with that for Nagawicka Lake color 12, Fig. 37 ; with that for
Mean Color 20, Fig. 45 ; or even that for Rudolf Lake, color 50,
Fig. 48 ; these show contrast. On the other hand that for Ade¬
laide Lake, color 28, Fig. 47, indicates a similar effect of a large
amount of suspensoids in a water of higher color.
Lake George, Fig. U6
Lake George, with its color rating of 24, is an excellent ex¬
ample of a water in which absorption is due mainly to color and
water, suspensoids having only a minor effect. The curves for
total and for color are close together in the short-wave spectrum,
and total absorption is unusually small in the middle spectrum.
The pronounced U-shape of the total curve is another indication
of the small effect of suspensoids.
The color of the filtered water was 16 ; the transmission ordi¬
nates were recomputed for color 24, as is stated on p. 112, so as
to bring the color curve up to that of the settled water. Filtra¬
tion removed almost all of the suspensoids, as is shown by the
low absorption at 7000 A. Filtration also made little or no
change in the composition of the mixture of color materials, as
is indicated by the uniform course of the suspensoid curve
through the spectrum.
Suspensoids had an unusually small action, considering the
high color of the water. The mean percentage is about 14, in
James & Birge — Lake Waters and Light
137
Fig. 46. Lake George. 3-component analysis of absorption of light.
Curves and lettering as in Fig. 37. The settled water was rated as color 24 ;
that of the filtrate was 16. See Tables X, XI. The color curve of the fil¬
trate was recomputed for color 24, as stated for Fig. 47. The water had
been in cold storage for 9 months; color was not reduced by storage, but
most suspensoids had settled out, as in Muskellunge, Fig. 41. In this matter
George differed from Little Papoose, Fig. 19, and Horsehead, Fig. 34;
lakes in which color of water and form of filtrate curve were greatly
modified by settling.
the short-wave half of the spectrum, declining to means of about
11 and 9 per cent in the other half.
Lake Adelaide. Fig. U7'
The diagram for Lake Adelaide, color 28, shows that this
water holds much the same relation to that of Lake George as
the water of Marl Lake has to that of Muskellunge Lake; sus-
pensoid absorption is much greater. The form of the curve for
total absorption, as well as its position, indicates a large amount
of suspensoids. The same fact is indicated by the wide spread,
4078-6000 A, between the color curve and that for total absorp¬
tion, as compared with their relation in Lake George.
The color rating of the filtered water was 18; this was re¬
computed for color 27, a sufficiently close match for the color 28
found in the settled water. The curve for suspensoids is evi¬
dently affected by color, as was that of Marl Lake. The mean
value for suspensoids alone would probably not exceed 50 per
cent absorption. It is primarily an effect due to organic sus-
138 Wisconsin Academy of Sciences, Arts, and Letters
4000 A 5000 6000 7000 8000
Fig. 47. Adelaide Lake. 8-component analysis of absorption of light.
Curves and lettering as in Fig. 37. The water of Adelaide is related to that
of George much as that of Marl and of Muskellunge are related. Its color
is somewhat higher and color absorption is greater, but the main difference
between the waters is in the effect of suspensoids. The slope of curve S
shows that some action of color accompanies that of suspensoids. Suspen¬
soids in Adelaide are mainly organic; in MJarl they are mainly inorganic,
as its name implies. Filtration of Adelaide reduced color to 18; it was
recomputed for 27, by extracting the square root of the percentile trans¬
mission found at each point of observation, and then multiplying the two
factors together.
pensoids, which are producing about the same effect on light as
that of the inorganic particles in the water of Marl Lake.
Rudolf Lake . Fig. J>8
The settled water of Rudolf Lake, color 50, is compared with
the filtered water at color 48, a sufficiently close match. Filtra¬
tion reduced color to 24, and ordinates for the color curve were
recomputed for twice that rating. The even course of the sus-
pensoid curve shows that filtration made little change in the
composition of color materials.
In this water total absorption is greater than in that of Lake
George and less than in Lake Adelaide; in both of these lakes
color rating was about one-half that of Rudolf Lake. The dif¬
ference between Rudolf and George is mainly due to greater ab¬
sorption by color; this is indicated by the position and by the
form of the color curve. At 5000 A the color in the water of
George absorbs about 50 per cent of incident radiation and that
James & Birge — Lake Waters and Light
139
Fig. 48, Rudolph Lake. 3-component analysis of absorption of light.
Curves and lettering as in Fig. 37. Filtration reduced color-rating to 24;
the color curve was recomputed for color 48. Note the separation of curves
T and W at 7400 A, including a band of local selective absorption at that
wave-length. Figs. 37-48 form a series, showing 3-component analysis of
light absorption in waters whose color rating ranges from 0 to 50, and
therefore covers most of the larger and deeper lakes.
in Rudolf absorbs about 80 per cent. The color curve in Rudolf
is beginning to be concave downward at wave-lengths less than
than 5000 A; a condition necessarily associated with higher
colors, when plotted on the percentage basis, and one which be¬
comes more marked as color increases.
The suspensoid curve showed a very great apparent absorp¬
tion at 4480 A, the shortest wave-length for which it could be
computed. Its position is marked by a dot in the diagram. This
situation not infrequently appears when the attempt is made to
assign to several factors small differences between high per¬
centages of absorption. A very small error or a small alteration
in color composition makes a great percentile change in sus¬
pensoid absorption. It will be noticed that local ups and downs
in the suspensoid curve are frequent and noticeable in Rudolf
Lake. This characteristic is likely to increase with the color of
the water.
Factor curves have been computed for waters of higher
colors, and with success ; but they are likely to show irregular-
140 Wisconsin Academy of Sciences, Arts, and Letters
ities due either to causes like those just mentioned or to the
effects of computation in magnifying the inevitable accidental
errors in the color curve, which is assumed as a standard when
suspensoid absorption is computed. In Oxbow Lake, for instance
color must be computed up from filtered color 10 to that of 65 in
the settled water ; and this was found to yield a low absorption
at certain wave-lengths. So with Tadpole Lake, whose settled
color is 130 while that of the filtrate is 46 ; here the percentile
transmission for color is the cube of that observed; and with
this decrease of transmission ordinates come marked Irregular¬
ities. The same situation is present with the water of Mary
Lake, color 109, filtered color 34. If such waters are to be thus
analyzed, careful and detailed study must be made in each case.
Helmet Lake . Fig . 49
The water of Helmet Lake had the highest color found in the
present study; in two samples the color was not altered by fil¬
tration. In a third sample it was reduced from 236 in the set¬
tled water to 212 in the filtrate. This small reduction of color
was accompanied by a large increase of transmission of radia¬
tion, much larger than could be attributed to the change in color.
Fig. 49. Helmet Lake. 3-component analysis of absorption of light.
Curves and lettering as in Fig. 37. In one sample of water from Helmet
Lake filtering reduced color from 236 to 212 and made a noteworthy in¬
crease in transmission of light. Fig. 49 is given to show that the methods
used in analysing absorption at lower colors apply also in this extreme
case.
James & Birge — Lake Waters and Light
141
From these curves therefore has been computed a 3-component
analysis of the action of the water on light. Its object is rather
to show possibilities of applying these methods in an extreme
case, than to present exact quantitative results.
The factor curves for settled and filtered waters are directly
compared ; no attempt is made to correct the color curve for the
small reduction in color. Fig. 49 shows the result; the suspen-
soid curve is quite as good as could be anticipated. There was
in it a marked notch at 7200 A, which has been smoothed out.
The absorption indicated by the suspensoid curve is about 60 per
cent, 5800-7000 A, with the usual changes in the rest of the
spectrum. v
In a water of so high color the factor curves necessarily indi¬
cate a great absorption, much greater than would be found if
their ordinates are reduced so as to indicate the contribution
which each curve makes to the total actual absorption, as is done
for Nagawicka Lake in Fig. 37. Absorption ordinates, 6000-
7000 A, will be reduced to about 70 per cent of their indicated
percentage, and those in the remainder of the spectrum must be
reduced even more; at 7500 A to 60 per cent of that indicated
for factor curves.
The remaining diagrams of this section illustrate methods of
securing approximate determinations of suspensoid absorption
in waters in which the regular absorption curves can not be
used as they have been in the cases already described. No very
significant value need to be attached to these diagrams, as fur¬
nishing a quantitative account of what is going on in the waters.
They are examples of methods of testing the objectivity of color
ratings and the similarity of color effects on radiation in waters
whose rating is alike. They indicate the regular presence of a
non-selective factor in the total absorption of radiation which is
effected by the several waters ; and they also indicate that this
factor is present in definite and measurable amounts.
Use of Mean Color Curve. Fig. 50
Columbian Lake
Columbian Lake is one of the waters shown in Fig. 21. The
settled water has the color 6 and that of the filtrate is rated as
trace. Recomputation of the color curve, as was done in the
142 Wisconsin Academy of Sciences, Arts, and Letters
4000 A 5000 6000 7000 8000
Fig. 50. Columbian Lake. Use of mean color curve in 3-component an¬
alysis. Curves and lettering as in Fig. 37. The color of lake waters of
Groups II and III is associated with suspensoids and colloids and is often
reduced to zero or trace by filtration. Color curves of filtrates of such
low colors can not be directly recomputed for the higher color of the settled
water. The water of Columbian Lake filtered from color 6 to trace. The
attempt is made to secure a 3-component analysis, using the color curve
for mean color 6, Fig. 42. The result is satisfactory; but the method can
be applied only when it yields a non-selective suspensoid curve.
Note the marked band of local absorption near 6000 A, and compare
with that found in Horsehead Lake, color 16, Fig. 34.
case of Lake George and other water, is not possible when the
color rating of the filtrate is so small. The experiment was tried
of comparing the curve of total absorption in the settled water
of Columbian Lake with the mean curve for color 6; the same
as that shown in Fig. 41. The result is a very even suspensoid
curve, at about the right percentile level for waters of color 6.
The curve has an unusual feature in the apparent presence of
a marked band of selective absorption at 6000 A. Small indi¬
cations of such a band are often found, but one so strongly
marked is uncommon. An excellent example may be seen in
Fig. 34, Horsehead Lake; it is in the curve for settled water,
color 16.
Comparison of Two Waters of the Same Color. Fig. 51
Helmet and Mary Lakes
It is obviously impossible to work out the presence of sus¬
pensoid absorption in the dilutions from high colored waters as
James & Birge—Lake Waters and Light
143
4000 A 5000 6000 7000 8000
Fig* 51. Helmet and Mary lakes. Use of two waters in 3-component
analysis. The 5 per cent dilution of the filtered water of Helmet had a
color rating of 14 ; that of the 10 per cent dilution of the unfiltered water
of Mary had the same color; see Figs. 25, 26. The water of Mary had also
a higher absorption. The factor curve of Miary may be treated like a factor
curve from a settled water and that of Helmet as a color curve from the
same water. The resulting curve shows that the greater absorption of the
water of Mary is due to a non-selective factor, a suspensoid curve.
this can be done for undiluted waters. But the presence among
the dilutions of those from the filtered water of Helmet Lake,
enables the student to make comparisons with the dilutions from
unfiltered waters.
Table XII shows that the 5 per cent dilution of Helmet Lake
has the color 14, and that the 10 per cent dilution of the unfil¬
tered water of Lake Mary has the same rating. Absorption in
the unfiltered water is higher than that in the filtrate. The
original color of the water of Helmet was 264 and that of Mary
was 109.
The factor curves for the two waters are compared in Fig.
51: the effect of water on radiation has been eliminated; the
curve for Mary is taken as the general factor of a settled water
and that of Helmet as the color curve of a filtrate, both waters
having color 14. Computation shows that the suspensoid curve
thus obtained is a normal one, with absorption values of 20 to
30 per cent. The diagram shows the presence of a non-selective
factor which accounts for the higher absorption in Mary; and
it appears that the effect on radiation of color 14 is about the
same in the two waters.
144 Wisconsin Academy of Sciences, Arts, and Letters
100 r - T - - - 1 - 1 - -r - . - r - , . . I - 1 - r - -T - - - - - - - - -
80 - = - -
*0— \N— ' -
40 - - - - -
X^VSVn,^C0LL0ID 10
20 - [_^> , - - - -
% ■ " — -L ^
4000 5000 6000 7000 8000
Fig. 52. Curves from dissolved and colloid colors in filtrates compared
in 3-component analysis. Some filtrates show a high absorption, apparently
due to colloids associated with their color. A group of such waters of
color 10 is compared with a similar group of waters whose color is dissolved.
The result shows the presence of a non-selective factor, which adds to ab¬
sorption but does not raise the color-rating.
Dissolved and Colloid Colors. Fig. 52
Frequent reference has been made in this report to colors
whose materials are in solution, while other colors are attached
to or connected with colloids. No sharp or clear distinction can
be made between these types of color, which necessarily grade
into each other. A selection has been made of colors in filtered
waters which seem to be especially good representatives of the
two types, and mean color curves for them are plotted in Fig. 52.
The colloid colors show the higher absorption. Computation of
a third factor curve to account for the difference shows that this
curve has a non-selective character. It apparently represents
the action of the colloids, which add to absorption by the waters
but do not increase their color.
Percentile Absorption and Transmission
Curves for Mean Colors
Fig. 53 gives mean percentile curves for color absorption
derived from the data for filtered waters contained in Table XI ;
most of the mean curves for total absorption are shown in Fig.
15; those for colors higher than 46 are from the dilutions of
Helmet Lake, given in Fig. 25. Curves for colors 4, 6, 10, 20 are
James & Birge — Lake Waters and Light
145
Fig. 53. Curves showing action on light of color alone, in spectral
region 4000-7000 A. Curves for colors 0-30 are mean results from data in
Table XI; those for colors 4, 6, 10, 20 are repeated from Figs. 40, 42, 44,
45. Curves for higher colors are from single waters. Color 46 is from Tad¬
pole Lake; those for colors 55-180 are from dilutions of Helmet; compare
Fig. 25. Note that all curves are general factor curves for filtrates of the
given color; see p. 122. The form of curves approaches a straight line
near color 30 ; and at higher colors its concavity is reversed. This becomes
more strongly marked as color increases.
found in Figs. 40, 42, 44, 45. Curves are carried to 7000 A only
because of irregularity of results between 7000 A and 8000 A.
The diagram shows curves of the typical forms; those for low
and medium colors have an upward-facing concavity; they be¬
come nearly straight line curves between colors 46 and 55 ; those
for still higher colors have a downward-facing concavity.
The curves of this diagram are intended to show the absorp¬
tive effect of color alone as found in a meter of water, that of
water and of suspensoids are eliminated. They furnish a pic¬
ture whose general effect is correct, but in which details are
subject to much variation and are by no means accurate. Espe¬
cial attention should be given to the numerous irregularities
shown by the curves in the spectral region 6000-7000 A. Those
for colors 0, 4, 6 cease to add to the absorptive effect of water
before they reach 7000 A; the general course of the curves for
146 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 54. Percentile transmission curves for effect of color, at 7 wave¬
lengths 4000-7000 A. Ordinates are percentile transmissions; abscissas
are the number giving the color rating. At color zero there is no color ab¬
sorption for any wave-length; that is, there is 100 per cent transmission.
Color acts most strongly on the short-wave spectrum; for 4000 A trans¬
mission becomes zero near color 30 ; that is, absorption is 100 per cent, as is
shown in Fig. 53. At 4500 A transmission does not become zero until
color 75 is reached. For colors above 80 there are few cases and percentages
are correspondingly uncertain.
colors 10 and 20 would indicate a similar termination. Fig. 48
gives a curve for color 48 in Rudolf Lake ; it has been computed
from observed color 24. It should be compared with the curve
for color 46, from Tadpole Lake, in the present diagram. Simi¬
lar comparisons will bring out the kind of agreement and differ¬
ence found in waters of like color rating.
In Fig. 54 the data for Fig. 53 have been used to prepare
transmission curves for 7 wave-lengths of the spectrum, from
4000 A to 7000 A. Abscissas represent color rating and ordi¬
nates represent percentile transmission. For colors of 80 and
below the results have a fairly adequate basis in observation;
beyond that color, cases are so few that the curves are little
better than estimates.
James & Birge—Lake Waters and Light
147
At color 0 there is 100 per cent transmission, or zero absorp¬
tion, for all wave-lengths. Color acts most efficiently on the
short-wave spectrum. Transmission is zero near color 30 for
radiation of 4000 A; that for 4500 A reaches zero between
colors 70 and 80; radiation of 5000 A continues to pass until
colors are reached well beyond 125, but no great dependence can
be given to results that call for such high colors.
SUMMARIES
Introduction
1. This paper is Part I of a report on the absorption and
transmission of light by distilled water and by the waters of
Wisconsin lakes. The present Part deals with percentile ab¬
sorption, as determined at numerous wave-lengths of the spec¬
trum; and it gives numerous diagrams plotted from these data.
Part II, now ready for the press, will apply these determina¬
tions to a solar energy curve and will discuss the changes which
lake waters effect in quantity of energy and in the form and
composition of the energy spectrum. It will also present the
transmission through lake waters of total incident solar energy
and that of the spectral color bands.
2. The data on which this report on lake waters is based are
contained in Chapter III. Tables X, XI, XII state percentile
absorption of radiation, as determined at 21 or 22 wave-lengths
of the visible spectrum, by a stratum of water one meter in
thickness. Each series of determinations covers the spectrum
from 4078 A to 8000 A, and in some cases is extended to 3650 A.
The tables contain 171 such series made on waters from 49 lakes.
They were examined as settled waters, as filtrates, and as dilu¬
tions.
3. The report exhibits the results of these observations pri¬
marily through diagrams containing percentile absorption
curves. The ordinates of these curves are percentile absorptions
and the abscissas are wave-lengths of the visible spectrum. The
general story of the report may be gathered from the diagrams
which accompany each chapter.
Chapter I
1. A full account is given of the construction and operation
of a monochromator, designed for determination of absorption
148
James & Birge—Lake Waters and Light
149
of light by lake waters, in the spectral region 3650-8000 A.
2. The methods are stated of observing, and of calculating
and presenting the results of the observations.
Chapter II
1. Apparatus and methods are described for the examination
of distilled water. Special methods are described for the prep¬
aration of the water and for the preservation of its purity.
2. Table I shows the results, stated both as coefficients of
extinction and as percentages of absorption. The minimum ab¬
sorption found was 0.49 per cent at 4730 A; at 3650 A it was
3.65 per cent. These determinations were made from water
contained in tubes which were silver plated inside. Absorption
was greater by water in tubes coated inside with paraffin or
ceresin. See Table X for determinations from water in such
containers.
3. The results reached in the present report are compared in
detail with those reported by 9 other observers. Table II states
the percentile absorption as found for each 100 A of the visible
spectrum. The comparison shows wide variations in determina¬
tions made by equally good observers.
4. Table II and Figs. 3-6 make clear the complex agreements
and differences between these observers. A few examples may
be stated: at 7000 A the lowest determination is 36.3 per cent,
that of the present report is 45.0 per cent ; at 6500 A the mean
of 8 determinations is 24.9 per cent, the maximum is 31.2 per
cent and the minimum is 18.9; at 5300 A the mean of 7 is 2.4
per cent, the maximum is 4.5 and minimum is 0.3 per cent; at
3650 A the present report gives 3.65 per cent, Sawyer finds 21.6
per cent.
5. These differences are so great that they cannot be ascribed
to the nature of the method or the instruments employed ; they
seem to depend on differences in the water which was examined,
and these, in turn, seem to come from causes not yet fully under¬
stood. The results stated in this report for absorption in fil¬
tered lake waters are in general harmony with those here re¬
ported for distilled water.
6. A brief discussion is given of the effect of temperature of
water on its capacity for absorption of light. This is based on
150 Wisconsin Academy of Sciences, Arts, and Letters
the work of Ganz, Collins, and Baldock; no observations were
made on this subject for the present report.
7. The results found by these students are stated in Tables
V, VI, VII. Those of Collins and Baldock constitute a continu¬
ous series, covering the spectral region 5500-8000 A. They are
the basis for the diagram, Fig. 7, which compares absorption by
water at 25°-26° with that at 90°-96°.
Chapter III
1. Summaries for Chapter III are necessarily very imperfect,
since the chapter contains the data which are fundamental in
this study of lake waters. It also attempts to summarize the
chief points in the significance of these data. It states the action
on light of such waters, as due to substances present in the lakes
and as modified by settling, by filtering, and by diluting the
waters. There are 4 sections : 1. General conditions in lake
waters. 2. Tables of data and discussion of effects of treatment
of waters. 3. Mean curves for percentile absorption of light,
and classification of lakes. 4. Examples from individual lakes.
2. This study has not examined the effect on light of gases
contained in the water or that of inorganic salts in solution.
Section 1. Pp. 45-59
1. This section discusses the various substances contained in
lake waters and which affect the transmission of light. These
are divided into two groups, Suspensoids and Colors. Suspen-
soids are minute particles or masses which obstruct the passage
of light ; they are usually organic or colloid ; when present in
numbers they cause turbidity ; they may be removed in part by
settling, in larger part by filtration ; color is often associated
with them.
2. The term Colors is used in this report to designate certain
products of decomposition of organic matters in lake waters,
ranging in color from light yellow to dark red or nearly black.
They may be dissolved and in such condition as to pass a filter,
or they may be connected with or part of suspensoids and re¬
movable by the filter. These are called colloid colors. In any
lake the observed color is due to a mixture of numerous and
various color substances.
James & Birge—Lake Waters and Light
151
3. The standard of color is the platinum-cobalt or U.S.G.S.
standard. Colors are stated as cardinal numbers, 2, 6, 16, 54,
etc. ; each number represents the number of mg/1 of platinum
in the solution whose color best matches that of the water ex¬
amined. The uses, value and defects of the standard are dis¬
cussed. The sources, amounts, and types of colors in lake waters
are presented and discussed, and the action of lake waters on
incident light is stated in general terms. Figs. 9, 10, 11 belong
here.
Section 2. Tables VIII-IX
Figs. 12, IS; pp. 59-66
1. This section contains the statistical material on which the
report is based. Tables VIII, IX show the position and nature
of the lakes whose waters have been examined, and the effects
produced by filtration. Tables X, XI, XII contain the detailed
data of percentile absorption of light at 21 (or 22) wave-lengths
of the spectrum, and for settled, filtered, and diluted lake waters.
In these tables the waters are arranged in order of increasing
color.
2. The general effect of the various treatments of lake waters
on their absorption of light is discussed; in Figs. 12, 13 the
spread of single determinations of absorption in waters of the
same color is presented.
Section 3. Figs. 1U-18; pp. 67-78
1. Figs. 14, 15, 16 contain percentile absorption curves de¬
rived as color means from Tables X, XI, XII, for settled, filtered,
and diluted lake waters. The forms of such absorption curves
are discussed, and also the effects on them of various treatments
of lake waters and of increase of color.
2. Lake waters are classified in three groups, based on their
types of color material as brought out by filtration and its effect
on percentile absorption of radiation.
Group I. Water whose color is largely extractive and dissolved.
Group II. Waters whose color is largely colloid and therefore re¬
movable by filtration.
Group III. Extreme cases of Group II; waters whose color rating is
low and is reduced by filtration to 0 or T.
152 Wisconsin Academy of Sciences, Arts, and Letters
3. Percentile absorption, as determined for the present re¬
port, is compared with that found in Wisconsin lakes by Pieten-
pol some 15 years earlier. The general character of the lake
waters and of their action on light seems to be fairly constant.
Comparison is also made with determinations made on German
lakes by von Aufsess and on a Minnesota lake by Erikson. Their
results can be interpreted by the results from the larger series
of Wisconsin lakes.
Section 4. Figs. 19-28 ; pp. 78-98
1. This section contains 8 diagrams, Figs. 19-26, which illus-
strate the effect of settling, filtering, and dilution on light ab¬
sorption by the waters of individual lakes. They show the be¬
havior of the waters from the three Groups defined above, when
subjected to these treatments. This behavior is discussed for
each diagram.
2. Figs. 27, 28 contain absorption curves for settled or fil¬
tered lake waters whose effect on light was read to 3650 A. They
show the same characteristics as do those for the smaller spec¬
tral range, but in a more extreme form.
Chapter IV
1. In this chapter there is attempted an elementary analysis
of the action of lake waters on light. The total effect of a water
is divided between two or more factors ; and approximate quan¬
titative values are assigned to them. Three methods are pre¬
sented of such distribution of total action; each yields curves
of a different type ; the results are given in diagrams rather than
in numerical tables. In all cases the unit for the action of water
is a stratum one meter in thickness.
2. The ordinates for remainder curves are derived from those
for a percentile absorption curve by subtracting the absorption
ordinates of pure water. The resulting curve shows a relatively
high percentile absorption in the short-wave spectrum, its
amount depending primarily on the color of the water. Ab¬
sorption becomes less in the long-wave spectrum and is small or
negligible, 7400-8000 A. Remainder curves also show for any
lake water the percentile absorption which is added to that of
water by all other factors combined. In this computation ab-
James & Birge — Lake Waters and Light
153
sorption by pure water is given its full observed value. Figs.
29, 30 show remainder curves of settled and filtered waters,
3. In Figs. 31, 32 remainder and factor curves are shown for
the same waters that appear in Figs. 29, 30. Marked irregular¬
ities appear in factor curves, 7000-8000 A; these are discussed
on pp. 122-123.
4. Factor curves are based on percentile transmission of radi¬
ation, the complement of percentile absorption. Transmission
ordinates for any curve plotted in the diagrams are measured
downward from the 100 per cent line, taken as zero. Three types
of such curves are used: the general factor curve, the color
curve, and the suspensoid curve. The methods of computing
ordinates for such curves are given in Table XIV, for the water
of Nagawicka Lake.
5. The ordinates for the general factor curves are those per¬
centile transmissions which, multiplied by those of pure water,
will give as a product the percentile transmissions observed in
the water examined. This method gives rise to 2-component
analysis of absorption, shown in Figs. 33-36 and pp. 114-123.
6. The general factor curve of a settled lake water may be
further analyzed and its transmission ordinates may be assigned
to two components, color and suspensoids, these with water fur¬
nish the transmission ordinates for a 3-component analysis of
total absorption of radiation by a settled lake water. In 3-com¬
ponent analysis of any water the factors are :
Efc' ■
A. Water as water. Highly selective in its action on light; maximum
7400-8000 A; minimum (1-2 per cent) 4000-5500 A.
B. Color. Highly selective. Maximum in short-wave spectrum, amount
dependent on grade of color; minimum in long-wave spectrum,
amount negligible for low colors.
C. Suspensoids. Action on light comparatively non-selective ; amount
dependent on quantity of suspensoid present in the water.
D. At any wave-length of the spectrum the product of the percentile
transmissions for these three curves will equal that of the water
examined, as found at that point.
E. In Figs. 37-49 and accompanying text 3-component analysis of ab¬
sorption of radiation in various waters is illustrated and discussed.
7. Any factor curve in these diagrams may be taken as a per¬
centile absorption curve, showing the percentile effect which its
factor or group of factors would have if operating alone on the
154 Wisconsin Academy of Sciences, Arts, and Letters
water under examination. The presence of the other factor or
factors reduces its actual effect by reducing the quantity of
radiation on which it can operate. The set of factor curves for
any water may be computed as additive absorption curves , as is
done for Nagawicka Lake, Fig. 38, pp. 128-129.
8. Factor curves for color and for suspensoids may be com¬
puted in certain cases where the regular methods cannot be ap¬
plied. Three such cases are shown in Figs. 50, 51, 52 and accom¬
panying text.
9. Mean factor curves for absorption by color are given in
Fig. 53 for the spectral region 4000-7000 A and for colors 0-180.
Corresponding transmission curves are given in Fig. 54.
MICROSTRATIFICATION OF INLAND LAKES*
Lester V. Whitney
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 68.
Introduction
It has been known for a long time that inland lakes of suffi¬
cient depth undergo a large scale thermal stratification in sum¬
mer and winter. In summer, a temperate lake stratifies into an
epilimnion, a thermocline and a hypolimnion. The epilimion
consists of the upper water which is kept in active circulation
by the wind and which has practically a uniform temperature
throughout. Immediately below this stratum is the thermocline,
which is characterized by a rapid decrease in temperature. The
rate of temperature change is sometimes as much as 5° C. per
meter or more. Below the thermocline is a body of relatively
more stagnant water known as the hypolimnion in which the
temperature declines gradually toward the bottom. These three
strata usually show marked chemical and biological differences.
In addition to this general thermal stratification of the wa¬
ter, there is also a microstratification in the thermocline and
hypolimnion. This microstratification was discovered by means
of the transparency meter shown in Figure 1. This apparatus
shows the relative transparency of the water at different depths ;
readings were taken in 10 Wisconsin lakes in 1936 and this re¬
port is based on the results obtained in these studies.
The transparency meter is similar to that used by Hans Pet-
tersson* 1 in Gullman Fjord in 1933 and 1934. Pettersson also
measured the variation of density with depth and obtained
curves similar to those for temperature and density of inland
lakes. He took readings at much greater depths than is possible
on small inland lakes such as those in Wisconsin; readings are
indicated on his graphs about every meter where the transpar-
* This investigation was supported by a grant from the Brittingham
Trust Fund.
1 H. Pettersson. Jour, du Cons. 10 Cl): 48. 1935. Also Rap. et Proc.-Verb. d. Reunion!
101: 3-7. 1936.
155
156 Wisconsin Academy of Sciences , Arts, and Letters
ency changed rapidly and about every 5 meters where the change
was less rapid. The present report shows that readings must
be taken at much smaller depth intervals in order to determine
the limits of the various strata in inland lakes.
Apparatus
In taking readings with the transparency meter shown in
Figure 1, a beam of light from an ordinary 3 candle power auto¬
mobile bulb was focused on the photoelectric cell, type RCA 868.
Insulated wires from the supporting cable connected the light
source to batteries in the boat and another set of wires carried
the photoelectric current, induced by light falling on the photo¬
cell, to an amplifier where its effect was multiplied enough to
give a reading on a microammeter. As the apparatus was low¬
ered into the water, variations in the amount of light reaching
the photocell from the automobile bulb produced variations in
the microammeter readings. Curves of relative transparency at
different depths were then drawn as shown in Figures 3 to 9.
Water-tight containers housing the light source and the pho¬
tocell were clamped one meter apart on a galvanized iron pipe
0.9 cm. in diameter. Flat glass windows about 3 cm. in diameter
were inserted in each housing. A condensing lens immediately
Fig. 1. Underwater transparency apparatus. Parallel beam from light
source passes through 1 m. of water to reach photocell window. As ap¬
paratus is lowered in horizontal position, layers of varying transparency
are detected by photocell response.
Whitney — Micro stratification of Lakes 157
in front of the light source focused a beam of parallel light on
the photocell window. The small diameter of the light beam
made it possible to detect differences in transparency within a
few centimeters in depth, a degree of accuracy found necessary
in many cases. The apparatus, which was lowered in a horizon¬
tal position, was light enough in weight to be supported by the
cable carrying the wires connected to the amplifier and batteries
in the boat above.
A box, approximately 50 X 23 X 30 cm., housed the ampli¬
fier and the dry cells used for the light source. It was found
that two sets of dry cells in parallel furnished a steady current
for the 3 c.p. light source. The amplifier box and the basket
containing the connecting cable were carried easily in an ordi¬
nary rowboat.
The Amplifier Circuit
Figure 2 is the circuit diagram of the amplifier. Light falling
on the photocell causes a proportional deflection of the micro¬
ammeter. When light strikes the caesium oxide coated semi-
cylindrical plate of the photocell, electrons are ejected from it
and attracted to the positive staff; then they flow through the
batteries, through the resistances R8, R7, Ri, Re, and back
through the photocell. This electron current in passing through
the resistance Ri raises the voltage on the grid of the amplifying
Fig. 2. Amplifying circuit.
158 Wisconsin Academy of Sciences, Arts, and Letters
tube, and there results an increase in plate current. Before the
grid voltage is raised, however, adjustments can be made so
that none of the plate current flows through the microammeter ;
instead, it all flows from the amplifying tube, through the milli-
ammeter, through the resistance R2 and back through the bat¬
teries. Under these conditions, the circuit acts precisely as a
balanced Wheatstone bridge; the resistances of the amplifying
tube and R2 act as one arm and the resistance R3 and R4 as the
other arm. When the voltage of the grid is raised, because of
light falling on the photocell, the bridge is effectively unbal¬
anced, and current, which is proportional to the light on the
photocell, flows through the microammeter.
The sensitivity can be increased by increasing the resistance
Ri. When this resistance is made larger there is more of a
voltage change on the grid for the same photoelectric current, or
for the same light intensity on the photocell. Resistance Ri can
be varied from 30,000 to 30 million ohms by means of a multiple
pole switch ; thus the sensitivity of the amplifier can be changed
by a factor of a thousand. In actual practice, this full range of
sensitivity of the amplifier cannot be used, without special cor¬
rections because of the nonlinearity of the photocell at high light
intensities.
It is sometimes desirable to measure small variations in a
light of strong intensity, or to balance out the effect of an un¬
wanted light such as daylight. This can be done by adjusting
the potentiometer R7 and the resistance R8 (for fine adjust¬
ments). This adjustment makes possible an arbitrary change
in the voltage of the grid, and therefore a voltage change of the
grid, due to light falling on the photocell, can be counterbal¬
anced.
Methods of Taking Readings . The first few sets of readings
were taken at night, but this was soon found to be unnecessary
as daylight could be allowed for as already indicated. When
using this method, a separate balance for daylight had to be
made at each depth since the intensity of daylight steadily de¬
creased with increasing depth. This necessitated turning on and
off the artificial light at each depth and it was most important
that the light return to the same intensity. It was found to do
so within detectable error.
Whitney — Micro stratification of Lakes
159
Daylight could also be allowed for by taking an entire set of
readings with the light on, and then subtracting readings of
daylight alone. This method has the disadvantage of mixing up
the water due to the apparatus being raised and lowered through
it, a factor not to be ignored as will be evident later. A further
disadvantage is the fact that daylight intensities may change
while readings are being taken. Of course such change is usu¬
ally negligible on clear days, but the first method was the better.
Beyond 10 m. it was usually found that daylight could be ig¬
nored.
Results
Figures 3 to 9 are a series of curves showing the way in
which the relative transparency changed with depth in the lakes
studied. Distances to the right are proportional to microam¬
meter deflections resulting from light penetrating 1 m. of water
at the various depths. Quantitative significance cannot be at¬
tached to these curves because of the combined selective action
of the light source and photocell ; further, absolute values shown
are not comparable between one curve and the next as it was
necessary to increase the intensity of the artificial light on some
lakes. However, for any one curve transparency values are
comparable since a constant light source was maintained for
each series of readings.
The curves show some interesting characteristics. Changes
in transparency are often sudden and large in magnitude. With¬
in short distances it is not uncommon for the transparency to
drop to half or less than half the value immediately above. A
layer of water only a few centimeters thick is often found with
a transparency greatly different from that immediately above
or below. The lower water of some lakes seems to be made up
of layers and the number of these layers is sometimes surprising.
In spite of the odd appearance of some of the curves, at
least two sets of readings, checking each other, were taken in
each case. In some instances, such as in Mud Lake, four or five
sets were taken at places in the lake somewhat removed from
each other. Although there is good reason to believe that the
pattern of a transparency curve would not be the same over an
entire lake, such a curve is repeatable at stations within a radius
of 15 to 80 m. or more. More will be said about the changes
160 Wisconsin Academy of Sciences, Arts, and Letters
taking place in transparency curve patterns in the discussion of
Lake Mendota.
All the readings on the northern Wisconsin lakes were taken
during the last two weeks of August 1986. No attempt was
made to follow any one lake in detail; instead, several lakes
were measured in order to get some idea of the different types
of transparency curves that existed and to discover similarities
between them if possible. The only similarities emerging so far
seem to be the constancy of the transparency in the epilimnion,
at least one sudden drop to a minimum transparency just below
the epilimnion (with few exceptions) and the tendency for a
precipitate decrease in transparency immediately above the
bottom. As yet, there are insufficient data to classify lakes ac¬
cording to similar patterns ; on the contrary, each pattern so far
found seems to suggest an individual story.
Northern Lakes
South Trout Lake. The first readings were taken in South
Trout Lake on the night of August 18, 1936 (Figure 3). A de¬
crease in transparency began at 10.5 m. and reached a minimum
at about 13.8 m. After a rise to practically its former value, the
transparency showed little change in the next 10 m. although
there was a slight drop at about 20 m. At 26 m. there was a
sudden decrease, and the final drop began at 28 m. Constant
values of transparency between 27 and 28 m. were obtained
several times.
North Trout Lake. On the following night, readings were
taken on North Trout and they show several differences from
those of South Trout (Figure 3). These differences were not
altogether unexpected since North Trout is separated from
South Trout by a narrow channel and is practically a separate
lake. There was a minimum below the epilimnion at about 12.5
m. The narrow band of relatively high transparency at 14.7 m.
was missed in the first set of readings when half-meter intervals
were used. In the next two series, when readings were made
every quarter meter, the band was clearly evident. The mini¬
mum of transparency between 18 and 19 m. was repeated on
several sets of readings.
Whitney — Micro stratification of Lakes
161
Fig. 3. Transparency curves for South and Nortfh Trout Lakes,
August 18 and 19, 1936.
Crystal Lake. As might be expected in a lake as clear and
low in color as this one, only small changes in transparency In
the lower water were found (Figure 4). The amplifier was set
at maximum sensitivity for reading small differences in a light
of strong intensity and scale readings were confined between
830 and 950 as indicated on the curve. Two minima were
found, at 11.00 and 13.5 m., and a maximum at 12.8 m. The
final drop in transparency was extremely sudden.
Muskellunge Lake. In this lake all the transparencies below
the epilimnion were lower than those above. Two minima and
two maxima are evident between 10 and 14 m. The cut-off at
the bottom was sharp.
162 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 4. Transparency curves for Crystal, Muskellunge, Mud, Helmet,
and Silver Lakes.
Mud Lake. Mud Lake shows some of the most extreme vari¬
ations in transparency within short distances. This is a small
lake, about 38.5 ha. in area, and the water is moderately colored,
or about 37 on the platinum-cobalt scale. The thermocline was
quite near the surface. Readings were necessary every 10 cm.
Care had to be taken to avoid tipping the boat, as it was easy to
change the microammeter reading 30 per cent or more merely by
shifting one’s position in the boat or by leaning one way or an¬
other. Four sets of readings, two with the boat at one station
and two at another, were taken. At each location a set of read¬
ings was taken with the apparatus descending and another with
it ascending. With the apparatus descending, readings repeated
themselves, including even the minor rise in transparency at
4.4 m. (Figure 4). In the ascending series this detail was not
found. Apparently this difference was due to a slight mixing
of the water as the supporting pipe moved through the water
ahead of the light beam. On other occasions it was found that
raising and lowering the apparatus was sufficient to obliterate
some of the finer details of the stratification.
Helmet Lake. This lake, similar in appearance to Mud Lake
but about 4 m. shallower, was the most highly colored (265) of
all the lakes studied. The first maximum and minimum are
Whitney - — Micro stratification of Lakes
163
higher than those in Mud Lake, and there follows only a single
broad maximum centering at about 3.5 m. There was no addi¬
tional fine detail. The last 4 m. above the bottom showed practi¬
cally a constant transparency.
Silver Lake. In general, this lake showed relatively low
transparencies below the epilimnion. Three rather well defined
minima were found at 8.8, 13.3, and 16.3 m. Maxima centered
about 10.0, 14.5, and 16.6 m. As usual, the bottom cut-off was
sharp. When readings were taken between 12 and 14 m., some
rather sudden fluctuations in the mircroammeter readings took*
place. These were believed due to fish swimming across the light
beam.
Weber Lake . This lake, Figure 5, has an area of 15.6 ha. and
a maximum depth of 13.5 m. Notice the sharp minima at 7.8
and 12.4 m. The maximum immediately above the bottom is
also quite sharp. Two other minima at 8.4 and 9.7 are evident.
Nebish Lake . No sudden drop in transparency was found
immediately below the epilimnion in this lake, but instead there
was a region of higher transparency about 3 m. thick. This is
the only lake in which there was such a marked exception to the
general rule. Net plankton catches showed that the 5-10 m.
stratum of Nebish was well populated with Daphnia pulex and
the greater transparency in this stratum might be due to the
Fig. 5. Transparency curves for Weber, Nebish, and White Sand lakes.
164 Wisconsin Academy of Sciences , Arts , and Letters
fact that these water-fleas removed phytoplankton and the or¬
ganic detritus from this stratum for their food.
White Sand Lake. Two sets of readings are shown for this
lake which were taken at stations about 15 m. apart. The two
curves give some idea of how different sets of readings compare
with each other. The two curves are similar in general appear¬
ance, a maximum occurring just below 10 m. and a minimum a
little below 13 m. But in addition, there is surprising agree¬
ment in the fine detail. Just above 10 m. there are two minor
minima and maxima. Between 10.5 and 13 m. there is evidence
on both curves of two more minor minima and maxima. Be¬
tween 14 and 17 m. small variations occur on both curves, varia¬
tions one would be tempted to ignore as experimental error if
only one set of readings had been taken, but which assume signi-
cance when they repeat themselves. The depths at which these
minima and maxima occur do not always agree and such differ¬
ences may correspond to real differences in the lake. However,
as the supporting cable was calibrated only in meters and small
distances had to be measured with a meter stick sometimes used
over the side of the boat, there were chances of small errors in
the depth measurements between one set of readings and the
next. In this lake special care was taken to note the sudden¬
ness of the bottom cut-off. In the last 30 cm. the cable was
lowered first 5 cm. at a time, and finally 1 cm. at a time. The
drop from the maximum reading shown on the curve to zero
took place within 2 cm. After the cut-off, the apparatus was not
resting on firm bottom but it sank about a half meter farther
into the mud. It is well known that the bottoms of most lakes
are covered with a soft muddy ooze but it was somewhat sur¬
prising to find such sharp lines of demarkation between rela¬
tively clear water and the practically opaque mixture of mud
and water covering the bottom.
Lake Mendota
After the return to Madison from the Limnological Labora¬
tory at Trout Lake, a study was made of Lake Mendota. In
the latter part of September the apparatus was reconstructed so
that readings of light scattering could be made. A separate
source of light was mounted along the side of the photocell
housing so that a strong beam from a 32 c.p. bulb could be sent
Whitney — Micro stratification of Lakes 165
out ahead of the photocell. Two 6 volt storage batteries were
required for this light. Readings of scattered light intensities
were made late in September and early in October.
The supporting pipe was also drilled with holes and con¬
nected to a rubber tube and pump for drawing water samples.
With several openings in the horizontal pipe, water is drawn
from a narrower stratum than would be the case were there but
a single opening. It is an advantage to have the water sampler
a part of the light measuring apparatus, because the microam¬
meter can be watched and used as a guide at the same time that
the water is drawn.
Changes in Transparency Patterns . It was expected that
the region of irregular transparencies would be confined to deep¬
er water as the epilimnion became thicker in the colder fall
weather. Figure 6 illustrates this fact. Curves were obtained
at Station 2 on September 3, September 26, and on October 8.
The first minimum was found at progressively lower depths,
first appearing at 11.5 m., then at about 15 m., and finally at a
little over 16 m. On the week end of Oct. 10 the fall turnover
took place. The temperature dropped below freezing and strong
winds made the lake too rough to venture on for the next few
days. After this overturn the transparency of the lake was
practically the same from top to bottom.
On Sept. 3 the transparency below the first minimum re¬
turned to a value as large or larger than above, but on Sept. 26
the transparency in the lower water was relatively much smaller
than above. On Oct. 8 the transparency at 16 m. dropped so
suddenly to practically zero that it was thought a mistake in
depth measurements had been made and the bottom had been
reached. In order to read the lower transparencies at all, the
sensitivity of the apparatus had to be multiplied by 20.
Readings were taken on Sept. 26 and Oct. 8 at Station 1 in
shallower water. Between these dates the first minimum dropped
from 13 m. to 16 m. in depth. Notice that on these two curves
the maximum following is quite pronounced, whereas on the
deep water curves the corresponding maximum is almost negli¬
gible.
Comparing the two curves for Sept. 26, a marked difference
in the depth of the first minimum is evident. On this day there
was an off shore wind and it was expected that the thermocline
166 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 6. Transparency curves for Lake Mendota.
in the shallower water near shore should be closer the surface.
On Oct. 8, with the lake much calmer, there was less difference
in depth between first minima for the two curves. A compari¬
son between the four curves taken Sept. 26 and Oct. 8 shows
more similarity between curves taken at the same station than
between curves taken on the same day.
Scattering. Several sets of transparency and scattering'
readings were taken, and Figure 7 shows a typical pair. Notice
that the transparency readings in the lower water again were
so small that the sensitivity had to be made about 20 times lar¬
ger than for readings taken above.
The more scattering particles there are in water, the lower
the transparency should be and the higher the scattering. The
curves for the lower water show some of the expected antisym-
metrical characteristics. For example, the transparencey mini¬
mum at 19.3 m. is matched by a scattering maximum ; the trans¬
parency maximum at 20.2 m. is matched by a less pronounced
scattering minimum. However, the sudden and permanent de¬
crease in transparency at about 17 m. is not accompanied by a
permanent rise in scattering. The scattering does increase at
Whitney - — Micro stratification of Lakes
167
Fig. 7. Transparency and scattering curves for Lake Mendota.
this depth, but it rises only to a maximum and then falls again.
In the lower water the scattering is less on the average than
that above. In some other sets of readings, there have been
instances in which large changes in transparency are accom¬
panied by no detectable changes in scattering. A large increase
in the amount of highly colored dissolved material in a lake
would reduce the transparency but might also reduce scattering.
Hence transparency and scattering curves might be antisym-
metrical or similar depending on the amounts and kinds of ma¬
terials immersed in the water at different depths. A comparison
of such curves yields more information than either type taken
alone.
Organic content . In the latter part of September, water
samples were taken in an effort to discover correlations between
transparency maxima and minima and organic content. One
liter samples of water were centrifuged and loss in weight meas¬
ured when converting the centrifuged material to ash. Results
so far obtained are inconclusive although some evidence points
168 Wisconsin Academy of Sciences, Arts, and Letters
to an increase in organic content at the first minimum below the
epilimnion. Visual inspection of several water samples taken
early in September, however, revealed large numbers of Daph-
nia at the first minimum below the epilimnion, and practically
none in the maximum following.
Bacterial counts. Bacterial counts were made on one set of
water samples ; although this evidence is too meager to warrant
generalizations, interesting differences were found. On Sept.
26 at a station between Picnic Point and Second Point where
the water is about 17.5 m. deep, water samples were drawn
from three levels. Since all the water had to come through
the same rubber hose, 10-15 liters were drawn from each level
before taking any for bacterial counts. None of the water went
through the pump; it was delivered directly from the hose to
the sampling bottle. The total capacity of the 40 m. hose was
about one and one-half liters; thus the 10-15 liters drawn
through it first should be sufficient to flush it out.
Water was drawn from the 10 m. level in the epilimnion,
from the 16 m. level at the first transparency minimum, and
from the 16.4 m. level at the following transparency maximum.
Four dilutions of each sample were poured in five replicates.
After three weeks incubation at 28° C. counts were made. Large
distinct, typical colonies were found on all plates. However, in
addition to the typical flora, the plates from the 16 m. trans¬
parency minimum, even in the highest dilution, showed a very
large number of pinpoint colonies. Neither in the epilimnion
nor at the 16.4 m. maximum were these pinpoint colonies found.
Counts were made of the appropriate dilutions, including con¬
servative estimates of the number of pinpoint colonies. Rela¬
tive numbers of colonies beginning at the epilimnion were ap¬
proximately in the following ratios :
Depth Relative Count
10.0 m., epilimnion . 10
16.0 m., transparency minimum . 200
16.4 m., transparency maximum . 1
As already mentioned, general conclusions cannot be drawn
from these data, but, typical or not, this one set of samples sug¬
gests that the first transparency minimum has a far richer bac¬
terial flora than the other levels tested.
Whitney — Micro stratification of Lakes 169
Winter Readings
In December 1936, and in February and March 1937, trans¬
parency readings were taken on Lake Mendota under the ice.
In February and March, temperatures were read with a Wheat¬
stone Bridge resistance thermometer which was attached to the
transparency apparatus. An audio amplifier and earphones were
used in the resistance thermometer circuit instead of a galva¬
nometer; under the best conditions temperatures could be read
to Vso of a degree Centigrade. Bacterial counts were also made.
The first transparency readings (Figure 8) were taken on
December 22, and there was little change with depth. The lake
opened again a few days later, and then froze over once more
on January 5, 1937.
Readings taken on February 26 showed definite stratification.
There were transparency maxima at about 12 and 20 m., and a
minimum at 16.5 m. The change in transparency from the top
to 12 m. was gradual. However, on March 1 the upper water
had become quite opaque, and there was a rapid increase in
transparency near 4 m. Between February 26 and March 1 a
thaw took place, and quantities of muddy water flowed into the
lake from all the shores. This muddy water spread out over
Fig. 8. Transparency curves and bacterial counts for Lake Mendota
in winter.
170 Wisconsin Academy of Sciences, Arts, and Letters
the entire lake in a thin layer under the ice. It was found at
Station 2, which was about at the middle of the lake.
The last two curves of Figure 8 are for March 17 at Stations
1 and 2. The opaque layer under the ice was present at both
places. Between 3 and 7.5 m. at Station 1, and between 3 and
12 m. at Station 2 the transparency was constant. The deeper
water at both stations was more transparent and showed small
differences with depth. There was a maximum in transparency
just below 15 m. at both locations.
Temperatures. Figure 9 shows the transparency and tem¬
perature curves for March 17. The regions of constant trans¬
parency at both stations are also regions of constant tempera¬
ture. There are never sudden changes in transparency without
some changes in temperature with depth. The rise in tempera¬
ture above 3 m. can be readily explained. The opaque layer of
water under the ice absorbs practically all the energy coming
from the sun and sky; consequently the temperature of this
layer is raised. One should expect some mixing action to take
Fig. 9. Transparency and temperature of Lake Mendota in winter.
Whitney — Micro stratification of Lakes 171
place because the warmer water has a greater density than the
cold water below. The warm water layer, immediately under
the ice, has been observed in the past by Dr. Juday. He found
that the effect was most marked when the ice was clear and the
sun was shining brightly. The water in contact with the ice was,
of course, nearly at 0.0° C. ; but the temperature rose rapidly
and fell again until a depth of about 3 m. was reached. Below
this depth, the temperature increased slowly toward the bottom,
with a more rapid rise below 14 m.
Bacterial counts. Bacterial counts were made from water
samples drawn on March 1 and on March 17. The numbers
shown opposite these curves (Figure 8) give the numbers of
bacteria colonies per cubic centimeter of water. Small differ¬
ences in transparency do not seem to be followed by consistent
differences in bacterial counts, but, in general, where the trans¬
parency is small the bacterial count is large.
The curve for March 1 does not show much difference in
transparency between 1.5 and 2.5 m. However, readings taken
with a sensitive setting of the photocell amplifier indicate that
the transparency at 2.5 m. was about four times that at 1.5 m.
The curve for Station 1, March 17, shows the best correla¬
tion between bacterial counts and transparencies. If the bac¬
teria, or if whatever is associated with bacteria, are the pri¬
mary cause of transparency changes, a semi-logarithmic graph
of numbers of bacteria plotted against relative transparency
should be approximately a straight line. The open circles on
Figure 10 are points from Station 1. They constitute a fairly
regular curve and indicate that bacteria, or material associated
with bacteria, may be responsible for much of the transparency
changes. The solid dots are values from Station 2, March 17.
They are not as regular as the other points, but they fit the gen¬
eral picture.
Discussion of Results
The transparency and scattering apparatus, which consists
of artificial light sources and a photoelectric cell, has proven to
be a sensitive detector of the microstratification existing in in¬
land lakes. This stratification exists both in summer and in
winter, and causes of it require further investigation.
172 Wisconsin Academy of Sciences, Arts, and Letters
It may prove impractical to attempt an explanation of all
the fine details of any particular transparency curve, but it
should be possible to discover the general factors at work which
cause the stratification. Some progress in this direction has
already been made. The fact that the greatest changes in trans¬
parency usually occur in regions of relatively large temperature
change may indicate that a lake acts as a filter or sorting ma¬
chine in which particles of different sizes, shapes, and densities
slowly sink and accumulate at levels of similar densities as de¬
termined by temperature and other physical conditions. Pos¬
sibly, quantitative relationships can be worked out. Data ob¬
tained in late fall and in winter show some regularity between
bacterial counts and transparency. The evidence on the num¬
bers of Daphnia — though meager and inconclusive — indicates
plentiful food supplies and other conditions which are favorable
to organisms in particular zones.
At present, more data need to be obtained before a satisfac¬
tory explanation can be given of the microstratification phe-
o-1 -2 A .6 -SIP 2 4 6 8 10 _ 20 40 60 100
Fig. 10. Semi-logarithmic graph of transparency and bacterial count
for Lake Mendota, March 17, 1937. Open circles for Station 1, solid dots
for Station 2. Transparency values are relative only, and are proportional
to response of the photocell. Bacterial count is given in numbers of colonies
per cubic centimeter of water.
Whitney — Micro stratification of Lakes
173
nomenon. More information on the nature of the stratification
might be obtained by the use of a series of light filters. The use
of filters should also make it possible to measure the spectral
extinction coefficients, and to correlate them with measurements
of the transmission of solar radiation. To understand the con¬
ditions which bring about the microstratification of lake waters,
more data of all kinds are needed to discover the differences
which exist between one stratified layer and the next. Further
work on the correlation of chemical and organic content, bac¬
terial counts, conductivity and temperature measurements is
being planned.
Summary
1. The transparency and scattering properties of water were
measured with an apparatus consisting of a light source, photo¬
cell, and amplifier.
2. Transparency measurements were taken on 10 northern
Wisconsin lakes; transparency and scattering measurements
were taken on Lake Mendota. In the thermocline and hypolim-
nion differences in transparency were found, great changes often
taking place over a distance of but a few centimeters.
3. Readings on Lake Mendota showed that with the lowering
of the thermocline in the fall, the transparency of the hypolim-
nion became relatively much smaller than that in the epilimnion.
4. Scattering and transparency curves were anti symmetrical
in some regions but similar in other regions.
5. Organic content and bacterial counts were correlated with
the transparency curve patterns.
6. Readings taken under the ice on Lake Mendota showed
stratified layers which correlated with temperature readings and
bacterial counts.
CONTINUOUS SOLAR RADIATION MEASUREMENTS IN
WISCONSIN LAKES*
Lester V. Whitney
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 69.
Introduction
The Wisconsin Geological and Natural History Survey has
carried on work for several years on the relations existing be¬
tween sunlight penetration and the waters of small lakes. Four
reports have been published and other material is on hand.
Numerous observations have been made on the quantity of light
energy delivered to the surfaces of the lakes; on its transmis¬
sion to various depths; and on the solar energy spectrum as
found in air and as modified by its passage through lake waters
of various types. In general, observations in the past have been
made in the direct sunshine and during the middle hours of the
day.
The Survey is also investigating the quantitative utilization
of solar energy by lake algae and other aquatic plants in the
manufacture of organic matter. These studies call for quanti¬
tative determinations of light in lake waters under a wider
range of conditions than studied heretofore, and records for the
entire day are needed. Such records show the effect of different
elevations of the sun, of diffuse light as met with on cloudy days,
of color or turbidity of lake water, and of varying surface of
the lake in wind and calm. The following report deals with the
continuous solar radiation records obtained during the summers
of 1935 and 1936.
Apparatus
Automatic Double Recorder. The Survey began a study dur¬
ing July and August 1935 at the Trout Lake Limnological Labor¬
atory with apparatus which furnished continuous automatic rec-
* This investigation was supported by a grant from the Brittingham
Trust Fund.
175
176 Wisconsin Academy of Sciences, Arts, and Letters
ords of solar energy both in air and in water at various depths.
With such records, the total solar energy delivered to the surface
of the lakes, and its transmission to various depths, could be
found under all conditions normally present in summer. The cen¬
tral part of the apparatus was a Cambridge Double Recorder
which measured continuously the solar energy received by sev¬
eral thermopiles or Photox cells. In 1935 records were made on
four lakes of different color and transparency ; these were Trout,
Crystal, Boulder, and Muskellunge. The records were checked
and standardized with records of solar energy made by an auto¬
matic Kipp Solarimeter with a Richard Recorder. The Kipp
Solarimeter in turn was calibrated in terms of the standard
solarimeter used by the United States Weather Bureau in Madi¬
son, Wisconsin. In 1936 the work was confined to Trout Lake,
but Photox cells as well as thermopiles were used, continuous
records being taken at greater depths than before.
The apparatus was set up so that it could be readily moved
from lake to lake. The recorder itself was supported by a
strong wooden frame solidly bedded on the sandy shore of the
lake. It was covered by a tent, but for special protection a large
wooden case covered the recorder and its frame. The purpose
of the case was to protect the recorder if a wind storm should
carry the tent away. Happily no such storm occurred; the re¬
corder functioned perfectly throughout both summers.
It should be added that the recorder was received by the
Survey in August 1934. The general arrangement of recorder
and thermopiles was worked out and the apparatus set up on
Trout Lake at that time by Mr. Donald Kerst. The instrument
came at so late a date in August that records for only a few
days could be made.
The recorder itself consists essentially of two galvanometers
so arranged that a continuous record can be made on a rotating
drum by arms which extend out from the galvanometers. Be¬
low the ends of these arms and just above the drum is an inked
thread. A chopper bar descends once a minute, pressing the arms
and thread against the paper, leaving dots on the record. The
galvanometer swings freely between readings. Just after the
dots are made, automatic switches are thrown, connecting the
galvanometers to other thermopiles or Photox cells. At the
same time a different colored thread is moved under the chopper
Whitney — Solar Radiation Measurements 177
bar. Thus a continuous record can be made of four light re¬
ceivers, which may be placed at different depths under water or
in air. Variable resistances are used in the circuits to allow for
differences in energy received. The resistances are adjusted
from time to time so as to have as large a deflection as possible.
Thermopiles. In the experiments of 1935, only thermopiles
were used ; three were of the Moll type and one was constructed
by Mr. Foerst of the Physics Department. Those used under
water were the Moll type and were fastened to a pipe frame
which hung from a buoy anchored about 180 m. from shore.
They were connected with the recorder by insulated cables. In
these small lakes this distance gives sufficient depth of water
and places the receivers outside of any turbidity of the water
due to contamination by wash from shore. These lakes are little
affected by shore wash since their margins are composed of
clean sand. The transmission of light through water at the
buoy was found to be the same as that in the center of the sev¬
eral lakes. Since the under water thermopiles were fastened
rigidly to the pipe frame, they were always in the same position
relatively to each other. When light sensitive receivers are sus¬
pended separately, there are errors in the calculated values of
transmission due to their not being exactly parallel to each
other. The thermopile for air records was set up either on the
beach near the recorder tent or on a platform supported by the
buoy.
Since a thermopile really responds to temperature differences
(set up between its elements), it is important that such differ¬
ences be not caused by varying temperature in the water sur¬
rounding it. At the thermoclines of small lakes the tempera¬
ture of the water may change so rapidly with increasing depth
that the readings are in error. The readings will also be in error
if the thermopile is immersed in warm and in cold water in
rapid succession. Such a condition can exist on a windy day
at the bottom of the epilimnion where cold water waves may be
washing over the thermopile. In one instance, variations in the
readings caused by such temperature changes in Helmet Lake
were from 10 to 20 times as great as readings due to light in¬
tensity. Of course such readings were discarded. The case of
Helmet Lake is exceptional; in general, lake waters are suffi-
178 Wisconsin Academy of Sciences f Arts , and Letters
ciently constant in temperature for readings to be entirely re¬
liable.
Photox Cells. In 1936, Westinghouse Photox cells, type SW-
30, were purchased. These are copper oxide photo-voltaic cells
requiring no external batteries for their operation. Their color
sensitivity follows closely that of the human eye, and they have
a high current response. Continuous records were made in Trout
Lake at depth of 7 m. with a Photox cell covered by a filter which
transmitted but of the light incident upon it. Greater depths
could have been recorded had the cables extending to the buoy
been long enough to reach the deeper water farther from shore.
The Photox cells were calibrated in terms of thermopile read¬
ings. The response of a thermopile is proportional to the total
radiant energy incident upon it; but the response of a Photox
cell is selective with wave-length, and not at all linear for in¬
tense light. To minimize the effect of selectivity with wave¬
length, Photox cells were never used at less than 3 m. in depth.
The filtering action of 3 m. of water reduces daylight radiation
to the visible range of wave-lengths. To obtain a linear response,
the total light intensity on the Photox cells was reduced by dense
filters. Continuous records were made of Photox cells and
thermopiles at the same depth for the final calibration ; readings
were found to be proportional when filters transmitting V25 or
less were used.
The filters were made with the materials at hand at the Trout
Lake Limnological Laboratory, and consisted of flat glasses
coated with varying densities of lamp black and shellac. These
filters were not neutral as the filter factor for daylight was dif¬
ferent from that for artificial light; however, for the purpose
intended they performed satisfactorily.
The characteristic curves for the Photox cell show that the
smaller the external resistance used in series with the cell, the
less important are changes in temperature, and the more linear
are the readings for a given range of light intensities. The ex¬
ternal resistance, consisting of the cable, the galvanometer, and
the Ayrton shunt, was at first about 250 ohms ; later it was re¬
duced to but 40 ohms.
Correction for Glass Covers . It is desirable to obtain read¬
ings proportional to the energy falling on 1 sq. cm. of the earth's
Whitney- — Solar Radiation Measurements 179
surface, or on 1 sq. cm. at some depth in a lake. An uncovered
thermopile parallel to the earth’s surface would give such read¬
ings to a high degree of accuracy, but a glass cover reduces
the energy falling on the thermopile elements, different amounts
for different angles of the incident rays.
Fig. 1. Relative thermopile readings with change in angle of incident
light. Circles are experimental readings; solid line shows computed values
proportional to F2cos0, and dotted line is cosine curve which shows values
proportional to vertical component of radiation.
The dotted curve of Figure 1 is the cosine curve and shows
how an uncovered thermopile would respond to a constant light
source at different angles. Readings which followed this curve
could be accepted without correction. However, readings of a
thermopile with a glass cover are proportional to F2cosQ, where F
is the fraction of light transmitted across a single glass-air
boundary. F is computed from Fresnel’s formula, and changes
in value with 0. The solid line shows values proportional to
F2cos0; the circles are experimental points obtained with a
thermopile turned at various angles to the sun. These readings
were taken at Trout Lake at a bright spot among the trees where
180 Wisconsin Academy of Sciences, Arts, and Letters
the sky contribution was negligible. The agreement of the ex¬
perimental points with the solid line is good.
The solid line follows the cosine values (which require no
correction) quite closely until past 50° ; therefore readings of
solar energy may be accepted without any correction when the
zenith angle of the sun is under 50°. The error in accepting a
reading at 60° is not large, but beyond this angle the percentage
difference between the two curves is too great to ignore. Read¬
ings in air were not used beyond this value of zenith sun angle.
Light which is incident on a thermopile in air is subject to
two air-glass reflections; light incident upon a thermopile in
water is subject to one water-glass and one air-glass reflection.
If glass has an index refraction of 1.53, the reflection of perpen¬
dicular light from a glass-air boundary is 4.5 per cent ; the cor¬
responding reflection from a glass-water boundary is 0.5 per
cent. Tests were made which checked these values. Thin cover
glasses were placed in the path of a direct beam of light to a
pyrheliometer, and readings were taken with and without films
of water between them. Since the cover glasses were extremely
thin, the absorption of light in the glass itself was neglected.
The results showed that for light of normal incidence, readings
taken under water should be reduced 4.0 per cent when com¬
pared to those taken in air. Calculations also showed that when
readings of diffuse light, or light from an evenly clouded sky,
were taken under water, they should be reduced by about 6.5
per cent when compared to similar readings in air. These dif¬
ferences are small and a reduction of 5.0 per cent should be satis¬
factory for average conditions.
Standardization of Readings. The reading caused by light
falling on a thermopile is directly proportional to the thermo¬
electric power of the thermopile elements, inversely proportional
to the total resistance in the circuit, and proportional to the sen¬
sitivity of the meter employed. A table of equivalent deflections
was worked out, since different cables, meters, and light re¬
ceivers were used in combination at various times. The equiva¬
lent deflections for the Photox cells were obtained by direct com¬
parisons with the thermopile records. Comparisons were made
with the Kipp Solarimeter for converting readings to calories
per sq. cm. per minute. During the summer of 1935, when the
Cambridge Double Recorder and the Kipp solarimeter were
Whitney — Solar Radiation Measurements 181
sometimes a half-mile or more apart, check comparisons were
made by using one of the Moll thermopiles connected to a high
sensitivity meter as an intermediate.
Results
In 1935 more readings were taken on Trout Lake than on
any other. Preliminary experimenting with apparatus was done
there. The thermopiles were first supported on individual plat¬
forms which were hung by chains from the buoy; this did not
prove satisfactory and an improved method of fastening both to
a rigid frame was developed. If the faces of the light receivers
are not rigidly fixed with respect to each other, calculated trans¬
missions cannot be depended upon. Ideally, the receivers should
be parallel and maintained in a horizontal plane with respect to
the earth at all times. This is impossible because of the motion
of the waves; however, if the receivers are fixed with respect
to each other, errors caused by their not being strictly parallel
are not serious. In this case, an asymmetry between morning
and afternoon readings appears, but it is regular and can be
allowed for.
Figures 2 and 3 are reproductions of original records made
by the Cambridge Double Recorder with about one-third of the
total number of dots recorded. The breaks in the curves of Fig¬
ure 2 indicate when the resistances in series with the thermo¬
piles were changed. For galvanometer 1, curve A is the record
from the thermopile above water, curve B the record from the
thermopile at a depth of 1 m. For galvanometer 2, curve A is a
duplicate of the one above, and curve B the record from the
thermopile at 4 m. The sudden drops of the A curves at the
ends of the day are due to the shadowing of the air thermopile
by trees on shore. Notice that the points for air are quite regu¬
lar, those for 1 m. quite irregular, and those for 4 m. more regu¬
lar again. The irregularity at the shallower depth is due to
wave action ; at greater depths the effect of individual waves is
less because the cone contributing irregular illumination extends
over a larger surface area.
Figure 3 of Boulder Lake, Aug. 22, 1985, is a record of a
typically cloudy day. Because of the high absorption in this
lake, the thermopiles were placed at only % m. and 1 m. in depth.
For galvanometer 1 all points are for the thermopile at % m. ;
182 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 2. Cambridge Double Recorder record for Crystal Lake August 15,
1935. A' represents results in air without resistance and A with resistance
in circuit. B' represents results at 1 m and 4 m. without resistance and
B with resistance in circuit. The factors by which the readings must be
multiplied to convert them to cal. per sq. cm. per min. are as follows:
Whitney — Solar Radiation Measurements 188
for galvanometer 2, curve A is for the air thermopile, and curve
B for the thermopile at 1 m.
Transmissions were calculated for various hours of the day
from such records. Averages of points were taken when waves
were running and transmissions were not appreciably different
from those for still water. In some instances when the sun was
Fig. 3. Cambridge Double Recorder record for Boulder Lake August
22, 1935. The depths of the thermopiles and the factors by which the read¬
ings must be multiplied to convert to cal. per sq. cm. per min. are:
Factor Depth Factor Depth
0.0034 Vz m. Galv. 2 Curve A 0.0294 Air
B 0.0040 1 m.
Galv. 1
184 Wisconsin Academy of Sciences , Arts , and Letters
covered and uncovered frequently by clouds, energies repre¬
sented by individual points were added up for a period of about
a half-hour and average transmissions calculated. Such values
could not be weighted highly because of the greater probable
error of individual points; however, these values did not differ
materially froip others found.
The first transmission curves drawn, shown on Figures 4 to
7, are averages from many records. In any one day there are
variations brought about by such natural causes as haziness,
waves, and intermittent clouds; it was thought that experi¬
mental errors would average out and regular trends become more
apparent.
T i i r~ r
TROUT LAKE
O CLEAR
® CLOUDY
"50 5
-40
64%
Transm/ss/on of / m. of Wa ter
-20
AM 7
Transm/ss/oaz From Surface To / m.
TIME
II 12
-1— _ L—
_L
-J/2_
5 PM
Fig. 4. Transmission for Trout Lake, July 1935.
The results for Trout Lake (Figure 4) were based on rec¬
ords for nine days; five of which were clear, two were partly
clear, and on two the sky was evenly clouded. Results for Cry¬
stal Lake (Figure 5) were based on records for eight days, clear
and cloudy weather being about equally divided. Results for
Boulder and Muskellunge lakes (Figures 6 and 7) were based on
one clear and one cloudy day for each. After the work on Cry¬
stal Lake the time remaining was short, but it was found that
one good clear day record and one good cloudy day record were
sufficient. As it happened, records for both Boulder and Mus-
Whitney — Solar Radiation Measurements
185
kellunge lakes were obtained on five consecutive days, cloudy
and clear weather obligingly alternating. Transmissions for
the first meter of water could not be calculated for the early
morning or late afternoon because of the shadowing of the air
thermopile by trees on shore.
CRYSTAL LAKE
OCLFAR
• CLOUDY
-60
-70
$ — — to — -5 - % - 2 '•"""s - — ~qT — 2 — ~~rrdOX-
TRANSMIJ3ION OF / M. OF WAT£R
-60 y
IT
UJ
-50
•40 §s - g-
-30
AM 7 8
_o_
o
9
I
- - * - 8 - 0 - § - k - *0%-
Transmission From Surface To /m-
TIME
10 II 12 I 2 3 4 5 pm
i i i i I i i
Fig. 5. Transmission for Crystal Lake, August 1935.
These curves show the usual differences between lakes and
the expected differences in transmission between the first meter
and the deeper water ; however, they do not show definite varia¬
tions with the angle of the sun or definite differences between
clear and cloudy days.
In Crystal Lake the transmission of the first meter was 40
per cent, after this about 80 per cent. Trout and Muskellunge
lakes are similar to each other ; Trout Lake showed slightly more
transmission in the first meter, but Muskellunge Lake showed
slightly more transmission after the first meter. Boulder Lake
transmits but 11 per cent in the first half-meter, and 31 per cent
per half-meter after that ; 31 per cent per half -meter is 9.6 per
cent per meter.
These average curves seem to indicate that the transmissions
over the course of a day are approximately constant, and that
the transmission for a clear day is about the same as for a
cloudy day. The second half-meter of Boulder Lake may pos-
186 Wisconsin Academy of Sciences , Arts, and Letters
Fig. 6. Transmission for Boulder Lake, August 1935.
sibly be an exception to the latter statement, but the absolute
values of the readings on this lake were small and significance
cannot be attached to still smaller differences. In many cases
the transmissions were somewhat higher near the noon hour of
a clear day; on the other hand, transmissions near the begin-
MUSKELLUNGE LAKE
i i
O CLEAR
• CLOUDY
- 66 %
Transmission or /m. of Wat or
O
QC
•40£
-30
-20
AM 7
I
; — f — — s — 2s%
Transmission From Surface To / m ■
TIME
6 9 10 II 12 l 2 3 4 5PM
i i i i i i
Fig. 7. Transmission for Muskellunge Lake, August 1935.
Whitney — Solar Radiation Measurements 187
ning or end of the day were often as high as the average for the
entire day. It is evident that total differences in transmission
are small and it is possible that averaging the results for many
days might tend to obscure rather than bring out regular differ¬
ences.
Theoretical Calculations. Theoretical calculations were made
to give a more definite idea of the differences in transmission
which should be expected for various angles of the sun and for
various percentages of sun and sky energy.
If the sky contributed nothing, the problem would be simple ;
in that case all that would be necessary would be a calculation
based on the various path lengths of the sunlight in water. The
following set of figures shows what should be expected for two
different values of transmission:
Table I
Transmission of direct sunlight.
If the transmission of a lake with the sun directly overhead
were 80.0, then with the sun at 60° from the zenith the trans¬
mission should be 74.6; if the transmission with the sun over¬
head were 60.0 the transmission at 60° should be 51.1. When
the sun is 60° from the zenith, light rays under water must
travel about 1.31 m. to reach a depth of 1 m.
The fact that the sky contributes, changes the nature of the
results. To allow for the effect of the sky, it is first necessary
to calculate the mean length of path for diffuse radiation, which,
by definition, is that distance which a single beam of light would
travel and be diminished by the same per cent. If we know this
mean path, the problem of diffuse radiation is reduced to the
problem of a single ray of light ; all the diffuse light energy may
be considered as concentrated in a single beam which travels the
mean distance.
In the calculations, the sky was assumed to be of uniform
brightness ; this assumption is not strictly true but it introduces
no great error. The sky was divided into ten-degree zones near
188 Wisconsin Academy of Sciences , Arts , and Letters
the zenith, but into five-degree zones near the horizon because
of the rapid change of the Fresnel reflection at large angles.
The Fresnel reflection from a smooth water surface is but 2 per
cent with the sun overhead, 6 per cent for the sun 60° from the
zenith, 35 per cent for the sun 80° from the zenith, and 100 per
cent for the sun at the horizon. In computing energies, the fac¬
tors allowed for were: (1) the different areas of sky contribut¬
ing to each zone, (2) the diminishing vertical intensities in pro¬
portion to the cosines of the zenith angles, (3) the changing
Fresnel reflections, and (4) the different path lengths under
water. Each zone was considered separately; energies were
calculated for the levels just above the water surface, just below
the surface, and for meter intervals down to a depth of 6 meters.
Then the total energies for each level were added together. The
following table gives a sample of the results when the transmis¬
sion of the water was assumed to be 70 per cent:
Table II
Mean path lengths of diffuse radiation.
When the total energy striking the water surface from all the
zones of the sky was taken as 100 per cent, the total energy
from all the zones just under the water surface added up to 93.2
per cent. The total energy reaching 1 m. was 61.1 per cent of
the original amount, and so on down to 6 m. where but 7.6 per
cent of the original energy remained. Notice that approxi¬
mately 7 per cent of the total diffuse light energy is lost by re¬
flection in going through the water surface. The energy pres¬
ent at 1 m. is 65.4 per cent of that just below the surface, the
energy at 2 m. is 65.5 per cent of that at 1 m., and the energy at
6 m. is 66.1 per cent of that at 5 m. In other words, for a lake
which transmits 70 per cent of a single beam per meter, the
transmission for diffuse light ranges from 65.4 to 66.1 per cent
in the first six meters. A single beam of light in this water
Whitney — Solar Radiation Measurements 189
would have to travel 119.2 cm. to be reduced to 65.4 per cent of
its original intensity; it would have to travel 116.0 cm. to be
reduced to 66.1 per cent of its original intensity. Poole and At¬
kins (1926) calculated a mean path of 119 cm. for diffuse light,
a result consistent with those found here.
The mean path length diminishes, or the transmission in¬
creases as the light penetrates more deeply. This is because of
the more rapid absorption of the slanting rays and is similar to
what happens when light composed of different colors passes
through water ; the more penetrating wave lengths become pre¬
dominant.
Calculations were also carried through based on transmis¬
sions of 75 per cent and 80 per cent, but the mean path lengths
for the various depths were found to be substantially the same.
Hence the calculations of mean depths may be applied to a band
of wavelengths with varying transmissions; in other words, to
the total radiant energy spectrum penetrating water.
In the calculations for the mean path of combined sun and
sky energy, 118.5 cm. was used as the average path for diffuse
light for all depths and all values of transmission coefficients.
Errors resulting from using this average are negligible. An
average distribution of sun and sky energy given by Kimball
(1919) was used:
The energy reaching any depth can be considered as being made
up of two parts: (1) that part resulting from the general sky
radiation calculated on the basis of a path length of 118.5 cm.,
and (2) that part from the direct sun which is subject to the
Fresnel reflection and path length characteristic of the particu¬
lar angle of sun elevation. The transmissions and mean path
lengths for combined sun and sky were carried through for sev¬
eral values of transmission coefficients. As before, it was found
that the effective path length was practically independent of the
transmission coefficient used, and differed but slightly with
depth. These small differences were ignored. Figure 8 shows
mean path lengths and Figure 9 the corresponding transmissions
for different sun angles. Notice that the mean path length
190 Wisconsin Academy of Sciences , Arts , and Letters
reaches a maximum, and the transmission a minimum near 74°
zenith sun angle. At the beginning and end of the day the values
approached are those characteristic of a diffuse sky.
More complete calculations of mean path lengths were also
made for different angles of the sun and different ratios of sun
Fig. 8. Mean path lengths of total radiation for various angles of
the sun, assuming Kimball's distribution of sun and sky energy.
Whitney — Solar Radiation Measurements
191
and sky energy. The results of these calculations are summar¬
ized in the following tables :
Table III
Mean path lengths of combined sun and sky energy.
The values in this table are to be used when the percentage of sky
energy is measured with a thermopile which has a flat clear glass cover.
Path lengths are in centimeters.
Zenith Angle of Sun
Table IV
Mean path lengths of combined sun and sky energy.
The values in this table are to be used when the percentage of sky
energy is measured either with an uncovered thermopile or with a thermo¬
pile which has a hemispherical glass cover. Path lengths are in centimeters.
Zenith Angle of Sun
192 Wisconsin Academy of Sciences, Arts, and Letters
The percentage of sky energy may be obtained experimental¬
ly merely by taking a reading with the thermopile shaded from
the direct rays of the sun. The difference between Table III
and Table IV is due to the fact that the measured percentage of
sky energy will be slightly different if a thermopile with a flat
glass cover is used instead of one with a hemispherical cover.
However, the differences between the two tables are practically
negligible excepting for large zenith angles.
Fig. 9. Transmissions of total radiation for various angles of the sun,
assuming Kimball’s distribution of sun and sky energy.
Figure 10 shows Table IV in graphical form. If the zenith
angle of the sun were 40°, and the sky energy were 20 per cent
of the total, the mean path length would be 115.0 cm.; if the
zenith angle were 60° and the sky energy 70 per cent of the
total, the mean path would be 122.4 cm. Both tables agree for
these values.
These tables furnish a convenient means of converting read¬
ings, taken under various conditions of sun and sky, to a com-
Whitney — Solar Radiation Measurements
193
mon standard : that is, to zenith sun and 100 per cent sunlight.
The tables could not be used for comparison with the continuous
readings of 1935 and 1936 because the percentages of sky en¬
ergy were not measured. However, the results agreed qualita¬
tively so well with calculations based on Kimball’s distribution
of sun and sky energy, that Tables III and IV were made up
based on the same method of calculation.
Figure 9 shows the variations in transmission which should
be expected for different angles of the sun, when the percentage
of sky energy for a typically clear day follows that given by
Kimball. Transmissions for Crystal Lake should be similar to
the upper curve; transmissions for Trout Lake should be simi¬
lar to the middle curve. These curves show about 7 and 10 per
cent differences in transmission respectively. When one con¬
siders that there are usually variations in the readings due to
waves and other natural causes, that the sun was never closer
than 25° from the zenith (noon at Trout Lake) when readings
were taken, and that readings are quite small after the zenith
angle of the sun is more than 60°, it is quite understandable
that only minor differences in transmission were found with
thermopiles one meter apart. As a matter of fact, in most cases
the theoretical differences in transmission were of about the
same order of magnitude as experimental errors.
The relative importance of experimental errors can be de¬
creased by increasing the distance between light receivers. Sup¬
pose there were a real variation in transmission over the course
of several hours from 0.65 to 0.62, a variation of about 4.5 per
cent. With receivers one meter apart, this variation would be
within the range of usual experimental error. But if the re¬
ceivers were two meters apart, the corresponding transmissions
registered experimentally would be 0.422 and 0.384. These val¬
ues differ by approximately 9 per cent. With receivers four
meters apart the corresponding transmissions would differ by
about 18 per cent. Random experimental errors should then
become relatively unimportant. Of course, if the absolute read¬
ings become very small at the greater depths, the percentage
error in readings will increase, and there may be no net gain.
It is apparent that conclusive comparisons between experimen¬
tal results and theory can only be made on days when conditions
are steady, or with more sensitive receivers placed farther apart.
194 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 10. Mean path lengths of total radiation for various zenith sun
angles and various percentages of sky radiation. Dotted portions of lines
indicate conditions never met with experimentally. For example, one always
finds more than 10 per cent sky radation when the zenith angle of the
sun is 70°. I
Comparisons of Theory with Results. In 1936 a series of
readings were made on Trout Lake with Photox cells. These
cells did not arrive until late in July, and some difficulty was
experienced in calibrating them because of their nonlinearity at
large light intensities. As already mentioned, filters were used
Whitney — Solar Radiation Measurements 195
which transmitted }&> or less of the incident light. In the time
remaining there were few clear days with steady sunshine, but
several records were obtained which showed appreciable differ¬
ences in transmission over the course of a day. It was found
that on the clearest days, the variation in transmission over the
course of the day was the greatest. In the following table the
percentage variation in transmission is listed :
Table V
Per cent differences in transmisson.
On August 6 the Photox cells were 3 m. apart, and the trans¬
mission for this distance varied from about 0.24 at noon to about
0.184 in the early morning and late afternoon. The difference
between these transmissions is about 24 per cent. Had the
Photox cells been 1 m. apart the transmissions would have been
about 0.62 and .57 ; the difference between these transmissions
is about 8 per cent.
Notice that the greater the total energy at noon, the greater
the difference in transmission over the course of a day. More
energy at noon is the result of a clearer sky and a greater per¬
centage of energy from the sun. Figure 10 shows that under such
conditions the mean path length is shorter and consequently
the transmission is greater. Differences in transmission over
the course of a clear day would also be greater. Quantitative
conclusions cannot be drawn from the above data because per¬
centages of sky radiation were not measured; however, the
qualitative agreement with theory is good. For completely cloudy
days no differences in transmission are to be expected and on
records of such days no measurable differences were found.
Figure 9 shows that an upturn in transmission should occur
at the end of the day. Not enough Photox records were ob-
196 Wisconsin Academy of Sciences, Arts, and Letters
tained which were good at the beginning and end of the day to
check this feature; however, a few of the best thermopile rec¬
ords from both years showed good agreement in this respect.
Figure 11 shows experimental transmissions for these days
when the sky was steady and the points quite regular. The
theoretical values shown in comparison were arbitrarily made
to agree with the experimental values at noon; the other theo¬
retical points were calculated on the basis already indicated.
The actual distribution of sun and sky energy was undoubtedly
different from Kimball’s values which were used; nevertheless,
the qualitative agreement is as close as could be expected.
Fig. 11. Comparisons of theoretical and experimental transmissions
over the course of a day. Experimental values are based on the best ther¬
mopile records for both 1935 and 19(316. Horizontal lines at the ends of
the theoretical curves indicate diffuse sky values which are approached at
the beginning and the end of the day.
On the curve for Crystal Lake, August 15, 1935, the experi¬
mental points are different between morning and afternoon;
there is about a 5 per cent difference between the minima. This
difference could have been due to a slight tilt of the lower ther¬
mopile toward the east. However, the general form of the ex¬
perimental curve is in agreement with theory; at both ends of
the day the transmissions show a rise. The averages of morning
and afternoon points are also shown. The other curves, which
Whitney — Solar Radiation Measurements 197
are for Trout Lake, also agree satisfactorily both as to the total
variation in transmission and as to the position of the minima.
The horizontal lines drawn at the ends of the theoretical curves
are the diffuse sky values, and are practically an average of
values over the course of the day.
Discussion
The continuous records obtained with the Cambridge Double
Recorder in 1935 and 1936 have shown the effect of different
elevations of the sun, of clear and cloudy sky, and of waves on
the penetration of light into the waters of small lakes. Although
thermopiles have their limitations because of their relative in¬
sensitivity when compared to photocells, they have proven satis¬
factory down to depths of four meters for continuous records in
the clearest waters; and because of their non-selective proper¬
ties they have proven invaluable as a standard for total radia¬
tion measurements when using other types of light sensitive
instruments. Photox cells were used for continuous radiant
energy measurements at much greater depths. However, care
had to be taken when using them to avoid errors resulting from
their selective wave-length response, and their non-linearity
with large light intensities.
When the average results for the four lakes, Trout, Crystal,
Boulder, and Muskellunge were plotted, each lake showed its
characteristic transmission, but differences throughout the
course of a day appeared to be within experimental error. The
average for cloudy days was about the same as the average for
clear days. Waves cause a spread of points, but the average
seemed about the same as for smooth water.
Theoretical calculations were then made which showed that
for a typically clear day, with allowance being made for both
sun and sky radiation, the transmissions should reach a maxi¬
mum at noon and drop to a minimum a short time after dawn
and a short time before sunset. Higher values of transmission,
characteristic of a diffuse sky, should be found at sunrise and at
sunset. : rrW<;']
Under the conditions that measurements were taken in the
Trout Lake region, the calculated percentage differences in
transmission over the course of a day were of the same order of
magnitude as the usual experimental variations. The calculated
198 Wisconsin Academy of Sciences, Arts, and Letters
diffuse sky values were about the same as the averages of clear
day values. It was evident that only the best individual records
with steady sky conditions, or records obtained with receivers
several meters apart, could be used to check calculated trans¬
missions. The greater the separation of the receivers, the less
the relative importance of experimental errors.
In 1936 readings were obtained with Photox cells separated
by gaps of 3 and 4 meters; results showed definite differences
in transmission over the course of a day and the greatest differ¬
ences occured on the clearest days. A comparison of the best
thermopile records of both 1935 and 1936 with calculated values
showed satisfactory agreement both as to total differences in
transmission and as to the general form of the transmission
curve over the course of a day.
Since the calculation of mean path lengths based on Kim¬
ball's distribution of sun and sky energy over a typically clear
day gave results in agreement with experiment, tables were
made up based on the same method of calculation listing mean
path lengths for various angles of the sun and various percent¬
ages of sky radiation. By use of these tables the transmission
corrected for 100 per cent zenith sun can be computed from any
experimental measurement.
A word of caution should be added. All the work was done
in the epilimnion, and all computations were based on the as¬
sumption of optically homogeneous water. Other experimental
work by the author on the Microstrati fi cation of Inland Lakes
(1937) has shown that there are marked changes in the optical
properties of water in the thermocline and below. Strata of
water are found, often but a few centimeters thick, which have
transparencies greatly different from the water above or below.
The question of the selective absorption of the first meter and
the effect of scattering particles have not been taken up here.
Scattering should tend to make the directional effect of light
less important in deeper water; on the other hand, the greater
distances that slanting rays of any sort must travel (be they
scattered or direct) tends to maintain the relative importance of
the more vertical beams. Work by the author on the Trans¬
mission of Solar Energy and the Scattering Produced by Sus-
pensoids in Lake Waters (1938) shows that the ratio of light
received from the sun and sky to that scattered upward is prac-
Whitney — Solar Radiation Measurements 199
tically constant down to the thermocline. It seems safe to con¬
clude, therefore, that the methods outlined here of calculating
transmissions and energies delivered to different depths for vari¬
ous conditions of sun and sky are usable without appreciable
error from 1 m. to the thermocline.
This investigation was carried on under the direction of
Dr. Birge and Dr. Juday, and the writer wishes to express his
appreciation to them for their suggestions and assistance during
the course of the work.
Summary
1. Continuous records of solar radiation were made with a
Cambridge Double Recorder on several northern Wisconsin
lakes for the measurement of transmission under various condi¬
tions of sun, sky, and water surface.
2. Thermopiles were used as light receivers in 1935 and Pho-
tox cells were used in conjunction with thermopiles in 1936. The
greater sensitivity of the Photox cells permitted their use at
greater depths and showed more positively the differences in
transmission over the course of a day.
3. Theoretical calculations were made which predicted a
maximum transmission at noon on a clear day, minima in the
early morning and late afternoon, and an increase in transmis¬
sion to values characteristic of a diffuse sky at the ends of the
day. These calculations were based on data given by Kimball
of the distribution of sun and sky energies for various zenith
sun angles.
4. Experimental transmissions from the continuous records
agreed satisfactorily with calculated values, both as to the total
change in transmission over the course of the day and as to the
general form of the transmission curve.
5. Reference tables were constructed, calculated on the same
basis, giving mean path lengths of total radiation for any com¬
bination of sun angle and percentage sky energy. These tables
may be used to convert any experimental measurement of trans¬
mission to standard conditions of 100 per cent zenith sun.
200
Wisconsin Academy of Sciences, Arts, and Letters
Literature
Kimball, H. H., 1920. Variations in total and luminous solar radiation
with geographical position in the United States. Monthly Weather
Review. 47:769-793. Washington.
Poole, H. H. and Atkins, W. R. G., 1926. On the penetration of light into
seawater. Jour. Mar. Biol. Assoc 14: 177-198
Whitney, L. V. 1937. Microstratification of inland lakes. Trans. Wis.
Acad. Sci., Arts and Let. 31 : 155-173.
Whitney, L. V. 1937. Transmission of solar energy and the scattering
produced by suspensoids in lake waters. Trans. Wis. Acad. Sci., Arts
and Let. 31: 201-218.
TRANSMISSION OF SOLAR ENERGY AND THE
SCATTERING PRODUCED BY SUSPENSOIDS
IN LAKE WATERS*
Lester V. Whitney
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 70.
Introduction
The transmission of solar energy and the scattering pro¬
duced by suspensoids were measured for a number of northern
Wisconsin lakes in the summer of 1936. Readings of solar en¬
ergy were taken as deep as possible with a thermopile, and then-
overlapped and continued with a photoelectric cell. Energy
curves were plotted and transmissions calculated. A table was
made up listing the total transmissions corrected for 100 per
cent zenith sun and the scattering powers of the suspensoids.
The method of correcting for zenith sun is described in a paper
by the author on the Measurement of Continuous Solar Radia¬
tion in Wisconsin Lakes (1938) .
All measurements of scattered light intensities were taken
with the photocell face down ; the thermopile was not sensitive
enough for these readings. The usual procedure was to take a
series of measurements with the thermopile and photocell face
up and then to follow immediately with the photocell face down.
Simultaneous readings were also made of sun and sky intensity
incident on the water surface; a solarimeter maintained in a
horizontal position in the boat was used for this purpose. Ratios
of scattered light intensities to direct light intensities from sun
and sky were calculated for the various depths.
The scattering coefficients of the suspended particles and the
fractional amounts scattered compared to the amounts of light
purely absorbed, were computed from the ratios of light inten¬
sities and from the total extinction coefficients. These extinc-
This investigation was supported by a grant from the Brittingham Trust Fund.
201
202 Wisconsin Academy of Sciences, Arts, and Letters
tion coefficients were determined in the usual way from the
transmission measurements.
Apparatus
The photoelectric cell and amplifier which were used in this
study have been described in a paper on the Microstratification
of Inland Lakes (1938). In that work, the selective response of
the photocell with wave-length was not a material factor, but in
the measurement of total energies this factor has to be taken
into account. Fortunately, water has a filtering action on light
and radiation is confined to narrower bands of wave-length in
the deeper water. This fact makes it possible to obtain readings
with a selective receiver which are approximately proportional
to total energies. It was found that thermopile and photocell
readings became proportional after the first 4 or 5 meters ; this
was evidence that beyond these depths photocell readings were
approximately proportional to energies.
There is another type of nonlinearity which must be avoided
when using a photocell: at large light intensities the response
of a photocell does not increase in proportion to the incident
energy. This trouble may be avoided by using a diaphragm over
the photocell window to reduce the total light intensity. When
readings were taken overlapping those with the thermopile, a
diaphragm was used which admitted but 2 per cent of the total
light. The lower surface of the window over the photocell was
ground rough to scatter the light more uniformly over the sen¬
sitive receiving surface.
Range of Measurements . Measurements have been taken
with thermopiles to depths where the solar radiation has been
reduced to about 0.01 per cent of that incident on the surface
(Birge and Juday 1931). However, the measurement of such
small energies with a thermopile requires a galvanometer too
sensitive to be used in a boat ; a long cable must be employed to
connect the thermopile with the galvanometer. The photocell
and amplifier were capable of measuring 0.001 per cent of the
incident radiation under ordinary conditions and the entire ap¬
paratus was light enough in weight and small enough in size
to fit conveniently into a rowboat. Under the best conditions,
measurements of 0.00023 per cent ( a little over two-millionths)
Whitney— -Transmission & Scattering of Solar Energy 203
of the incident radiation were made. The microammeter, which
was ordinarily used with the amplifier, was the less sensitive of
the two Rawson instruments used in former solar energy work ;
the usual motion of the boat did not disturb its readings.
Results
Measurements were taken on 16 northern Wisconsin lakes
and curves were drawn of the vertical components of solar radi¬
ation, and of the vertical components of scattered light directed
upward. Energies were plotted against depths on semi-logari¬
thmic paper. On this type of paper a straight line indicates a
constant absorption coefficient and for many lakes the experi¬
mental points fell on practically a straight line. In some lakes
there were marked changes in slope and one may conclude that
sudden changes also took place in the optical properties of the
water. As a rule, the curves of scattered energies from below
were nearly parallel to the solar energy curves and this fact
indicated that the ratio of energies was constant with depth.
This result is reasonable for optically homogeneous water and
will be justified in the discussion on scattering phenomena.
Transmissions . In Figures 1 and 2 the solid lines are the
solar energy curves for 8 of the lakes on which measurements
were taken. They show the percentages of energies uncorrected
for zenith sun; in all cases the energy incident on the water
surface is taken as 100 per cent. The curves are similar to those
found in the past for these lakes.
Figure 1 shows lakes which have medium or relatively high
transmissions. After the first meter, the transmission for Crys¬
tal Lake is about 80 per cent and is practically constant to the
bottom. The transmission for Weber Lake is about 72 per cent
down to 8 m. ; from there on there is a small decrease. In Mus-
kellunge Lake there is a sudden drop in transmission at 10 m. ;
above this level the transmission is about 70 per cent; at 11 m.
it is only 40 per cent, and below this it is somewhat higher. The
region where the sudden change takes place is approximately at
the thermocline and is where a precipitate drop in transparency
was found in the microstratification study. The transmission of
Trout Lake shows less change with depth ; it is 65 per cent for
the first 10 meters, and about 68 per cent below 16 meters. This
result is also consistent with the microstratification study;
204 Wisconsin Academy of Sciences , Arts , and Letters
PER CENT O.OOI .005 .01 .02 .05 0.1 .2 .5 WO 2.0 5.0 10.0 20 50 100
Fig. 1. Solar energy curves for Crystal, Weber, Muskellunge and Trout
lakes. Solid lines show percentages of sun and sky energy at various
depths; dotted lines show percentages of scattered light as measured with
the receiver face down.
transparencies below the thermocline in Trout Lake were as
high as, or slightly higher than, those above.
All the microstratification work was done at a later date
than this work on transmission and scattering and conditions
may have changed somewhat in the interval; however, there
were several instances, such as those just cited, which showed
that, in general, the results of the two studies were consistent.
Figure 2 shows curves for lakes which have lower transmis¬
sions. Transmissions for Little John and Boulder lakes were
quite regular and nearly parallel to each other ; both were about
Whitney — Transmission & Scattering of Solar Energy 205
PER CENT O.OOI 0.01 0.! 1.0 10 »00
Fig. 2. Solar energy curves for Little John, Boulder, Mud, and
Little Star lakes.
33 per cent. The series of readings for Mud Lake were taken
with a very steady sun. A regular curve, which was nearly
straight, was obtained down to 11 m. where the limit of the
apparatus was reached. The bottom was only 2 m. deeper. One
might expect the transmission of Mud Lake to be less regular in
view of the sudden changes in transparency found in the micro¬
stratification study, but solar energy is subject to absorption by
all the water above. Readings were not taken at such small
intervals of depth in this study and the fine detail was not evi¬
dent. In the microstratification study the sudden changes in
transparency were confined to thin layers, but the average trans-
206 Wisconsin Academy of Sciences, Arts, and Letters
parency remained nearly the same from the top of the lake to
the bottom.
Little Star Lake showed the smallest transmission of all the
lakes measured. The transmission was only about 1 per cent
and the limit of readings was reached after only a little over
2 m. Readings were taken every half meter, but they were quite
regular. The high absorption in this lake was due to a luxuri¬
ant growth of algae.
Table I shows the percentage transmissions for all of the
lakes corrected for 100 per cent zenith sun. Two or more values
are listed for some of the lakes, in which the transmission
changed with depth. Total extinction coefficients are given in
the third column and these are connected directly with the meas¬
urements of intensities through the equation
L = Lekx
L and L are the intensities of radiation at two levels x meters
apart; e is the natural logarithmic base and k is the total ex¬
tinction coefficient. When the separation of levels is but 1 m.,
Ii/L = e'k = Transmission
The highest transmissions, about 82 per cent, are shown for
Crystal and Diamond lakes. Next is a large group of lakes with
transmissions ranging from 65 to 75 per cent; these include
Muskellunge, Nebish, Silver, Star, Trout, White Sand, and
Weber lakes. Allequash and Long lakes had transmissions of
about 50 per cent. Boulder, Little John, Mud, and Ruth lakes
were similar, with transmissions around 40 per cent. Little Star
Lake is listed as having a 2 per cent transmission corrected for
zenith sun. This value represents a 100 per cent increase in
transmission over that actually measured, but this lake is an
extreme case and the correction is probably only approximate.
Only one set of readings was made ; readings taken under other
conditions of sun and sky would be needed to verify the accuracy
of the corrected transmission.
Scattering. The transmission per meter, and the extinction
coefficient can be determined directly from intensity measure¬
ments. The percentage of light scattered as compared to that
absorbed and the scattering coefficient can also be determined
from intensity readings, but in not so direct a manner. When
Whitney — Transmission & Scattering of Solar Energy 207
light passes through a small element of volume of lake water it
is partially absorbed and partially scattered, and there is ab¬
sorption and scattering both by the water itself and by the sus¬
pended particles.
In the equation for the diminution of light intensity with
depth
Ix = I0e_kx
the extinction coefficient a may be written :
k = a^ — J— Sw/2 — J— a i. — |— Sp/2
aw and Sw are the absorption and scattering coefficients for pure
water ap and Sp are the absorption and scattering coefficients for
the suspended particles. Only half the scattering coefficients are
included in the expression for k since, on the average, it is as¬
sumed that half the scattered light is downward, and, while it is
no longer part of the primary beam, it continues in the same
general direction.
Sp is a measure of the light scattered by particles from an
elementary volume; a? a measure of the light absorbed. The
quantity scattered depends on the kind of particles, the number
of them per unit volume, and on the wave-length of the light.
The fraction of light scattered compared to that absorbed, or
Sp/(aP + Sp), should be independent of the number of particles
per unit volume and should depend only on the kind of particles
and on the wave-length of the light.
Scattering coefficients and the percentages of light scattered
by suspended particles were computed from measurements of
direct and scattered light intensities. An examination of the
mechanism of scattering showed that these quantities are given
by the following equations :
sp = F(b)kR
Ps = 100
Sp
ap — J— Sp
Ps = 100
F(b)kR
k[l+ F(b)R/2] -0.03
F(b) is an expression involving the mean path length of the
sun and sky light, values of which are listed in the appendix.
For a mean path length of 118 cm., F(b) equals 5.9 and this
208 Wisconsin Academy of Sciences, Arts, and Letters
value was used in computations for Table I. R is the ratio of
the reading of light intensity with the receiver face down to
that with the receiver face up and 0.03 is the numerical value of
the extinction coefficient of pure water. The percentage of light
absorbed by the particles,
Pa = 100 - Ps,
was also computed and listed in Table I. The fraction of light
scattered is approximately equal to F(b)R since the denomina¬
tor in the expression for P» is almost equal to k. Usually
F(b)R/2 is small compared to 1, and 0.03 is small compared to
k. The derivation of these equations requires some detailed
mathematics which is given in the appendix. In the derivation,
the scattering by pure water is assumed negligible as compared
to that by suspended particles. It is further assumed that the
transmission does not change appreciably with depth, since from
1 m. below the surface to the bottom of the epilimnion, curves of
intensity drawn on a semi-logarithmic scale are approximately
straight.
The equation derived for the scattering coefficient may be
written
R = sp/F(b)k,
or for a mean depth of 118 cm.,
R = sP/5.9k.
We should expect the ratio R to be approximately constant
where the water is optically homogeneous, or where the absorp¬
tion and scattering coefficients of the suspended particles are
constant. In most lakes R is about constant; this is shown by
the fact that the experimental curves of scattered and direct
light are nearly parallel on a semi-logarithmic scale. (Figs. 1
and 2).
In Table 1 the third and fourth columns give the extinction
coefficients and the ratios R, from which values of s, P-, and Pa
were computed. For some lakes, values are shown for several
depths. However, the equations for these quantities were de¬
rived on the assumption that the water be optically homogeneous
for some distance below the receiver. For this reason, most
values have been calculated for about the middle of the epilim-
Thcse results are different from those obtained by Williams (1929) in Puget Sound.
His transmissions, based on light scattered upward, were higher, than those found for di¬
rect radiation.
Whitney — Transmission & Scattering of Solar Energy 209
nion; calculations were made at other depths to obtain only a
qualitative idea of changes taking place.
Table I.
Transmission and scattering properties of lake waters .
a is the extinction coefficient , R the ratio of scattered to direct light inten¬
sities, SP the scattering coefficient, Ps the percentage of light scattered by
suspensoids, and Pa the percentage absorbed. The depth closest to the middle
of the epilimnion is indicated in italics.
210 Wisconsin Academy of Sciences, Arts, and Letters
The table shows a range of about 3 to 24 in the percentage
scattering. For example, when suspended material in Mud Lake
removes light from a direct beam, 3.3 per cent of that removed
is scattered in all directions and 96.7 per cent is absorbed; in
Little John Lake the scattering is 23.6 per cent, the absorption
76.4 per cent. Crystal and Trout lakes are between these ex¬
tremes and the scattering is about 13 per cent.
Values of the scattering coefficient Sp, range from 0.022 for
Crystal Lake to 1.06 for Little Star Lake. A smaller correction
to the transmission of Little Star Lake for zenith sun would re¬
sult in a still higher scattering coefficient, and would leave the
percentage scattering practically unaltered. About 50 times as
much light is scattered by particles from an elementary volume
in Little Star Lake as is scattered from a similar volume in
Crystal Lake. The scattering coefficient for Trout Lake is 0.045
and, if any differences due to the spectral composition of the
light be ignored, the figures for Trout and Crystal lakes indi¬
cate that there are about twice as many scattering particles per
unit volume in Trout Lake as in Crystal Lake. Of course, when
scattering coefficients or scattering percentages are given, the
figures apply to the entire mixture of suspended material and to
the kind of light present.
Muskellunge Lake, especially in the epilimnion, presents an
interesting case; its total absorption is low, but its percentage
scattering is practically as high as in Allequash. Its low extinc¬
tion and scattering coefficients probably indicate a lower con¬
centration of particles, but the particles scatter to about the
same extent as those in Allequash. Of course this fact is no
proof that the particles are the same kind.
No attempt was made in this work to determine the effect of
differences in the spectral composition of light on the scattering
coefficients and on the percentages of scattering. In many cases
such differences are undoubtedly important. Values of sP for
Boulder and Muskellunge lakes are about the same, but values of
P* differ by nearly a factor of three. Thus there is about the
same amount of light scattered from elementary volumes in each
lake, but, of the light scattered and absorbed, the suspended
particles in Muskellunge scatter about three times as much as
particles in Boulder. In these lakes the difference in the spec¬
tral composition of the light reaching the epilimnion probably
plays an important role; it is to be expected that the longer
Whitney — Transmission & Scattering of Solar Energy 211
wave-lengths, which predominate in Boulder Lake, should be
scattered less. Further investigations should be made with a
series of color filters to distinguish between differences in scat¬
tering due to the spectral composition of the light and differ¬
ences due to the kinds of suspended particles.
Discussion
Measurements of total transmission and the scattering prop¬
erties of suspensoids were made on 16 northern Wisconsin lakes
in the summer of 1936. The photoelectric cell and amplifier
made possible measurements at greater depths than before. With
the photocell face up, measurements were made of solar energies
down to depths where only a little over two-millionths of the
energy incident on the surface was present. Readings were
taken of scattered light with the receiver face down. A thermo¬
pile was used for the measurement of direct solar intensities in
the upper water and the photocell and thermopile readings were
overlapped.
When intensities were plotted on a semi-logarithmic scale,
the direct and scattered light readings were found to give lines
nearly straight and practically parallel to each other. A straight
line indicates a constant transmission; when lines are parallel
it indicates that there is a constant ratio between the direct and
scattered light intensities. Theory shows that under average
conditions this ratio should be given by the equation
R = s/5.9k
where s and k are the scattering and extinction coefficients re¬
spectively. Hence, where the optical properties of the water
are constant, as in the epilimnion, the ratio of scattered to direct
light intensity should remain constant.
Transmissions, total extinction coefficients, ratios of scat¬
tered to direct light intensities, scattering coefficients, and the
percentages of scattered energies compared to that absorbed
were tabulated for all the lakes measured. The extinction co¬
efficients were obtained directly through the relationship
Transmission = e~k
and the scattering coefficient and the percentage scattering com¬
puted by the equations
212 Wisconsin Academy of Sciences , Arts , and Letters
s = 5.9kR
The derivation of these equations is given in a mathematical
appendix.
There was a wide variation in the transmissions, the values
ranging from 2 per cent or less to 82 per cent. In some lakes,
such as Allequash and Muskellunge, the percentage scattering
was as high as 23; in others, such as Mud and Ruth, the per¬
centage was nearly as low as 3. Similarities in scattering per¬
centages do not necessarily mean similarities in the kinds of
suspended particles, but they do mean that on the average the
percentage of scattering per particle is about the same. Dif¬
ferences in scattering percentages may be due to the different
spectral composition of the light reaching the epilimnion, as well
as to the different kinds of particles. Further work is needed
to distinguish between these factors.
Whitney — Transmission & Scattering of Solar Energy 2lS
Summary
1. Measurements of solar energy were taken on 16 northern
Wisconsin lakes with a thermopile and a photoelectric cell.
Transmissions were determined for depths where but 2 to 10
millionths of the energy incident on the surface remained.
2. Readings were also taken with the photoelectric cell in¬
verted. These measurements were approximately proportional
to the percentage of light energy scattered upwards by sus¬
pended particles.
3. Equations were derived which permitted the computation
of the scattering coefficients, and scattering percentages. The
scattering coefficient is a measure of the light scattered per unit
volume of the scattering material ; the scattering percentage is
a measure of the average scattering per particle compared to
absorption and is independent of the number of particles per
unit volume.
4. A table was made up listing the transmissions and scatter¬
ing properties of the different lake waters ; transmissions ranged
from 2 to 82 per cent, the ratios of scattered light to direct light
intensity ranged from 1/20 to 1/180 and the scattering per¬
centages ranged from 3 to 24 per cent.
Literature
Birge, E. A., and Juday, C., 1932, Solar radiation and inland lakes. Trans.
Wis. Acad. Sci., Arts and Let. 27 : 523-562, Madison.
Whitney, L, V. 1938. Microstratification of inland lakes. Trans. Wis. Acad.
Sci., Arts and Let. 31: 157 Madison.
Whitney, L. V. 1938. Measurement of continuous solar radiation in Wis¬
consin lakes. Trans. Wis. Acad. Sci., Arts and Let. 31 : 177 Madison.
Williams, 1929. Horizontal and upward intensities of light in Puget Sound,
Publications of Puget Sound Biol. Sta., Vol. 7 : 129-135.
214 Wisconsin Academy of Sciences , Arts , and Letters
APPENDIX
DERIVATION OF SCATTERING EQUATIONS
The amount of energy removed from a beam of monochro¬
matic light in passing through a thin layer of water is propor¬
tional to a constant characteristic of the water (or to the extinc¬
tion coefficient), to the intensity of the beam, and to the thick¬
ness of the absorbing layer. This statement may be expressed
by the equation :
dlx = — klxdx (1)
The integration of this equation leads directly to the exponen¬
tial law of diminution of light intensity :
lx = Le~kx (2)
Io is the intensity or energy per sq. cm. per min. at the zero level,
I* is the energy per sq. cm. per min. at a distance x below this
level, e is the natural logarithmic base, and k the extinction co¬
efficient.
In equation 1, if I* and dx were both unity, k would be nu¬
merically equal to dL. Hence the extinction coefficient is the
rate of energy removed per unit light energy in the primary
beam, per unit distance traveled by the primary beam.
Part of the energy which is removed from the primary beam
is removed by absorption, and part by scattering in all direc¬
tions. Let us assume that half the scattered energy is directed
upward, the other half downward. Equation 1 may be rewritten
as the sum of several parts. Part of the energy is removed by
absorption and scattering of light by the water itself ; part by
the absorption and scattering due to the suspended particles.
The scattering of pure water is considered negligible compared
to that of the particles and equation 1 may be stated :
dL = — kw Ldx— aLdx - - Ldx. (3)
dl. = -<k. + a + — |— )I«dx (4)
The first term on the right side of equation 3 gives the energy
removed due to the water itself, the second term the energy
Whitney —Transmission & Scattering of Solar Energy 215
removed due to absorption by the suspended particles, and the
last term the energy removed due to scattering by the suspended
particles. is the extinction coefficient for water, a the absorp¬
tion coefficient of the suspended particles, and s the scattering
coefficient of the suspended particles, s/2 rather than s is used in
these equations because it is assumed that half the energy is
scattered back in the direction of the main beam. A comparison
of equations 1 and 4 shows that
k = kw + a + s/2 (5)
a and s are equal to the energy absorbed and scattered per unit
thickness per unit light intensity by the suspended particles.
We may now obtain an expression for the total scattered
light energy incident on a light receiver placed face downward
i
Io
x
'Light Sea tiered up
Fig. 3. Appendix
216 Wisconsin Academy of Sciences, Arts, and Letters
5K\. .
at a level x — 0. As a simplified, preliminary problem we shall
assume that the primary beam of light is vertical, and that the
light which is scattered backwards travels straight up instead
of in all directions (Fig. 3). The light receiver used may be
considered to have unit cross section. The energy scattered up¬
ward at a depth x from an element of volume 1 sq. cm. in area
and dx in thickness is
|l*dx = — |— Le-^dx. (6)
This amount of energy will be reduced in accordance with the
exponential law (equation 2) in traveling toward the receiver,
and therefore the energy incident upon the receiver will be
di = (-|— Ioe-tedx)e-to=-|-Le-2tedx. (7)
The total energy reaching the receiver is the sum of all such
small increments from all the levels below the receiver :
io = / di = ”2-/ e-2kxdx =-~ ( 8 )
The integration is carried to 00 because the bottom of the lake
is assumed to be far enough off that its contribution is negligible
If the receiver is placed face upwards at the same level it re¬
ceives energy Io, and the ratio of energies received from above
and below is
R = io/L = s/4k. (9)
Solving for the scattering coefficient we have
s = 4kR (10)
The percentage of light scattered compared to the total light re¬
moved by scattering and absorption by the particles is by defi¬
nition
100-
a + s
From equation 5 and from measurements on kw we have:
(ID
a = k —
kw = k
2 2
and substituting this value of a into equation 11 ;
- 0.03,
(12)
Whitney — Transmission & Scattering of Solar Energy 217
Ps
100-
af/2 + k - 0.03
Substituting for s from equation 10 :
(13)
Ps = 100
4kR
k (1 + 2R) - 0.03
(14)
Both the scattering coefficient and the percentage scattering,
equations 10 and 14, are now expressed in terms of the measure-
able quantities k and R. The extinction coefficient k may be
computed directly from transmission measurements, and the
ratio R is simply the ratio of the reading with the receiver face
down to that with the receiver face up.
In the simplified problem just considered it was assumed that
the direct beam (from the sun) came from overhead, and that
the scattering was either straight up or straight down. The
problem becomes more complicated when account is taken of the
fact that the incident beam may penetrate water at an oblique
angle, and that scattering takes place in all directions. A fur¬
ther complication is introduced when sky light as well as direct
sunlight is to be included.
The obliquity of the main beam, and the inclusion of sky
light may be taken care of by using the mean depth in the ex¬
ponential equation (2). The mean depth is by definition the
average length of path which light must travel to reach a depth
of 1 meter. When sunlight and sky light are to be included, the
values calculated for the mean depth b must be such that the
equation
I* = Ioe"kbI (15)
is valid. The mean depth b depends on the angle of the sun and
on the percentage of sky energy. A discussion of the calcula¬
tion of mean depths, and a table of values is given in the paper
by the author on the measurement of continuous solar radiation
in Wisconsin lakes. The differential equation corresponding to
the exponential law above is :
dl* = — kblxdx = — kbIoe"kbx dx, (16)
and this expression is similar to equation 1. Since bdx is the
distance traveled by the light in going through the layer of
thickness dx, the extinction coefficient (k) has the same defini¬
tion as before: it is the energy removed per unit light energy
218 Wisconsin Academy of Sciences, Arts , and Letters
in the primary beam, per unit distance traveled by the primary
beam. If we define I* as the energy falling on a horizontal sq.
cm. per min. (instead of the energy per sq. cm. perpendicular to
the beam) , then the element of volume dV, in which the energy is
removed, is numerically equal to the thickness of the absorbing
layer dx. Equation 16 accordingly becomes:
dL = — kbLe_kbxdx = — kbLe~kbxdV. (17)
We may now write an expression for the energy scattered
from the primary beam (or beams) similar to the last term of
equation 3. Recall that in equation 3 but half the scattered light
was included. The total energy scattered in all directions from
an element of volume is
sblxdV — sbLe~kbxdV. (18)
The scattering coefficient, s, has the same significance as before ;
it is the energy scattered in all directions from a primary beam
of unit light energy traveling unit distance.
Fig. 4. Appendix,
Whitney — Transmission & Scattering of Solar Energy 219
We may assume that scattering particles act like reflecting
spheres to a first approximation. It can be easily proven that
parallel light incident on a perfectly reflecting sphere is re¬
flected with equal intensity in all directions; in the following it
is assumed that scattering particles act in a similar way. The
energy scattered from a small element of volume dV, in the solid
angle AT, will be equal to the total energy scattered in all direc¬
tions times AT/47r, or
sbIoe“kbxATdV
4?r
We may now obtain an expression for the total scattered
energy incident on a receiver 1 sq. cm. in area placed face down
at the level x = 0. Contributions to this total come from all the
particles in the solid hemisphere belowr the receiver. Consider
the contribution from an annular ring, at a depth x below the
receiver (Figure 4). The total energy directed toward the re¬
ceiver from this ring is proportional to dV, the volume of the
ring; to sbIoe~kbx/4*', the energy scattered per unit volume per
unit solid angle ; and to AT, the solid angle subtended by the re¬
ceiver at a point on the ring. As before, the receiver is assumed
to have unit area. Figure 4 shows that
dV — 27rrCdx = (27rxtane) (xde/cos2©) dx
dV = (2^x2sin©/cos3©)d©dx, (20)
and
at = Cos3©/x2 (21)
The energy directed toward the receiver from the annular ring
is:
(sbIoe“kbx/47r) (27rx2sin©d©dx/cos3e) (cos3©/x2)
<*bTo
— — ~ — e kbxsin©d©dx. (22)
This energy must travel through the distance x/cos© before
reaching the receiver; hence the energy incident upon the re¬
ceiver is
e~kbxsin©d©dx e-
kx
COS0
sbL
2
r i + b cos gn
e-kx L cos q J sinededx.
(23)
£20 Wisconsin Academy of Sciences , Arts, and Letters
We may assume that the response of the receiver is propor¬
tional to the total energy incident upon it. This assumption is a
reasonable one because of the very small change in the Fresnel
reflection from a water-glass boundary, and because the inner
surface of the glass cover is assumed, ground rough. Whatever
energy strikes the inner surface of the glass is assumed to be
effective on the receiving surface of the photocell.
The total scattered light energy on the receiver is the sum of
the contributions from all the annular rings at all the levels to
the bottom of the lake. The bottom is assumed to be sufficiently
far off that its contribution is negligible, and therefore the inte¬
gration is carried to oo :
ohTo \ i + b cos e^i
io = — — J J e_kx ^ cos 0 Jsmededx. (24)
Integrating with respect to x we obtain
io
sin o cos Q ,
1 + b cos ©
(25)
This integral is a special case of formula 255 in Pierce, p. 36:
/sir
a
sin © f(cos ©)
de
+ b cos ©
where z = a + b cos ©.
Performing the integration,
- -if1
z — a
dz
lo =
sic fb-loge(l + b)]
2k L b J-
(26)
Experimentally, one measures the ratio R = io/Io, io being the
reading with the receiver face down, and L with the receiver
face up. Substituting in the value of io from equation 26 :
R =
"b — loge ( 1 + b)
2b
and solving for s,
= t^
2b
log. (1 + b)
s = F(b)kR,
kR
(27)
(28)
where
F(b) =
2b
b — loge (1 -f b)
Whitney — Transmission & Scattering of Solar Energy 221
The following values of F(b) were calculated for various path
lengths :
Mean path length b: 1.0 1.1 1.2 1.3 1.4 1.5
6.52 6.14 5.83 5.57 5.35 5.15
F(b):
An average value of F(b) equal to 5.9, based on a mean path
length of 1.18 m., was used for computing all values of the scat-»
tering coefficients and scattering percentages in the present
work. The change in F(b) with mean path length is not large,
and for purposes of comparison an average is satisfactory. Using
this average the equation for s is :
s = 5.9 kR,
and from equations 11 and 12
(29)
(30)
By definition the percentage absorption is
P* = 100 - P8
Values based on these equations are listed in Table I.
(31)
MINERAL CONTENT OF THE LAKE WATERS OF
NORTHEASTERN WISCONSIN
C. Juday, E. A. Birge and V. W. Meloche
From the Department of Chemistry, University of Wisconsin, and the
Limnological Laboratory of the Wisconsin Geological and Natural History
Survey. Notes and reports No. 73.
Introduction
The general glacial character of the Highland Lake District
of northeastern Wisconsin has been discussed in previous re¬
ports, so that only a brief reference need be given to it here.
The lakes of the District are situated in glacial material which
ranges in depth from about 40 to 70 m. (129-234 ft.) and which
consists primarily of sand, gravel and boulders of various sizes.
In some areas the glacial deposit also contains a considerable
amount of clay. In general the deposits are poor in carbonates
and a large part of those present in the original material has
been leached out since the close of the glacial period. This is
true particularly in the areas consisting of outwash material;
the morainal deposits seem to have a somewhat larger amount
of carbonates.
The scarcity of calcium and magnesium carbonates is well
illustrated by the small amounts of these substances found in
the waters of a large percentage of the lakes and also by the
general acidity of the sandy soil. In many cases it would re¬
quire as much as 5 metric tons of lime per hectare to correct
the soil acidity for agricultural purposes.
I. SILICA
Introduction
Quantitative analyses of the silica were made on the surface
waters of 519 lakes. Silica is an important raw material re¬
quired by diatoms and the estimation of the amount available
in the various waters constituted one of the standard chemical
determinations during the time that a general survey of the
lakes in the District was being made.
223
224 Wisconsin Academy of Sciences , Arts , and Letters
The colorimetric ammonium molybdate method of Dienert
and Wandenbulcke (1923) was used for the analyses. Picric
acid was used for the standard solutions and 25.6 mg. of this
acid were regarded as equivalent to 50 mg. of Si03. (Robinson
and Kemmerer 1930).
Data
Surface waters . The results obtained for the surface waters
of the 519 lakes are summarized in Table I. The amount of
silica found in the various waters ranged from none to a maxi¬
mum of 25.6 mg/1 in one lake, namely Muskellunge Lake near
Me Naughton. Second in rank was the Inkpot, a spring-fed
lakelet, with a maximum of 20.0 mg/1. The table shows, how¬
ever, that the surface waters of 267 lakes had less than 1.0 mg/1
of titratable silica; that is, more than 51 per cent of the lakes
contained this small amount of silica. Adding to these the lakes
in the next three groups gives a total of 424 lakes, or 81.7 per
cent of the total number, which had less than 4.0 mg/1 of silica,
and 452 of them, or 87.0 per cent, has less than 5.0 mg/1.
Vertical distribution. The vertical distribution of the silica
was studied in 66 lakes. The results for 8 lakes are given in
Table I
Silica content of the surface waters of 519 lakes. The various lakes
have been grouped at one and two milligram intervals , except the last group
which covers a wider range. The number and percentage of lakes in the
various groups is indicated as well as the mean quantity of silica. Tr.
equals trace.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 225
Table II. There was very little increase in the amount of silica
with increasing depth in some cases, such as Clear and Crystal
for example. A much more marked increase in the lower water
was noted in Anderson, Silver and Trout lakes on the other
hand. In these lakes the increase in the quantity of silica with
increasing depth was correlated with a decrease in the dissolved
oxygen and an increase of free carbon dioxide.
In comparison with these results on Wisconsin lakes, Yoshi-
mura (1932b) found 22.2 mg/1 of silica in the surface water of
Lake Busyu and 22.4 mg/1 at the bottom (20 m.) . Three springs
in the vicinity of the lake ranged from 20.7 to 24.7 mg/1. In 5 East
African lakes, Beadle (1932) found a minimum of 4.0 and a
maximum of 20.0 mg/1 of silica in the surface water.
II. IRON
Introduction
Quantitative studies of the iron content of the waters of the
northeastern lakes were made in 1930, 1932 and 1934. The de¬
terminations were made on unfiltered samples so that the re¬
sults represent the total iron content of the water, including that
which was present in the plankton and other particulate ma¬
terial as well as that which was in true solution. In 1934 de¬
terminations of the ferric and ferrous iron as well as of total
iron were made; these observations were confined to lakes in
which there was very little or no dissolved oxygen in the lower
water and to various samples of ground waters.
Methods
The thiocyanate, ferricyanide and alpha, alpha-dipyridil
methods for iron determinations were examined for these
studies. Speed, precision and convenience were desired because
the determinations were made at the Trout Lake field laboratory.
A common procedure described in Standard Methods of Wa¬
ter Analysis of the American Public Health Association (1936)
includes the determination of total iron by the thiocyanate meth¬
od, ferrous iron by the ferricyanide method and finally ferric
iron by subtracting the ferrous from the total iron. This pro¬
cedure was used in the early studies of the iron content of lake
waters, but it was soon noted that an irregularity existed in the
226 Wisconsin Academy of Sciences, Arts, and Letters
ferricyanide procedure for ferrous iron as described in Stand¬
ard Methods. Not only is the acid concentration too high for the
estimation of small concentrations of ferrous iron, but since
ferroferricyanide shares the properties of colloids, it is affected
not only by too much acid but also by moderate concentrations
of salts. Since the lake waters of northeastern Wisconsin are
fairly soft, the salt interference was not important, and by prop¬
er adjustment of the pH of the samples, the ferricyanide method
could be used for ferrous iron; by proper adjustment is meant
an excess of 0.5 ml of 5 M H2S04 in a 50 ml sample of water.
Most of the latest iron determinations were made by modifi¬
cations of the above procedures. The final methods are as fol¬
lows :
Ferric iron. The samples were acidified by adding 1 ml of
3 N HC1 to 50 ml of lake water. “With the iron samples in readi¬
ness, add 5 ml of the thiocyanate solution to the sample and to
the standards, mix and compare immediately.” (Standard Meth¬
ods, Amer. Public Health Assoc. 8th ed., p. 75, 1936). In Stand¬
ard Methods the water sample is evaporated to dryness before
the determination of the total iron ; this was found to be unnec¬
essary for most of these lake samples and the shorter method
indicated above was used.
Total iron. This was determined by using procedure A of
Standard Methods, p. 74.
Ferrous iron. As already mentioned, the ferricyanide meth¬
od was used during the early part of this study. This procedure
was checked, but later it was replaced by the subtraction method,
namely, Total iron — Ferric iron — Ferrous iron. When the
ferricyanide method is used, the acid should not be present in
excess of 0.5 ml of 5 M H2S04.
Quantitative Results
Surface waters. The iron content of the surface waters of
74 lakes was determined ; the quantity ranged from a trace to a
maximum of 2.00 mg/1. The surface samples of 25 of these
lakes yielded only a trace of iron ; in 18 other lakes the amount
ranged from 0.02 to 0.09 mg/1 and an additional 20 lakes fell
between 0.10 and 0.19 mg/1. This leaves only 11 surface waters
Juday, Birge & Meloche — Waters of N. E. Wisconsin 227
with more than 0.20 mg/1. A maximum of 2.00 mg/1 was found
in Allen or Brazell Lake, a shallow drainage lake with highly
colored water. The Inkpot was second in rank with 1.20 mg/1 ;
the other 9 lakes in this group fell between 0.20 and 0.35 mg/1.
Amounts ranging from a trace to 0.35 mg/1 have been re¬
ported by Birge and Juday (1911) for the surface water of other
Wisconsin lakes, by Kemmerer, Bovard and Boorman (1923)
for lakes in Idaho and Washington, by Minder (1929) for
Stausee Waggital and by Ruttner (1932) for tropical lakes in
Java, Sumatra and Bali. On the other hand, Whipple (1927)
found 0.6 to 1.1 mg/1 of iron in the surface water of Lake Cochi-
tuate and Yoshimura (1931) noted 1.0 to 2.5 mg/1 in the surface
waters of some Japanese lakes.
Vertical distribution . Serial observations covering the en¬
tire depth of 10 lakes were made. The results obtained in some
of these series are given in Table II. In the oligotrophic lakes,
such as Clear and Crystal, the increase in iron content with in¬
creasing depth was not very marked, although the bottom sam¬
ple of Clear Lake yielded eight times as much iron as the sur¬
face. In lakes of the eutrophic type, there was a more marked
increase with depth; Anderson, Nebish and Silver lakes are
good examples of this type. These increases in iron in the lower
water were correlated with minimal quantities of oxygen. Part
of the iron in the lower stratum came from the organic material
which sank into this region and decomposed there and part of
it, probably the larger part, came from the bottom deposits
where the ferric iron was reduced to the ferrous state under
anaerobic conditions and passed into solution. The iron content
of the bottom deposits of these lakes is rather high, constituting
from 1.2 to 1.5 per cent of the dry weight when expressed in
terms of Fe.
The largest amount found in the lower water of any of the
lakes was noted at 18 m. in Silver Lake on August 13, 1932,
namely, 6.9 mg/1 ; second in rank was 6.2 mg/1 at 19 m. in An¬
derson Lake on August 17, 1932. The 14 m. sample taken in
Nebish Lake on August 23, 1932, was third with 3.2 mg/1. The
vertical distribution of the total iron is shown graphically for
Silver and Trout lakes in Figures 1 and 2.
These results occupy an intermediate position among those
that have been reported for the bottom water of other lakes.
228 Wisconsin Academy of Sciences , Arts , and Letters
Table II
Vertical distribution of silica , iron, manganese, dissolved oxygen and
free carbon dioxide in 8 lakes. The results are given in milligrams per liter
of water.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 229
Table II Continued
Ruttner (1932) found a maximum of only 2.3 mg/1 of iron in
the bottom water of the various tropical lakes which he investi¬
gated, while Minder (1929) obtained 8.0 mg/1 at 28 m. in
Stausee Waggital. A maximum of 8.4 mg/1 was reported by
Birge and Juday (1911) in a previous investigation of Wiscon¬
sin lakes. In the bottom waters of Lake Cochituate, Whipple
(1927) found a range from 0.7 in April to 13.2 mg/1 in October.
Yoshimura (1931) obtained a maximum of 56.0 mg/1 in the
bottom water of a Japanese lake and 100.8 mg/1 in the lower
water of Takasuka-numa.
At the end of the summer stagnation period, Einsele (1936)
found a stratification of the iron in Schleinsee; the quantity
ranged from zero at the surface to 1.8 mg/1 at the bottom (11
m.) . It was found only in the lower stratum where the dissolved
oxygen did not exceed 1.4 mg/1. There was a correlation be¬
tween the vertical distribution of the iron and that of the phos¬
phate.
In the lakes of the Suwalki region in Poland, Stangenberg
(1936) found that the iron content of the surface waters ranged
from zero to a maximum of 0.76 mg/1; a larger amount was
found in the surface waters of shallow lakes than in those of
deep lakes. Also much larger amounts of iron were found in
the lower than in the surface waters of eutrophic lakes.
Ground waters. Total iron determinations were made on the
waters of 123 wells and 5 springs. These wells and springs were
230 Wisconsin Academy of Sciences , Arts , and Letters
located on the shores of 49 different lakes situated in various
parts of the lake district. The quantity of iron in the well wa¬
ters ranged from a trace to a maximum of 104.00 mg/1; the
latter amount was found in a sample taken from the well of the
Big Arbor Vitae Resort located on Arbor Vitae Lake. The next
largest amount was noted in a sample from the Stover Cottage
well on Clear Lake near Woodruff, namely 15.00 mg/1, while the
third in rank was the Norris well on Blue Lake with 5.60 mg/1.
The quantity of iron exceeded 1.00 mg/1 in only 24 of the 123
wells, however ; the mean quantity in the 99 wells with less than
1.00 mg/1 was 0.20 mg/1, so that the large majority of the well
waters may be regarded as having relatively small amounts of
iron. This mean, on the other hand, was considerably larger
than that of the surface waters of the lakes.
The total iron content of 4 spring waters ranged from 0.05
to 0.12 mg/1, while the fifth one yielded 3.60 mg/1.
The well and spring waters show the iron content of the
deeper strata of the ground water. Similar observations were
made on the upper stratum of ground water. Some holes were
dug down to the ground water level on the shores of Weber
Lake ; 8 holes were dug at distances of 25 cm. to 7 m. back from
the edge of the water in the lake. The total iron content of the
water which seeped into these holes ranged from 0.10 to 0.35
mg/1 ; the mean for the 8 samples was 0.21 mg/1.
Ferrous and Ferric Iron
During the summer of 1934, observations were made on 6
lakes for the purpose of ascertaining how much of the iron found
in the lower water was in the ferrous state and how much was
in the ferric state. Similar observations were also made on a
number of well waters.
Lake waters. Table III gives the results obtained in some of
the serial observations on lakes. A large proportion of the iron
found in the lower water of these lakes was in the ferrous state.
The analyses showed also that the proportion of ferrous iron
increased with the advance of the summer season; that is, a
larger proportion of ferrous iron was found in August than in
July. The results also showed that the ratio varied in the. differ¬
ent lakes. In Silver Lake for example, the iron found in the
Juday, Birge & Meloche — Waters of N. E. Wisconsin 231
lower water consisted of two-thirds ferrous and one-third ferric
iron; in Nebish Lake on the other hand, about 95 per cent of
the iron in the bottom water was ferrous and only 5 per cent
ferric. A similar result was obtained for the bottom water of
Muskellunge Lake where only about 5 per cent of the total iron
was in the ferric state.
Table III
Results obtained in serial observations on ferrous, ferric and total
iron. The amounts are indicated in milligrams per liter of water. Tr.
indicates trace.
The large proportion of the ferrous iron in the lower water
was due to the scarcity of dissolved oxygen in that stratum as
indicated in Table III. Under this condition, the ferric iron is
reduced to the ferrous state and readily passes into solution. As
already indicated, the bottom deposits in the deeper water of
these lakes contain a rather large percentage of iron and they
are undoubtedly the chief source of the iron content of the lower
water.
Ground waters. Ferrous and ferric iron determinations were
made on the waters of 46 wells. Results of some of these analy¬
ses are given in Table IV ; they have been selected for the pur¬
pose of showing the wide ratios in the various wells. In some
cases the amounts of ferrous and ferric iron were almost or quite
232 Wisconsin Academy of Sciences, Arts, and Letters
equal, as in the Latimer and Meade wells of Trout Lake and in
the North Camp well of Muskellunge Lake. In other cases from
90 to 100 per cent of the iron was in the ferrous state, as in the
South Camp well of Muskellunge Lake, the Sayner Hotel well
of Plum Lake, and the Sand Beach and The Point wells of Trout
Lake. In the sample from the Novak well of Trout Lake, only
about 10 per cent of the iron was in the ferrous state. The
ratios of ferrous to ferric iron in the various wells do not seem
to be correlated with the dissolved oxygen or the carbon dioxide
content of these well waters.
Table IV
This table shows the range of the ferrous and ferric iron in a number
of well waters, together with the oxygen and carbon dioxide content. The
results are stated in milligrams per liter.
III. MANGANESE
Manganese plays a more or less important role in the growth
of plants and it is also used to a certain extent by some animals.
Some analyses were made, therefore, to determine the mangan¬
ese content of some of the lake water.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 233
Methods
The manganese content of lake waters was determined by
the persulfate method using the procedure described in Standard
Methods of Water Analysis, A. P. H. A., p. 84, 1936. Manganous
sulfate was used as the color standard.
Data
Observations were made on the manganese content of the
waters of 8 northeastern lakes in the summer of 1932 ; they con-
Fe and Mn
0.0 0.2 0.4 0.6 08 1.0 1.2
Fig. 1. Vertical distribution of total iron, manganese and dissolved
oxygen in Trout Lake, August 22, 1932.
234 Wisconsin Academy of Sciences , Arts , and Letters
sisted of 9 series of samples covering the entire depth of these
lakes.
All of the series showed a more or less marked increase in
the quantity of manganese with increasing depth. Only a very
small amount was found in the surface water ; it ranged from a
minimum of 0.003 mg/1 in 3 of the lakes to a maximum of 0.023
mg/1 in the surface of Trout Lake. In the bottom water, the
amount varied from a minimum of 0.013 mg/1 in Crystal Lake
to a maximum of 1.2 mg/1 in Trout Lake. The results obtained
in some of these series are given in Table II. The increase with
depth was smallest in the oligotrophic lakes, such as Clear, Crys¬
tal and Weber. A more marked increase was noted in the lower
water of Nebish Lake and still greater ones in Silver and Trout
lakes. The maximum amount was obtained in the bottom water
of Trout Lake, namely 1.20 mg/1.
Figures 1 and 2 show the vertical distribution of manganese
and iron in Silver and Trout lakes. In both cases the major part
of the increase in manganese was found near the bottom and not
distributed throughout the hypolimnion. These results on Wis¬
consin lakes are similar to those obtained by Yoshimura (1931)
on Japanese lakes; he found a slight stratification in all of the
lakes that he tested and the most marked increase in manganese
came near the bottom. On the other hand, some of his results
were quite different; he found a maximum of 8.0 mg/1 in one
lake in comparison with a maximum of 1.2 mg/1 in Trout Lake
and the quantity of manganese was about one-fifth as large as
that of iron in his lakes. In the Wisconsin lakes, the ratio of
these two substances showed a rather wide variation, with the
amount of iron usually from five to ten times as large as that of
the managanese. Trout Lake, however, was an exception; in
this case the quantity of iron and manganese was about the same
down to a depth of 25 m. and below this depth the amount of
manganese was larger than that of iron ; the former was almost
four times as large as the latter at the bottom (32 m.) .
Kemmerer, Bovard and Boorman (1923) reported that no
manganese was found in the surface waters of three Idaho lakes
and one in Washington. Lake Pend Oreille in Idaho, however,
yielded 0.003 mg/1 which corresponds to the minimum amount
found in the surface waters of Wisconsin lakes.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 235
fe 1.0 a.0 3.0 4.0 5.0 6.0 70
Fig. 2, Vertical distribution of total iron, manganese and dissolved
oxygen in Silver Lake, August 13, 1932, Compare with Fig. 1.
Wiebe (1930) found that the manganese content of filtered
samples of Mississippi River water varied from a minimum of
0.044 mg/1 to a maximum of 0.128 mg/1 between May and Octo¬
ber 1929. These amounts are considerably larger than those
noted in the surface waters of Wisconsin lakes.
IV. CALCIUM
Method. The procedure used for the determination of cal¬
cium in lake waters or in lake water residues was a modification
of the Me Crudden method, which was described by Meloche and
Setterquist (1933). In this method the calcium oxalate, which
is precipitated, is separated by means of a centrifuge and finally
titrated in a hot acid solution by means of 0.01 N permanganate.
Surface waters. Quantitative determinations of the calcium
were made on the surface waters of 358 lakes. Of this number,
128 have neither an outlet nor an inlet and thus belong to what
has been called the seepage type ; their waters are very soft be¬
cause they are derived chiefly from precipitation. The other
230 have outlets and, therefore, belong to the drainage class of
lakes. The quantity of calcium ranged from a minimum of 0.13
mg/1 to a maximum of 18.8 mg/1. The smallest amount was
236 Wisconsin Academy of Sciences, Arts, and Letters
found in Keego Lake, a typical bog lakelet belonging to the
seepage type, while the largest amount was obtained in Little
Martha Lake, one of the Mercer chain, which is a spring fed
lake of the drainage type.
The various lakes have been separated into groups on the
basis of the calcium content of the surface waters and the results
are given in Table V. The largest number of lakes falls in the
0.0-0. 9 mg/1 group, namely 76 or 21 per cent of the total num¬
ber included in the table. The 1. 0-1.9 mg/1 group ranks second,
with 50 lakes or 14 per cent of the total number. These two
groups are made up chiefly of seepage lakes which have very
soft waters; there are 72 seepage lakes in the first group and
38 in the second. Only 19 of the 128 seepage lakes, in fact,
yielded more than 1.9 mg/1 of calcium and 8 of them fell in the
2.0-2.9 mg/1 group, with 3, 4 and 3 respectively, in the following
three groups. The maximum quantity of calcium found in the
seepage type was 6.1 mg/1 in Forest Lake. The 8.0-8.9 mg/1
group was third in rank, with 31 lakes all belonging to the drain¬
age type. The two groups between 16.0 and 17.9 mg/1 were not
represented and only 2 lakes fall in the 18.0-18.9 group.
In general the drainage lakes receive water from larger
basins than the seepage lakes and these inflowing waters contain
larger amounts of carbonates. Some of the drainage lakes, how¬
ever, do not have inlets, but they serve as the sources of small
streams during high water. Thus, while they are classed as
drainage lakes, they depend chiefly upon precipitation for their
water supply; as a consequence their waters are very soft and
they fall in the low calcium groups. The first three groups, for
example, contain 4, 13 and 9 drainage lakes, respectively, which
is 11 per cent of the total number of drainage lakes included in
the table; that is, the surface waters of 26 drainage lakes
yielded less than 2.0 mg/1 of calcium. On the other hand, only
12 of the 358 lakes yielded 13.0 mg/1 or more, so that the sur¬
face waters of 220 lakes, or a little more than 61 per cent of the
total number, contained between 2.0 and 12.9 mg/1 of calcium.
Table V also shows that the calcium content of the surface wa¬
ters of 237 lakes, or 66 per cent of those on which determina¬
tions were made, did not exceed 6.9 mg/1.
The distribution of the lakes in the various groups is shown
graphically in Figure 3. This diagram illustrates clearly the
Juday, Birge & Meloche — Waters of N. E. Wisconsin 237
large percentage of lakes falling in the two groups with less than
2.0 mg/1 of calcium. The groups between 4.0 and 8.9 mg/1 show
the next highest percentage of lakes. The 18.0-18.9 mg/1 group
is not represented in the diagram.
Fig. 3. Distribution of the lakes in the various calcium groups, based
on the percentage of the total number of lakes included in Table V.
Vertical distribution . Observations on the vertical distribu¬
tion of the calcium were made on 11 lakes ; the number of sam¬
ples used for each series ranged from 2 to 7, depending chiefly
upon the depth of the various lakes. The results obtained for
7 lakes are shown in Figure 4. The curves for Clear and Nebish
lakes (A and B) are the means of two series of determinations,
while those for Muskellunge, Silver and Trout lakes (D, E, F)
are the means of three series of readings on each lake ; only one
series was taken on each of the other 2 lakes. The lakes repre¬
sented in this diagram cover the range from somewhat more
than 2.0 mg/1 in Clear Lake to 10.0-15.0 mg/1 in Presque Isle
and Trout lakes.
The curves show only a small difference in calcium content
between surface and bottom in some cases, such as Clear and
Little Long (A and C) ; in Clear Lake the surface water yielded
2.1 mg/1 and the sample at 27 m. (1 m. above the bottom) 2.5
mg/1, while the readings in Little Long were 4.2 mg/1 at the sur-
238 Wisconsin Academy of Sciences , Arts, and Letters
Fig. 4. Vertical distribution of the calcium in 7 lakes. Curve A= Clear;
B=Nebish; C=Little Long; D == Miiskellunge ; E^Silver; F^Trout; G=
Presque Isle.
face and 4.3 mg/1 at 17 m. In Trout Lake the difference was
somewhat larger; the quantity ranged from 11.7 mg/1 at the
surface to 12.5 mg/1 at 30 m. A similar result was obtained on
Big Lake, which is not included in the figure, where the calcium
content was 14.7 mg/1 at the surface and 15.3 mg/1 at 17 m.
(1 m. above the bottom) . In Muskellunge Lake, the surface was
6.4 and 19 m. 7.5 mg/1, while the surface was 2.2 mg/1 in Nebish
Lake and 14 m. was 3.5 mg/1.
The largest difference was noted in Presque Isle Lake where
the surface sample yielded 10.6 mg/1 of calcium and the bottom
(28 m.) 14.9 mg/1; this is a difference of 4.3 mg/1. The greater
part of this difference was noted in the upper 10 m. ; at the lat¬
ter depth the calcium content was 13.9 mg/1, or 3.3 mg/1 more
than the amount at the surface.
The comparatively small increase in the calcium content of
the bottom stratum during the summer is due chiefly to the small
amount- of this material that is found in the bottom deposit of
the deeper water. In 24 lakes the quantity of calcium in these
deposits ranged from a minimum of 0.21 per cent to a maximum
Juday f Birge & Meloche — Waters of N. E . Wisconsin 239
1.74 per cent of the dry weight. In 7 of the 12 lakes on which
vertical series of calcium were taken, the percentages of calcium
in the bottom deposits ranged from a minimum of 0.24 per cent
in Crystal Lake to a maximum of 1.36 per cent of the dry weight
in Lake Mary. The bottom water of some of these 7 lakes con¬
tained from 5.0 to 15.0 mg/1 of free carbon dioxide, which made
conditions favorable for the solution of the calcium in the bottom
deposits, but the scarcity of this element in the mud prevented
any marked increase in the calcium content of this stratum of
water. Owing to the presence of free carbon dioxide in the
epilimnion of most of these lakes and also owing to the relatively
small amount of calcium generally present in this stratum, there
is no precipitation of calcium in the upper water resulting from
the photosynthetic activities of the phytoplankton and the large
aquatic plants, hence the upper stratum does not make any con¬
tribution of calcium to the lower water as a result of these activ¬
ities. Similar conditions have been noted by Yoshimura (1932a)
in Japanese lakes.
Calcium and Bound Carbon Dioxide
The relation between the bound carbon dioxide and the cal¬
cium content of the waters of these lakes was discussed in a
previous report (Juday, Birge and Meloche 1935). The results
showed that an increase of 2.0 mg/1 in bound carbon dioxide
was correlated with an average increase of 1.1 mg/1 in the cal¬
cium content of the water. When these lakes are grouped on
the basis of their calcium content, an increase of 1.0 mg/1 in
calcium is accompanied by an increase of approximately 1.5
mg/1 in bound carbon dioxide as shown in Table V. The relation
between calcium and bound carbon dioxide is shown graphically
in Figure 5. The solid line curve shows the mean quantity of
bound carbon dioxide in the various calcium groups ; the upper
broken line shows the maximum and the lower broken line the
minimum quantity of bound carbon dioxide in the several groups.
These maxima and minima are given in Table V. The largest
difference between maximum and minimum is found in the 11.0-
11.9 mg/1 group where the range is from 15.3 to 28.0 mg/1, a
difference of 12.7 mg/1 ; differences of more than 10.0 mg/1 are
indicated for 5 other groups. The smallest range is shown by
the 0.0-0.9 mg/1 group.
240 Wisconsin Academy of Sciences , Arts , and Letters
Table V
Calcium content of the surface waters of 858 lakes and its relation
to the hound carbon dioxide. The results are stated in milligrams per liter .
While there is a wide variation in the amount of bound car¬
bon dioxide in each calcium group, the means of the various
groups show a regular and consistent increase correlated with
the increase of calcium. Similar correlations have been noted
in some of the other characteristics of these lake waters, such
as the relation between specific conductance and the calcium and
magnesium content of the surface waters (Juday and Birge
1933) and that between the quantity of oxygen consumed and
the amount of organic carbon present (Juday and Birge 1932).
When a group consists of 15 or more lakes, its mean is gen¬
erally consistent with the means of other groups having a corre¬
sponding number of lakes.
Calcium Content of Well Waters
Calcium determinations were made on the waters of 49 wells
situated on the shores of 17 lakes ; the results obtained for some
these lake and well waters are given in Table VI. The largest
number of well waters was obtained from the shores of Trout
Lake, namely 24 ; next in order was Muskellunge Lake with 6.
In the 24 wells situated on the shores of Trout Lake, the quantity
of calcium varied from a minimum of 1.6 mg/1 in the Kern well
to a maximum of 59.4 mg/1 in one well at Rocky Reef ; a second
Juday, Birge & Meloche — Waters of N. E. Wisconsin 241
0 2 4 6 8 10 12 14 16
Fig. 5. Relation of the calcium content of the surface waters to the
bound carbon dioxide content. The solid line curve represents the mean
carbon dioxide content of the different calcium groups. The upper broken
line curve shows the maximum quantity of bound carbon dioxide in each
group and the lower broken line curve the minimum amount.
well at Rocky Reef yielded 35.6 mg/1, which was the second lar¬
gest amount found in the well waters. The mean quantity for
the 24 well waters is 13.8 mg/1 as compared with 11.7 mg/1 in
the surface water of the lake. In 15 well waters, the amount of
calcium was smaller than that of the surface water of the lake,
while in 9 samples it was larger. The well with maximum cal¬
cium content had five times as much as the surface water and
that with minimum had only one-seventh as much as the surface
water.
In the 6 wells situated on the shores of Muskellunge Lake,
the quantity of calcium varied from 3.2 to 11.7 mg/1, while the
surface water of the lake yielded 6.4 mg/1. The mean calcium
content of the 6 well waters is 5.8 mg/1, or 0.6 mg/1 below that
of the surface water of the lake. While Muskellunge Lake is
classed as a drainage lake, it does not have a permanent outlet ;
242 Wisconsin Academy of Sciences , Arts, and Letters
Table VI
The calcium content of lake and of corresponding well waters. The
results are given in milligrams of Ca per liter.
the outlet stream functions only at high stages of the water. A
small stream of water flowed out of the lake in 1930, but there
was no overflow from 1931 to 1936 inclusive.
The calcium content of the 4 wells situated on the shores of
Black Oak Lake ranged from 4.6 mg/1 to 14.4 mg/1, while that
of the surface water of the lake was 5.9 mg/1. The mean of the
4 well waters is 7.8 mg/1, or approximately 2.0 mg/1 larger than
that of the surface of the lake. Black Oak Lake also belongs to
the type which has intermittent outlets.
The 4 well waters from the shores of Clear Lake had a larger
calcium content than the surface water of the lake ; the quantity
ranged from 3.0 to 7.5 mg/1 in the wells as compared with 2.1
mg/1 in the lake water. The mean quantity found in the 4 wells
Juday, Birge & Meloche — Waters of N. E. Wisconsin 243
is 5.2 mg/1. Clear Lake belongs to the seepage type, so that it
has very soft water.
Only one sample of well water was obtained from each of
13 lakes. The calcium content of the well water was larger
than that of the surface water of the corresponding lake in 9
instances and smaller in the other 4. The largest percentile
difference was noted in Little Bass Lake where the surface water
of the lake yielded 0.6 mg/1 of calcium and the well water 7.8
mg/1; thus the well water yielded thirteen times as much cal¬
cium as the lake water. The largest quantitative difference was
found at Tenderfoot Lake where the surface water contained
11.8 mg/1 and the well water 27.5 mg/1, a difference of 15.7
mg/1. The next largest difference in this group was noted at
Lake Mamie, with 5.3 mg/1 in the surface water and 18.1 mg/1
in the well water.
In the 4 samples in which the lake water had more calcium
than the well water, the maximum difference was obtained at
Mercer Lake where the surface water yielded 11.2 mg/1 and the
well water 3.6 mg/1. The surface water of Forest Lake yielded
6.1 mg/1 of calcium and the well water 5.8 mg/1, while a spring
located on the shores of Little Papoose Lake yielded 7.6 mg/1
and the surface water of the lake 12.5 mg/1. In Little John Lake,
the calcium content of the surface water was 9.4 mg/1 and that
of the well water was 8.5 mg/1. Three of the lakes in which
the calcium content of the surface water was smaller than that
of the well water belong to the seepage type and 6 to the drain¬
age class ; in the group where the calcium of the surface water
exceeded that of the well and spring water, one belongs to the
seepage and 3 to the drainage type.
The general results obtained in these analyses show that the
well waters from the shores of lakes whose surface waters have
2.0 mg/1 of calcium or less yield more calcium than the lake
waters, but well waters from the shores of lakes with 3.0 mg/1
or more may contain either a larger or a smaller amount of
calcium than the corresponding lake waters. There was no direct
correlation between the depths of the wells and the calcium con¬
tent of their waters. On Trout Lake for example, the maximum
quantity of calcium was found in a well that was only 5 m. deep
and the minimum in another that was 6 m. in depth.
244 Wisconsin Academy of Sciences, Arts, and Letters
Calcium Content of Upper Ground Water
Samples of ground water were obtained along the margins of
three seepage lakes (Crystal, Long and Weber) by digging shal¬
low holes along the margins and letting the ground water seep
into them. These holes were sunk into the beach at distances of
25 cm. to 12 m. from the edge of the water in the lake. This
method was followed during the summer of 1934, but in 1935
a well-point was also driven into ground to a depth of 25 cm. or
more and samples were taken from it. The samples of water
secured by these two methods represented only the upper stra¬
tum of ground water surrounding the lake. The well waters
represented deeper strata since they came from depths of 4 m. to
25 m.
The calcium content of the upper ground water taken from
these seepage holes and from the well-point showed a wide vari¬
ation. In some cases the amount was substantially the same as
that of the surface water of the lake; in other instances the
ground water samples yielded twelve to seventeen times as much
calcium as the surface water of the lake. Most of these ground
water samples yielded only two to four times as much calcium
as the surface water of the lake. The results show that the
ground water problem in the vicinity of the lakes is a complex
one which needs a more extended investigation. No definite
conclusions can be drawn from the data now in hand.
V. MAGNESIUM
Method . Some early determinations of magnesium were made
by the usual macro silicate procedure in which magnesium am¬
monium phosphate is precipitated and ignited to magnesium
pyrophosphate. Later determinations were made by the pre¬
cipitation of the magnesium oxyquinolate and the subsequent
titration by permanganate. This procedure proved to have sev¬
eral undesirable features and its use was discontinued. All re¬
cent magnesium determinations have been made by the micro
precipitation and centrifuge separation of magnesium ammo¬
nium phosphate; the phosphate was then dissolved and its con¬
centration determined by the ceruleomolybdate method. Some
determinations were also made by the titan yellow method, but
due to the enhancement of the magnesium color by the presence
Juday, Birge & Meloche — Waters of N. E. Wisconsin 245
of calcium, this method was applied only to samples of relatively
soft water which were low in calcium.
Surface waters. Quantitative analyses of magnesium were
made on the surface waters of 231 lakes; of this number 59
were seepage and 172 were drainage lakes. The quantity of
magnsium in these waters varied from a minimum of none to a
maximum of 6.5 mg/1, but the great majority of them yielded
relatively small amounts. The first part of Table VII gives a
summary of the results obtained in the analyses where the vari¬
ous lakes are grouped on the basis of the magnesium content of
their surface waters.
Table VII
Magnesium content of the surface waters of 231 lakes and its relation
to the calcium and the bound carbon dioxide content of the same waters.
The results are stated in milligrams per liter of water. The lakes are
grouped according to the magnesium content of the water.
The maximum number of lakes falls in the 2. 0-2.9 mg/1
group, namely 59, or a little more than 25 per cent of the total
number. The 3. 0-3. 9 mg/1 group ranks second with 52 lakes
and the 1.0-1. 9 group third with 44. The first four groups con¬
tain 196 lakes, or about 85 per cent of the total number. Adding
the 4.0-4.9 mg/1 group gives a total of 221 lakes, or more than
95 per cent of the total, with less than 5.0 mg/1 of magnesium.
In comparison with this, Table V shows that only 52 per cent
of the lakes had less than 5.0 mg/1 of calcium.
Thirty-seven of the 41 lakes in the 0.0-0.9 mg/1 group belong
to the seepage type and 18 of the 44 in the second group. This
leaves only 3 seepage lakes in the third group and one in the
fourth; all of the seepage lakes, therefore, yielded less than 4.0
mg/1 of magnesium. The drainage lakes which belong to the
246 Wisconsin Academy of Sciences, Arts , and Letters
first two groups represent bodies of water that have no inlets,
but which have temporary outlets at high stages of the water;
as a result they show some of the characteristics of seepage
lakes.
A maximum of 6.5 mg/1 of magnesium was found in the
surface waters of Spring and Sweeney lakes.
Magnesium and Calcium
Calcium determinations were made on all of the lakes in¬
cluded in the various magnesium groups. The analyses thus
show the quantitative relation between these two substances in
the surface waters of the 231 lakes. This relation is indicated
in the second part of Table VII. It shows that the mean quan¬
tity of calcium gradually increased with increasing amounts of
magnesium; that is, the mean calcium content rose from 1.2
mg/1 in the 0.0-0.9 group to 14.7 mg/1 in the 6.0-6.9 group.
The individual lakes, however, showed a wide range in the
ratio of calcium to magnesium. In some waters the two were
present in substantially equal amounts, while in a few samples
the amount of magnesium was somewhat larger than that of
calcium. In the great majority of the lakes, on the other hand,
the quantity of calcium exceeded that of magnesium; in some
samples the former was six times as large as the latter.
The small calcium and magnesium content of the waters of
the northeastern lakes in comparison with that of Lake Mendota
is shown in Table VIII. Lake Mendota, for example, has more
than sixty-five times as much calcium per hectare of surface as
Weber Lake, almost thirty-five times as much as Crystal, more
than six times as much as Muskellunge and almost twice as much
Table VIII
Comparison of the Ca and Mg content of the waters of five north¬
eastern lakes with that of Lake Mendota.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 247
as Trout Lake. Likewise the magnesium content of the north-
eastern lakes is very much smaller per unit area than that of
Lake Mendota. The surface water of Mendota contains about
sixty-five times as much magnesium per hectare as Weber, ap¬
proximately forty-four times as much as Crystal, nearly twenty
times as much as Muskellunge and more than five times as much
as Trout Lake.
Magnesium and Bound Carbon Dioxide
The mean bound carbon dioxide content of the surface wa¬
ters falling in the various magnesium groups showed a regular
increase with increasing magnesium content. It rose from 1.8
mg/1 in the 0. 0-0.9 group to 21.6 mg/1 in the 6. 0-6. 9 group.
(Table VII.)
The individual lakes, however, show a wide range in the
amount of bound carbon dioxide in the different magnesium
groups. The largest range was found in the 2. 0-2.9 group where
the minimum was 1.5 mg/1 and the maximum 24.3 mg/1 of bound
carbon dioxide; this represents a sixteenfold difference. The
water in the latter sample contained 15.4 mg/1 of calcium and
only 2.7 mg/1 of magnesium, so that the large bound carbon
dioxide content was due chiefly to the large amount of calcium.
The minimum quantity of bound carbon dioxide in this group,
namely 1.5 mg/1, was correlated with 1.4 mg/1 of calcium and
1.0 mg/1 of magnesium.
A fifteenfold difference between minimum and maximum
amounts of bound carbon dioxide was found in the 1.0-1. 9 mag¬
nesium group and a six-fold one in the 0.0-0.9 group. The limits
in the other groups did not exceed threefold.
Hardness
Natural waters that contain relatively small amounts of cal¬
cium and magnesium are regarded as “soft” and those that con¬
tain comparatively large amounts are called “hard”. Iron and
aluminum also contribute to the hardness of water, but they are
present in such small quantities in the northeastern lake waters
that they may be disregarded. The major portion of the hard¬
ness of these waters is due to carbonates because sulphates and
chlorides are present only in small amounts.
248 Wisconsin Academy of Sciences, Arts, and Letters
Several methods are used in stating the degree of hardness
of water, but the results of these analyses are given in the last
column of Table IX in terms of milligrams of CaC03 per liter of
water. These results are based on the analyses of the 231 lake
waters on which both calcium and magnesium determinations
were made. The mean quantity of calcium and magnesium has
been used for the hardness computation. The table shows that
the hardness increased from a minimum of 6.5 mg/1 of CaC03
in the 0.0-0.9 calcium group to a maximum of 65.7 mg/1 in the
16.0-18.9 group.
Table IX
Calcium and magnesium content of the surface waters of 231 lakes.
The lakes are grouped on the basis of the calcium content of the surface
waters. Results are given in milligrams per liter of water. The hardness
of the water is indicated in milligrams of CaCOz per liter.
Various standards have been proposed for the purpose of
indicating the general degree of hardness of waters. In one
classification, waters with a calcium carbonate hardness of 30-
50 mg/1 are considered “very soft”, 50-100 mg/1 “moderate” and
100-300 mg/1 “hard”. The results given in Table IX show that
215 of the northeastern lake waters fall below the 50 mg/1 stand¬
ard, and therefore, belong to the “very soft” group in this classi¬
fication, while all of the others, 16 in number, belong to the
“moderate” group. Furthermore, 115 of these lake waters, or
50 per cent of the total number, fall below the 30 mg/1 minimum,
so that they may be regarded as “extremely soft”.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 249
One degree of hardness in the French system of classifica¬
tion represents 10 mg/1 of CaC03 ; on this basis 65 of these
lakes do not exceed one degree of hardness and none of them
exceeds 7 degrees. In the German system, one degree of hard¬
ness is equal to 10 mg/1 of CaO, which is equivalent to 17.9 mg/1
of CaC03 ; according to this standard, 75 of the lakes have less
than one degree of hardness and none of them exceeds 4 degrees.
Results of Other Investigators
A relatively small number of lakes with such small quantities
of calcium and magnesium have been reported by other investi¬
gators. Delbecque (1898) lists 4 French lakes in which the
amount of calcium ranged from 2.0 to 4.0 mg/1 and the magnesi¬
um from 0.6 to 1.5 mg/1. Halbfass (1923) records 4 lakes with
0.7 to 1.0 mg/1 of calcium and 0.4 to 0.6 mg/1 of magnesium.
Pesta (1929) lists 5 elevated alpine lakes in which the calcium
falls between 0.7 and 2.0 mg/1 and the magnesium between 0.11
and 1.5 mg/1.
Yoshimura (1932) found 2.6 mg/1 of calcium at the surface
and 3.4 mg/1 at 20 m. in Lake Rusyu; he also determined the
calcium content of 83 other Japanese lakes. In 40 of these lake
waters, the quantity of calcium ranged from zero (6 lakes) to
5.0 mg/1, while 29 fell in the 5-10 mg/1 group. These Japanese
lakes, therefore, are like those of Wisconsin in that a large per¬
centage of them are poor in calcium.
Ueno (1934) obtained only 2.0 mg/1 of calcium in the surface
water of a Japanese lake situated in the Nikko mountain range ;
he also lists 4 other lakes in this region which fall between 3.7
and 5.0 mg/1. There was a marked seasonal variation in the
calcium content of some of the lakes which he investigated. In
one lake the calcium decreased from 11.5 mg/1 in June to 7.5
mg/1 in October and in another from 6.5 mg/1 in June to 3.5
mg/1 in October. Such marked seasonal changes have not been
found in any of the Wisconsin lakes.
Ohle (1934) made chemical analyses of the waters of a con¬
siderable number of north German lakes. He lists two lakes
which yielded less than 2.0 mg/1 of calcium and two others which
had between 2.0 and 3.0 mg/1; all of those with less than 10.0
mg/1, he regards as calcium poor. With respect to the ratio
250 Wisconsin Academy of Sciences, Arts, and Letters
between calcium and magnesium, Ohle records one lake which
had 75.5 mg/1 of calcium and only 2.7 mg/1 of magnesium; this
represents a twenty-eightfold difference.
In comparison with the above records, 126 out of 358 Wiscon¬
sin lakes included in this report had less than 2.0 mg/1 of calcium
and 186 of them yielded less than 5.0 mg/1. The surface water
of one lake yielded only 0.13 mg/1 of calcium and two others only
0.16 mg/1. The surface waters of 85 northeastern lakes con¬
tained less than 2.0 mg/1 of magnesium and none of them ex¬
ceeded 6.5 mg/1. A large percentage of the lakes of northeastern
Wisconsin, therefore, may be regarded as very poor in calcium
and magnesium.
VI. FLUORINE
Observations were made on the fluorine content of some of
the northeastern waters during the summer of 1935. The meth¬
od devised by Willard and Winter (1933) was used for these
determinations. The waters of 7 lakes, 4 springs and 13 wells
were analyzed. The quantity of fluorine found in the lake wa¬
ters ranged from a minimum of 0.07 mg/1 to a maximum of 0.51
mg/1; in the spring waters, the range was from 0.12 to 0.25
mg/1 and in the well waters from 0.10 to 0.45 mg/1. All of these
results show that only a small amount of fluorine is present in
the lake waters and ground waters of this Lake District.
VII. CHLORIDE
Observations were made on the chloride content of the sur¬
face waters of 484 lakes. Of this number, 211 were seepage
lakes and 263 were drainage lakes. The silver nitrate method,
with potassium chromate as indicator, was used for the deter¬
minations; details of this procedure are given by Standard
Methods of Water Analysis, p. 34, 1936.
The results obtained for the two types of lakes were tabu¬
lated separately, but the means for the different groups were
substantially the same ; so they were combined in compiling the
data for Table X. Only a single observation was made on the
surface waters of 292 lakes, while 2 to 12 determinations were
made on each of the others. For those that were visited more
than once, the mean of the different readings was used in com¬
piling the table.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 251
Data
Surface samples. In general the amount of chloride in the
surface waters of these lakes was small ; it ranged from a mini¬
mum of only 0.1 mg/1 to a maximum of 4.5 mg/1. The largest
amount was noted in Maple Lake, a small body of water with a
maximum depth of 2 m., which is situated at the edge of the
village of Three Lakes ; this quantity is more than twice as large
as that noted in any other lake, and the location of this lake sug¬
gests that it may be due to a certain amount of pollution. The
next largest amount was 2.2 mg/1, but only 3 lakes out of the
474 yielded 2.0 mg/1 or more. Taking this group of lakes as a
whole, therefore, it may be said that their waters contain only
small amounts of chloride.
Table X
Chloride content of the surface waters of U7J* lakes of northeastern
Wisconsin. The number and percentage of lakes in the various groups are
indicated , together with the mean quantity of chloride expressed in milli¬
grams per liter.
The table shows that 142 lakes, or 30 per cent of the total,
fall in the 1.00-1.19 mg/1 group, while the 1.20-1.39 mg/1 group
was second with 93 lakes and the 0.80-0.99 mg/1 group was third
with 88. These three groups together account for 323 lakes, or a
little more than 68 per cent of the total number.
Annual variations. In 38 of the 182 lakes that were visited
more than once, the readings obtained for the chloride content
of the surface waters were the same in the different years, or
252 Wisconsin Academy of Sciences , Arts , and Letters
they did not differ by more than 0.2 mg/1. A few lakes belong¬
ing to this group are given in Table XI. Some of them yielded
the same readings in three years, such as Crab Lake, while the
difference was only 0.1 mg/1 in Shishebogama in the same period
of time. Fishtrap Lake showed a difference of only 0.2 mg/1 in
four years. Larger differences were noted in the surface sam¬
ples of the other lakes ; a few of them are included in Table XII
in order to show the range of variation. In some cases the an¬
nual differences were as much as threefold or fourfold, such as
Bird, Mann, Trout, Weber and Cardinal Bog. In the majority
of the surface samples, however, the annual differences were
less than twofold in amount and many of them were as small as
0.3 to 0.4 mg/1.
Table XI
Lakes in which the chloride content of the surface water was the
same or substantially the same in the different years in which they were
visited .
Juday, Birge & Meloche — Waters of N. E. Wisconsin 253
Seasonal variations. Rather marked seasonal variations in
the amount of chloride in the surface water were noted in some
of the lakes, while others showed no difference whatever. Some
30 of these lakes were visited in May 1927 and then again during
the following summer ; some of these seasonal results are given
in Table XIII. There was no change in the chloride content of
the surface waters of some of the lakes, such as Mann and Weber
for instance, and the difference was only 0.1 mg/1 in Trout Lake,
which is within the limit of error of this colorimetric method.
Table XII
Range of annual variation in chloride content of surface waters of
northeastern lakes.
254 Wisconsin Academy of Sciences , Arts, and Letters
Similar small differences were found in Little John and Bass
lakes. In Big St. Germain Lake the amount of chloride found in
the July sample was three times as large as that in May, while
in Tomahwak Lake the difference was almost fourfold. In For¬
est and Little Tomahawk lakes the difference was less than three¬
fold and it was less than twofold in Big Carr and Clear lakes.
Vertical distribution. Series of samples covering the entire
depth were taken in 64 lakes. The various series consisted of 2
to 8 samples each, depending upon the depth of the water ; the
maximum depths of these lakes ranged from 3 m. to 35 m.
In general the chloride was uniformly distributed from sur¬
face to bottom as shown in Anderson, Clear, Fence and Trout
lakes (Table XIV) . In some cases a somewhat smaller amount
of chloride was found in the lower water than at the surface ;
Big, Long and Mary lakes represent this type. On the other
hand a few lakes showed a larger amount of chloride at the bot¬
tom than at the surface; Day and Weber lakes represent this
kind of distribution. These variations in vertical distribution
are not great enough, however, to be of any marked significance,
especially since the colorimetric method used in making these
determinations is not a very exact one owing to the difficulty in
determining the end point.
VIII. SULPHATE
A quantitative determination of the sulphate content of the
residues obtained from the evaporation of the surface waters of
234 lakes was made. As already indicated, these waters yielded
relatively small amounts of residue owing to their softness ; it
was necessary, therefore, to use micro-chemical methods for the
sulphate determinations.
Method. A sample of about 10 mg. of residue was weighed
in a platinum boat and the organic matter was destroyed by
ignition. The boat and residue were then placed in a 10 ml
platinum evaporating dish and after the residue was treated
with 1 ml cone. HC1, the boat was washed with a minimum of
distilled water and removed. The solution was then evaporated
to dryness and baked for one hour at 110° C. The dehydrated
material was moistened with HC1 (1-1) and the solution was
decanted through a 3 cm. filter paper. The dish was washed
Juday, Birge & Meloche — Waters of N. E . Wisconsin 255
Table XIII
Range of seasonal variation in chloride content of surface waters of
northeastern lakes .
thoroughly with 5 per cent HC1 and the washings decanted
through the funnel. The hot filtrate, which should not exceed
8 ml, was treated with a hot 10 per cent solution of barium chlor¬
ide; the solution was kept hot until the precipitate had settled,
usually about 12 hours. The precipitate was washed three times
by decantation through a micro platinum Gooch crucible; 2 ml.
were used for each washing. Finally the precipitate was trans¬
ferred to the crucible and washed three to five times with dis¬
tilled water. The crucible and contents were dried at 110° and
256 Wisconsin Academy of Sciences , Arts, and Letters
Table XIV
Verticle distribution of the chloride in 8 of the northeastern lakes.
then ignited to constant weight. The increase of weight was
calculated to sulphate.
Data
The quantity of sulphate found in the surface waters of these
lakes ranged from a minimum of 0.75 mg/1 in Diamond Lake to
Juday, Birge & Meloche — Waters of N. E. Wisconsin 257
a maximum of 7.86 mg/1 in the Muskellunge Lake which is sit¬
uated in Town 40 N., Range 9 E. The former lake has neither
an inlet nor an outlet and thus belongs to the seepage type, while
the latter has an outlet and, therefore, belongs in the drainage
class.
The results of the analyses are summarized in Table XV.
Both drainage and seepage lakes are included in this table. The
difference in the amount of sulphate in the two types of lakes
was not great enough to make it worth while to separate them.
While the minimum quantity was found in a seepage lake, the
lakes falling in the next group, namely 1. 0-1.9 mg/1, are half
seepage and half drainage. The groups with the largest amounts
of sulphate were chiefly drainage lakes, however; that is, the 4
lakes in the 7. 0-7.9 group belong to the drainage class, as well as
all of those in the 6. 0-6.9 group, while 7 of those included in the
5. 0-5. 9 group are listed as seepage lakes.
Table XV
The 23U lakes on which sulphate determinations were made on the
surface water are grouped at one milligram intervals in this table. The
number and percentage of lakes in the different groups are indicated ,
together with the mean quantity of sulphate (SOi) in the various groups.
The table shows that the great majority of the lakes yielded
between 2.0 and 5.9 mg/1 of sulphate; the four groups falling
between these two limits include 190 lakes or a little more than
81 per cent of the total number included in the table.
No sulphate determinations were made which covered the
entire depth of any of the lakes, but analyses of the bottom sam¬
ples of 19 lakes were made ; the results obtained on some of them
are given in Table XVI. In most cases the amount of sulphate
proved to be larger in the lower water than at the surface as
258 Wisconsin Academy of Sciences, Arts , and Letters
shown for Anderson, Black Oak, Muskellunge, Silver, Toma¬
hawk and Weber lakes. On the other hand, there was very little
difference between surface and bottom in Big Carr, Day and
Trout lakes, and in a few cases the quantity found in the lower
water was smaller than that in the surface water as in Brago-
nier, Ike Walton and Long lakes.
Yoshimura (1933) found a maximum of 474 mg/1 of sul¬
phate in a Japanese lake which received water from a volcanic
region.
Table XVI
Sulphate content of the surface and bottom waters of some of the
lakes of northeastern Wisconsin. The amount is indicated in milligrams
per liter.
Juday , Birge & Meloche — Waters of N. E . Wisconsin 259
IX. NITROGEN COMPOUNDS
Introduction
The waters of inland lakes contain various kinds and
amounts of nitrogenous materials which belong to two general
classes, namely, organic and inorganic nitrogen compounds. The
former owe their existence chiefly to the activities of the aquatic
plants and the latter constitute the main source of the nitrogen
utilized by the plants in the manufacture of protein materials,
but some of the amino acids that are present may be used directly
in this process. In the nitrogen cycle, the organic nitrogen com¬
pounds in turn decompose and form ammonia which is then
oxidized to nitrites and finally to nitrates. Some of the am¬
monia may be used directly in protein synthesis, but nitrates
probably constitute the chief source of nitrogen used for protein
building by the aquatic flora.
The organic nitrogen content of the northeastern lake waters
is fully discussed in another paper so that it is not necessary to
give it further consideration in this report.
A quantitative study of the three forms of inorganic nitro¬
gen, namely free ammonia, nitrite and nitrate, was made for
three summers on a considerable number of lakes situated in
northeastern Wisconsin. While most of the determinations were
limited to surface waters, series of samples covering the entire
depth of a number of lakes were taken during this period, more
especially in those having depths of 12 m. or more.
The three kinds of inorganic nitrogen were determined ac¬
cording to procedures given in Standard Methods of Water
Analysis of the American Public Health Association, Sixth Edi¬
tion. The distillation method was used for the free ammonia,
the sulphanilic acid procedure for nitrite and the phenoldisul-
phonic acid method for nitrate nitrogen.
Ammonia Nitrogen
Surface waters . Quantitative determinations of free ammo¬
nia were made on 43S surface samples which were taken from
276 lakes. Of this number, 174 samples, or 40 per cent of the
total number, did not yield any ammonia, while 150 more, or 34
per cent, contained only a trace; thus 74 per cent of these sur¬
face samples yielded very little or no ammonia nitrogen. Of
260 Wisconsin Academy of Sciences , Arts , and Letters
the remaining samples, 99 were recorded as having 0.01 mg/1,
14 contained 0.02 mg/1 and one 0.03 mg/1 of ammonia nitrogen.
The surface waters of this entire group of 276 lakes, there¬
fore, may be regarded as being very poor in ammonia nitrogen.
The quantities were much smaller than those reported by various
investigators for other lakes. Birge and Juday £1911) found
0.14 mg/1 in the surface water of Garvin Lake and from 0.10
to 0.17 mg/1 in Lake Mendota. Domogalla, Juday and Peterson
(1925) reported a variation from a minimum of 0.01 mg/1 in
late June to a maximum of 0.26 mg/1 in late February for the
surface water of Lake Mendota; in a later paper, Domogalla,
Peterson and Fred (1926) give a range from 0.07 mg/1 in July
to 0.24 mg/1 in October.
According to Yoshimura (1932a), the ammonia nitrogen con¬
tent of the surface waters of 9 eutrophic lakes varied from 0.015
to 0.600 mg/1, while in 9 oligotrophic lakes it ranged from zero
to 0.035 mg/1. In Takasuka Pond, the amount ranged from zero
in July and August to 0.31 mg/1 in March (Yoshimura 1932b).
In the German lakes studied by Ohle (1934), the ammonia nitro¬
gen varied from a trace to a maximum of 0.877 mg/1 ; the largest
amounts were found in the highly colored waters.
In the lakes of northeastern Wisconsin, however, there was
no correlation between ammonia content and the color of the
water. In those samples that yielded 0.02 and 0.03 mg/1 of
ammonia, the color ranged from zero to a maximum of 127 on
the platinum-cobalt scale. Neither was there any direct corre¬
lation with the centrifuge plankton in this group of lakes; the
dry organic matter of the centrifuge plankton varied from 0.44
to 2.01 mg/1 in these samples.
Series . During the progress of this investigation, 144 series
of ammonia determinations were made on 57 lakes. The largest
number (15) was taken on Trout Lake, next in rank were 5 lakes
on which 5 series each were obtained. The number of samples
in each series ranged from 2 to 6 depending upon the depth of
the various lakes.
The lakes fall into two classes on the basis of the ammonia
content of the water of the hypolimnion. Very little or no am¬
monia was found in the lower strata of lakes that were oligo¬
trophic in character, while the lower water of eutrophic lakes
Juday, Birge & Meloche— Waters of N. E. Wisconsin 261
yielded fairly large amounts of ammonia nitrogen. Table XVII
shows the results obtained on both types of lakes. The lakes
included in this table represent various depths and also waters
with varying degrees of hardness/ ranging from very soft to
those with about the maximum quantity of carbonates found in
this lake district. They also show the range of variation in the
quantity of inorganic nitrogen.
Those in which no ammonia nitrogen was found at any depth
are represented by Clear and Crystal lakes. These two bodies
of water have neither an inlet nor an outlet and the water is
very soft. The crop of both phytoplankton and of large aquatics
is small so that the amount of decomposable material reaching
the lower strata of water is correspondingly small. The bottom
water of these two lakes is usually well supplied with dissolved
oxygen in late summer; in Clear Lake the amount rarely falls
below 3.0 mg/1 and in Crystal it has always been above 6.0 mg/1
in the August series.
Trout Lake shows a relatively small amount of ammonia
nitrogen in the lower water in late summer ; the quantity ranged
from 0.02 to 0.05 mg/1 in late August at depths of 30 to 32 m.
Catfish Lake yielded a similar amount at 22 m. in the series
taken on July 31, 1928. Larger amounts were found in the lower
water of Nebish and Presque Isle lakes, which were followed by
Black Oak, Bragonier, Little Tomahawk, Pallette, Adelaide and
Big lakes.
The largest amount of ammonia nitrogen was observed at
21 m. in Lake Mary on July 11, 1928, namely 4.0 mg/1. This
lakelet has an area of 1.2 ha. and a maximum depth of 22 m.
It is dystrophic in character and the lower water contains no
dissolved oxygen for several months each year. A sample of
water taken at 20 m. on May 7, 1927, yielded 2.16 mg/1 of am¬
monia nitrogen, which indicated that there was no vernal over¬
turn of the water in that year since the ice does not disappear
until late April or early May. A further indication of the ab¬
sence of the vernal overturn was the fact that the lower water
had no dissolved oxygen on May 7.
The next largest amount of ammonia nitrogen was noted at
11 m- (1 above the mud) in Wildcat Lake on August 24, 1926,
namely 1.5 mg/1. Then followed Muskellunge Lake with 0.96
262 Wisconsin Academy of Sciences , Arts, and Letters
Table XVII
Ammonia , nitrite and nitrate nitrogen in lake waters of northeastern
Wisconsin. The quantity of nitrogen and of dissolved oxygen is indicated
in milligrams per liter of water.
Juday , Birge ■ & Meloche — Waters of N. E. Wisconsin 263
Table XVXX continued (Nitrogen)
mg/1 at 18 m. on August 20, 1926, and Big Lake with 0.88 mg/1
at 17 m, on August :.3, 1927.
These ammonia results are similar in quantity to those that
have been reported for other lakes, with the exception of the
amount in the epilimnion of these northeastern lakes. Birge and
264 Wisconsin Academy of Sciences , Arts , and Letters
Juday (1911) found 0.14 mg/1 of ammonia nitrogen at the sur¬
face of Garvin Lake on October 10, 1905, and 3.04 mg/1 at 9 m.
In Lake Mendota also, the amount ranged from 0.17 mg/1 at the
surface to 1.35 mg/1 at 22 m. on September 25, 1905. The
amounts reported for the bottom waters of these two lakes are
similar to those found in the lower water of Lake Mary and
Wildcat Lake.
In Takasuka Pond, Yoshimura (1930) obtained 0.03 mg/1 of
ammonia nitrogen at the surface, none at 4 m. and 2.20 mg/1 at
6 m. on July 4, 1930. Ruttner (1931) found a wide range in
the ammonia content of the tropical lakes which he studied.
Usually the epilimnion yielded only traces, but the quantity in
the lower water ranged from a trace in Toba Lake to 4.70 mg/1
in Lamongan Lake.
With the exception of the upper water, the results obtained
in Lake Mary were much the same as those recorded by Ohle
(1934) for Schwarzsee where the ammonia varied from 0.88 mg/1
at the surface to 4.30 mg/1 at the bottom (5.1 m.). In Pinnsee
he found 0.06 mg/1 at the surface and 1.53 mg/1 at 8.2 m. ; the
latter is substantially the same as that obtained in the lower
water of Wildcat Lake.
Ammonia determinations were made on two lakes in early
May of 1927 and again the following July and August. The
ammonia content of the water at 30 m. in Trout Lake changed
from a trace on May 5 to 0.02 mg/1 on August 20. At 20 m. in
Lake Mary, the amount was 2.16 mg/1 on May 7 and 1.30 mg/1
on July 29. Some early and late summer readings were also
made on several lakes in the same year and they serve to show
the changes that took place as the summer season advanced. The
ammonia content of the water at 18 m. in Muskellunge Lake in
1927 was 0.28 mg/1 on June 29 and 0.96 mg/1 on August 20. In
1926 the amount was 0.02 mg/1 at 16 m. in Silver Lake on July 8
and 0.24 mg/1 at this depth on August 19 ; it was 0.80 mg/1 at
11 m. in Wildcat Lake on July 11 and 1.50 mg/1 on the following
August 24, while Blue Lake showed 0.24 mg/1 at 12 m. on June
30 and 0.41 mg/1 some six weeks later on August 13.
With the exception of Lake Mary, the second and later bot¬
tom samples in these lakes showed a larger ammonia content
than the earlier ones. That is, there was a more or less marked
increase in the quantity during the interval between the two
Juday, Birge & Meloche — Waters of N. E. Wisconsin 265
samplings. Undoubtedly this increase was due chiefly to the
accumulation of ammonia derived from the decomposition of
nitrogenous organic material, but some of it may have been due
to variations in sampling. While both sets of samples were
taken in substantially the same locality in each lake, the sampl¬
ing stations were not definitely marked so that the second sample
may have been taken a short distance from the first where the
water had a slightly different ammonia content. An attempt
was always made to secure the deepest sample just one meter
above the bottom in order to avoid the sediment, but slight vari¬
ations in this distance above the bottom may account for part of
the difference. The lower stratum of water is also subject to
shifting through slow currents induced by movements of the
upper strata; thus, while the later sample may be taken at the
same station and at the same distance above the bottom as the
earlier one, it does not represent exactly the same water from
which the earlier sample was obtained because of the shifting
which has taken place in the meantime.
In view of the marked constancy of the temperature of the
lower two or three meters of water, however, it does not seem
probable that the movements of this lower stratum result in any
marked vertical mixing of the water therein. Also, since all of
the changes indicated an increase of ammonia, it may be con¬
cluded that this increase in quantity is due chiefly to the accumu¬
lation of ammonia in this stratum as a result of decomposition
and not to variations in the water from which the samples are
taken due to a shifting of the bottom water or to a slight differ¬
ence in the depth at which the sample is obtained. Slight differ¬
ences in the location of the sampling station, the shifting of the
lower water and variations in the distance above the bottom may
be responsible for part of the difference, but these factors prob¬
ably play only a minor role, or none at all.
The decrease in ammonia at 20 m. in Lake Mary can not be
readily explained from data in hand. It seems probable that it
was not due to a shift in the lower water during the intervening
period because the lake is small and the surface water is dis¬
turbed very little by wind; thus there ought to be little or no
movement of the lower water resulting from this factor. The
bacterial population at a depth of 20 m. was found to be rather
large by Miss Bere (1933), namely one to two million per cubic
266 Wisconsin Academy of Sciences , Arts, and Letters
centimeter, and the decrease was probably due to the utilization
of the ammonia by these organisms in their metabolic processes.
The changes in the ammonia content of the lower water at
other seasons of the year have not been followed in the north¬
eastern lakes, but it seems probable that they do not differ essen¬
tially from those that have been observed in other lakes. Domo-
galla, Juday and Peterson (1925) found that the ammonia ni¬
trogen varied from a minimum of 0.007 mg/1 at a depth of 20 m.
in Lake Mendota in July to a maximum of 0.76 mg/1 in March.
Domogalla, Fred and Peterson (1926) noted an increase in the
quantity of ammonia nitrogen in the bottom water of Lake Men¬
dota during the winter ; it rose to a maximum of 0.56 mg/1 about
the middle of March. A sharp decline to 0.19 mg/1 followed in
April during the vernal period of circulation ; this was succeeded
by a gradual increase during the summer which rose to a maxi¬
mum of 0.69 mg/1 in September, with a marked decrease in Octo¬
ber at the beginning of the autumnal period of circulation.
In some of the northeastern lakes, comparable amounts of
ammonia nitrogen were found in the bottom water on approxi¬
mately the same dates in different years, while in others there
were marked differences. The record for Muskellunge Lake il¬
lustrates the former type; the amounts in this lake at a depth
of 19 m. were 0.28 mg/1 on June 29, 1926, 0.24 mg/1 on July 2,
1927 and 0.26 mg/1 on June 27, 1928. On the other hand, the
quantity was 0.72 mg/1 at 16 m. in Star Lake on August 2, 1926
and 0.24 mg/1 on August 3, 1928, which represents a threefold
difference.
Nitrite Nitrogen
Nitrite determinations were made on 504 surface samples
from 307 different lakes. Of this number 369 samples contained
no nitrite nitrogen whatever and 125 more yielded only a trace;
thus only 10 surface samples had measurable amounts of nitrite.
In 5 of these samples, the amount of nitrite nitrogen was 0.001
mg/1 and in 4 it was 0.002 mg/1; the other sample contained
0.004 mg/1, but it was taken in a lake which was subject to some
sewage pollution at the time. These low nitrite results indicate
that there was a rapid transformation of ammonia to nitrates,
so that there was no accumulation of nitrites in the surface
waters.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 267
In the surface water of Lake Mendota, Domogalla, Juday and
Peterson (1925) found that the nitrite nitrogen ranged from
none in September to 0.018 mg/1 in January. Atkins (1980) did
not obtain any nitrites in the upper 20 m. of the English Chan¬
nel in August. On the other hand, Verjbinskaya (1932) observed
that the nitrites appeared in the Barents Sea in spring concur¬
rently with the onset of the phytoplankton, increased during the
summer months and decreased again in autumn, completely dis¬
appearing toward November.
Series. In this investigation, 138 sets of nitrite observations
were made which covered the entire depth of 58 lakes. The
largest number of series was taken in Trout Lake, namely 11 ;
in the other lakes the number ranged from one to 5. In 121
series the quantity of nitrite nitrogen did not exceed a trace at
any depth, while in 17 series measurable amounts were found in
the lower water. In the latter cases, the quantity of nitrites
varied from 0.001 mg/1 to 0.012 mg/1. The maximum amount
(0.012 mg/1) was noted in the samples from 20 m. and 26 m. in
Presque Isle Lake on August 11, 1927 ; on August 9, 1928, the
28 m. sample from this lake yielded 0.008 mg/1. (Table XVII).
The sample from 21 m. taken in Tomahawk Lake on August 8,
1928 also yielded 0.008 mg/1. Next in rank were Papoose Lake
with 0.006 mg/1 of nitrite nitrogen at 18 m. and Star Lake with
0.005 mg/1 at 16 m. These accumulations of nitrites were found
only in lakes where the lower water contained very little or no
dissolved oxygen.
At a depth of 20 m. in Lake Mendota, Domogalla, Juday and
Peterson (1925) observed that the nitrites ranged from 0.004
mg/1 in April and May to 0.060 mg/1 in March. Atkins (1930)
obtained as much as 0.038 mg/1 of nitrite nitrogen below a depth
of 25 m. in the English Channel in August. Yoshimura (1932c)
states that the amount was so small in Takasuka Pond through¬
out the year that it might be regarded as negligible ; a maximum
of 0.003 mg/1 was obtained in November.
Nitrate Nitrogen
Surface samples. Quantitative determinations of nitrate ni¬
trogen were made on 818 surface samples taken from 472 lakes ;
the number obtained from each lake varied from one to 10. In
268 Wisconsin Academy of Sciences , Arts, and Letters
the great majority of the lakes that were visited more than once,
the different samples were taken in different years. A summary
of the results obtained is given in Table XVIII where the various
lakes are grouped according to the amount of nitrate nitrogen
found in them. In the case of lakes that were visited more than
once, the mean of the different determinations was used in ascer¬
taining the group to which they belonged.
Table XVIII
Nitrate nitrogen content of surface waters of northeastern lakes. The
lakes are grouped on the basis of the quantity of nitrate nitrogen. The
means of the various groups are indicated in milligrams per liter of water.
The maximum number of lakes falls in the 0.015-0.019 mg/1
group, namely 153, or a little more than 32 per cent of the total
number; the mean quantity for this group is 0.016 mg/1. The
0.010-0.014 group ranks second with 118 lakes and the 0.020-
0.024 group third with 111 lakes. Thus the quantity of nitrate
nitrogen in the surface waters of these lakes fell between 0.010
mg/1 and 0.024 mg/1 in 382 cases, or in 81 per cent of the total
number of lakes examined.
The nitrates exceeded 0.024 mg/1 in 65 lakes and the surface
water of 11 of them yielded 0.036 mg/1 or more. Of the latter
number, 7 were drainage and 4 were seepage lakes ; 10 of them
contained rather large amounts of plankton at the time that the
samples were taken. The dry organic matter in the centrifuge
plankton ranged from a minimum of 1.02 mg/1 to a maximum of
4.18 mg/1, but in the eleventh member of the group the quantity
of centrifuge plankton was only 0.62 mg/1. Nine of these lakes
are so shallow that the water is kept in complete circulation dur¬
ing the summer, so that the various dissolved substances are
evenly distributed from surface to bottom at that time.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 269
Two seepage and 3 drainage lakes yielded the smallest
amounts of nitrate nitrogen, namely 0.004 mg/1. The quantity
of centrifuge plankton in them ranged from 0.42 mg/1 to 2.36
mg/1 of dry organic matter; the depth of these 5 lakes varies
from 3 m. to 28 m. In 17 lakes the nitrate nitrogen content of
the surface water was less than 0.010 mg/1. The mean quantity
found in the surface waters of the 464 lakes was 0.018 mg/1.
Series. In the serial observations, 184 sets of samples were
taken on 64 lakes ranging in depth from 5 to 35 m. The number
of samples in each series varied from 2 to 7, depending upon the
depth of the lake. In some lakes the quantity of nitrate nitro¬
gen was substantially uniform from surface to bottom; this
group is represented in Table XVII by Bragonier, Crystal, Little
Tomahawk and Pallette lakes. In most of the lakes, however,
there was a more or less marked increase in the amount of ni¬
trates in the lower water ; this type of distribution is shown by
most of the lakes included in the table. The difference between
surface and bottom nitrate nitrogen content ranged from less
than twofold in Mary and Nebish lakes to ninefold in Catfish
Lake and more than twelvefold in Trout Lake.
0.01 0.02 0.03 0.04 0.05 G.0*
Fig. 6. Vertical distribution of the nitrate nitrogen in Black Oak Lake
on August 24, 1928 ; in Clear Lake on August 7, 1928; in Lake Mary on
July 11, 1928; and in Presque Isle Lake on August 9, 1928. Note the de¬
crease in the nitrate nitrogen in the bottom sample of Presque Isle Lake;
compare with Adelaide and Catfish lakes in Fig. 7.
270 Wisconsin Academy of Sciences, Arts, and Letters
The vertical distribution of the nitrates is shown graphically
in Figures 6 and 7. Figure 6 shows 4 lakes in which the differ¬
ence between surface and bottom nitrate nitrogen did not exceed
fourfold. Black Oak and Clear lakes had smaller amounts of
nitrates than the other two, but they show the characteristic in¬
crease in the hypolimnion. In Lake Mary, there was a some¬
what larger amount of nitrate nitrogen at the surface than at
5 m. and 10 m., but below the latter depth there was a distinct
increase in quantity.
In Presque Isle Lake, the 20 m. sample yielded almost four
times as much nitrate nitrogen as the surface sample ; the great¬
er part of the increase was noted between 10 m. and 20 m. The
quantity found at 28 m. was somewhat smaller than that at 20
m. ; this may be attributed to the activity of nitrate reducing
bacteria at 28 m. Similar results have been reported by Yoshi-
mura (1982b) for Lake Busyu.
Figure 7 represents 4 lakes in which the nitrate nitrogen
showed a more marked increase in the hypolimnion ; it has been
necessary to plat them on a different scale on account of the
Fig. 7. Vertical distribution of the nitrate nitrogen in Adelaide Lake
on July 17, 1928; in Catfish Lake on July 31, 1928; in Long Lake (Phelps)
on August 21, 1929; and in Trout Lake on August 25, 1928. Compare
with Fig. 6.
Juday, Birge & Meloche — Waters of N. E. Wisconsin 271
larger quantity in this stratum. All of them yielded about the
same amount of nitrates in the epilimnion, but there was a sharp
rise below a depth of 5 m. in Adelaide Lake and of 10 m. in the
others. Two of the bottom samples of Trout Lake contained the
largest amounts of nitrate nitrogen noted in any of the series,
namely 0.040 mg/1 on August 25, 1928 and 0.195 mg/1 on August
24, 1934. In 5 other series of observations on Trout Lake, the
nitrates in the bottom samples ranged from 0.052 mg/1 on July
1, 1931 to 0.110 mg/1 on August 27, 1931.
The 22 m. sample of Catfish Lake yielded a smaller amount
of nitrate nitrogen than that taken at 18 m., thus showing the
same phenomenon observed in Presque Isle Lake. A much more
marked nitrate reduction in the lower water was found in Ade¬
laide Lake where the nitrates were only one-third as large at
21 m. as at 10 m. (Table XVII and Figure 7).
The bottom samples of 6 lakes that are not included in the
table and diagrams yielded between 0.060 and 0.100 mg/1 of
nitrate nitrogen; they were Nokomis (0.075 mg/1), Papoose
(0.090 mg/1), Silver (0.060 mg/1), Star (0.060 mg/1), Turtle
(0.070-0.100 mg/1), and Two Sisters (0.073 mg/1). This group
of lakes thus occupies an intermediate position between those
represented in Figures 6 and 7.
Series were taken on 31 lakes in two or more years. The
results showed that the variation in the quantity of nitrate nitro¬
gen in the surface water of each of these lakes ranged from less
than twofold to almost sixfold. The mean difference for the
entire group was approximately twofold; that is, the mean of
the minimum quantities found in the surface waters of the vari¬
ous lakes was 0.010 mg/1, while the mean of the maximum
amounts was 0.021 mg/1.
Similar variations in the quantity of nitrates were also noted .
in the bottom water of these lakes in the different years. In 11
of the 31 lakes, the maximum amount of nitrate nitrogen found
in the bottom water in the different years was less than twice as
large as the minimum and in 13 other cases the difference was
less than threefold. In all except 7 lakes, therefore, the annual
variations in nitrate content were less than threefold in the
bottom water. In one lake the difference between maximum
and minimum in the different years was eightfold ; there was
also one sevenfold difference, one sixfold and two fivefold.
272 Wisconsin Academy of Sciences , Arts , and Letters
The increase in the nitrate content of the water of the hypo-
limnion in summer is due chiefly to the decomposition of the
nitrogenous organic matter which sinks into this stratum from
the upper water and decomposes there. To this is added a cer¬
tain amount of nitrate derived from the decomposition of the
organic matter in the bottom deposits, which constitutes from
25 to 40 per cent of the dry weight of this material. Through
the activity of denitrifying bacteria, the quantity of nitrate in
the bottom water of some lakes was found to be smaller than
that a few meters above this region.
In comparison with the results obtained on the northeastern
lakes, the nitrate content of the surface water of Lake Mendota
ranged from a minimum of 0.008 mg/1 in September to a maxi¬
mum of 0.230 mg/1 in April as reported by Domogalla, Juday
and Peterson (1925) ; at a depth of 20 m. it varied from 0.016
mg/1 in December to 0.366 mg/1 in February. The maxium
amounts at both the surface and the bottom of Lake Mendota
were considerably larger than any obtained in the northeastern
region.
Strom (1933) found that the nitrate nitrogen content of the
surface waters of 5 Norwegian lakes ranged from none to 0.040
mg/1 in August, while the amount in the bottom waters varied
from 0.070 mg/1 to 0.180 mg/1. In two of these lakes he found
a smaller amount at the bottom than in samples taken some dis¬
tance above the bottom. Thus the ranges which he gives fall
within the limits noted for the Wisconsin lakes.
Most of the surface samples taken by Rakestraw (1932) in
the neritic waters of the Gulf of Maine yielded between 0.010
mg/1 and 0.025 mg/1 of nitrate nitrogen and more than 80 per
cent of the lakes of northeastern Wisconsin fall within these
limits.
While the nitrate nitrogen content of the waters of the north¬
eastern lakes is similar in quantity to that of other bodies of
water, the total inorganic nitrogen is generally considerably
smaller; this is especially true in the epilimnion where photo¬
synthesis takes place chiefly and thus where the demand is great¬
est for inorganic nitrogen in the manufacture of organic nitro¬
gen compounds. This is due chiefly taf the scarcity or total
absence of ammonia nitrogen in the upper stratum and to the
Juday, Birge & Meloche — Waters of N. E. Wisconsin 273
absence of nitrite nitrogen in the upper water of practically all
of the lakes.
Summary
In the surface waters of 519 lakes, the amount of silica
ranged from none to 25.6 mg/1; more than 50 per cent of the
waters had less than 1.0 mg/1 and about 87 per cent of them less
than 5.0 mg/1. In some cases there was little or no increase in
silica with increasing depth in^the summer period of stratifica¬
tion, while in others there was a marked rise in the lower water.
The iron content of the surface waters of 74 lakes varied
from a trace to 2.0 mg/1 ; 3.0 to 6.0 mg/1 were noted in the lower
water of some of the lakes, especially where the bottom stratum
had a considerable amount of free carbon dioxide and little or no
dissolved oxygen. Under anaerobic conditions most of the iron
was in the ferrous state.
Manganese was found only in very small amounts (0.003-
0.023 mg/1) in the surface waters, but there was a definite in¬
crease with increasing depth in some of the lakes.
The calcium content of the surface waters of 358 lakes ranged
from 0.13 to 18.8 mg/1; 52 per cent of them had less than 5.0
mg/1 and 86 per cent of them less than 10.0 mg/1 of Ca. The
magnesium found in the surface waters of 290 lakes varied from
0.1 to 6.5 mg/1; 36 per cent of them had less than 2.0 mg/1 and
81 per cent less than 5.0 mg/1. There was a wide range in the
ratio of calcium to magnesium; in two cases the surface sam¬
ples contained twice as much magnesium as calcium, in others
the two elements were present in about equal amounts, while
in still others the quantity of calcium was two to six times
as large as that of magnesium. A large percentage of the waters
contained unusually small amounts of calcium and magnesium,
so that they may be regarded as very soft, some of them as
extremely soft. From a quantitative standpoint, the bound car¬
bon dioxide was more closely correlated with the amount of cal¬
cium than with that of magnesium.
The quantity of fluorine ranged from 0.1 to 0.5 mg/1 in the
24 lake, spring and well waters that were analyzed.
The chloride content of 474 surface waters was small, rang¬
ing from 0.1 to 4.5 mg/1; in general it was substantially the
same from surface to bottom in the various lakes.
274 Wisconsin Academy of Sciences , Arts , and Letters
The surface waters of 234 lakes contained 0.7 to 7.8 mg/1 of
sulphate; in some lakes the amount of sulphate was about the
same at surface and bottom, but in others there was a somewhat
larger quantity at the bottom than at the surface.
The surface water of 74 per cent of the lakes in summer con¬
tained no ammonia nitrogen or only a trace; an additional 22
per cent yielded 0.01 mg/1. Little or no ammonia nitrogen was
found in the lower water of a number of lakes, while varying
amounts were noted in the bottom stratum of others, especially
in lakes which had little or no dissolved oxygen in this stratum.
With five exceptions no nitrite nitrogen was present in the sur¬
face waters, but small amounts were observed in the bottom
stratum where little or no dissolved oxygen was found.
The nitrate nitrogen ranged from 0.004 to 0.070 mg/1, with
a mean of 0.018 mg/1, in the surface waters of 472 lakes. In
some cases the nitrates were rather uniformly distributed from
surface to bottom in summer, while in others there was a more
or less marked increase in the lower water. Nitrate reduction
took place in the bottom strata of some of the lakes.
Literature
Atkins, W. R. G. 1930. Seasonal changes in the nitrite content of seawater.
Jour. Mar. Biol. Assoc. United Kingdom. 16:1515-518.
Beadle, L. C. 1932. Scientific results of the Cambridge Expedition to East
African lakes, 1930-31. IV. The waters of some East African lakes in
relation to their fauna and flora. Jour. Linn. Soc. London. 38:157-211.
Bere, Ruby. 1933. Numbers of bacteria in inland lakes of Wisconsin as
shown by the direct microscopic method. Int. Rev. ges. Hydrobiol. u.
Hydrog. 29:248-263.
Birge, E. A. and C. Juday. 1911. The inland lakes of Wisconsin. The
dissolved gases and their biological significance. Bull. No. 22. Wis.
Geol. and Nat. Hist. Survey. 259 pp. Madison.
Delbecque, A. 1898. Les Lac Francaise. 436 pp. Paris.
Dienert, F. and F. Wandenbulcke. 1923. Sur le dosage de la silice dans
les eaux. Bull. soc. chim. 33:1131-1140.
Domogalla, B. P., C. Juday and W. H. Peterson. 1925. The forms of nitro¬
gen found in certain lake waters. Jour. Biol. Chem. 63:269-285.
Domogalla, B. P. and E. B. Fred. 1926'. Ammonia and nitrate studies of
lakes near Madison, Wisconsin. Jour. Amer. Soc. Agron. 18 : 897-911.
Domogalla, B. P., E. B. Fred and W. H. Peterson. 1926. Seasonal varia¬
tions in the ammonia and nitrate content of lake waters. Jour. Amer.
Water Works Assoc. 15:369-385.
Juday, Birge & Meloche— Waters of N. E. Wisconsin 275
Einsele Wilhelm. 1936. Ueber die Beziehung des Eisenkreislaufs zum Phos-
phatkreislauf im eutrophen See. Arch. f. Hydrobiol. 29:664-686.
Halbfass, W. 1923. Grundzuge einer vergleichenden Seenkunde. 354 pp.
Berlin.
Juday, C., and E. A. Birge. 1932. Dissolved oxygen and oxygen consumed
in the lake waters of northeastern Wisconsin. Trans. Wis. Acad. Sci.,
Arts & Let. 27: 415-485.
Juday, C., and E. A. Birge. 1933. The transparency, the color and the
specific conductance of the lake waters of northeastern Wisconsin.
Trans. Wis. Acad. Sci., Arts. & Let. 28:205-259.
Juday, C., E. A. Birge and V. W. Meloche. 1935. The carbon dioxide and
hydrogen ion content of the lake waters of northeastern Wisconsin.
Trans. Wis. Acad. Sci., Arts & Let. 29 : 1-82.
Kemmerer, G., J. F. Bovard, and W. R. Boorman. 1923. Northwestern lakes
of the United States: Biological and chemical studies with reference
to possibilities in production of fish. Bull. Bur. Fish. 39:51-140.
Minder, L. 1929. Chemische Untersuchungen am Stausee Waggital. Verh.
Intern. Ver. Limnol. 4:454-461.
Ohle, W. 1934. Chemische und physikalische Untersuchungen nord-deutscher
Seen. Arch. Hydrobiol. 26: 386-464, 584-658.
Pesta, O. 1929. Der Hochgebirgssee der Alpen. Die Binnengewasser. Bd.
8. 156 pp. Stuttgart.
Rakestraw, W. 1932. Phosphorus and nitrogen in the neritic waters of the
Gulf of Maine. Int. Rev. ges. Hydrobiol. u. Hydrog. 27:151-160.
Robinson, R. J., and G. Kemmerer. 1930. Determination of silica in min¬
eral waters. Trans. Wis. Acad. Sci., Arts and Let. 25:129-134.
Ruttner, F. 1931. Hydrographische und hydrochemische Beobachtungen auf
Java, Sumatra und Bali. Arch. Hydrobiol. Suppl. 8 : 197-454.
Standard methods of water analysis. 1936. Amer. Pub. Health Assoc.
309 pp. New York.
Stangenberg, Marjan . 1936. Eisenverteilung in den Seen des Suwalki-
Gebiets wahrend des Sommers. Arch, d’ Hydrobiol. et dTchthyol. 10:
48-75.
Strom, Kaare M. 1933. Nordfjord lakes. A limnological survey. Skr.
N. Vidensk.-Akad. Oslo. I. M. N., 1933, No. 8.
Ueno, M. 1934. Limnological characteristics of the Nikko mountain range.
Zool. Soc. Japan. 1934:324-337.
Verjbinskaya, V. 1932. Observations on the nitrite changes in the Barents
Sea. Jour, du Cons. 7:47-52.
Wiebe, A. H. 1930. The manganese content of the Mississippi River water
at Fairport, Iowa. Science. 71:248-249.
Willard, H. H., and O. B. Winter. 1933. Volumetric method for determina¬
tion of fluorine. Ind. & Eng. Chem. Anal. ed. 5: 7-18
Yoshimura, S. 1931a. Seasonal variation of iron and manganese in the
water of Takasuka-numa, Saitama. Jap. Jour. Geol. & Geog. 8:269-279.
Yoshimura, S. 1931b. Contributions to the knowledge of the stratification
of iron and manganese in lake water of Japan. Jap. Jour. Geol. &
Geog. 9:61-69.
276 Wisconsin Academy of Sciences, Arts, and Letters
Yoshimura, S. 1932a. Contribution to the knowledge of the nitrogenous
compounds and phosphate in the lake waters of Japan. Proc. Imp-
Acad. 8: 94-97.
Yoshimura, S. 1932b. Limnological reconnoissance of Lake Busyu, Hukui,
Japan. Sci. Rep. Tokyo Bunrika Daigaku. 1: 1-27.
Yoshimura, S. 1932c. Seasonal variation in content of nitrogenous com¬
pounds and phosphate in the waters of Takasuka Pond, Saitama,
Japan. Arch. f. Hydrobiol. 24: 155-176.
Yoshimura, S. 1933. Kata-numa, a very strong acid water lake on volcano
Katanuma, Miyagi Prefecture, Japan. Arch. f. Hydrobiol. 26: 197-202,
THE ESTIMATION OF MAGNESIUM IN LAKE WATER
RESIDUES
V. W. Meloche and Katherine Pingrey
From the Department of Chemistry, University of Wisconsin, and the
Limnological Laboratory of the Wisconsin Geological and Natural History
Survey. Notes and reports No. 83.
Introduction
In the study of the northern lakes of Wisconsin conducted
by the Wisconsin Geological and Natural History Survey, it has
been necessary to determine the magnesium content of the wa¬
ters. Although a satisfactory macro gravimetric method is
known, the softness of the waters makes it necessary to use a
micro procedure. While the gravimetric procedure which de¬
pends on the separation of magnesium ammonium phosphate
and its subsequent ignition to the pyrophosphate can be used for
micro samples, such a procedure is time consuming and requires
apparatus which is not easily adapted to the field laboratory. A
volumetric method exists which depends on the separation of
the 8-hydroxy-quinolate and its solution and titration by 0.01 N
permanganate or br ornate. Still another method in the field of
colorimetry depends on the formation of magnesium hydroxide
and the subsequent adsorption of titan yellow to give a pink
color. The titrametric methods have proven unsatisfactory due
to the concentrations of magnesium available and the unsatis¬
factory nature of the end points. The colorimetric procedure
likewise has definite limitations in that calcium enhances the
color due to magnesium, and the background color of the dye
requires the use of a filter or adherence to the Nessler principle
of comparison. With this in mind we attempted to adapt the
ceruleomolybdate method for phosphorus to the micro estima¬
tion of magnesium.
x The work reported in this paper was supported in part by funds from the Wisconsin Alumni
Research Foundation.
277
278 Wisconsin Academy of Sciences, Arts, and Letters
Reagents
Molybdate reagent: 10 grams of ammonium molybdate were
dissolved in 100 cc of distilled water. This solution was added
to 300 cc of sulfuric acid (1-1).
Stannous chloride: 2.15 grams of stannous chloride were dis¬
solved in 20 cc of concentrated hydrochloric acid. After dilution
to 100 cc, a small piece of tin was added.
Standard phosphate: 4.394 grams of anhydrous potassium
dihydrogen phosphate were dissolved in water and diluted to one
liter. One cc of this solution contained 1 mg of phosphorus.
A solution containing 0.001 mg of phosphorus per cc could be
prepared from the stock solution by appropriate dilution.
Other reagents:
25% sodium acetate solution
4% ammonium oxalate solution
5% hydrochloric acid
1:1 ammonium hydroxide
10% microcosmic salt solution
2% (by volume) ammonium hydroxide wash solution
Procedure
The principle upon which this procedure is based is the re¬
moval of calcium by the modified McCrudden method of Meloche
and Setterquist, the destruction of the oxalate and ammonium
ions, and the estimation of the magnesium in the residue by pre¬
cipitation of magnesium ammonium phosphate and the subse¬
quent determination of the phosphorus by the Deniges colori¬
metric method.
A 10-15 mg. sample of a lake water residue is weighed on a
micro weighing glass, transferred dry to a 15 cc centrifuge tube,
moistened with a small drop of water, and dissolved in one or
two drops of concentrated hydrochloric acid. It is sometimes
necessary to heat slightly in order to dissolve all of the residue.
The solution is diluted to about 2-3 cc and one drop of bromcre-
sol purple is added. The solution is made alkaline with 1:1 am¬
monium hydroxide and then just acid with 5% hydrochloric,
adding one drop of acid in excess.
Meloche & Pingrey — Estimation of Magnesium 279
Two cc of 4% ammonium oxalate are added and the solution
allowed to stand ten minutes; the solution is then diluted to
10 cc and centrifuged for 5 minutes. The supernatant liquid is
decanted into a small evaporating dish and the precipitate of
calcium oxalate is washed twice with 2-3 cc portions of aqua
ammonia, centrifuging and pouring the supernatant liquid into
the evaporating dish as above. About one-half gram of am¬
monium chloride is added to the solution and the solution is evap¬
orated under a reflector and a 60 watt electric light bulb until
the volume is approximately 1-2 cc. About 1 cc of nitric acid is
added and evaporation continued to dryness. The residue is
moistened with one to two drops of nitric acid and again evap¬
orated to dryness. The dish is now placed in a muffle furnace at
200° C for 10 minutes. The residue is dissolved with water,
adding a drop of acid if the residue goes into solution with diffi¬
culty. Transfer to a 15 cc centrifuge tube. One cc of 10%
microcosmic salt solution is added and then 4 cc of 1:1 aqua
ammonia. The solution is heated to 60-70°C for 30 minutes to
allow complete precipitation.
The solution is diluted to 12 cc and the tube is centrifuged
for 10 minutes. If any particles of precipitate cling to the top
of the liquid, they are removed with the aid of a vaselined rod.
The particles are removed from the rod by dipping it into a
small test tube which contains a drop of concentrated hydro¬
chloric acid. The supernatant liquid in the centrifuge tube is
decanted and the tube drained by inverting on a filter paper for
one minute. The precipitate is washed once with a 5 cc portion
of 2% aqua ammonia, the liquid decanted and the tube drained
as before.
The precipitate is dissolved in 1-2 drops of hydrochloric acid
and transferred to a 500 cc volumetric flask. The content of the
test tube containing the particles dissolved from the stirring rod
is added to this flask and the whole solution is diluted to the
mark with distilled water. An aliquot part, either 1 or 5 cc is
used for colorimetric comparison with the phosphorus stand¬
ards in Nessler tubes of 50 cc capacity. The aliquot of the sam¬
ple is placed in one tube and a set of standards is prepared by
using amounts of the standard solution varying from each other
one-quarter of a cc. The aliquot of the sample should be so cho¬
sen that it is not necessary to use more than 5 cc of the standard.
280 Wisconsin Academy of Sciences, Arts, and Letters
When more than 5 cc of standard is used, it is difficult to distin¬
guish differences in color intensity.
The tubes containing the sample and standards are filled to
the mark with distilled water. Two cc of the molybdate reagent
and four drops of the stannous chloride solution are added to
each of the tubes and they are allowed to stand for 10 minutes
in order to allow the color to develop.
Calculation
The number of cc of standard phosphate solution used in the
comparison standard multiplied by .001 is equivalent to the milli¬
grams of phosphorus in the aliquot of sample. If this value is
multiplied by 0.784 the phosphorus is converted to the equiva¬
lent in terms of magnesium in the aliquot. If the aliquot was
one-fifth of the total sample, multiply by 5 and the result will be
the total magnesium in the sample. If one wishes to convert
this value to per cent, divide by the weight of the sample and
multiply by one hundred. It should be noted that we do not re¬
port results in per cent by weight but in parts per million. This
is possible in our case because the volume from which the residue
was obtained is always recorded.
Results
In the development of the procedure, known solutions of pure
magnesium chloride were analyzed. The results are shown in
Table I.
Table I
Precipitation for onehalf hour at 60-70 °
From these data it is observed that on a solution of pure
magnesium chloride it does not seem to make a great deal of
difference whether the magnesium is precipitated in the cold or
Meloche & Pingrey — Estimation of Magnesium
281
hot; the cold precipitation merely takes a longer time. One
encounters quite a different situation, however, in the analysis
of lake water residue, for here we have a rather high contamina¬
tion with calcium. In the event that the calcium is removed by
the oxalate procedure, the filtrate containing magnesium also
contains considerable oxalate ion. In the micro procedure de¬
scribed, it seemed important to destroy this contamination.
Table II shows the type of agreement which may be expected
when calcium is removed by the oxalate procedure. The oxalate
which contaminates the magnesium solution is removed by the
use of nitric acid, evaporation to dryness and subsequent heating
of the residue in a muffle furnace at 200°C.
Table II
282 Wisconsin Academy of Sciences , Arts , and Letters
In Table III, we have presented some typical results of the
analysis of lake water residues as well as the comparison with
results obtained by the 8-hydroxyquinoline method (KMn04).
Results which have been published by the Survey on the mag¬
nesium content of waters of Wisconsin lakes have been obtained
by several methods. A few of the earliest results were obtained
by the use of the standard macro magnesium ammonium phos¬
phate method. Later the 8-hydroxyquinoline method (KMn04)
was used for a rather large number of samples. Realizing the
limitations of this method, we checked many results by means
of the colorimetric method described above.
As has been mentioned, the use of the titan-yellow method
has been studied. Since we have referred to the interference of
calcium in this method, it would appear that the separation of
calcium as oxalate and the destruction of the oxalate ion would
provide a magnesium solution on which the titan-yellow method
could be used. Since we have not had success with this modifi¬
cation, we have not used it in any of our regular magnesium
work. However, promising results have been obtained by the
use of the photelometer and it may be possible that the titan-
yellow method can be successfully adapted to lake work.
Summary
1. A method is described for the determination of magnesium
in lake water and lake water residues which involves the fol¬
lowing points :
a. The separation of calcium by means of the modified Me-
Crudden method.
b. The precipitation of magnesium ammonium phosphate
and separation by means of the centrifuge.
c. And finally, the solution of the precipitated magnesium
ammonium phosphate and the indirect estimation of the
magnesium by means of the colorimetric molybdate esti¬
mation of the phosphorus in this precipitate.
2. It was pointed out that while the titrametric 8-hydroxyquin¬
oline method and the colorimetric titan-yellow method had
been used, the method described was selected because of its
flexibility and relative precision.
Madison, Wis.
Meloche & Pingrey — Estimation of Magnesium
288
Literature
1. Stankov and Miholic, J.A.C.S. 52, 200, (1930).
2. Benedetti Pichler and Schneider, Microchemie Emich Festschrift 1,
319, (1930).
3. Hough and Fichlen, J.A.C.S. 52, 4752 (1930).
4. Deniges, G., Determination quantitative des plus faibles quantites de
phosphate dans les produits biologiques par le methode ceruleomolyb-
dique. Compt. Rend. Soc. Biol. Paris 84 (17); 875 (1921). Also C. R.
Acad, des Sci. and Prac. Agric. 2 (4) ; 822.
5. Denis, J.B.C. 52, 411, (1922).
6. Kolthoff, Chem. Weekblad. 24, 254, (1927).
7. Juday, Birge, Kemmerer and Robinson, Tr. Wis. Acad. Sci. 23, 233,
(1928).
8. Meloche and Setterquist, Tr. Wis. Acad. Sci. 28, 291, (1933).
9. MbCrudden, F. C., J.B.C. 7, 83, (1909) ; 7, 201, (1909) ; 10, 187, (1911).
10. Juday, Birge and Meloche, Tr. Wis. Acad. Sci. 31, 223, (1938).
SODIUM AND POTASSIUM CONTENT OF WISCONSIN
LAKE WATERS AND THEIR RESIDUES.
D. Lohuis, V. W. Meloche and C. Juday
From the Department of Chemistry, University of Wisconsin, and
the Limnological Laboratory of the Wisconsin Geological and Natural
History Survey. Notes and reports No. 84.
In another paper a procedure will be described by which
total alkali, i.e., sodium and potassium, can be determined by
means of the polarograph. The present paper will be devoted to
the presentation of data which were obtained in the examination
of a group of seepage and drainage lakes of northeastern Wis¬
consin. Since the methods used in the collection and prepara¬
tion of samples have been described elsewhere in this series, it
is not necessary to include that information here.
Drainage Lakes
Distribution of Total Alkali : Table I, Figure 1, 2
None of the waters of this group represent very great hard¬
ness, the highest bound C02 being 31.5 ppm. Nevertheless, there
is considerable variation under this limit and it is important
that the alkali results be presented not only on a percentage
basis but also in parts per million. The spread in total alkali,
expressed in parts per million as potassium, runs between 1.06
and 4.76 for the 61 lakes listed in Table I. There are 8 results
which fall between 1 and 2 ppm; 15 which fall between 2 and
2.9 ppm and 34 which fall between 2.9 and 4 ppm. Three lakes
have alkali concentrations which are slightly over 4 ppm and
only 2 lakes have alkali concentrations which approach 5 ppm.
This study was supported by a grant from the Graduate Re¬
search Committee of the University of Wisconsin. The results
for the majority of the lakes in the drainage group fall between
2 and 4 ppm, the greater number of these falling between 2.5
and 3.5 ppm.
Another comparison may be made in which the per cent of
total alkali in the lake water residues is used. In these cases a
285
286 Wisconsin Academy of Sciences, Arts, and Letters
Table I
DRAINAGE
Lohuis, Meloche & Juday — Sodium and Potassium 287
ppm Alkali as K
Fig. 1. Drainage Lakes , ppm of total alkali reported as potassium
plotted against the number of lakes.
known volume of water was evaporated to dryness and the resi¬
due was prepared for analysis by drying in a vacuum desiccator
at 60°C. The results for this comparison are also shown in
Table I. The per cent of total alkali in the residues prepared
from drainage lakes varies between 2.76 and 10.0. The greatest
number of the residues contain between 4 and 7 per cent of total
alkali; forty eight of the 61 lakes have residues which contain
between 4 and 6 per cent alkali.
Color and Total Alkali : Table I, Figure 3, 4
The highest color exhibited by any lake in this drainage group
was 277, the total alkali for this lake being 2.94 ppm. The lake
waters which show the lowest color show a relatively wide vari¬
ation in total alkali content. In the color range 0 to 28, samples
show a variation in total alkali from 1 to 4.88 ppm. In spite of
these extremes, there is a definite limit. Most of the lakes of the
drainage group contain water which is not very highly colored.
Of the 61 different samples which represent 58 different lakes,
288 Wisconsin Academy of Sciences , Arts, and Letters
‘/Alkali asK
Fig. 2. Drainage Lakes, percent of total alkali reported as potassium
plotted against number of lakes.
29 fall in the very narrow range of 4 - 50 color and 2.8 - 3.8 ppm
total alkali. Of the remaining lakes, 7 are practically 100 in
color and these vary from 1.99 to 3.91 ppm total alkali. Six lakes
show total alkali concentrations in the range, 4 to 4.81 ppm.
Four of these have colors of about 25 and the other 2 have colors
of about 80. Eight lakes show alkali concentrations of less than
2 ppm, five of these colors of about 10 or less, 2 colors of about
50 and 1 a color of 110.
A comparison of the percentage composition of the residues
prepared from the lake waters and the color of the waters shows
equally marked limits. With the exception of a few extreme
values, most of this drainage group of lakes are limited to the
small range of 4 - 6.7 per cent total alkali and less than 100
color. Thirty three, or more than half of the group, have resi¬
dues which show alkali percentages of 4 to 6 and the color of the
water under 50. The sample which shows the highest alkali
content comes from a low colored water and the sample which
Lohuis, Meloche & Juday — Sodium and Potassium 289
Fig. 3. Drainage Lakes , color reported in terms of the platinum cobalt
standard of “Standard methods” plotted against per cent of total alkali
shows the lowest alkali content comes from the highest colored
water.
Bound CO 2 and Total Alkali : Table I, Figure 5, 6
In this instance there seems to be a general trend and one
observes an increase in ppm of total alkali with the increase of
bound C02. An examination of Figure 5 shows 2 ppm alkali
for 4 ppm of bound C02, 3 ppm alkali for 14 ppm bound C02
and 3.5 ppm alkali for 22 ppm bound C02. It is not surprising
that such a relationship exists for, since the bound C02 indicates
the relative hardness of the lake water for most cases, the harder
the water the more dissolved mineral will be found per milliliter.
In a stricter interpretation one must admit that many variables
exist which affect this picture. However, when one notices the
small spread in the alkali concentrations, and the low magnitude
of the concentrations, some variables may be overlooked because
their direct effect on the alkali concentration will be extremely
small. Furthermore, the effect of certain variables such as tem¬
perature, free C02, acidity, sulfate, and mineral availability are
290 Wisconsin Academy of Sciences, Arts, and Letters
y}
o
o
50
100
150
Color
200
250
Fig. 4. Drainage Lakes, color reported in terms of the platinum cobalt
standard of ‘ Standard Methods” plotted against per cent of total alkali
in the residues of the waters.
5
4
V)
o
a 3
<
£ 2
Q_
Q_
I
4 12 20 20
ppm Bound COz
Fig. 5. Drainage Lakes, ppm of bound carbon dioxide plotted against
ppm of total alkali which is reported as potassium.
Lohuis, Meloche & Juday — Sodium and Potassium 291
included in a relationship which includes variations in bound
C02 and are therefore included when the bound C02 is compared
to alkali. When the results are expressed in terms of per cent
total alkali in the lake water residue and these are compared to
ppm of bound C02 for the respective water samples, the trend
described above disappears. If there is any generalisation which
one may make on the basis of this comparison, it is that the
spread in the per cent of alkali in the water residues taken from
soft water lakes in this group is greater than the spread of re¬
sults in the harder waters. However, it must be noted that a
rather large proportion of this group of lakes falls in the limits
of 4 to 6 per cent total alkali and 7 to 20 ppm bound C02. The
residue which contained the highest per cent of alkali was ob¬
tained from a soft water lake and the highest bound C02 repre¬
sents a lake water whose residue contained only 8.83 per cent
total alkali, a relatively low alkali content.
Sid fate and Total Alkali : Table I, Figure 7, 8
A comparison of ppm of sulfate and ppm of total alkali shows
the results to be quite scattered. In general it may be seen in
6
*6
o
a
X
2
10 20 30 40
ppm Bound CO 2
Fig. 6. Drainage Lakes , ppm of bound carbon dioxide is plotted against
per cent of total alkali in the lake water residues.
292 Wisconsin Academy of Sciences, Arts, and Letters
Figure 7 that the waters which contain less than 2 ppm if alkali
have less than 4.5 ppm of sulfate and most of the samples which
have more than 5 ppm sulfate have more than 2 ppm of alkali.
However, about 17 lakes which have a total alkali concentration
of between 2 and 4 ppm have a sulfate concentration of B - 4.5
ppm. Two lakes which show an alkali concentration of over
4 ppm have only slightly more than 3 ppm of sulfate. The high¬
est alkali concentration is in a lake which has the high sulfate
content of 6.58 ppm.
ppm Sulfate
Fig. 7. Drainage Lakes, ppm of sulfate (S04) plotted against ppm of
total alkali expressed as potassium.
When the per cent of total alkali in the lake water residues
is compared to the ppm of S04 for the waters from which the
residues were prepared, an even distribution is observed. The
percentage of alkali varies between 4 and 7 for samples which
vary between 2 and 7.2 ppm of S04.
The Ratio if Na/K : Table I, Figure 9.
Before these results are described it will be necessary to men¬
tion the analytical procedure used in the determination of the
Lohuis, Meloche & Juday — Sodium and Potassium 293
ppm SO4
Fig. 8. Drainage Lakes , ppm of sulfate (SO*) plotted against per cent
of total alkali in the lake water residues.
sodium and the potassium and discuss its influence on subsequent
interpretations.
In the determination of total alkali by means of the polaro-
graph, the sodium and potassium are electrolyzed together. It
has not been possible as yet to produce a separate wave for each
element because the reduction potentials of the two lie too close
together. As will be mentioned in another paper, the procedure
which was used involved measuring the height of the wave due
to sodium + potassium, determining the sodium on a separate
sample by the zinc uranyl acetate method, and getting the po¬
tassium by subtracting the sodium from the total alkali wave.
It has been customary in many cases in the past to report sodium
and potassium and ignore the less common lithium, rubidium
and caesium. A preliminary polarographic study of lake water
residues shows that this cannot be done in all cases because an
appreciable concentration of lithium was shown to be present.
Since lithium interferes with the zinc uranyl acetate determina¬
tion of sodium, this will make some of the sodium results a little
294 Wisconsin Academy of Sciences, Arts, and Letters
high. Until more lithium results are available, it will be neces¬
sary to accept the sodium results, realizing that when they are
corrected, the sodium result will be slightly lower and the po¬
tassium in these cases will be slightly higher.
Realizing the general effect of this correction and also that it
will probably affect only a small proportion of the samples de¬
scribed here, we may now consider the results in Figure 9. In
12 3 4
Ratio Na/K
Fig. 9. Drainage Lakes. The ratio of sodium to potassium Na/K calcu¬
lated from the percent sodium and potassium in lake water residues
plotted against the number of the lakes.
this group of drainage lakes we are reporting work on 60 lakes.
There are 7 lakes which fall in the range 0.1 to 0.5 Na/K, 11
lakes in the range 0.5 to 1.0 Na/K and 21 lakes between 1.0 and
1.5 Na/K. It is plain that a ratio near 1 :1 is favored for between
0.5 and 1.5 Na/K we find 32 lakes or slightly more than half of
the total number examined. When a correction is applied for
lithium, some of the sodium results will be still lower and the
ratios will be lower. It is impossible to say at present which
lakes this will affect.
Lohuis, Meloche & Juday — Sodium and Potassium 295
Seepage Lakes
Results are reported on only 29 seepage lakes. This would
ordinarily be considered a rather small number of samples upon
which one might establish a trend. In any event we will report
the variations as they occurred.
Distribution of Total Alkali: Table II, Figure 10, 11
The highest total alkali reported was 3.09 ppm while the
lowest was 0.68 ppm. Five lakes had a total alkali of about 2.5
ppm and 7 lakes had values around 0.9 ppm. Thirteen lakes
have a total alkali between 1 and 2 ppm. When the results are
reported in terms of parts per million, one looks for a relation¬
ship which involves concentration. Since the water in seepage
lakes is exceedingly soft, it may be said that the magnitude of
the results reported is in keeping with the observed softness of
the water.
When the alkali results are reported in terms of the per cent
of total alkali in the lake water residues, it is observed that
12 3 4
ppm Alkali as K
Fig. 10. Seepage Lakes, ppm of total alkali reported as potassium
and plotted against number of lakes.
Alkali as K is Number of Lakes
rv . CL. * _ *
% Alkali as K.
11. Seepage Lakes, the per cent of total alkali in the lake water
aes reported as potassium and plotted against number of lakes.
75 150 225 300
Color
Fig. 12. Seepage Lakes , color in terms of the platinum cobalt standard
of “Standard Methods” is plotted against ppm of total alkali expressed as
potassium.
Lohuis, Meloche & Juday — Sodium and Potassium 297
Table II
SEEPAGE
although the values range from 3.4 to 11.8 per cent, 20 of the
residues have total alkali of 4.8 to 7.3 per cent. Twelve of these
are about 7 per cent and 8 are about 5.5 per cent.
Color and Total Alkali : Table II, Figure 12, 13
With the exception of two lakes, the colors for the group are
under 70. Fourteen lakes which have alkali concentrations of
0.68 to 2.0 ppm are all under 15 in color. Two more lakes which
have colors less than 15 have alkali concentrations of about 3
ppm. There are two relatively highly colored lakes included in
this group, Tadpole with a color of 101 and Helmet with a color
of 268. The alkali concentration in Tadpole is 2.75 ppm while
that in Helmet is 2.45 ppm.
Here again, when the per cent of total alkali in the lake water
residue is compared with color of the original water, interesting
deviations as well as expected groupings are observed. For the
two highly colored lakes, Helmet water residue contains 3.44 per
cent total alkali and Tadpole contains 4.85 per cent. Twenty one
298 Wisconsin Academy of Sciences, Arts, and Letters
if*
o
75
<50 225
Color
3 00
Fig. 13. Seepage Lakes , color in terms of the platinum cobalt standard
of “Standard Methods” is plotted against per cent of total alkali in the
water residues.
of the thirty lakes contain between 5 and 8 per cent total alkali
in their water residues. Bird Lake gives residues which con¬
tain the highest per cent of alkali, 11.80, and the color for this
lake is only 9. The residue which contains the lowest per cent
of total alkali is Jag and the color of the water is 18. Men¬
tion may be made of certain lakes, the waters of which show
little or no color. Residues from Anderson lake contain 5.13 per
cent alkali, Diamond — 5.6 per cent, Pauto — 7.15 per cent, and
Stormy— 6.45 per cent. From this we see there are extremes in
all directions, high color and low alkali, low color and high alkali.
Bound C02 and Total Alkali : Table II, Figure 14, 15
The two figures which compare the distribution of total alkali
to bound C02 merely illustrate the relatively small range in
which the waters of the seepage lakes are grouped. All but 5 of
these lakes have a bound C02 of less than 6 ppm. Seventeen of
the lakes have alkali limits between 0.68 ppm and 1.77 ppm while
the bound C02 varies between 1.2 and 4.1 ppm. The few scat-
3
00o°
,1 , 1-1 I _ L. _ _ 1 .... 1 1.1 _ l _ _ I , - L,
I 2 3 4 5 6 7 8 9 |6 » I 12
ppm Bound COz
Fig. 14. Seepage Lakes , ppm of bound carbon dioxide plotted against
ppm of total alkali which is expressed as potassium.
!2
1 i
10
9
ID 8
O
“5 7
6
O'
5
4
3
Fig. 15. Seepage Lakes , ppm of bound C02 plotted against per cent
of total alkali in the lake water residues.
I 2 34 56 789 10 II 12
ppm Bound COz
300 Wisconsin Academy of Sciences , Arts , and Letters
tered samples which have the hardest waters of the group of 30
lakes have bound C02 between 8 and 10.5 ppm and alkali values
between 0.89 and 2.42 ppm.
When the alkali results are reported in terms of per cent,
more than 15 lakes which have a bound C02 content of less than
5 ppm have residues which contain between 4.5 to 8 per cent of
total alkali. The five lakes which have the hardest water of the
group have residues which contain 4.5 to 5.5 per cent total al¬
kali.
Sulfate and Total Alkali : Table II, Figure 16, 17.
Here again the limited number of lakes examined and the
relatively low mineral content of their waters prevents a more
critical presentation of the results obtained. When the ppm of
total alkali is plotted against ppm of sulfate, it is observed that
all but 5 of the lake waters fall in the range 0.7 to 2.0 ppm alkali
and 1.0 to 5.11 ppm of sulfate. Twelve lakes fall in the more
restricted range of 0.7 to 2.0 ppm alkali and 2 to 4.3 ppm sulfate.
ppm SO4
Fig. 16. Seepage Lakes , ppm of sulfate (SO*) plotted against ppm of
total alkali which is expressed as potassium.
Lohuis, Meloche & Juday — Sodium and Potassium 301
CO
o
ppm Sulfate (SO4)
Fig. 17. Seepage Lakes , ppm of sulfate plotted against per cent of total
alkali in the lake water residues.
In Figure 17, per cent of total alkali is plotted against ppm
of sulfate. The residues for the most part contain a rather high
per cent of total alkali, 5 to 7.5 per cent, and are quite evenly
distributed between lakes of low and high sulfate content. This
is also true of the residues which contain the low percentages of
total alkali.
The Ratio of Na/K : Table II, Figure 18.
In considering the ratio of sodium to potassium in samples
taken from seepage lakes, the ratio Na/K is over 1.5 for only
3 lakes. Twenty three lakes have a ratio of near or less than
one. As observed in the similar paragraph on drainage lakes,
a possible correction for the interference of lithium will only
lessen the results for sodium and therefore the values for the
ratio will be even lower.
Total Alkali of Drainage and Seepage Lakes
Parts per million : Figure 1, 10.
In making this comparison it must be remembered that the
lakes in the drainage group although not considered hard water
302 Wisconsin Academy of Sciences, Arts, and Letters
Ratio Na/K
Fig. 18. Seepage Lakes. The ratio of sodium to potassium, Na/K, is
plotted against the number of lakes in a given range.
lakes contain more dissolved mineral in general than do the lakes
in the seepage group which we examined. The bound C02 for
the drainage lakes varies between 2.5 and 30.5 ppm while the
bound C02 for the seepage lakes varies between 1.2 and 10.5 ppm
with most samples falling between 1 and 4 ppm. The total alkali
for the drainage lakes varies between 1 and 5 ppm and most
lakes contain around 3 ppm. For the seepage lakes the total
alkali varies between 0.68 and 3.09 ppm; 21 lakes have less than
2 ppm and 17 lakes have less than 1.5 ppm alkali.
Per cent total alkali in residues : Figure 2, 11.
The per cent of total alkali in residues of the waters from
drainage lakes varies between 2.76 and 10.0 and most of the
residues contain 5 to 6 per cent while most of the lakes of the
drainage group fall in the range of 4 to 7 per cent. It is inter¬
esting to note that while there are extremes in both groups
the per cent of alkali in residues from the seepage lakes falls for
the most part in the limits set for the drainage group, most resi¬
dues in the seepage group containing between 4.5 and 7.5 per
Lohuis, Meloche & Juday ■ — Sodium and Potassium 303
cent total alkali. It is true that the distribution in the seepage
group is somewhat different than it is in the drainage group.
The residue containing the lowest per cent of alkali is in the
drainage group and the residue containing the highest per cent
of alkali is in the seepage group.
Conclusion
The total alkali has been determined on a group of drainage
and seepage lake waters and their residues. The total alkali
was determined by means of the polarograph, sodium by the
zinc uranyl acetate method and potassium by the difference be¬
tween the two. Since total alkali was expressed as potassium,
the sodium values were converted to potassium equivalents be¬
fore the subtraction could be made.
The results were tabulated with color, free and bound carbon
dioxide, pH and sulfate and the following comparisons were
made :
1. Distribution of total alkali in drainage lakes on the basis
of ppm and per cent of alkali in the residues.
2. Variation of alkali with the color of the water.
3. Variation of alkali with the bound carbon dioxide of the
water.
4. Variation of alkali with the sulfate content of the water.
5. Variation of the ratio, Na/K, for the various lakes in the
drainage group examined.
These comparisons were duplicated for seepage lakes. Finally,
the total alkali and the ratio, Na/K, were compared for drainage
and seepage lakes.
One can best summarize the results which were obtained by
examining the figures supplied with the text. However, it
might be well to emphasize a few points regarding the abundance
of total alkali in the water and the ratio of sodium to potassium.
Although the drainage lakes are not hard water lakes, they do
contain more dissolved material per liter than do the seepage
lakes. This relationship also holds for total alkali. When one
prepares residues by evaporation of the water, an analysis of
these residues on a per cent basis shows that, in general, the
304 Wisconsin Academy of Sciences , Arts , and Letters
residues from the seepage lakes contain a higher proportion of
total alkali. In regard to the ratio between sodium and potas¬
sium, the drainage lakes for the most part show a ratio of nearly
1:1 with a possible lithium correction lowering this ratio to less
than one, i.e., the sodium would be lowered. For the relatively
small number of seepage lakes which were examined, the ratio
of sodium to potassium is for the most part less than 1:1 and
any possible lithium correction would make this value still lower,
i.e., expressed as Na/K the values would be less than one.
Studies are now in progress which concern the composition
of bottom muds of the lakes. We hope to extend the compari¬
sons of total alkali content of the waters when this information
is available. Lithium analyses are now being made on the same
residues used in this study and should corrections be necessary
they will be published in a supplementary report.
Madison, Wisconsin.
Laboratory of Limnology
GEOLOGY AND GROUND WATER OF THE TROUT LAKE
REGION, VILAS COUNTY, WISCONSIN*
Carl Fries, Jr.
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 81.
Introduction
The material for this report was gathered under the direction
of the Wisconsin Geological and Natural History Survey in an
effort to obtain information on ground water movement and the
relation of the geology to the chemistry of the waters of several
lakes in northeastern Wisconsin. The problem had been raised
as to whether these lakes represent perched water tables or
whether there is interchange of lake and ground waters.
The area studied lies in southern Vilas County, Wisconsin,
and detailed work was limited to the southern half of T.41N.,
R.7E. The field work was done during July and August of 1936
near the Limnological Laboratory of the Wisconsin Geological
and Natural History Survey at Trout Lake, Wisconsin. The ge¬
ology was determined from road cuts, test pits, and well cuttings.
The test pits were about three feet wide, five feet long, and four
to twelve feet deep, permitting examination of the material in
place. The well cuttings were used for determining the nature
of the deeper drift.
The elevations of some of the lakes were determined under
the direction of the writer by students of the Armour Institute
of Technology in training in the summer camp at Trout Lake.
Other elevations were obtained by the writer with a telescopic
alidade, plane table, and engineer’s rod. The chemical analyses
of water samples were made by chemists of the Laboratory. In¬
formation on well depths and water analyses was obtained from
the Survey records. Samples of well cuttings were taken from
wells drilled during the summer. Numerous samples of soil and
glacial deposits were collected and examined for size, assortment,
and presence of calcareous material.
This investigation was supported by a grant from the Brittingham Trust Fund.
305
306 Wisconsin Academy of Sciences, Arts, and Letters
Three samples of clay were analyzed quantitatively for con¬
tent of calcium and magnesium. These came from the till in the
Winegar moraine, the till in a drumlin west of Trout Lake, and
a deep well east of Crystal Lake. The first two were screened
through a 150 mesh screen, and the third was used as found.
The material was dried at 100°c., and a five gram sample of each
was stirred for five minutes in 30 cc. of 1 :1 HC1. The solutions
were neutralized and filtered, and the calcium precipitated as an
oxalate and titrated with standard permanganate solution. The
magnesium was then precipitated as magnesium ammonium
phosphate, filtered, ignited, and weighed. The calcium and mag¬
nesium were calculated as carbonates and percentages were fig¬
ured by weight. This procedure seems satisfactory for the pur¬
pose at hand.
Helpful criticism and suggestions were offered by W. H.
Twenhofel, E. F. Bean, C. Juday, and It. R. Shrock. Aid in in¬
terpretation of glacial history was kindly offered by F. T.
Thwaites.
Geology of the Area
The bedrock underlying the area studied seems to be entirely
crystalline (Thwaites 1929). Inasmuch as crystalline rock ex¬
cludes practically all free water, except that held in joints, a
study of the ground water in this region is largely limited to a
study of the overlying unconsolidated drift. This drift is largely
composed of gray and pink granites, pink rhyolite, gabbro, ba¬
salt, gneiss, and schist. Lesser quantities of red sandstone, red
slate, red iron formation, and quartzite are present. There is an
occasional pebble of chert, but no limestone or dolomite was
found near the surface. The known maximum depth of the drift,
found by well drilling in section 7, T.41N.,R.7E. exceeds 225
feet.
Moraines
The Muskellunge moraine (Fig. 4) is an east-west trending
ridge of discontinuous, irregular knobs. These knobs are com¬
posed of stony, ill-assorted material which contains many large
boulders and very little clay. Much of the moraine is kame de¬
posit of interbedded sand and stony gravel. The upper two and
one-half feet of the material is light tan in color and is underlain
Fries— Geology and Ground Water
307
by two feet of dark brown-red, oxidized drift. Below the zone
of weathering, of which the depth is quite irregular, the color of
the drift is light brown-red. This is not a calcareous drift.
Steep-sided kettles are numerous. Few of these contain water,
probably due to their comparative height and the porosity of the
underlying material; they are not well enough sealed to hold
water above the general water table. The highest part of the
Muskellunge moraine is the knob on which the fire tower stands.
V
-.8'^f -1.7
Fig. 1. Location of test pits around Weber Lake indicating relation be¬
tween lake level and ground water table. The arrows indicate the direction
of the dip of stratified sands in the test pits. Figures at the left of the
arrows indicate the water table level relative to the lake level on August 31,
1936.
308 Wisconsin Academy of Sciences , Arts, and Letters
This is 1847 feet above sea level and two hundred feet higher
than the sand plain north of the moraine in which Crystal and
Weber lakes lie.
The Boulder moraine (Fig. 4) is even more discontinuous
than the Muskellunge. Like the latter, it seems to be a typical
recessional moraine composed chiefly of kames. It is lower in
elevation, and much of it either is covered by outwash or was
eroded away by water which drained from the melting ice to the
north. North Trout Creek lies in a broad valley which cuts
through the moraine.
The Winegar moraine is a broad, continuous ridge which ex¬
tends across the northern portion of the county and is about six
miles north of the Boulder moraine. It differs from the two
southern moraines in that it is composed chiefly of till which is
brighter red in color and contains a much larger percentage of
clay. Analysis of the fine material from the till of this moraine
which passes through a 150 mesh screen indicates that it con¬
tains 7.45% CaC03 and 0.51% MgC03 by weight. It is probable
that this is a terminal moraine of the fourth Wisconsin (Man¬
kato) substage (Leverett 1929). It may possibly cover a reces¬
sional third Wisconsin (Cary) moraine of red sandy till
(Thwaites 1937). Striae on many outcrops of bedrock north of
Winegar indicate that the ice which brought the red clay till
moved S 10°-20°W.
Ground Moraine
The region east of Trout Lake, adjoining Allequash and Little
John lakes (Fig. 4), is composed of material and has topography
similar to that of the Muskellunge and Boulder moraines. The
elevations of the hilltops are unequal and considerably higher
than the surrounding outwash plains. Small patches of till are
present, though much of the material is partially washed. The
till, like that in the moraine, is low in clay content and is light
brown-red in color. A large esker trending N 30°E extends
from Allequash Creek, near Trout Lake, for a distance of about
a mile.
Drumlins
A group of drumlins lies to the west and northwest of Trout
Lake. (Fig. 4). These hills rise high above the level of the
Fries — Geology and Ground Water
309
outwash plains and have their long axes trending about S 35° W.
A large road material pit in a drumlin on Highway 51, west of
Trout Lake, shows the material to be unstratified and unassorted
drift. The weathered zone is like that of the till in the moraines.
Many of the rounded stones of the coarser-grained igneous rocks
are so weathered that they can be easily crumbled by hand. Gla¬
cial striae are excellently preserved on the large boulders. Analy¬
sis of the fine material which passes through a 150 mesh screen
indicates that it contains 6.50% CaC03 and 0.47% MgC03 by
weight, only slightly less than in the till of the Winegar moraine.
Fig. 2. Composite diagram illustrating material and structure of de¬
posits surrounding Weber and Crystal lakes. The letter A represents sand
blackened with organic matter; B unstratified, light-gray, leached sand;
C unstratified fine- to coarse-grained, dark brown sand containing numerous
pebbles; D fine- to coarse-grained, light yellow sand containing pebble beds,
well defined stratification with topset overlying foreset beds; E zone of iron
concentration, banded, mottled, bright rusty-yellow sand, partially cemented ;
F laminated, fine- to coarse-grained, uncemented, pale yellow sand; 0 fore¬
set beds, fine- to coarse-grained, light yellow sand without pebbles, well de¬
fined stratification; T topset beds; W water table.
310 Wisconsin Academy of Sciences , Arts , and Letters
Color and calcareous content suggest that there is a genetic re¬
lationship between the two drifts. Only one striation indicating
direction of ice movement has been found south of the Winegar
moraine. This is on a bedrock ledge in section 34, T.43N.,R7E.
and trends about S 50°W. (Thwaites 1929).
A unique feature of the drumlin observed is a well defined
cleavage concentric to its surface. The layers are one-sixteenth
to one-eighth inch in thickness. The cleavage surfaces curve up
over and down under the pebbles and bowlders. All observed
material below the active zone of weathering shows this pseudo¬
bedding. The till is extremely hard packed, stands in vertical
faces, and is easily separated into thin layers along these cleav¬
age surfaces. Alden (1905) has described this feature in till of
drumlins in southern Wisconsin. Many of the pebbles and bowl¬
ders in the till are coated with iron oxide which gives them a
dull blue-gray sub-metallic luster. This coating is not present
on pebbles in the outwash.
Outwash
The greater part of the surface material south of the Wine-
gar moraine is water-sorted, stratified sand and gravel. This
material offers a minimum of resistance to ground water move¬
ment. It forms southwestward sloping plains of varying eleva¬
tions which were determined by the drainage outlets available
at the time of its deposition, and the majority of the lakes lie in
depressions in this outwash. The plains exhibit erosional and
ice contact terraces. The former are well developed along Little
Trout River ; ice contact terraces border Stephenson Creek. The
highest plain, against the south edge of the Muskellunge mo¬
raine, lies at an elevation of about 1700 feet. A similar high
plain, bordered by ice contact terraces on the south, lies north of
Trout Lake. A lower plain, about 1655 feet in elevation, extends
east-west just north of the Muskellunge moraine. A third plain,
at about 1620 feet, surrounds Trout Lake and extends westward
along Little Trout River.
Deep Drift Deposits
Information on deep drift deposits was obtained from well
drillers and from examination of a few samples from wells
drilled during the summer. The description of the material is
Fries— -Geology and Ground Water
311
given below in tabular form. The wells are located by number
in Figure 4.
Well No 6
East end of Crystal Lake. Elevation 1650 feet, 5 feet above water table.
Thickness Depth
Description (feet) (feet)
Yellow sand . . 30 0- 30
No information . ; . 50 30- 80
Laminated, rubbery, calcareous, gray clay . 8 80- 88
Gray sand with thin laminae of gray clay . 7 88- 95
Gray, sandy gravel . 7 95-102
Coarse, well washed, gray gravel, plentiful
water supply . 5 102-107
Quantitative analysis of the clay from this well shows it to
contain 31.65% CaC03 and 2.52% MgC03. These percentages
contrast significantly with the clays of the upper drifts. The
sand and gravel below the clay contain a much higher percent¬
age of dark-colored, ferro-magnesian minerals than does the
surface material. This gray gravel also contains pebbles of
dolomite, whereas none was found in the upper drift.
Well No. 5
Between Crystal and Mluskellunge lakes. Elevation 1650 feet,
5 feet above water table.
Thickness Depth
Description (feet) (feet)
Well driller reports material similar to that in No. 6 98
Well No. U
Just south of Allequash Creek along the shore of Trout Lake.
Elevation 1622 feet, 8 feet above water table.
Thickness Depth
Description (feet) (feet)
Yellow, gravelly sand . . . . 10 0-10
Poorly sorted, yellow to brown sand, silt and gravel 25 10-35
Gray, sandy gravel, good water supply . 5 35-40
The gray, sandy gravel contains occasional pebbles of dolo¬
mite and is composed of a higher percentage of ferro-magnesian
minerals than is the yellow sand and gravel above it.
Well No. 3
Near Trout Lake shore at Laboratory. Elevation 1620 feet, 5 feet
above water table.
Thickness Depth
Description (feet) (feet)
Brown to yellow sand and gravel . 35 0-35
Gray, sandy gravel . . . 22 35-57
Elevation Elevation
312 Wisconsin Academy of Sciences , Arts, and Letters
Well No 2
One-eighth mile north of Stephenson Creek along shore of
Elevation 1620 feet, 6 feet above water table.
Thickness
Description (feet)
Brown to yellow sand . 35
Reddish-brown till . 40
Water-bearing, coarse gravel . 10
Trout Lake.
Depth
(feet)
0-35
35-75
75-85
Well No. 1
Point Campsite, about 300 feet from Trout Lake shore. Elevation 1640 feet,
25 feet above water table.
Thickness Depth
Description (feet) (feet)
No information . 200 0-200
Rubbery, gray clay . ? 200- ?
Coarse, water-bearing gravel . ? ? -223
In all of the records of these deep wells it is notable that
gravel deposits underlie either clay, till, or outwash of a differ¬
ent character. The deep drift is gray in color, contains pebbles
of dolomite, and is composed of more basic rock than is the sur¬
face drift. It will be shown that the character of the water
coming from the deeper horizons reflects the changed composi¬
tion of the material.
Fig. 3. Diagrams illustrating the irregular drop of the water table
between lakes. Dots represent elevation of water table in test pits.
313
Fries — Geology and Ground Water
Material Around Weber and Crystal Lake Basins
Weber and Crystal lakes lie in a flat, sandy plain north of
the Muskellunge moraine. (Fig. 4). The material of this plain
is principally sand and silt which shows fair sorting. The loca¬
tions of test pits dug around Weber Lake are shown in Figure 1.
The arrows indicate the direction of dip of the stratified sands.
The lower beds dip at an angle of 30° to 35° and these are over-
lain by nearly horizontally stratified sands. (Fig. 2). No strati¬
fication is observable in the weathered zone extending down-
Xy/\ Moraine. 1=3 Out wash and Thm Till fl 1118 Drum I'm.
ES3 Out wash Plain. © Deep Well - Hypothetical Boundary
Fig. 4. Glacial features of Trout Lake region. After F. T. Thwaites,
1929.
314 Wisconsin Academy of Sciences, Arts, and Letters
ward for two and one-half to three and one-half feet from the
surface, probably due to disturbance by roots and chemical alter¬
ation. The top two to three inches of the material grades from
black, through gray to brown. The remainder of the weathered
zone is dark brown in color and contrasts with the light yellow
unaltered sands below. Bright yellow and brown, irregular
bands of iron concentration are present at or just above the
water table, four to eight feet below the surface. Test pits
around Crystal Lake reveal the same type of material and struc¬
ture. Between the two lakes the sands dip toward Weber Lake.
Two test holes were made with the use of a small casing and
bailer near the west shore of Crystal Lake in about four feet of
water. The depth of the holes was from three to five feet. Neith¬
er clay nor gravel was found, but only gray to yellow sand con¬
taining small quantities of organic matter. This seems to be
leached beach sand derived from the shores by erosion during
periods of high water. With the apparatus available it was
not possible to go deeper.
It is possible that this region was a large lake which received
melt water from blocks of ice to the north, south, and east. The
present basins were preserved by stagnant masses of ice. Drain¬
age from the lake was toward the west around the southern side
of the huge ice mass which must then have occupied the Trout
Lake basin. Inasmuch as the bedding of the sands around Crys¬
tal and Weber lakes is apparently undisturbed, it is suggested
that glacial drainage continued a short time after the ice blocks
in the basins had melted, thus causing partial fills by sands.
This would account for absence of till near the surface, absence
of a till seal around the sides of the basins, and absence of lake
clays.
Summary
Study of the materials in well cuttings, test pits, and road
cuts indicates that three types of drift overlie the crystalline bed¬
rock of this region. The deep drift is gray in color and contains
a high percentage of dark-colored minerals. In it are pebbles
of gray dolomite and deposits of gray, calcareous clay. This
drift is found only in deep wells and is not known to be at the
surface anywhere in the area. It is possibly of Tazewell age or
older.
Fries — Geology and Ground Water
315
A second type of drift is brown-red in color. It contains
very little clay. The percentage of dark-colored minerals is
small and dolomite seems to be absent. The fine material pass¬
ing a 150 mesh screen is low in carbonates. From the informa¬
tion available, it appears that its thickness is extremely vari¬
able. This drift occurs at the surface south of the Winegar
moraine. It is probably of Cary age.
A third type of drift contains a high percentage of clay and
is bright red in color. Calcareous content is low. This drift is
present at the surface in and northward from the Winegar mo¬
raine. It is probably of early Mankato age. (Leverett 1929).
The evidence is sufficient, the writer feels, to conclude that
these drifts represent three separate advances of glacial ice;
the first may have come from nearly east, and the second and
third came from the northeast and the north-northeast.
The lakes in the Trout Lake region do not lie in basins whose
sides are sealed by till or clay. Sand and gravel are mixed with
small quantities of silt and clay, thus permitting slow movement
of water.
Chemistry of Lake and Ground Waters
Samples of water taken from lakes, wells, and pits were
analyzed at the Survey Laboratory at Trout Lake throughout
the summer. These analyses indicate the parts per million of
bound carbon dioxide, calcium, magnesium, iron, and silicon di¬
oxide which are in solution in the water. Figure 6 shows the
locatons from which these water samples were taken; the fig¬
ures represent the parts per million of calcium plus magnesium.
Lake Waters
All the lakes in the area are classified as soft water bodies.
The bog lakes which are isolated from drainage, regardless of
elevation, are by far the softest, and their waters are almost as
pure as though distilled. The seepage lakes — Crystal, Long,
Geneva, Weber, Ruth, Little John, Jr., and Muskellunge — form a
second group with slightly more mineral matter in solution.
The drainage lakes — Allequash, Little John, Mann, and Trout —
have the greatest quantity of dissolved minerals and form a
third group. (Table I).
316 Wisconsin Academy of Sciences , Arts , and Letters
e Test pit. — — Drainage divide.
Fig. 5. Water table elevations in test pits and lake elevations.
The bog lakes are almost totally sealed from the surrounding
ground water by accumulated organic matter on their sides and
bottoms. They receive a small amount of run-olf, and precipi¬
tation in the form of rain and snow. Thus but little mineral
matter is brought into them. The abundant organic matter
growing in these lakes undoubtedly uses some of that supplied
and locks it up in the bottom deposits.
The lakes with neither inlets nor outlets, but which are not
yet sealed by organic matter, apparently receive and deliver
water through their basins. The rate or extent of flow can not
Fries— Geology and Ground Water
317
be determined, and that it takes place is purely hypothetical.
Further evidence rests in water table and lake elevations. At
any rate, these waters are, in general, harder than the bog lakes.
The drainage lakes are notably harder than the other types
of lakes. It must be inferred that these lakes receive a plentiful
supply of water by seepage, inasmuch as some of them do not
have inlets, yet they supply constant outlet flows.
Table I
Chemical analyses of different hypes of lake waters. The results are
stated in milligrams per liter of water.
Ground Water at the Water Table
The ground water at the water table is related to nearby
lake waters in mineral content. Crystal Lake has 1.7 parts per
million of calcium plus magnesium in its water. The ground
water at the east end of the lake contains 2.3 parts. Samples of
ground water around Weber Lake contain 2.2, 2.5, 5.5, 2.2, 2.2,
1.9, 1.3, 3.6 parts, and the lake contains 1.7 parts. Figures for
Little John Lake are 18.0 for the lake and 21.1 for the ground
water at the north end of the lake near its outlet. Mann Lake
contains 19.7 parts and the ground water varies from 6.3 to
15.9. The figure for Mann Creek is 20.6 and for the ground
water nearby 20.0. (Fig. 6),
318 Wisconsin Academy of Sciences, Arts, and Letters
0 Well Water.
a Surface Water. {■ Figures are parts per million
0 Test Pit. J Ca plus Mg.
Fig. 6. Calcium and magnesium content of waters from wells, test
pits, and the surface of the lakes.
■ -v^
From the figures quoted above it is apparent that the ground
water is similar to the adjacent lake waters, with the exception
of the bog lakes which are in a special category. This similarity
suggests that there is interchange of lake and subsurface water,
Fries— Geology and Ground Water
319
even though there are no visible surface inlets or outlets. We
may conclude, then, that the lakes do not lie in sealed basins.
Well Water
The greatest number of the water analyses were made on
samples taken from wells. The depth of only a small percentage
of the wells is known. These are chiefly the wells which have
been drilled within the past few years. In all known cases the
shallow wells are notably softer than the nearby deep wells. At
the east end of Crystal Lake, a very shallow well has 1.6 parts
of calcium plus magnesium per million, while a well with a
depth of 107 feet, within 30 feet of this shallow well, has water
which contains 15.6 parts. A similarly deep well at the north
side of the lake tests 14.8 parts.
On the southwest shore of Trout Lake, a shallow well was
deepened during the summer from 22 to 140 feet. The hardness
of the water increased from 13.1 to 27.6 parts of calcium plus
magnesium. The wells at Rocky Reef on the southeast shore of
Trout Lake are peculiarly hard, even though they are shallow.
This may possibly be due to surface contamination by sewage,
inasmuch as wells of similar depth nearby are much softer. A
225 foot well on the Point between the north and south portions
of Trout Lake contains 92.0 parts of calcium plus magnesium
per million. This is the hardest water known in the region.
Although wells vary in hardness at the same depth from
place to place, in every known case the hardness increases with
depth, discounting possible surface contamination. The solvent
action of water increases with depth, but the enormous increase
in hardness in the well waters in the region cannot be thus ex¬
plained. Evidence indicates that it is due to material which con¬
tains more limestone or dolomite than does the surface drift.
This increase is explicable if a calcareous drift of different age
and origin underlies the surface deposits.
Water Table Elevations
Lakes
The Muskellunge moraine forms the drainage divide between
the Manitowish River to the north and the Wisconsin River to
the south. (Fig. 4). It is apparent from the map that there is
320 Wisconsin Academy of Sciences , Arts , and Letters
a regular decrease in elevation of the lakes from the divide
northward and westward. It seems probable that Trout Lake
receives drainage from the south, east, and north by both run¬
off and seepage. The lakes are thus connected and form parts
of a sloping water body whose flow is retarded and regulated by
the nature of the surficial deposits, which in this region are
chiefly sand and gravel with varying small quantities of silt.
Bog lakes No. 3 and No. 4 are several feet higher than would be
expected. (Fig. 5). These lakes are small and lie in basins
lined with many feet of organic matter which serves to nearly
seal them. Thus they do not drop as rapidly as the water table
during the dry season of the year when the latter is being low¬
ered by seepage and plant life. Readings taken early in the
spring and late in the autumn would more clearly reveal the
relation between the bog lakes and the water table.
Water Table
Figure 5 shows both water table and lake elevations. A cur¬
sory study of these elevations strikingly brings out the relation
between the two. It is shown that the water table slopes gradu¬
ally between two lakes of different elevation, indicating drainage
from one to the other. (Fig. 3). It is also noticeable that al¬
though the water table is lower than a lake on one side, yet it is
higher than the lake on some one of the other sides. Figure 1
illustrates relative readings taken early and late in the summer
around Weber Lake. In every case the reading of the water
table taken in early July was higher than that taken at the end
of August. The early reading showed the water table to be
higher at the east end of the lake. In late August, however, all
readings showed the water table to be below lake level. The
lake was raised by rains during the last two weeks in August.
All of this water was probably absorbed by the dry sands
around the lake before reaching the water table. Thus the effect
was noticeable on the lake level while the water table continued
to lower by seepage. There is an apparent lag in equalization
between the two.
The foregoing evidence from elevations of water table around
the lakes studied leads us to the conclusion that the basins in
which they lie are not sealed; they do not represent perched
Fries — Geology and Ground Water 321
water tables ; they are integral parts of the ground water of the
region.
Conclusions
The region under consideration has been found to be one of
surficial stony, sandy, non-calcareous drift. This drift contains
little clay. A lower, calcareous drift of different age and origin
is shown to underlie this at varying depths. It contains lenses
of calcareous clay, sand, and gravel.
Upper ground water is notably soft, but there is an increase
in hardness with increase in depth, undoubtedly due chiefly to
the character of the lower drift. The hardness of the lake wa¬
ters increases relatively to their distances away from the divide,
since the ground water must pass through greater quantities
of drift. The hardness of the water in Mann Lake is explicable
if the calcareous drift lies higher than lake level under the mo¬
raine to the south.
The seepage and drainage lakes in the region, although in ice-
formed kettles, are not sealed by till and do not represent
perched water tables. The only evidence of any seal from ground
waters lies in the organic matter which lines the basins of the
bog lakes. Evidence points to exchange of lake and ground
water, slight in the case of seepage lakes and pronounced in the
drainage lakes. The water table does not descend away from
the divide at a uniform gradient ; it is characterized by irregular
drops. This can be expected only where materials are hetero¬
geneous as they are in this region.
Suggestions for Future Research
Little is known of the glacial history of this region prior to
the retreat of the last ice. For a better understanding of the
character of the ground water it is suggested that the number,
age, and origin of the different drifts be studied, in conjunction
with their lithological composition. Tracing of moraines and
location of bedrock outcrops in the neighboring territory may
aid in revealing direction of movement of glacial ice. Arrange¬
ments should be made with well drillers to secure samples taken
at five foot intervals from all new wells. Detailed analysis of
mechanical and mineralogical composition of these samples
322 Wisconsin Academy of Sciences , Arts, and Letters
might reveal the exact source of the minerals in solution in the
waters, as well as aid in deciphering glacial history.
In the study of ground water movement it is suggested that
observation wells be established around several of the lakes, ex¬
tending below the water table and so constructed that measure¬
ments of elevation can be made frequently throughout the year.
Gauges should be established on some of the lakes and frequent
checks made on changes of level. Precipitation data can be cor¬
related with these results. Dyes or chemical solutions may be
introduced into the ground in an attempt to gain positive evi¬
dence of movement. This may be found effective particularly
where the slope of the water table is greatest and movement is
at its maximum.
Further study on the problem of lake basin seal is suggested.
Inasmuch as the material surrounding the lakes contains no evi¬
dence of any seal it is suggested that careful analysis be made
of material taken from test holes into their bottoms in an effort
to locate the position of the till which dropped from the melting
ice blocks which formed them. This study should include analy¬
sis and determinations of thickness of organic matter and sedi¬
ments in the deeper parts of the lake basins.
Literature
Alden, W. C. 1905. The drumlins of southeastern Wisconsin. U. S. Geol.
Survey. Bull. No. 273. Washington.
Leverett, Frank. 1929. Moraines and shorelines of the Lake Superior Basin.
U. S. Geol. Survey. Prof. Paper No. 154 A, pp. 29-30. Washington.
Thwaites, F. T. 1929. Glacial geology of part of Vilas County, Wisconsin.
Trans. Wis. Acad. Sci., Arts and Let. 24: 109-125.
Thwaites, F. T. 1937. Outlines of glacial geology, pp. 82-84.
THE DISTRIBUTION OF HETEROTROPHIC BACTERIA
IN THE BOTTOM DEPOSITS OF SOME LAKES*
Arthur T. Henrici and Elizabeth McCoy
From the Departments of Bacteriology of the University of Minnesota
and the University of Wisconsin and the Limnological Laboratory of the
Wisconsin Geological and Natural History Survey. Notes and reports
No. 74.
Introduction
In recent years limnologists have given more attention than
formerly to the role which bacteria, through the chemical trans¬
formations that they bring about, play in the economy of lakes ;
and since it has been learned that the concentration of bacteria
in bottom mud is vastly greater than that in the water above,
attention has been directed toward a study of the bacterial flora
of the bottom deposits of lakes. As yet this knowledge is very
fragmentary, due to some extent to a lack of interest on the part
of bacteriologists and to the laborious character of the work in¬
volved, but more particularly to a lack of suitable apparatus for
the collection of samples. Further, a lack of interest in the bacte¬
riology of bottom deposits may be attributed to a lack of knowl¬
edge of the conditions under which the bacteria are growing,
i.e., of the chemistry and physics of the deposits themselves.
Nevertheless there is a growing realization that such inves¬
tigations may throw considerable light upon the problem of lake
ecology. From our knowledge of conditions in lakes we may
surmise in a general way what takes place in the bottom de¬
posits, and a large part of the published literature dealing with
the bacteria of lake bottoms consists of such speculation rather
than of deductions from observed data. References to the earlier
literature may be found in Naumann’s monograph (Naumann,
1930).
It is probable that the greater part of the organic matter of
lakes is derived from the plankton, though it is obvious that this
* This investigation was supported in part by a grant from the Wisconsin Alumni Research
Foundation.
323
324 Wisconsin Academy of Sciences , Arts , and Letters
proportion will vary greatly with different lakes. Upon the
death of the plankton organisms, the more soluble or more readily
decomposable organic substances are rapidly leached or fer¬
mented out. The remainder settles to the bottom, probably carry¬
ing with it an attached mass of bacteria already active in their
work of decomposition.*
The surface layer of the bottom deposit therefore contains,
especially in eutrophic lakes, a high concentration of organic
matter, and a large number of bacteria. Assuming that the
organic matter is sufficiently decomposable, these conditions
favor an enormous consumption of oxygen and apparently even
in shallow, unstratified lakes, conditions are completely an¬
aerobic a few millimeters below the mud-water surface (Grote,
1934). It is, of course, fundamental to the problem to know
something regarding the nature of this organic matter, and its
decomposability by various types of microorganisms ; our knowl¬
edge of this matter is, however, meagre. Black (1929) has fur¬
nished some information on the proportion of organic matter
in the bottom deposits of various lakes. Allgeier, Peterson, Ju-
day, and Birge (1932) have reported on the decomposability of
the organic matter in several samples from Lake Mendota, and
concluded that this organic matter consisted mainly of lignin
and of protein-like substances. Steiner and Meloche (1935)
found that 30 to 48 per cent of the total organic deposit is lignin.
Speranskaja (1935) reported on the organic matter of various
lakes from the neighborhood of Moscow. By analyses made fol¬
lowing Waksman’s scheme for soil, he found that lakes in which
oxygen disappeared from the hypolimnion during stagnation
showed a lower ash content and a higher value for total nitrogen
than lakes which contained dissolved oxygen throughout the
year. In the autumn the bottom deposits contained more nitro¬
gen, sugar, hemicelluloses and cellulose than after winter stag¬
nation. The following G/N ratios are of interest : plankton,
5.59; bottom deposit in the fall, 8.38; at the end of winter,
10.60. Lignin-like complexes formed the largest component in
the organic analyses.
* It has been shown by Waksman, Reuszer, Carey, Hotchkiss and Renn (1933) that in the
sea a large proportion of the bacteria are attached to diatoms or other plankton organisms, and
unpublished data of our own indicate that this is also true of lakes; the blue-green algae especially
develop a coating of attached bacteria. In this connection it is noteworthy that bottom deposits
exhibit a marked ability to adsorb bacteria, and that the bacteria of bottom deposits are especially
susceptible to adsorption by bottom muds (Rubentschik, Roisin and Bieljansky, 1936).
Henrici & McCoy — Heterotfophic bacterid 325
Under at least serili-, if riot total, anaerobic conditions, and
usually at low temperatures, this organic matter undergoes a
slow bacterial decomposition. While undoubtedly a very com¬
plex process involving many species of bacteria arid a great
chain of chemical reactions, it would appear thkt eventually this
organic matter is largely liberated into the water as hydrogen,
methane, and carbon dioxide (Allgeier, et. al. 1932). The meth¬
ane and hydrogen are then oxidized by the activities of autotro¬
phic bacteria. The formation arid oxidation of these gases ap¬
parently plays an important part in the oxygen regimen of
lakes, especially under winter conditions. (Kusnetzow, 1935).
The above summary may serve to emphasize the importance
of the bottom bacteria to limnology, arid also the inadequacy of
our present knowledge. In addition to their role in the decom¬
position of organic matter, the bacteria of lake deposits also may
take part in transformations of inorganic substances as sulphur
(Galliher, 1933), iron (Naumaftn, 1930), calcium (Williams and
McCoy, 1934), and silica (Brussoff, 1933). A knowledge of the
bacteriology of bottom deposits is important, therefore, not only
to limnology, but also to sedimentary petrology. It might be
mentioned further that such knowledge may eventually prove to
be of importance in explaining the origin of petroleum deposits,
and in the agricultural utilization of drained lands.
Direct Microscopic Studies
Grote (1934) has calculated, from the oxygen-consuming
power of bottom deposits, that in the top centimeter of a rich
mud there must be 8 billion bacteria per ccm., or M25 of the total
Volume of the mud. Karsinkin and Kusnetzow (1931), applying
a modification of Winogradsky's method for the direct micro¬
scopic enumeration of soil bacteria, found 1,368,800,000 per ccm.
at the 0-3 mm. level ; 310,800,000 at 150-153 mm., arid 197,600
at 300-303 mm. in the “braunschlamm” of Lake Glubokoje. The
numbers at the surface level were found to increase progres¬
sively from May to September. Later, Ifusrietzow, (1935) ap¬
plied the method of Germanow (1932) which attempts to sep¬
arate the bacteria from the mud by elution with salt solution.
Counts as high as 5,923,000,000 per ccm. were found. There was
no appreciable decrease with depth to 16 cm. Larger numbers
were found in bottom deposits of “gas-emittirig,> lakes than are
326 Wisconsin Academy of Sciences , Arts, and Letters
present in Lake Glubokoje, which does not give off appreciable
amounts of methane.
We have been unable to convince ourselves that the direct
microscopic methods are useful for quantitative work. We en¬
countered all of the difficulties met in attempts to enumerate soil
bacteria by such methods, and in addition further trouble due to
the much larger amount of stainable organic debris and a much
greater tendency for the bacteria to occur in large clumps so
closely packed together that they can not be counted. We have
obtained the clearest microscopic picture by suspending some of
the mud in Amann’s fluid (lactic acid, glycerine, phenol, cotton
blue) examined wet under a coverslip. This fluid tends to make
the debris and the colonies more transparent, and stains the bac¬
teria a deep blue without coloring much of the debris. All that
we have learned from such observations is that the bacteria are
very abundant at the surface, and decrease rapidly with depth;
below 15 cm. they are hard to find. One count from station 3,
Lake Alexander, taken April 23, 1933, gave 44,175,000 per cc.
at the surface and 1,905,000 at 40 cm. depth. These values may
be compared with plate counts given in Table V. When making
these counts one continually has difficulty in deciding whether a
given body is a bacterium or a bit of debris, and in seeing all
of the bacteria in dense clumps, which produces a strong feeling
of lack of confidence in the observations. Most of the bacteria
appear to be spherical or rod forms in clusters ; filamentous types
are seldom seen. Photomicrographs of several representative
fields are shown in Figure 1.
A more satisfactory method may be found in the Cholodny
(1930) technique, which consists in placing the slides directly in
the mud and observing the bacteria which grow upon the glass.
Karsinkin (1934) has made several observations by this method,
and found that more bacteria grow upon the slides immersed in
the upper layers of mud than in the water over them. We have
made a number of attempts to study the bacteria of bottom de¬
posits by this method, but with only moderate success. It is
difficult to place the slides at a known depth save in shallow
water. If attached to a taut line, they are continually raised
and lowered by wave action ; if on a slack line, they sink to an
unknown depth. In several cases, a number of 50 x 75 mm.
slides have been fastened end to end, and. the entire train fas-
Fig. 1. Photomicrographs of representative fields of bottom deposits
suspended in Amann’s fluid with cotton blue, showing bacteria of several
types
Henrici & McCoy — Heterotrophic Bacteria > 327
tened to a framework of copper-plated steel rod, which could be
pushed into the mud in shallow water to such a depth that the
upper slide was still visible in the water. The position of the
mud-water line could be easily observed, and the depths of the
fields measured on the slides. The slides were stained with crys¬
tal violet and counted according to methods described for the
enumeration of water bacteria (Henrici, 1933, 1936). In the
following table the results of two such observations are given.
Table I
Distribution of bottom bacteria by modified Cholodny method ,
Lake Alexander.
It will be seen that in both instances the bacteria which grew
upon the slides in the mud are much fewer than those which
attached themselves to the glass in the water immediately above.
This is more striking when considered in relation with plate
counts. The cultivable bacteria in the water are counted in hun¬
dreds per ccm., those in the mud in millions.. It is noteworthy
that there is no significant change with depth.
The above examples were chosen because they gave the high¬
est counts observed. In most of our attempts to use this method
the bacteria have been too few to count. It is therefore apparent
that this method, satisfactory as it has proven in the study of
water bacteria, is of little value for the bottom deposits. Either
there are very few bacteria in the mud capable of attachment to
the glass, or (and this seems more likely) the competition with
the vast surface supplied by the solid matter of the bottom de¬
posit allows only a small number to grow upon the slides. The
morphological types found upon these slides have been almost
exclusively minute rod forms irregular in arrangement.
328 Wisconsin Academy of Sciences , Arts , and Letters
Collection of Samples
Samples of bottom deposits may be collected by dredges,
borers, and tube samplers. Profile samples can be obtained only
with borers and tube samplers, since dredges necessarily mix
up the sample. Borers can be used only in very shallow waters.
Practically, therefore, we are limited to samples obtained with
tube samplers. These are very unsatisfactory, but better than
none.
Various types of tube samplers, their uses and limitations,
have been described by Naumann (1930), and by Lundqvist
(1925). All work on the same principle. A tube of glass or
metal is arranged with a check valve at one end. It is dropped,
driven, or pushed into the mud, cutting a cylindrical core or
“wurst' \ On pulling up the sampler, the valve closes, preventing
water from washing out the sample from above, while the vis¬
cosity and density of the mud packing the lower opening pre¬
vents it from dropping out. The samplers we have used have
been based upon this principle, and have been changed during
the course of the work, mainly by increasing the size. They
have all been made from materials obtainable at a plumber's
supply store.
In Figure 2, there is presented a drawing of the sampler now
used. This consists of three parts. The portion designated I is
the sampler tube, A, and the check valve B. The tube A is a
piece of inch brass pipe, threaded at one end, and ground to a
sharp edge at the other, approximately 75 cm. in length. A series
of threaded holes, about 9 mm. in diameter, are drilled along one
side, 2 cm. between centers. In earlier models they were spaced
3 cm. apart. These are fitted with slotted, threaded plugs which
may be removed with a screw driver, and allow the insertion of
pipettes for the removal of samples at various levels. The check
valve is a standard plumber's check valve, of cast bronze, also
1 1/2 inch size.
Parts II and III may be used alternatively for work in shal¬
low water or deep water respectively. Part II consists of an¬
other length of IV2 inch pipe, C, 75 cm. long, threaded at both
ends. Over this is slipped a cylinder (D) of cast lead, about
20 cm. long and with walls about 3 cm. thick. This is provided
with two eyelets to which a rope may be fastened. The internal
Henriei & McCoy— Bet erotrophic Bacteria
829
ism
Fig. 2. Apparatus for collecting core samples of bottom deposits. See
text for explanation.
diameter of this lead cylinder is a few millimeters greater than
the outside diameter of the brass pipe, so that it will slide freely
up and down. A 2-inch reduction coupling, E, is placed in re¬
verse on the upper end of the brass pipe C to serve as a stop for
the lead cylinder.
In use, the lower end of pipe C is threaded into the upper end
of the valve B, and the whole is lowered gently until the tube A
enters the mud in a vertical position. Now by alternately lifting
330 Wisconsin Academy of Sciences , Arts , and Letters
and dropping the lead cylinder D, the apparatus is driven into
the mud.
For use in deep waters, part II is replaced by part III, com¬
posed of another piece of 1^2 inch brass pipe, 50 cm. in length,
threaded at one end. To this three vanes of sheet brass are
soldered as shown, and a bail is fastened at the top for attaching
the rope. The lower end of this apparatus is threaded into the
upper end of valve B, and after it is certain that the rope is
securely fastened, the entire equipment is dropped overboard,
the vanes serving to maintain a vertical position. The velocity of
the apparatus serves to drive it a sufficient distance into the mud
in depths of 6 meters or over.
It will be seen that this apparatus differs from those pre¬
viously described only in the utilization of standard plumbing
supplies. The use of holes closed with screw plugs for remov¬
ing the samples is an adaptation of an idea of Klein and Steiner
(1929), who closed the holes with plugs of melted paraffin. In
earlier models we simply applied a strip of one or two inch
adhesive tape, which could be stripped off to remove the samples,
but found that this tended to leak about the edges if the tubes
were transported some distance to the laboratory. We have not
used the screw plugs sufficiently to be sure that they are a real
improvement over the adhesive tape.
In use the valve B is fastened to one of the two alternative
heads, and sample tubes are screwed into the lower end of the
valve as needed. A number of tubes (A) may be carried on a
single collecting trip. These are plugged with cotton at both
ends and sterilized in the autoclave if adhesive plaster is used,
in the hot air oven if brass plugs are used. The cotton plugs are
removed just before dropping the apparatus, and after a sample
is obtained the lower opening is immediately closed with a cork
stopper. The tube is then unscrewed and the upper end is simi¬
larly stoppered. The tubes must, of course, be maintained in a
vertical position at all times to avoid mixing of the contents.
After the tubes have been brought to the laboratory, the
brass plugs are unscrewed from the top down (or the adhesive
tape is slowly stripped off) until the water which runs out shows
the first trace of turbidity. A pipette is made ready, and a sam¬
ple removed from the first opening below this level, which is
arbitrarily designated the top of the mud (or 0 cm.). Samples
Henrici & McCoy — Heterotrophic Bacteria 331
are then removed at regular intervals from this level to the
bottom of the tube. Mud cores have been obtained as long as
50 cm., but usually they are less than 40. The pipettes are made
from 7 mm. glass tubing, cut square at the ends, about 20 cm. in
length. They are calibrated at the one cubic centimeter level, and
the lower end is not glazed in the flame but left sharp and of full
diameter; the upper end is plugged with cotton. The samples
are removed by aspiration. This may require rather strong
suction at the lower levels where the mud becomes quite dense,
and some trouble has been experienced in these lower levels from
the expansion of dissolved gases under suction, which tends to
break the column of mud. The 1 cc. samples in the pipettes are
then blown into 100 cc. dilution blanks, or if of stiff consistency,
a wire may be used to push the cotton plug into the tube, which
then serves as a plunger to expel the sample.
Profile samples do not provide a true picture of the bottom
deposit at various levels. Because of the viscosity of the mud,
and its tendency to adhere to the walls of the tube, it becomes
compressed as the tube is forced into the bottom, and the length
of the core obtained is thus less than the depth of the bottom
deposit included in it. This has been clearly shown by Lundqvist
(1925) who compared the cores obtained by his tube sampler
with samples obtained by a borer from the same station. The
various mud strata could be readily identified by the nature of
the included pollen grains. It was found that the cores obtained
by the tube sampler were compressed from II2 to % of the real
depth. The degree of “compaction” of the mud by tube samplers
has also been studied in varying types of bottoms by Wrath
(1936). It is clear that this compression will be greater with
smaller tubes, and will also vary with the size of the valve open¬
ing. We believe, however, that the apparatus described is about
as large as can be safely used from a small rowboat. Anything
larger would be awkward to handle and especially difficult to pull
out of the mud without a winch.
The compression of the mud will also vary with the consist¬
ency of different samples, so that even if a single apparatus were
used in collecting all of the samples, the results would not be
directly comparable. A further source of error may be a ten¬
dency for mud to adhere to the walls of the tube and thus be
mixed with material at lower levels. This has also been noted
332 Wisconsin Academy of Sciences , Arts , and Letters
by Wrath (1936) who experimentally studied the samples ob¬
tained by forcing tubes into artificial bottoms made of clays and
sand of differing color. He found that when his sampler was
forced through red clay, a film of this material adhered to the
tube and contaminated the material below. Such material is
however much stickier than the usual deposits of inland lakes,
and we believe that this has not been a serious factor in our own
observations. Samples from Lake Mendota show a very sharp
division into an upper layer of dark gyttja and a lower layer of
light marl, and there is no indication that appreciable amounts
of the dark material are carried into the lighter material below.
Moreover, Lundqvist (1925) has found a sharp stratification as
regards pollen inclusions in profile samples obtained with tube
samplers.
The error caused by compression of the tube samples cannot
be calculated or allowed for, and has been a very discouraging
element in the study of bottom deposits. It would be highly sat¬
isfactory if someone could develop an apparatus which would
collect undisturbed samples from the bottoms of lakes. Never¬
theless the tube sampler is the best apparatus available, and it
would be a mistake to postpone all investigations of bottom de¬
posits for lack of a better. After all, the relations of the various
layers in the sampler core must bear some relationship to the
strata in the lake bottom. If general laws govern the vertical
distribution of bottom organisms, and if a sufficiently large num¬
ber of samples are collected, the variations introduced by the use
of different samplers or by varying consistency in bottom de¬
posits may be treated as statistical errors which will cancel out,
and the laws will be revealed. We believe that a step in this
direction is shown by our own data.
Cultural Studies
Although a number of papers have been published relating
to the cultivation of particular types of bacteria, especially those
concerned with sulphur transformations, from bottom deposits,
a surprisingly small amount of work has been done on the gen¬
eral distribution of the ordinary heterotrophic bacteria in the
muds of lakes.
Klein and Steiner (1929) collected some profile samples of
bottom deposits from the lower lake at Lunz, in Austria. These
Henrici & McCoy — Heterotrophic Bacteria 333
studies were made incidental to more comprehensive studies of
the bacteria in the water. Nutrient gelatine was used for plate
counts ; lower counts were obtained when Nahrstoff-Heyden agar
was used. The occurrence of specific groups, as urea decompos¬
ing, nitrifying, denitrifying, sulphate-reducing, etc., was also
determined by the use of appropriate media. In all cases the
bottom deposits contained much larger numbers of bacteria than
the water above. Samples from the middle of the lake showed
higher counts than those from the shores. There was no defi¬
nite change in the numbers with depth in the mud. In two sam¬
ples the proportion of aerobes to anaerobes was determined in
gelatine cultures; aerobes were more numerous. Both aerobic
and anaerobic nitrogen-fixing organisms were present at all
depths. Urea-decomposing organisms were very abundant, but
only in the upper layer. Nitrifying and denitrifying organisms
were detected. Sulphate-reducing bacteria were obtained from
the mud, not from the water. Thiosulphate oxidizing bacteria
occurred both in the bottom and in the water.
Steiner (1931) reported the occurrence of both aerobic and
anerobic bacteria from bottom deposits capable of digesting cell¬
ulose and chitin.
Zih (1932) reported further work on the lower lake at Lunz.
He also found higher counts in the bottom mud than in the wa¬
ter, and noted qualitative differences between the bacteria culti¬
vated from the two habitats.
Williams and McCoy (1935) cultivated bacteria from the
bottom deposits of Lake Mendota. Sodium caseinate agar was
used for the general flora. Samples were taken only from the
tops and bottoms of cores about 2% feet long. Mean counts
from these levels were 5,200,000 per cc. and 90,000 per cc. re¬
spectively. Urea fermenting and iron precipitating bacteria
were also counted. By the extinction-dilution method various
physiologic groups were determined. Most of the known groups
of soil organisms were found, but nitrifiers and cellulose decom¬
posers were scarce or absent.
Duggeli (1936) reported an extensive series of observations
on the bacteria cultivated from the bottom mud from the “Rot-
see”, a lake near Luzern in Switzerland. This is apparently a
rather heavily polluted lake ; the bottom is a true “faulschlamm”,
giving off large amounts of gas, in which methane and nitrogen
334 Wisconsin Academy of Sciences, Arts, and Letters
are dominant, but hydrogen sulphide is prominent. The bottom
deposit contains considerable amounts of iron sulphide. Pro¬
file samples were not obtained. The dredge samples were diluted
and cultured on a variety of media suitable for various groups
of heterotrophic. bacteria. The numbers of bacteria cultivable
on Nahrstoff-Heyden agar varied from 51,000 to 1,440,000 per
wet gram. The numbers fluctuated markedly from place to
place and from time to time ; the reasons for these fluctuations
were not determined. Numerous species of common heterotro¬
phic types were identified.
Several papers have been published dealing with the occur¬
rence of bacteria in the bottoms of the seas. Waksman, Reuszer,
Carey, Hotchkiss and Renn (1933) noted larger numbers in bot¬
tom deposits than in the sea water, a decrease with depth of the
mud, and the occurrence of various physiologic groups. Waks¬
man, Hotchkiss and Carey (1933) studied the occurrence of
various bacteria concerned in the N-cycle and N-fixing organ¬
isms were abundant, while denitrification appeared unimportant.
Reuszer (1933) reported further on the distribution of hetero¬
trophic bacteria. A dilute glucose peptone medium was used for
plating. Bacteria in the bottom deposits were much more nu¬
merous than in the water, more numerous in mud than in sand.
There was noted a decrease with depth, greatest in the first 2.5
cm. “There appeared to be on the surface of the mud a very
thin layer containing organic matter in a less advanced stage of
decomposition supporting a bacterial growth much richer not
only in numbers but also in types of bacteria than the lower
layers.” The decrease with depth was not correlated with a de¬
crease in the organic carbon. The numbers in the surface layers
decreased with increasing distance from land. ZoBell and An¬
derson (1936) observed the distribution of heterotrophic bac¬
teria in marine deposits from the Pacific, using a dilute meat
extract peptone agar. Counts as high as 420,000,000 per wet
gram were encountered, counts in the millions being found in
bottom samples from water depths to 2,000 meters. These num¬
bers dropped sharply with increasing depth in the mud, the
greatest decrease occurring in the first few centimeters. An¬
aerobes were less abundant than aerobes, but decreased less with
depth, so that the ratio of anaerobes to aerobes increased with
depth.
Henrici & McCoy — Heterotrophic Bacteria
335
Cultural Methods
Since anaerobic conditions apparently prevail in the bottom
deposits of lakes, one would naturally think first of anaerobic
culture methods. These are, however, so cumbersome and un¬
suited for quantitative work that we have done very little with
them until recently. Aerobic and facultative forms are cer¬
tainly more numerous than strict anaerobes at all levels.
We have used a single plating medium for all of our work,
patterned after the medium of Fred and Waksman for counting
soil bacteria, but modified to provide a wider variety of organic
nutrients. Our medium contains 0.05% each of sodium casein¬
ate, peptone, starch, glycerol, and dibasic potassium phosphate,
1.5% of agar, in tap water. It was thought that such a mixture
might give higher counts by providing nutrients favorable to a
greater variety of bacteria, and comparative tests indicated that
this is true.
Flat-sided lotion or medicine bottles with screw caps were
used in place of petri plates. These make possible extensive
plating operations under field conditions where the use of Petri
dishes would be almost impossible. The bottles were of 120 cc.
or 180 cc. capacity, and contained 20 cc. to 30 cc. of medium.
After inoculation the bottles were placed on their sides so that
a flat thin layer of agar was formed. The caps were screwed
tight. This prevented evaporation during long incubation pe¬
riods, and also eventually led to conditions favorable for micro-
aerophilic, if not anaerobic, species. It was found that when
methylene blue was added to the agar it was completely decolor¬
ized in about a week if colonies developed to the number of
about 100 per bottle or more.
Cultures were incubated at “room” temperature, mostly in
unheated field laboratories where the temperatures varied wide¬
ly. The duration of incubation has varied somewhat according to
the temperature, from one to three weeks. Five replicate cul¬
tures were made from each dilution, and most of the counts
recorded are the means of these five, but occasionally bottles
were rejected because they were overgrown by molds or spread¬
ing bacteria, or because they differed markedly from their repli¬
cates.
The number of bacteria has been calculated in terms of cubic
centimeters rather than in grams, as is customary with soil sam-
336 Wisconsin Academy of Sciences , Arts, and Letters
pies. No systematic observations have been made on dry weights
of these samples, Ip three instances, however, 1 cc. samples
have been taken, dried to constant weight, and weighed. The
results are given in Table II. It will be seen that the water
content varied considerably*
Table II
Dry weight of 1 cc. samples from different depths.
Nomenclature of Bottom Deposits
We shall use the Swedish names of bottom deposits intro¬
duced by von Post, which have been precisely defined by Nau-
mann (1930), and which are now in international use. “Gyttja”
i$ the black mud,, rich in organic matter, which fills the basins
of most eutrophic and oligotrophic lakes. In profundal stations
it is of fine texture (“feindetritus gyttja”) derived mainly from
plankton. In shoreward stations with rooted plants it takes on
a coarser fibrous texture (“grobdetritus gyttja”) due to the re¬
mains of plants. It may be mixed with more or less calcium
carbonate (“kalkgyttja”) in lakes rich in lime. “Dy” is the
brown ooze so characteristic of dystrophic lakes, composed large¬
ly of organic matter which has been carried into the lake in
colloidal form from the surrounding bog lands and then precipi¬
tated.
Henrici & McCoy— Meter otrophic Bacteria
337
Fig. 3. A portion of Alexander Lake showing the location of stations,
and mean counts of bacteria from the surface water and from the surface
layer of mud.
338 Wisconsin Academy of Sciences , Arts, and Letters
Lake Alexander
Lake Alexander is a relatively shallow eutrophic lake situa-
ated in Morrison County near the center of Minnesota. It is
approximately 4’^ miles long and 2 miles across in the widest
portion, with a maximum depth of 15 meters and an estimated
mean depth of 9 meters. It is frozen usually from the middle of
November to the middle of April and does not stratify in the
summer.
No chemical studies have been made, but the water is quite
hard, and in some areas lime-loving aquatic plants, especially
Chara, are abundant. The shore line is irregular, presenting a
number of sheltered bays alternating with sand beaches and
reefs of boulders. There are three islands. The lake is very
productive; pike, wall-eyed pike and bass are relatively abun¬
dant, perch and sunfish very much so. The lake “blooms” quite
regularly in August. The bottom consists almost entirely of
gyttja; all samples tested have contained enough calcium car¬
bonate to effervesce visibly with the addition of acid. In some
littoral stations this becomes sufficient in amount to give the
deposit a light grey color and should perhaps be considered as
marl.
Bottom samples have been collected from 6 stations, two of
which are located in the open lake, three close to shore in shel¬
tered bays with abundant plant life, and one in a transitional
position. The location of these stations is indicated on the map
shown in Figure 3.
Station 1 is an open lake station, the depth being 7 meters.
The bottom is a soft feindetritus gyttja which continues with¬
out change in appearance or texture to as deep a level as samples
have been obtained. The following plate counts have been ob¬
tained from this station:
Table III
Lake Alexander , Station 1.
Henrici & McCoy — Heterotrophic Bacteria 339
Station 7 is also an open lake station, at a spot where a small
depression in the bottom reaches a depth of 13 meters. Two
samples from this station gave the following counts :
Table IV
Lake Alexander t Station 7.
Station 5 is located in a bay somewhat sheltered from the
main body of the lake. The water is 2 meters deep, the bottom
carpeted with Ceratophyllum, but no plants reach the surface.
The bottom is a black gyttja transitional in texture between that
of the open lake and that of the strictly littoral stations. Only
two samples have been collected. One, taken May 15, 1933, gave
1,861,000 bacteria per cc. from the surface level. Another taken
Sept. 2, 1933, gave 165,800 at the surface, and 1,175 at 32 cm.
Station 2 is located in the same bay, but at the very edge of
the water, where a dense mat of Ceratophyllum, Naias, Elodea
and various Potamogetons merges with a marsh of rushes, cat¬
tails and sedges. The bottom is a coarse-fibered gyttja with
much lime. Again but two samples were taken. On May 15,
1933, there were obtained 479,200 colonies per ccm. at the sur¬
face; on Sept. 2, 1933, counts from surface layer gave 78,000
per cc., while 2,480 were obtained from the 32 cm. level.
Station 3 is a littoral station in another bay, the water less
than 1 meter deep. Here the bottom is carpeted with Chara,
and Potamogetons are also abundant. This bay adjoins the same
marsh, but the vegetation is somewhat different, wild rice being
the dominant emergent plant at the edge. For some reason this
little bay is much richer in lime than other portions of the lake.
The bottom deposit is made up largely of the lime encrusted
remains of Chara. It is rather dark at the surface, but becomes
gradually paler with depth, so that at a level of 10 cm. or deeper,
it is almost pure marl. The following counts have been obtained :
340 Wisconsin Academy of Sciences , Arts , and Letters
Table V
Lake Alexander , Station 3.
Station 4 is located in another shallow bay very rich in vege¬
tation. Here the water in midsummer is so filled with pond
lilies, Potamogetons, Elodea and other aquatics that a boat is
pushed through it with great difficulty. Naias is the dominant
bottom plant. The bottom deposit is very black, coarse, fibrous
material. Two samples have been taken which gave the follow¬
ing results :
Table VI
Lake Alexander , Station 4.
The studies on Lake Alexander were the first which we have
undertaken. The sampler used was a small one, with 1 inch iron
pipe with perforated brass tube inserts having an inside diam¬
eter of about % inch. It did not penetrate readily. Partly for
this reason and partly for a lack of sufficient plating equipment,
the sampling is inadequate at the lower levels. The data from
the surface layer are, however, sufficient to suggest something
regarding the horizontal distribution of bacteria in the bottom
muds. The averages of the counts from the 0 cm. level are shown,
for the various stations, on the map (Figure 3), together with
averages for a number of plate counts from the surface water
obtained at the same stations and plated on the same medium.
It will be seen that the littoral stations, rich in plant life, show
much higher counts than those from the open lake stations and
Henrici & McCoy — Heterotrophic Bacteria
341
that the stations rank in the same order with regard to numbers
of water bacteria and number of bacteria in the bottom.
A number of colonies from the Lake Alexander plates were
isolated for pure culture observation. Colonies were transferred
to meat extract peptone agar, and about one-fifth of the colonies
transferred failed to grow on this medium. Ten to twelve col¬
onies were transferred from each sample, and an attempt was
made to choose these roughly in proportion to the abundance of
the various types of colonies on the plates. The analysis of
these cultures may be looked upon as indicating, though only
roughly, the distribution of the various types of heterotrophic
bacteria in the bottom deposits. The isolated cultures which
grew were transferred to the usual routine culture media, and
examined for morphologic and cultural characters.
The results of some of these pure culture studies are pre¬
sented in Table VII, where they have been analyzed in two ways.
Cultures from the open lake stations 1 and 7 are compared with
those from the littoral stations 3 and 4 ; and cultures from the
upper levels, 0 to 8 cm., are compared with cultures from the
lower levels, 20 to 32 cm. It is possible that the greater inci-
Table VII
342 Wisconsin Academy of Sciences, Arts, and Letters
dence of Coccaceae in the lower levels, and the greater incidence
of Bacillaceae in the profundal stations are significant, but aside
from these we can see no noteworthy differences in the distribu¬
tion of the various types.
Unfortunately, we have no similar data for the distribution
of different types isolated from the water at these same stations.
Plate counts of water samples were, however, made, and an ex¬
amination of the colonies on these plates did not indicate any
great difference between the bacterial floras of the two habitats,
save perhaps for a somewhat higher incidence of chromogens in
the water.
Lake Mendota
The characteristics of Lake Mendota have been thoroughly
described in numerous publications from the Wisconsin Geo¬
logical and Natural History Survey. The physical and chemical
Fig. 4. Map of Lake Mendota showing location of stations. Mean
counts of bottom bacteria in the upper 18 cm. were as follows : — Sta. A,
753,500 bacteria per cc.; Sta. 1, 249,580 per cc.; Sta. 2, 38,180 per cc.;
Sta. 3, 40,700 per cc.; Sta. 4, 35,650 per cc.; Sta. 5, 836,900 per cc.
Henrici & McCoy— Heterotrophic Bacteria
343
characteristics of its bottom deposits have been described by
Twenhofel (1933). Lake Mendota differs from the other lakes
we have studied especially in the high content of calcium. A
considerable amount of this is being continuously precipitated,
so that the bottom sediment consists very largely of a fine-tex-
tured marl. Over-lying this marl is a layer of soft black ooze of
varying thickness. The demarcation between the black ooze and
the lighter colored marl is fairly sharp in most samples which
we have collected, and this is the only evidence of distinct strati¬
fication of bottom deposits which we have observed.
Samples have been collected from six stations, whose posi¬
tions are shown on the map (Fig. 4). Station A was chosen for
its convenience to the laboratory. It is an open lake station, 20
meters deep. Stations 1 to 4 are in a line extending from deep
water toward the shore at depths of 18 m., 9 m., 5 m., and 1 m.
respectively. Station 5 is located in deeper water, 20 m. The
bottom deposits from stations A, 1 and 5 did not differ greatly
from each other; the division between black ooze and fine marl
occurred at levels 12 to 20 cm. from the top. As one approaches
shoreward, this division comes closer to the surface and at sta¬
tion 4 the black ooze is absent, being replaced by sand. This
difference is probably to be attributed to wave action which
tends to keep the fine black material in suspension in the upper
layers of water. There were no aquatic plants growing at any
of the stations from which bottom samples were taken.
The results of the counts obtained from Lake Mendota are
given in the following tables.
344 Wisconsin Academy of Sciences , Arts , and Letters
Table VIII
Lake Mendoia , Station A.
*These samples were plated using a medium which contained 0.05% glucose
instead of 0.05% starch.
Table IX.
Lake Mendota, , Station 5.
Henrici & McCoy — Heterotrophic Bacteria
345
Table X
Lake Mendota , Station 1.
346 Wisconsin Academy of Sciences , Arts , and Letters
Table XIII
Lake Mendota , Station 4.
In comparing the various stations in Lake Mendota, and the
various Wisconsin lakes with each other, we have chosen to
average the counts for the top 18 centimeters rather than those
from the surface layer, as was done with Lake Alexander. This
seems to be a fairer comparison, especially because the 0 cm.
count is not always the highest. Our method of selecting this
mark is rather arbitrary and results for the top few centimeters
vary greatly with different samples because of the differing
density of the bottom deposit at the place where it joins the
water. We have chosen 18 cm. as the depth to which to average
the counts because the great majority of our cores have extended
to this depth, while many have not extended deeper.
These averages are presented with the map shown in Figure
4. It will be seen that the distribution of bottom bacteria is
quite different from that in Lake Alexander; the counts from
profundal stations are higher than those from the littoral sta¬
tions. This may be explained by the difference in the form of
the lake basins. Lake Mendota has rather high banks, a smooth
shore line, and no islands to break the waves. The shores are
mostly sandy or gravelly, with aquatic plants abundant only in
a few bays. The shore at Station 4 is a sand beach, and there
were no rooted plants at the time the samples were taken. As
one approaches the shore from station 5 towards station 4, the
bottom deposit shows a steadily decreasing layer of gyttja above
the marl, which gradually disappears completely, being replaced
by sand. This may be explained by wave action, as has been
mentioned, which will tend to wash out the light, flocculent or¬
ganic matter and keep it in suspension until the water reaches
a depth sufficient for this material to settle out of the zone of
wave action.
We believe that we may assume such a relationship will be
found in lakes generally. The numbers of bacteria will be very
Henrici & McCoy — Heterotrophic Bacteria
347
high in those littoral stations which are protected by the shore¬
line from wave action, where aquatic plants are abundant, and
where the bottom deposits consist of coarse gyttja; very low in
those littoral stations where wave action prevents the accumu¬
lation of organic matter, and the bottom is sandy.
No pure culture studies have been made from Lake Mendota,
but some attempt has been made to distinguish the different
types of colonies, and in the data collected during 1936, the pro¬
portion of pigment-forming bacteria has been determined while
counting the colonies. This proportion did not show any signifi¬
cant variation from station to station, but does show some
change with depth, as is shown in Table XIV, where the per¬
centages of chromogens from all the 1936 data have been ar¬
ranged.
Table XIV
Proportion of chromogens in bottom deposits of Lake Mendota.
It will be seen that below 18 cm. the proportion of pigmented
bacteria is distinctly less than in the upper levels. Separate
counts of the chromogens have not been made from other lakes,
but it is our impression that this decrease of the chromogens
with depth in the mud is generally true. Of the colorless col¬
onies, spore-formers were the most abundant in plates of Lake
Mendota mud. Spreading types, especially B. mycoides, appeared
to be more abundant than in other lakes which we have studied,
excepting the bog lakes.
348 Wisconsin Academy of Sciences , Arts, and Letters
Lakes op Northeastern Wisconsin
During the summer of 1935 samples of bottom mud were
collected from several lakes in northeastern Wisconsin. These
lakes have also been very completely described in publications
from the Wisconsin Geological and Natural History Survey. No
attempt was made to survey different parts of the lakes; all
samples were collected from the deepest portion of the lake for a
comparison of the different types of lakes, which are so richly
varied in this region. Samples were taken with a smaller sam¬
pler than the one we have described, the sample tubes being but
one inch in diameter, but otherwise identical.
The lakes studied may be grouped according to their general
characters. Trout Lake, Weber Lake and Crystal Lake are oli-
gotrophic. Muskellunge Lake and Little John Lake are eutrophic,
but are not nearly so productive as either Lake Mendota or Lake
Alexander. Helmet Lake and Mary Lake are dystrophic. Boul¬
der Lake and Brazelle Lake are peculiar and cannot be readily
classified.
Oligotrophic Lakes
Trout Lake is the largest and deepest of the lakes that we
have studied. Samples were taken at a depth of 32 meters in
South Trout Lake. The bottom deposit consists of a very fine,
loosely packed, black gyttja with a distinct greenish hue. It
does not pack well in the sampler tube, and a number of attempts
to obtain samples failed. Counts from Trout Lake are pre¬
sented in Table XV.
Table XV
Trout Lake
Crystal Lake and Weber Lake are very much alike in their
general characters, but Weber Lake is somewhat more produc-
Henrici & McCoy — Heterotrophic Bacteria 349
tive as a result of fertilizing experiments conducted in recent
years. Crystal Lake is quite the lowest in productivity that we
have studied. Its water is very clear and very low in mineral
content. The bottom deposits of these two lakes are much like
those of Trout Lake, but have presented even more difficulty in
sampling because of the failure of the deposit to pack firmly in
the sampler tubes. Counts from these lakes are presented in
Table XVI.
Table XVI
A striking feature in the cultures from the bottom deposits
of these oligotrophic lakes has been the occurrence of large num¬
bers of colonies of green algae. These have been noted occa¬
sionally in samples from nearly all of the lakes studied, but they
have been more numerous in the northeastern Wisconsin lakes
than in Alexander or Mendota, and much more numerous in the
oligotrophic lakes than the others, occurring at all depths,
though more abundant near the surface. In some cases there
have been more colonies of algae than of bacteria.
Eutrophic Lakes
Big Muskellunge Lake and Little John Lake differ from those
just mentioned, especially in the occurrence of considerable
aquatic vegetation near the shores. The bottom deposits appear
the same in both, a fine black gyttja without any greenish color.
A sample from Big Muskellunge Lake taken August 27, 1936,
gave 31,600 bacteria per cc. at the 0 cm. level.
Dystrophic Lakes
Mary Lake and Helmet Lake are both typical sphagnum bog
lakes with dark brown water, high in organic matter. Mary
350 Wisconsin Academy of Sciences , Arts , and Letters
Table XVII
Lake is deeper and narrower. The bottom deposits are alike,
the peculiar brown ooze, or dy, of mucoid consistency composed
largely of colloidal organic matter precipitated from the bog
water.
Table XVIII
Boulder Lake and Brazelle Lake
Boulder Lake is a drainage lake occurring in the course of
the Manitowish River. It is very shallow. The water entering
the lake is dark brown in color, coming from extensive bog
lands. The effluent water is quite clear. Most of the organic
matter, therefore, is either decomposed or deposited in the lake.
The bottom deposit is a typical brown dy. The lake, however,
cannot be considered dystrophic ; it is highly productive.
Brazelle Lake is also a drainage lake occurring in the course
of a small brook, quite small and only 2 meters deep. The water
is dark brown and very turbid. This water comes apparently
Henrici & McCoy — Heterotrophic Bacteria
351
from sphagnum bogs in its upper reaches, but the lake itself is
surrounded by marsh land with cattails and rushes, presenting
some of the characters of a “niedermoor”. This lake is remark¬
able for the very high number of bacteria in the water, as de¬
termined both by plate counts and by direct microscopic counts.
The bottom deposit is brown in color and contains a large
amount of fibrous vegetable matter in various states of decom¬
position. In Naumann’s terminology, it would perhaps be desig¬
nated as a “grobdetritus gyttja” with some “dy”. While these
two lakes cannot be classified as dystrophic types, they are close¬
ly related to these types as far as their bottom deposits are con¬
cerned.
Table XIX
A sample from Boulder Lake taken August 17, 1936, gave 31,300
bacteria per cc. at the surface level ; another taken July 24, 1936,
gave 466,000.
Cultures from Mary, Helmet, Boulder and Brazelle lakes
have been striking in the relative lack of chromogens and the
great abundance of aerobic spore-formers. In Helmet and Mary
Lakes these have occurred almost to the exclusion of all others.
Comparison of Different Lakes
For comparison of the bottom bacteria from different types
of lakes we have chosen only data from profundal stations, i.e.,
stations 1 and 7 from Lake Alexander, and stations A, 1 and 5
from Lake Mendota ; samples from the northeastern Wisconsin'
lakes were all taken from the deepest part. We have averaged
the counts from all levels to a depth of 18 cm., for reasons which
have already been indicated. The results are shown in Table
352 Wisconsin Academy of Sciences , Arts , and Letters
XX, in which the lakes are arranged in the order of magnitude
of their bottom counts. For comparison counts are also given
of the number of bacteria in the lake water. These are also
averages of several samplings from all depths in the lakes,
nearly contemporaneous with the bottom samplings.
It will be seen that a classification of the lakes according to
the number of bacteria in the bottom deposits tends also to class¬
ify them according to types. This is, as with so many such
classifications, more clear at the extremes than in the middle.
The highly productive hard water lakes, Mendota and Alexander,
are sharply distinguished from the very soft water oligotrophic
lakes, Weber and Crystal. Of the northeastern Wisconsin lakes,
the dystrophic types with their very high organic content stand
higher than the others, but the variations in the counts from
Boulder Lake to Big Muskellunge are probably not great enough
to be significant. It will be noted that the bacterial counts from
water samples from the different lakes do not correlate well with
the counts from bottom deposits. Such a correlation was appar¬
ent in comparing the different stations in Lake Alexander.
Another comparison is interesting. The bacterial counts
from bottom deposits are much higher than those from the wa¬
ter. But the bottom bacteria, and presumably their activities,
are largely concentrated in the surface layer of mud, while the
water bacteria are fairly uniformly distributed at all depths,
and presumably are to some degree active throughout. A com¬
parison of the two habitats cannot be made without some con¬
sideration of the relative volumes involved. If the average counts
from bottom deposits to a depth of 18 cm. are multiplied by this
depth, one gets a figure for the total number of bacteria in a
column of bottom deposit 1 cm. square and 18 cm. deep; this
depth will include nearly all of the bacteria in the bottom. Simi¬
larly, by multiplying the average number of bacteria per cc. in
the water by the depth of the lake in centimeters, one gets a
comparable figure for the number of bacteria in the water. The
averages of these values for all of the lakes, excluding Mendota
and Boulder where water counts were not available, are : bottom
deposit, 865,000; water, 317,100. Even considering depth, there
are on the average more bacteria in the bottom than in all of the
overlying water. But this proportion varies markedly with the
different lakes. The figures are presented in the last three col-
Table XX
Comparison of different types of lakes.
Henrici & McCoy— Heter otrophic Bacteria
353
O ® g.S
3 M.S
u w
I-S'g
pg
- a.g
PQ
cife:
o ^
| a
°1
Pt-e
►» ©
Eh ffl
OH • 05 iH OJ 00 l> CO 03
1 16 1> ;d«HHCoo
o o * o
§00000
OflOWOO
• 03 © © co ca cl os
* eo i> © co t- to ®T
* QO rH CO 00 O ^ iO
00 ?H CO 03 03 t- I iH
•OOON^OOO
* 00 rH 0^03 00 00
* 03 © 03* C© <g 03 00
©H^eooi^eo
* GO fc- CO t© iH
mo • co oo to to c
. t- oo • tp m © to©
• to CO * os to r
* CS?
^ gg *c8 *e$ *c$ *ci *«$
ag* a 22222
c« O >» >4 >>£>»£* >»
WW00 000000
x: J2
ft ft
8 8.g
33-g
WW o
WWQ
ft.s 'S.s.a
.g g,S ©,3,5
— .a £ g< h g< g<
fte*. ©>• ft p g 3 8 8
O © o 43 S £
to ^ bH p£
o\3s3oo
o ©
MM
•©■o
XX
&&
354 Wisconsin Academy of Sciences , Arts , and Letters
umns of Table XX, the final column showing the quotient of the
total bottom bacteria divided by the total water bacteria. Un¬
fortunately no counts from water samples are available for
Lakes Mendota and Boulder. Brazelle Lake is, as in many other
respects, an exception ; it has an abnormally high count of water
bacteria. But the remaining lakes again appear to be classified
by the ratio of bottom bacteria to water bacteria. Those lakes
with a high content of organic matter (Alexander, Helmet and
Mary) have very high ratios ; the very oligotrophic lakes
(Weber and Crystal) have low ratios; while the intermediate
lakes have intermediate ratios. We are unable to explain this
relationship but believe that it will eventually prove significant.
It may possibly be determined by the relative distribution of
organic matter between the water and the bottom deposits. It
is noteworthy that the bog lakes, which have the highest ratios,
contain not only the highest amounts of organic matter, but prob¬
ably also less readily decomposable organic matter. Present in
the water in colloidal form, it tends to settle more rapidly than
it can decompose, probably carrying bacteria from the water with
it. A somewhat similar condition, less in degree, may be postu¬
lated for the highly eutrophic lakes. The production of large
quantities of plankton during a short space of time leads to the
development of great quantities of particulate organic matter
which again settles more rapidly than it can decompose. With
the less productive oligotrophic lakes, the quantity of particulate
organic matter is much smaller in proportion to the volume of
water and possibly therefore decomposes more completely before
settling. However, if this line of reasoning is correct, the propor¬
tion of organic matter in the bottom deposits should vary with
the organic matter and productivity of the lakes, which from the
data of Black, and of Steiner and Meloche, is apparently not the
case. Possibly when further data are available regarding the
kinds of organic matter present in the water and in the bottom
this will become clearer. There may possibly be also involved the
peculiar and interesting surface-volume relationship which ZoBell
and Anderson (1936) have recently found to play so important
a part in the growth of marine bacteria.
Henrici & McCoy — - Heterotrophic Bacteria
355
Vertical Distribution of Bottom Bacteria
We approach a discussion of the vertical distribution with
some caution, fully realizing that two unknown factors, contami¬
nation and compression of the bottom strata, enter into our fig¬
ures. It is quite obvious both from our own data and from the
work of others that bottom bacteria are most abundant in the
very surface layer of the mud and decrease with depth. Con¬
tamination, i.e., the carrying of surface material to lower depths
in the core by adhesion to the walls of the tube, will tend to
reduce the amount of decrease. Conversely, compression of the
column will tend to magnify the decrease with depth. These
two factors, antagonistic to each other, both depend upon the
adhesive qualities of the bottom deposit and perhaps will vary
to a similar degree in different deposits. Whether the tube
sampler can yield significant data regarding the vertical distri¬
bution of bottom bacteria can only be determined by trial.
For purposes of comparison of the vertical distribution,
counts have been converted to a percentile basis, the highest
count from each sample being considered 100, counts from other
levels being converted to a percentage of this value. Individual
curves for 39 samples, all in which cultures have been made from
3 or more levels, are presented in Figure 5. These will show the
general tendency as well as the degree of variation in the indi¬
vidual samples.
These curves show considerable irregularity, undoubtedly
due to a variety of factors, among which may be mentioned the
difficulty in determining precisely the top of the column, which
has led to some variations at the beginning of the curve ; and a
tendency for the bottom bacteria to occur in colonies in the bot¬
tom mud, which may in some samples give rise to abnormally
high counts. Such aberrant counts may be recognized by the
occurrence of a very large number of colonies all of the same
species. Undoubtedly a number of other factors, inherent both
in the mud and the method, have also influenced the counts.
In some of the first samples plated there was a clear break in
these curves at the 8-12 cm. levels and it was thought this might
be of some significance, since Naumann has so strongly empha¬
sized the division of the bottom deposit into an upper or “actual”
layer, and a lower, or “historical” layer. The division between
these layers, he claimed, was determined by the depth to which
356 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 5. Curves showing, on a percentile basis, the vertical distribution
of the bacteria in 39 core samples.
burrowing benthic animals, especially Tubificidae, penetrate.
That the bottom deposit changes with depth from recently pre¬
cipitated material to a more mineralized sediment is of course
obvious, but whether such a sharp division occurs is not so clear.
Naumann set the depth to which worms penetrate at about 10
Henrici & McCoy — Heterotrophic Bacteria 357
cm. We were therefore looking for some change in the distribu¬
tion of the bacteria at or near .this level. But as more samples
were collected, it became obvious that a sharp break in the curve
was exceptional, rather than the rule. An inspection of the va¬
rious curves shown in Figure 5 shows that the only feature com¬
mon to all is a decrease with depth, more extensive in the upper
layers. "jH|fl
A sufficient number of samples have been collected so that
the results may be averaged, and the fluctuations due to errors
of technique or to local chance variations may be eliminated by
the statistical principle of cancelling out. This has been done in
the graphs shown in Figure 6. Again the number of bacteria
have been converted to a percentage of the highest count for
each sample. Each dot represents a count from one level of a
sample. The encircled dots indicate the average percentage
value for the bacteria at the depth indicated. The curve to the
left has been plotted arithmetically, that to the right on semi-
logarithmic paper. No mathematical curve-fitting is necessary
to show that a logarithmic curve fits the data. The curves have
been fitted by simple inspection.
Fig. 6. Curves showing, on a percentile basis, the mean vertical dis¬
tribution of the bacteria in bottom deposits of several lakes. See text for
explanation.
358 Wisconsin Academy of Sciences, Arts, and Letters
It is clear that when chance variations are eliminated, the
vertical distribution of the bacteria in the bottom deposits is
very definite. The numbers decrease with depth at a perfectly
constant rate. Errors due to contamination or compression by
the sampling apparatus may have changed the slope of the curve
but could not alter the fundamental relationship.
Again, the significance of this relationship is quite obscure
to us. The curve is at once reminiscent of the logarithmic sur¬
vivorship curves of disinfectant experiments, which have been
discussed by Rahn (1932). But the significance of the logarith¬
mic survivorship curve has itself not been satisfactorily ex¬
plained. If this is a phenomenon if the same kind, then the
curve may be interpreted as indicating that the bottom bacteria
are active only at the very mud-water interface and are dying
below. If this is a survivorship curve, then the depths in the
mud may be converted to time. But what a tremendously long
time! ZoBell and Anderson (1936) have pointed out that con¬
ditions in the bottom deposits of the sea, namely low temperature
and absence of oxygen, are just those conditions which experi¬
mentally have proven most conducive to great longevity of bac¬
teria in a dormant state. The data of Reuszer (1933) and of
ZoBell and Anderson (1936) indicate that a similar relationship
between numbers of bacteria and depth in the bottom will be
found in marine deposits.
We are, however, hardly in a position to venture an opinion
as to the significance of these curves. We are satisfied to record
the facts, reserving interpretation for the future when more
data will be available. It is highly important that information
be obtained regarding the vertical variations in chemical com¬
position and physical properties, especially the concentration and
nature of the organic constituents. pH, oxidation-reduction po¬
tentials, etc., before any conclusions are drawn with regard to
the activities of the bacteria at varying depths. Fortunately,
current literature, especially the publications of Kusnetzow and
his associates (Kusnetzow, 1935; Kusnetzow and Kusnetzowa,
1935; Borutzky, 1935; Rossolimo and Kusnetzowa, 1935) indi¬
cate that such information is being collected.
Henrici & McCoy — Heterotrophic Bacteria
359
Summary
Profile samples of bottom deposits have been collected from
Lake Alexander in central Minnesota, Lake Mendota in southern
Wisconsin, and Lakes Mary, Helmet, Boulder, Trout, Brazelle,
Little John, Big Muskellunge, Crystal, and Weber in northeast¬
ern Wisconsin. These lakes present a wide range in character¬
istics. Bacterial plate counts have been made from different
levels in the samples. Limitations of the sampling apparatus are
recognized and discussed.
It has been found that counts from littoral stations are much
higher than those from profundal ones if the shoreward zone is
occupied by aquatic plants, lower in the case of a sandy beach.
Counts from profundal stations show a fair correspondence with
the productivity of the lakes, being highest in the hard water
eutrophic lakes (Mendota and Alexander) and lowest in the very
oligotrophic lakes (Weber and Crystal) ; dystrophic lakes gave
high counts.
A correlation was observed between the concentration of bac¬
teria in the bottom deposit and in the water when different sta¬
tions of a single lake (Alexander) were compared. No such
correlation was observed when different lakes were compared.
The calculated total number of bacteria in the bottom was much
greater than the total number in the water in the case of the
highly eutrophic and dystrophic lakes; lower in the case of the
oligotrophic lakes.
The plate counts show a marked decrease with depth in the
mud. When the counts are plotted against depth, there is a gen¬
eral tendency for the curves to drop markedly at first, more
slowly beyond. Statisical treatment of the entire series of ob¬
servations indicates that a logarithmic curve will best fit the
data. This is similar to a survivorship curve from a disinfec¬
tion experiment. This may indicate that bacterial activity at
the bottoms of lakes is carried on almost exclusively at the mud-
water level, the bacteria dying below\
Literature Cited
1. Allgeier, R. J., Peterson, W. H., Juday, C., and Birge, E. A. The
anaerobic fermentation of lake deposits. Int. Rev. ges. Hydrobiol.
26: 444. (1932).
360 Wisconsin Academy of Sciences, Arts, and Letters
2. Black, C. S. Chemical analyses of lake deposits. Trans. Wis. Acad.
Sci. 24:127. (1929).
3. Borutzky, E. W. Die vertikale Verteilung des Benthos in den Seeabl-
agerungen, etc. Arb. der Limn. Sta. zu Kossino, 20:148. (1935).
4. Brussoff, A. Ueber ein Kieselbakterium. Arch. f. Mikrobiol. 4 :1 (1933).
5. Cholodny, N. Ueber eine neue Methode zur Untersuchung der Boden-
mikroflora. Arch. f. Mikrobiol., 1:620. (1930).
6. Diiggeli, M. Die Bakterienflora im Schlamm des Rotsees. Rev. d’Hy-
drologie. 7:205-354. (1936).
7. Galliher, E. W. The sulfur cycle in sediments. Jour. Sed. Petrol.
5:51. (1933).
8. Germanow, F. Zur Methode der Bakterienberechnung im Boden. Proc.
and Papers, 2nd Int. Congress of Soil Sci. 5:239. (1932).
9. Grote, A. Der Sauerstoffhaushalt der Seen. Die Binnengewasser, Bd.
XIV. Stuttgart, 1934.
10. Henrici, A. T. Studies of freshwater bacteria. I. A direct microscopic
technique. Jour. Bact. 25, 277. (1933).
11. Henrici, A. T. Studies of freshwater bacteria. III. Quantitative as¬
pects of the direct microscopic method. Jour. Bact. 32: 265. (1936).
12. Karsinkin, G. S. and Kusnetzow, S. J. Neue Methoden in der Lim-
nologie. Arb. der Limn. Sta. zu Kossino. 13-14: 47. (1931).
13. Karsinkin, G. S. Zum Studium des Bakterialen Periphytons. Arb. der
Limn. Sta. zu Kossino. 17: 21. (1934).
14. Klein G. and Steiner, M. Bakteriologisch-chemische Untersuchungen
am Lunzer Untersee. I. Die bakteriellen Grundlagen des Stickstoff-und
Schwefelumsatzes im See. Oesterreich. Botan. Zeitschr. 78: 289.
(1929).
15. Kusnetzow, S. I. Microbiological researches in the study of the oxygen¬
ous regimen of lakes. Verh. der Internat. Verein. f. theoret. u. ange-
wandte Limnol. 7:562. (1935).
16. Kuznetzow, S. I. and Kusnetzowa, Z. I. Bacteriological and chemical
investigations on lake muds in connection with bottom emission of
gases. Arb. der. Limnol. Sta. zu Kossino. 19: 127 (1935).
17. Lunqvist, G. Methoden zur Untersuchung der Entwicklungsgeschichfee
der Seen. Aberhaldens Handbuch der biolog. Arbeitsmethoden, Abt.
IX, Teil 1:427. (1925).
18. Naumann, E. Einfiihrung in die Bodenkunde der Seen. Die Binnenge¬
wasser, Band IX. Stuttgart 1930.
Henrici & McCoy — Heterotrophic Bacteria
861
19. Rahn 0. The physiology of bacteria. Philadelphia, 1932.
20. Reuszer, H. W. Marine Bacteria and their role in the cycle of life in
the sea. III. The distribution of bacteria in the ocean waters and muds
about Cape Cod. Biol. Bull. £5:480 (193i3).
21. Rossolimo, L and Kusnetzowa, S. Die Boden-gasausscheidung als Faktor
des Sauerstoffhaushaltes der Seen. Arb. der Limn. Sta. zu Kossino.
17: 87. (1934).
22. Rubentschik, L., Roisin, M. B. and Bieljansky, F. M. Adsorption of
bacteria in salt lakes. Jour. Bact. 32:11. (1936).
23. Speranskaja, T. A. Angaben ueber die Untersuchung des organischen
Stoffs der Schlammablangerungen der Seen. Arb. der Limn. Sta. zu
Kossino. #0:67. (1935).
24. Steiner, M. Beitrage zur Kenntnis der Zellulose- und Chitinabbaues
durch Mikroorganismen in stehenden Binnengewassern. In “75 Jahre
Stella Matutina, Festchrift Band II”, p. 367. Feldkirch, 1931.
25. Steiner, J. F. and Meloche, V. W. A study of ligneous substances in
lacustrine materials. Trans. Wis. Acad. Sci. 29 : 389. (1935).
26. Twenhofel, W, H. The physical and chemical characteristics of the
sediments of Lake Mendota, a fresh water lake of Wisconsin. Jour.
Sed. Petrol. 3:68. (1933).
27. Waksman, S. A., Reuszer, H. W., Carey, C. L., Hotchkiss, M. and Renn,
C. E. Studies on the biology and chemistry of the Gulf of Maine, III.
Bacteriological investigations of the sea water and marine bottoms.
Biol. Bull. £4:183. (1933).
28. Waksman, S. A., Hotchkiss, M., and Carey, C. L. Marine bacteria and
their role in the cycle of life in the sea. II Bacteria concerned in the
cycle of nitrogen in the sea. Biol. Bull. £5:137. (1933).
29. Williams, F. T. and MlcCoy, E. On the role of microorganisms in the
precipitation of calcium carbonate in the deposits of fresh water lakes.
Jour. Sed. Petrol. 4:113. (1934).
30. Williams F. T. and McCoy E. The microflora of the mud deposits of
Lake Mendota. Jour. Sed. Petrol. 5:31. (1935).
81. Wrath W. F. Contamination and compaction in core sampling. Science.
84:563. (1936).
32. ZoBell, C. E. and Anderson, D. Q. Vertical distribution of bacteria in
marine sediments. Bull. Amer. Assoc. Petrol. Geol. #0:258. (1936).
33. ZoBell, C. E. and Anderson, D. Q. Observations on the multiplication
of bacteria in different volumes of stored sea water and the influence
of oxygen tension and solid surfaces. Biol. Bull. 71: 324. (1936).
34. Zih, A. Beitrage zur Bakteriologie der Lunzer Seen, mit einem Anhang
bei F. Ruttner. Int. Rev. d. Ges. Hydrobiol. #£:431. (1932).
THE SILICA AND DIATOM CONTENT OF
LAKE MENDOTA WATER
V. W. Meloche, G. Leader, L. Safranski and C. Juday
From the Department of Chemistry, University of Wisconsin and the
Limnological Laboratory of the Wisconsin Geological and Natural History
Survey. Notes and reports No. 82.
Introduction
In a study of the waters of various Wisconsin lakes it has
become important to know something about the seasonal change
in diatom content in relation to the utilization of silica. Birge
and Juday (1922) studied the net and nannoplankton of lakes
Mendota, Monona, Waubesa and Kegonsa and since that time
have extended that study to northern lakes of Wisconsin. The
work did not refer to vertical distribution but emphasized sea¬
sonal variations. In that part of the study which was devoted
to net plankton particular reference was made to variations in
organic matter, nitrogen, crude protein, ether extract, ash, com¬
position of the ash and finally to the organisms which were re¬
sponsible for the variations. These data together with those re¬
ported by Atkins and his co-workers (1923-24) will be discussed
in greater detail later in this paper.
The purpose of the present report is to describe the average
variation in silica content as compared to the variation in dia¬
tom content of the water of Lake Mendota. The vertical distri¬
bution will be described according to the results obtained at one
station.
Apparatus and Methods
The water samples were taken in Lake Mendota at a station
some 200 meters east of Picnic Point, where the water has a
maximum depth of 18.5 m. They were obtained by means of a
sampler which is illustrated in Figure 1. Earlier models of this
instrument have been described, but no description of the pres¬
ent one has yet been given. The instrument consists of a brass
tube 42 cm. long and 6.7 cm. in diameter. The ends are closed by
363
364 Wisconsin Academy of Sciences, Arts, and Letters
two large rubber stoppers. The large tube and the rubber stop¬
pers are supported by a small brass tube 1.3 cm. in diameter
and 58 cm. long which passes through the center of the large
one.
When the sampler is open as shown in Figure 1, the large
tube is supported from the under side of the upper stopper by
two chains; the jaws which hold the sampler open are attached
to the upper side of the upper stopper and they clamp over a
small brass ring at the upper end of the small brass tube. These
Fig. 1. Diagram showing the water sampler used in obtaining samples
at different depths for vertical distribution of diatoms. It is a modfiication
of the sampler previously described.
Meloche & Leader-Silica and Diatoms
365
jaws are activated by two bronze springs. The entire instru¬
ment is supported by a line which passes through the opening in
the small brass tube ; a knot near the end holds the sampler on
the line. The sampler is lowered to the depth from which a
sample is desired and a brass messenger is then sent down the
line which separates the jaws and allows the instrument to
close ; it has a capacity of 1250 ml.
For the diatom counts, 500 ml. of water were run through a
high speed centrifuge and the material was preserved in forma¬
lin until the counts were made. The various diatoms were enu¬
merated in a regular counting cell, using a Whipple micrometer ;
the magnification was 100 times and twenty counts were made
in order to obtain a mean for each sample. Reasonable checks
were obtained for several samples on which recounts were made.
Individual cells were counted in such forms as Cyclotella, but
the filaments of Melosira and the colonies of Fragilaria were
taken as units.
Silica. The samples for silica determinations were placed in
pyrex bottles and the analyses were made as promptly as pos¬
sible after they were taken. The procedure for the determina¬
tion of silica was the colorimetric molybdate method of Dienert
and Wandenbulcke (1923) . In the preparation of the picric acid
standard, the concentration of the picric acid was corrected to
conform with the work of Robinson and Kemmerer (1930), in
which it was shown that 25.4 mg. of dry picric acid were equi¬
valent to 50 mg. of silica. In accepting this value for standardi¬
zation of silica, a slight error is introduced in the examination of
water which has as low a silica content as one part per million
or less. Since the primary interest in the problem is the varia¬
tion of silica between samples, this slight error is not significant.
Results
Abundance of Diatoms
On each of the dates mentioned in Table I and Figure 2, a
single series of samples was taken. These samples were taken
at the station mentioned and represented different depths, sur¬
face, 3, 6, 9, 12, 15, and 17 meters. The samples for a single
day were examined and the results averaged. These are the
values which appear in Table I and Figure 2. The full black
366 Wisconsin Academy of Sciences , Arts, and Letters
Table I
Average number of diatoms and amounts of silica in the entire series
of seven samples taken on the different dates .
portions in the diagram represent the diatom counts per liter
and the dotted outline represents silica concentrations in parts
per million.
Although the diatom counts are shown to be somewhat irreg¬
ular, there is an obvious trend toward seasonal activity, the
peaks in January and April being somewhat less marked than
those observed in October and June. Among the factors which
might be expected to influence the growth and abundance of
plankton would be the earliness or lateness of the respective sea¬
sons and, coincidental with this, the physical and chemical fac¬
tors which show seasonal variation, i.e., temperature, density of
the water, free and bound carbon dioxide, oxygen, silica, etc.
Seasonal activity in the growth of Plankton was also observed
by Juday and Birge, and by Atkins et al. It is of particular in¬
terest to note some of the comparisons which can be made be¬
tween the results reported by Birge and Juday in Bulletin 64
and in results presented in the present report. While they dis¬
cussed the whole body of net plankton, this study was limited to
the diatoms. As has been previously described, Lake Mendota
is deep enough to become thermally stratified in summer ; this
has been discussed in detail in Bulletin 22. Most of the data on
net and nannoplankton in Bulletin 64 are reported in terms of
Meloche & Leader — Silica and Diatoms
367
AVERAGE DIATOM COUNT
ppm Si Of
P o J-J- ^
2 ? (n a
ll O'?
^ S S o'
“ o 2. §
^ ^ w
8 a 3 «
CT
P CD
a ^
^ £ Hb
! C+ 2
S' » 3
&< p
co P
1 p 2
H 2
<£ o o
c« rr"
(D 3
4 4 CL
S' § &
„ 3 £
fD CO
►4 S3
Fig. 2. Diagram showing the seasonal distribution of silica and diatoms.
The solid black bars represent the average number of cells and colonies
368 Wisconsin Academy of Sciences , Arts, and Letters
milligrams of dry organic matter per cubic meter of water, but
numerical results for diatoms are given in Figures 22, 26, 32
and 33. These diagrams show peaks for diatoms in April-May
and for September-October ; the autumnal crop of diatoms may
extend into November or even December, while fairly large in¬
creases may be found in July or August in the different years.
In the present study Figure 2 shows abundance peaks in
January, April, June and October. The maximum numbers were
found in October both in 1935 and 1936. The vernal maximum
was much smaller in 1936 than those reported by Birge and
Juday.
Vertical Distribution
The vertical distribution of the diatoms is shown in some
selected series in Table II and Figures 3-8. In the January
series, the diatoms were fairly uniformly distributed from sur¬
face to bottom; there was a slight maximum at the surface, a
minimum at 6 m. and another maximum at 17 m. The April
series represents conditions soon after the spring overturn of
the water, with the temperature substantially uniform from
surface to bottom; as a result of the circulation of the water,
the diatoms were quite uniform in their vertical distribution.
i:
Fig. 3. Vertical distribution of silica and diatoms on January 10, 1936.
Note the uniform distribution of both.
Meloche & Leader — Silica and Diatoms
369
Table II
Vertical distribution of silica and diatoms of several series of samples
taken at different seasons of the year.
370 Wisconsin Academy of Sciences , Arts , and Letters
Table II (Continued)
By June 4, the thermocline had begun to form; this series
shows a maximum number of diatoms at the surface, a minimum
near the thermocline and an increase thence to the bottom. The
thermocline was well established by June 16 and a large number
of diatoms was found in the epilimnion; the average for the
upper stratum was about a million diatoms per liter of water,
Meloche & Leader — Silica and Diatoms
371
while the average for the lower stratum was only about 50,000
per liter.
The September series showed a relatively small number of
diatoms at all depths. In October, however, a very large number
was found and they were fairly evenly divided between the upper
and lower water as a result of the autumnal overturning and
circulation of the water.
Silica and Diatoms
Table I and Figure 2 give the average silica concentrations
and diatom counts for a period of one year. The pH of the
water of Lake Mendota ranges from 8.8 in the upper water dur¬
ing the summer to 7.4 in the lower water during the summer
and winter periods of stratification. (Juday, Fred and Wilson
1924). The alkaline status of the water may have some influ¬
ence upon the solution of the silica, but a more important factor
in the distribution and availability of the silica is the spring and
autumn overturning and circulation of the water. In this proc¬
ess the silica which has accumulated in the lower water is
200,000
K
I
I
s
bf
5
180.000
160.000
140.000
120,000
100.000
80.000
60.000
40.000
20.000
rr
I
£
o
£
*0
€
O)
~ 1.6
1.4
12
i.O
.8
.6
.4
2
Fig. 4. Vertical distribution of the silica and diatoms on April 10,
1936 Also shows a uniform distribution.
372 Wisconsin Academy of Sciences, Arts, and Letters
brought into the zone of photosynthesis and thus made available
for the diatoms that are growing in this stratum.
Figure 2 shows a general decrease in the amount of silica
toward November and January, and a perceptible increase dur¬
ing the latter part of March and the early part of April. The
amount was relatively small during the summer, but when the
cooling and more complete circulation of the water took place
in September and early October, there was a marked increase.
Figure 2 shows that the marked increase in silica in September
was followed by a correspondingly large increase of diatoms in
October which was accompanied by a very considerable decrease
in the silica. The other increases in diatoms, however, are not
accompanied by such distinct changes in the amount of silica.
While there was a large change in the amount of silica between
April 10 and April 23, 1936, there was not an equally prominent
increase in the diatoms; the increase of the latter was only a
moderate one. On the other hand, the slight decrease in silica
between January 1 and January 10, 1936, was correlated with
only a slight increase in the diatoms. Undoubtedly several fac¬
tors are involved in the problem, and much more data will be
required to work out a more complete and adequate correlation.
For the greater part of the year, Lake Mendota water has a
silica concentration less than 0.4 ppm and the highest noted in
this study was 2.5 ppm. Slight variations in concentration
K
I
l
5
is
$
X
140.000
/ 20.000
100.000
80,000
60.000
40.000
20.000
!■!
n
JiL
- i.o
- .8
A X A
1.11
J5 rS
.6
- .4
- 2
Fig. 5. Vertical distribution of silica and diatoms on June 4, 1936.
Note the maximum amount of silica at the surf ace.
Meloche & Leader — Silica and Diatoms
373
within such a range are almost beyond the limits of the precision
of the colorimetric method.
Fig. 6. Vertical distribution of silica and diatoms on June 16, 1936.
Compare with Figure 5 and note marked increase in number of diatoms.
Vertical Distribution of Silica
Figure 3, January 10, represents a fairly even distribution
of silica from top to bottom, with a correspondingly even distri¬
bution of the diatoms. Fig. 4, April 10, represents a somewhat
higher concentration of silica (0.8 ppm), but no variation in
amount from top to bottom ; neither was there any striking vari¬
ation in the vertical distribution of the diatoms. This is to be
expected since it represents the spring period of the overturning
and circulation of the water.
Conditions presented in Figures 5 and 6 are very different,
since the lake was then thermally stratified. On June 4 a rela-
374
Wisconsin Academy of Sciences, Arts, and Letters
1
I
is
•3
<o
5
*
Fig. 7. Vertical distribution of silica and diatoms on September 18,
1936. Compare with Figure 8.
tively small number of diatoms was found at the various depths.
On June 16 the surface water contained 670,000 diatom cells
and colonies per liter, with more than 200,000 per liter at 3 m.
There was a very marked decline in the number at 6 m. and an
insignificant number below this depth. The silica concentration
was relatively high at 17 m., but this was well below the zone of
photosynthesis so that diatoms were unable to multiply at that
depth.
Figures 7 and 8 show the vertical distribution of diatoms
and silica on September 18 and October 9, respectively. A com¬
paratively small number of diatoms was found at all depths on
the former date, but a maximum of more than 280,000 per liter
was obtained at 9 m. on the latter date, with nearly as large a
number at 15 m. Most of the body of water was taking part in
the circulation on October 9, so that conditions were favorable
for the distribution of the diatoms to the lower depths. A much
smaller amount of silica was found at 15 and 17 m. on October 9
than on September 18, as well as in the 0-12 m. stratum. The
more extended circulation of the water on October 9 than on
September 18 also brought the various organic and mineral sub¬
stances that had accumulated in the lower strata in summer into
the upper water and made conditions favorable for the growth
of the diatoms.
Meloche & Leader — Silica and Diatoms
375
8
I
a
300.000
280.000
260.000
240.000
220.000
200.000
180.000
160.000
(40,000
120,000
100.000
80,000
60.000
40,000
20,000
%
'O
- Z
Fig. 8. Vertical distribution of silica and diatoms on October 9, 1936.
Compare with Figure 7.
Summary
1. Variations in the abundance of the diatoms were corre¬
lated with the seasons of the year.
2. Seasonal variations were also noted in the vertical distri¬
bution of the diatoms.
3. Variations in the silica concentration of the water ranged
from 0.4 to 2.5 ppm., a little more than a sixfold difference.
4. A definite decrease in silica concentration was correlated
with a marked increase in diatoms in October; definite correla¬
tions were not noted at other seasons of the year.
376 Wisconsin Academy of Sciences, Arts, and Letters
5. A uniform vertical distribution of both silica and diatoms
was found during the vernal and autumnal periods of circulation
of the water.
Literature
Atkins, W. R. G. 1929-30. Seasonal variations in the phosphate and sili¬
cate content of sea-water in relation to the phytoplankton crop. V.
1927-29. Jour. Miar. Biol. Assoc. 16: 821-852.
Birge, E. A. and C. Juday. 1911. The inland lakes of Wisconsin. I. The
dissolved gases and their biological significance. Bull. No. 22. Wis.
Geol. & Nat. Hist. Survey. 259 pp. Madison.
Birge, E A. and C. Juday. 1922. The inland lakes of Wisconsin. II. The
plankton. I. Its quantity and chemical composition. Bull. No. 64. Wis.
Geol. & Nat. Hist. Survey. 222 pp. Madison.
Juday, C., E. B. Fred and Frank C. Wilson. 1924. The hydrogen ion
concentration of certain Wisconsin lake waters. Trans. Amer. Micro.
Soc. 43: 177-190.
Robinson, Rex J. and George Kemmerer. 1930. Determination of silica in
mineral waters. Trans. Wis. Acad. Sci., Arts & Let. 25: 129-134.
PHOTOSYNTHESIS OF AQUATIC PLANTS AT
DIFFERENT DEPTHS IN TROUT LAKE,
WISCONSIN*
Winston M. Manning, C. Juday and Michael Wolf
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 67.
Introduction
This investigation is a continuation of the research on photo¬
synthesis which has been conducted by the Wisconsin Geological
and Natural History Survey at the Trout Lake Limnological
Laboratory during each summer since 1932. In the earlier work
(Schomer 1934; Schomer and Juday 1935; Curtis and Juday
1937) several species of green and blue-green algae were used in
studying the rate of photosynthesis at different depths in some
of the lakes of northeastern Wisconsin. Observations were made
in lakes with water ranging from very transparent (Crystal
Lake) to very highly colored (Helmet Lake). The rates of
photosynthesis were correlated with measurements of light in¬
tensities at the various depths in each lake.
The experimental procedure in all of this work has been
based on that used by Marshall and Orr (1928) in studying the
photosysthesis of marine diatoms at different levels in the ocean.
Emerson and Green (1934) and Barker (1935) have called at¬
tention to some of the limitations of the method.
The purpose of the present study is threefold: (1) To in¬
vestigate quantitatively the limitations of the experimental
method; (2) to determine for Chlorella, at least approximately,
the quantum efficiency of photosynthesis as a function of light
intensity; (3) to extend the observations of photosynthetic rate
to some aquatic plants other than algae.
General Procedure
All experiments described in this paper were carried out at
Trout Lake during the summer of 1936. In most respects the
* This investigation was supported by a grant from the Wisconsin
Alumni Research Foundation.
377
378 Wisconsin Academy of Sciences , Arts , and Letters
procedure was like that described in earlier reports (Schomer
and Juday 1935). The plant material to be used was suspended
in filtered Trout Lake water and transferred to round glass-
stoppered, calibrated bottles (average volume approximately
140 cc.). The concentration of the suspension, in cells per cc.,
was determined by counting under a microscope the number of
cells per unit volume. For each depth during a run, four of the
bottles were placed in a small wire basket, the two lower bottles
in each basket being painted black and covered with black cloth
bags to keep out light. The baskets were simultaneously sus¬
pended in the lake for a definite time (usually one to four hours)
and then immediately analyzed for dissolved oxygen, using a
modified form of the Winkler method. . In addition, from two to
six bottles of the suspended plant material had been analyzed for
oxygen at the beginning of each run. The change in oxygen
content of the black bottles gave a measure of the respiration
rate, while the difference in final oxygen content of the black and
clear bottles gave a measure of photosynthesis. This of course
involves the assumption that respiration is the same in light
and darkness.
The baskets were suspended in the lake from buoys which
were anchored in such a way that shading of the baskets could
not occur between 7 a.m. and 6 p.m. The depth of the water at
the buoys was 16-20 m. The greatest depth employed in this
work was 14 m.
Some of the experiments were made with square bottles in¬
stead of the round bottles. The square bottles did not have
ground glass stoppers and were closed by means of rubber dia-
phrams placed over the bottle mouths. The sides of these bot¬
tles, while not optically plane, were nearly enough so to permit
a fairly accurate measurement of the fraction of light absorbed
by suspensions of Chlorella. The volumes of the square bottles
were about the same as for the round ones, each side having an
area of about 9x4 cm. or 36 cm2.
Oxygen analyses on duplicate bottles usually agreed to better
than 0.2 mg. of oxygen per liter (or parts per million) except
when the amount of photosynthesis was sufficient to raise the
concentration of dissolved oxygen far above the usual saturation
value.
Manning , Juday & W olf— Photosynthesis 379
Plant material . A pure culture of Chlorella pyrenoidosa was
maintained. This culture was of the same strain as used for
the past several years in research on the quantum efficiency of
photosynthesis, carried out in the Botany Department at the
University of Wisconsin. Apparently it is a different strain
from that used by Schomer and Juday (1935) since the average
cell diameter in their cultures was about 8 microns, whereas in
our cultures it was about 4 microns. The amount of reaction
per algal cell was correspondingly less in our work.
A supply of the Chlorella was kept growing on agar slants
under nearly continual illumination. The light source was a 60-
watt Mazda lamp at a distance of about 30 cm. The nutrient
medium used for the agar slants had the following composition :
KN03 4.044 gm. ; MgS04 0.602 gm. ; KH2P04 1.702 gm. ; Bacto-
Peptone 23.0 gm. ; enough soil extract (garden loam + distilled
water) to make one liter. The pH was adjusted to 6.9 with
NaOH. From time to time the Chlorella was transferred from
the agar slants and grown in a liquid nutrient under natural
light as described in an earlier paper (Schomer and Juday
1935). Material was taken from these liquid cultures for the
photosynthesis experiments. After two or three weeks, the li¬
quid cultures usually became noticeably contaminated by a min¬
ute unicellular alga about one micron or less in diameter, prob¬
ably a species of Chroococcus. After reaching this stage the
cultures were discarded.
For the experiments on other species of plants, material was
collected from Trout and other nearby lakes and usually used
within 24 hours.
Composition of water. Trout Lake water, which was used as
the suspending medium in most cases, contained approximately
38 mg/1 of carbon dioxide during the summer of 1936. This
amount includes free, half-bound and bound carbon dioxide. The
carbonate-bicarbonate concentration was high enough to give a
moderate buffering action. For the maximum photosynthesis
observed in a single bottle, the dissolved oxygen content in¬
creased from 8.5 to 17.9 mg/1, a net increase of 9.4 mg/1. This
represents a utilization of 12.9 mg/1 of carbon dioxide in the
photosynthetic process, or about 34 per cent of the total carbon
dioxide present in the water. The reaction of the lake water in
this case would be increased from the normal of about pH 7.5 to
380 Wisconsin Academy of Sciences, Arts, and Letters
over pH 9.0 at the end of the run. In most of the experiments
the change in the reaction was much less than this.
The temperature of the surface water of Trout Lake varied
from 21° to 24° C. during the period of this investigation. Down
to a depth of about 8 m. (top of thermocline) , the temperature
differed very little from that at the surface, usually only one de¬
gree or less. In all series the rate of photosynthesis at 8 m. was
so far below the maximum for high light intensities as shown in
Figure 7 that the sharp decline in temperature in the thermo¬
cline probably had little effect on the observed rates.
Energy data. The amount of solar radiation reaching the
surface of the lake during the experiments was measured with a
self-recording solarimeter; the percentage of transmission in
the water and the spectral composition of the radiation at dif¬
ferent depths were determined with the pyrlimnometer. Details
concerning these instruments have been given by Birge and
Juday (1932).
Experiments with Chlorella
Before being used, the suspension of Chlorella was separated
from the liquid culture medium by centrifuging. The alga was
then suspended in filtered Trout Lake water for the experiments.
Using square and round bottles simultaneously, a comparison
of rates for different light intensities and for different algal
concentrations was made. The results of the two series of ex¬
periments are shown in Table I. The experiments covered a
tenfold range in light intensity and a sixfold range in algal con¬
centration. On August 14, the sky was heavily clouded; the
average intensity was less than one-tenth full sunlight. On
August 20 the sky was covered with light clouds during most of
the run. The rate of respiration per million cells appeared to
be considerably lower in the August 14 run, though the relative
uncertainty in the value was large because of the small oxygen
change. Checks between duplicate bottles were less consistent
in the case of the square bottles, probably because of the greater
difficulty in manipulating the rubber diaphram covers.
The figure 0.94 is the mean value of the ratio of rates for
square and round bottles. In obtaining this value, slightly
greater weight is given to the values obtained for the high algal
Manning, Juday & Wolf — Photosynthesis
381
Table I
Comparison of photo synthetic rates for square and round bottles. For
photosynthesis, the figures for A02 represent the difference between final 02
concentration in light and dark bottles , but for respiration A02 represents
the difference between final and initial 02 concentration. Infra-red is not
included in the figures for light intensity at the surface. In the run on
August 20, square bottles were not placed at 1, 3, 5 and 6 meters.
August 14, 1936.
Time: 4.03 hours (11:12 a.m. — 3:15 p.m.)
Cell concentration: 1,900,000 per cc.
Initial 02 concentration: 8.12±0.10 mg /l.
August 20, 1936.
Time: 4-00 hours (10:26 a.m. — 2:26 p.m.)
Cell concentration: 331,000 per cc.
Initial 02 concentration : 8. 96 ±0.10 mg /l.
382 Wisconsin Academy of Sciences, Arts, and Letters
concentration, since the total reaction per bottle was much
greater in this case.
In calculating values for quantum efficiencies, the total re¬
action in round bottles was always multiplied by the factor 0.94
to obtain the reaction corresponding to the square bottles, which
were necessarily used in measurements of light absorption.
Measurements were made of light absorption for several dif¬
ferent concentrations of Chlorella. The light source for these
measurements was an ordinary Mazda lamp, with a CuS04 solu¬
tion between the lamp and the algal suspension in order to re¬
move infra-red radiation. Measurements were made of both
total absorption and absorption for several different regions in
the visible spectrum. In the latter case a series of Jena glass
filters were used. These filters were previously used by Birge and
Juday (1931) in measuring the color distribution of light at
various depths in the Wisconsin lakes.
Absorption measurements were made in two ways. In the
first method a beam of nearly parallel light, of circular cross
section (diameter about one cm.), was passed through the bottle
containing the algal suspension and received by a photocell
mounted behind the suspension. The area of the photocell re¬
ceiver was only slightly larger than the cross-section of the inci¬
dent beam. The apparent absorption coefficient measured in this
way includes not only absorption, but also scattering, since any
light in the beam which was more than very slightly deflected in
passing through the suspension would have failed to strike the
photocell receiver. At high concentrations the extinction coeffi-
Manning, Juday & Wolf — Photosynthesis
383
cient decreases somewhat due to secondary scattering which
tends to restore some of the scattered light to the incident beam.
The apparent absorption coefficients obtained in this manner
do not vary from color to color so sharply as do the absorption
coefficients for the pure chlorophylls, because of the addition of
the relatively non-selective scattering coefficient to the absorp¬
tion coefficient. On this basis it is possible to correct for the
amount of light lost by scattering.
Table II
Absorption coefficients for different spectral regions.
The coefficients for pure chlorophyll are based on the results of Zscheile
(193U), but expressed in units of area as measured from his curves. The
observed coefficients are values for k in the expression I /Io=10~kcA, where
c is concentration in cells per cc. and d is path length in centimeters. The
value for pure chlorophyll in the violet is estimated from other sources ,
since Zscheile's curves do not extend that far.
The average absorption coefficients in different spectral
bands for pure chlorophylls a and b were determined from meas¬
urements of curves published by Zscheile (1934). These were
then compared with our observed values. A value of three was
assumed for the chlorophyll a/chlorophyll b ratio. The results
are given in Table II. The values in the last two columns of the
table are obtained by splitting the observed extinction coeffi¬
cients in such a way that the calculated absorption coefficients
vary in the same manner as those measured from Zscheile’s
data, while the scattering coefficients increase in a more or less
regular manner as the wave-length decreases. Any materially
different distribution between the last two columns which main¬
tains the proper ratio for absorption coefficients will result in
384 Wisconsin Academy of Sciences, Arts, and Letters
unreasonable fluctuations in the values for scattering coeffi¬
cients.
The upper curve in Figure 5 is calculated on the basis of
absorption coefficients determined by the method just described.
However, the method is very indirect and depends on the ques¬
tionable assumption that the absorption spectrum of chlorophyll
in the living cell is the same, except for the additional effect of
scattering, as it is when extracted. Consequently all other quan¬
tum efficiencies in this paper are calculated from absorption
measurements made in the following manner.
Table III
Light absorbed by suspensions of Chlorella.
The figures are given in terms of incident light as 100. Path length: 4.0 cm.
One whole side of the bottle containing the algal suspension
was illuminated with the light from a tungsten filament lamp. A
CuS04 filter was used as before. A large area thermopile (re¬
ceiver area about two cm2.) was used to measure light inten¬
sities, though a photo-cell might have been equally satisfactory.
The thermopile was moved from place Jo place behind each of
the three sides of the bottle which were not facing the light.
Intensity of light (per unit area) emerging from top and bottom
of the bottle was assumed to be equal to that of light emerging
from the two lateral faces. This assumption should be nearly
correct, since the whole front of the bottle is illuminated. The
difference between the amount of incident light and the total
emergent light should give directly the amount absorbed, thus
eliminating the scattering factor. The light scattered out the
side of the bottle facing the source is neglected, but the error
so introduced is small, since for particles of the size of Chlor¬
ella the scattering of light in the direction of the source is small
compared to that scattered in other directions.
Manning , Juday & Wolf — Photosynthesis
385
Fig. 1. Fraction of light absorbed by Chlorella (1 — I/I0) as a function
of cell concentration (10® cells per cc.). Curve OA represents observed ab¬
sorption; for highest point see Table III. S, 1, 3, 5, 7, 9 are calculated
absorption curves for the corresponding depths, in meters, in Trout Lake.
The results of absorption measurements made in this way
are given in Table III. In these, as in other absorption meas¬
urements, the incident light intensity was taken as that passing
through the bottle when filled with filtered lake water. Conse¬
quently, corrections for reflection were unnecessary.
In Figure 1 the curve marked OA is drawn from the data of
Table III. The energy distribution of sunlight is somewhat dif-
386 Wisconsin Academy of Sciences, Arts, and Letters
ferent from that of the light source used in the absorption meas¬
urements. Moreover, this distribution changes continuously
with depth in the lake. The remaining curves in Figure 1 are
calculated from the data on pure chlorophyll in Table II. Change
in color distribution with change in depth is shown in Table IV.
The figures in this table are taken from unpublished data ob¬
tained in 1935 by Drs. Birge and Whitney.
From the data of Figure 1 and Table IV, it is possible to cal¬
culate quantum efficiencies from our measurements of photo¬
synthesis in Chlorella. The values so obtained, of course, are
subject to the criticism that the measurements are not made
with monochromatic light. This objection is not a serious one
if we accept the correctness of the observation of Warburg and
Negelein (1923) that the quantum efficiency is nearly independ¬
ent of wave-length throughout the visible spectrum when cor¬
rection is made for absorption by pigments other than chloro¬
phyll.
Table IV
Color distribution in Trout Lake.
The use of an average light intensity may also introduce an
error, since on some days the momentary intensity may vary
over a wide range. However, no consistent difference has been
observed between results obtained on days when the intensity
was nearly constant and those obtained when intensity fluctua¬
tions were frequent and large. The use of an average wave¬
length in calculating the number of quanta per given quantity of
energy absorbed is a somewhat uncertain procedure, though the
possible error here is not very large. At depths of 10 m. and
more, the absorption curve for 9 m. is used (Fig. 1). This un¬
doubtedly produces an error in results at depths below 9 m., but
as may be seen in Table IV and Figure 1, the change in distribu-
Manning , Juday & Wolf — Photosynthesis 387
tion between seven and nine meters is small compared with
changes at lesser depths. Changes at greater depths are doubt¬
less much smaller, since over 80 per cent of the total radiation
at nine meters is in the green and yellow while the fraction in
the blue appears to have become nearly constant at ten per cent.
Because of the small total reaction, uncertainty in the analysis
of samples from below ten meters leads to a large experimental
error in the results obtained at such depths.
Somewhat more light than is indicated by the thermopile
readings may have been available for photosynthesis in bottles
which were placed below the surface. This is because a thermo¬
pile gives a true measure of energy intensity only for light
which strikes it at normal incidence, whereas in the bottles
which we used, light from the side as well as from above can be
effective in causing photosynthesis. Since the relative amount
of scattered light probably increases with increasing depth, the
error caused by this effect should also increase with increasing
depth. This would tend to make our quantum efficiency values
too high, especially those for low light intensities. Calculations
indicate that the error caused by this effect probably amounts to
10 per cent or less, since the sun was well above the horizon dur¬
ing all of our experiments.
The quantum efficiencies are calculated on the basis of mole¬
cules of oxygen released per quantum absorbed, but if the ratio
of oxygen released to carbon dioxide assimilated is nearly unity,
as usually assumed, then the efficiency figures will also represent
the number of molecules of carbon dioxide assimilated per quan¬
tum absorbed.
Because of the various sources of error mentioned above, the
absolute values for the quantum efficiencies may be in error by
as much as 20 or 25 per cent, especially the efficiencies for low
light intensities. However, the relative values, showing varia¬
tion as a function of light intensity and algal concentration, are
probably less in error. For experimental points shown in all
the accompanying figures, the possible error due to uncertain¬
ties in chemical analysis is shown by the length of the vertical
line drawn through the point.
Experiments at different concentrations. Ehrke (1931) used
a method similar to the one used in this work. He used marine
algae in his measurements of photosynthetic rate. Emerson and
388 Wisconsin Academy of Sciences , Arts, and Letters
Green (1934) criticize his results on the ground that the carbon
dioxide concentration in his vessel was being continually and
rapidly diminished by the process of photosynthesis ; this would
result in a continual fall in the rate of photosynthesis, so that
the rate, as measured over a period of several hours, would be
much less than the actual rate under natural conditions of more
constant carbon dioxide supply. The concentration of material
used in our experiments is very much lower than Ehrke used,
but it seemed desirable to see how much, if at all, our results
might be in error because of diminishing carbon dioxide tension.
For this purpose, two series of experiments were carried out,
in which the quantum efficiency for a single light intensity was
simultaneously determined for a series of algal concentrations.
The results are shown in Figure 2 where the quantum efficiency
is plotted against algal concentration ; for convenience the latter
is indicated on a logarithmic scale. For both runs all bottles
were placed at a depth of 2 m., but the light intensity on July
24 was about 75 per cent greater than that of July 27. The
quantum efficiencies are higher for the July 27 run than for that
of July 24, as would be expected from the results for the series
of light intensities shown in Figure 5. The July 24 run lasted
for one hour, while the July 27 run lasted three hours. The fig¬
ures beside each experimental point indicate total reaction in
mg/1 of oxygen produced. For the point in Figure 2 where the
oxygen produced is 11.56 mg/1, the net gain is actually only
about 9.0 mg/1, since respiration consumed about 2.5 mg/1 of
oxygen. Multiplying by the factor 44/32 gives approximately
the mg/1 of carbon dioxide consumed. As already stated, the
total carbon dioxide present in the water was about 38 mg/1.
The solid curves in Figure 2 show the results obtained in the
experiments, while the broken line curves are calculated, using
the observed values at low concentrations. The calculations are
based on the fact that, at intensities as high as these, the effi¬
ciency increases rapidly with decreasing intensities as shown in
Figure 5. As the algal concentration is increased, a larger and
larger portion of the algae is exposed to reduced intensities, due
to the shading by other algal cells between them and the light.
Consequently the total efficiency should increase as the algal con¬
centration increases, as shown in Figure 2.
Manning, Juday & Wolf — Photosynthesis
389
.009.
.008
.007
.006
.005
.004
.003
.002
.001
099 _
2, 0j_ I
— - ,
6.83 I
>0.97 -z^r
— In
4122 - 1ST
a 5.70
.2 3 .4 .6 .8 I 2 3 4 6 8 10
MULTIPLY BY I06
Fig. 2. Quantum efficiency as a function of cell concentration (cells per
cc.). A represents July 24, one hour, depth 2 m., intensity 180 XlO5 ergs/cm2/
sec. B represents July 27, 3.00 hours, depth 2 m., intensity 1.04 x10s ergs./
cm2/sec.
The points in both cases fit the broken line curves within ex¬
perimental error up to a concentration of about 106 cells per
cubic centimeter. Above this concentration the experimental
points fall definitely below the calculated curve. However, there
is little if any tendency for the quantum efficiency to decrease
with increasing concentration above 106 cells per cubic centi¬
meter. In the shorter of the runs, the efficiency is definitely
higher at a concentration of 7 X ID6 than it is at one of 3.5 X
106. For the longer run, there is only slight evidence of a rise
in efficiency between concentrations of 3.2 X 10° and 6.5 X 10°
Chlorella cells per cubic centimeter. Actually a horizontal
straight line could satisfactorily represent the results for the
longer run. It appears then that the tendency for increasing
efficiency with increasing concentration is just about balanced
by the inhibiting effect of continually diminishing carbon di-
390 Wisconsin Academy of Sciences , Arts, and Letters
oxide concentration, or by the closely related effect of inhibition
due to increased pH. In short runs, where the carbon dioxide
consumed is not too great, it appears that the tendency for the
quantum efficiency to increase rapidly at high algal concentra¬
tions is sufficient to overbalence slightly the opposite inhibiting
tendency.
At low light intensities, corresponding to depths of 9 or 10
meters, the tendency for efficiencies to increase at high concen¬
trations would be negligible, since the efficiency appears nearly
to have reached its maximum value (Fig. 5). On the other hand,
the total reaction rate is much less at such low intensities (Fig.
3), so the inhibiting effect is also reduced.
0| 2 3 4 5
■MULTIPl* BY 10*
Fig. 3. Reaction rate for Chlorella (A02, mg/1) as a function of light
intensity (ergs/cm2/sec) . July 25, 3.17 hours, cell concentration 3,250,000
per cc. j 4
Except for these two runs, we have used no concentration of
Chlorella higher than 3.25 X 10°. It may be concluded that al¬
though the inhibiting effect pointed out by Emerson and Green
(1934) can be detected in our work, it is not necessary to correct
for it provided that the opposite correction for the shading effect
is also neglected.
Experiments at different depths . Figures 3, 4 and 5 show
the results obtained from an experiment made at a series of
depths. The light intensities at each depth below the surface are
calculated from the surface intensity and from transmission
measurements made two days later at one meter intervals from
the surface to 14 m. In Figure 3 total reaction is plotted against
light intensity. The solid curve is in terms of incident intensity ;
the broken line curve is in terms of relative amounts of energy
Manning, Juday & Wolf — Photosynthesis
391
Fig. 4. Reaction rate for Chlorella as a function of light intensity;
identical with Figure 3 except that the intensity scale is logarithmic.
actually absorbed. The intensities for corresponding points on
the two curves are different except for the surface point because
the absorption coefficient decreases with increasing depth in the
lake (Fig. 1). Figure 4 is the same except that the light inten¬
sity scale is logarithmic. Here the curve is of the same type as
obtained when the rate of photosynthesis is plotted against depth
(Schomer and Juday 1935; Curtis and Juday 1937). In Figure
5 quantum efficiency is plotted against light intensity, the latter
scale being logarithmic. The upper curve represents efficiencies
calculated on the basis of absorption measurements made by the
first of the two methods described above (page 383). The lower
curve in Figure 5 is calculated from absorption measurements
made by the second method. Since the second method is much
more direct, the lower curve is probably more nearly correct.
The upper curve is included merely to show that the two differ¬
ent methods give results in reasonable agreement. The point at
14 m. is subject to large uncertainty because the total reaction
there was very small.
In Figure 6, curves are given for all runs made with Chlor¬
ella except those shown in Figure 2. The top curve in Figure 6
is for the run shown in Figures 3, 4, and 5. All the other curves
lie below it. The same figures for transmission in the lake were
used in all cases, though correction was always made for change
392 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 5. Quantum efficiencies for Chlorella as a function of light in¬
tensity (ergs/cm*/sec). The two curves are for the same experiment, but
calculated according to two different methods of measuring light absorption
as described in text p. 384. July 25 3.17 hours, cell concentration 3,250,000
per cc.
in transmission caused by changing elevation of the sun. The
five lower curves were all obtained from runs made at least two
weeks after July 27, when the transmission measurements were
made, and it is possible that the lake had become somewhat less
transparent at the time of the later runs. However, Secchi disk
readings gave little indication of such a trend. Depth (in meters)
at which the disk was just visible varied as follows: July 25,
4.8; August 6, 5.0; August 13, 5.0; August 19, 4.8; August 25,
4.7. At an intensity of 105 ergs/cm2/sec, corresponding in all
cases to a depth near the surface, the calculated efficiency is
nearly the same for four of the series of experiments.
In Figure 7 total reaction is plotted against light intensity
(logarithmic scale) for four of the runs with Chlorella. The
a02 scale has been arbitrarily adjusted for each curve in order
to permit comparison of their shapes. This figure shows con¬
clusively that high light intensity is not the only factor causing
Manning , Juday & Wolf— Photosynthesis
393
Fig. 6. Quantum efficiencies for Chlorella as a function of light in¬
tensity. Curve A, July 25, 3.17 hours, cell concentration 3,250,000 per cc.;
B, August 19, 1.03 hours, cell concentration 718,000 per cc. ; C, August 20,
4.00 hours, cell concentration 3811,000 per cc.; D, August 19, 4.00 hours, cell
concentration 718,000 per cc.; E? August 14, 4.05 hours, cell concentration
1,900,000 per cc.; F, August 18, 3.30 hours, cell concentration 1,210,000
per cc.
the lower reaction usually observed at the surface. In the curves
for August 18 and 20, the light intensity at the surface is less
than the intensity at 2 m. for the run of July 25, but though the
two-meter rate is practically a maximum for July 25, the surface
rates for August 18 and 20 are distinctly below the maximum.
However, in the run for August 14, where the surface intensity
was about one-twelfth of that for July 25, the maximum oc¬
curred at or very near the surface.
In Figure 8, which is similar to Figure 7, are shown the re¬
sults for a long and short run with Chlorella. These runs were
started at the same time, using the same concentration of cells.
One run was stopped after 1.03 hours, the other after 4.0 hours.
It happened that the average light intensity at the surface was
identical for the two runs, though the correction for changing
394 Wisconsin Academy of Sciences , Arts , and Letters
Fig. 7. Total reaction (A02, mg/1) as a function of light intensity
(ergs/cma/sec) for experiments with Chlorella. The A02 scale is arbitrary.
The heights of the curves have been made approximately equal in order to
facilitate comparison. To obtain actual values of A02, the readings on the
curves must be multiplied by the following factors: Ax2; Bx4; Cxi;
Dx0.67. A, August 14, 4.05 hours, cell concentration 1,900,000 per cc. ; B,
July 25, 3.17 hours, cell concentration 3,250,000 per cc. ; C, August 18, 3.30
hours, cell concentration 1,210,000 per cc.; D, August 20, 4.00 hours, cell
concentration 331,000 per cc.
solar elevation made the intensities at corresponding depths be¬
low the surface somewhat lower for the long run. In the figure,
a02 for the shorter run is multiplied by 4/1.03. Most of the
points for the short run lie, within experimental error, on the
long run curve. The apparently higher maximum for the 4-hour
curve may be due to growth.
Experiments with Other Algae
Figures 9, 10, 11 and 12 show the results of rate measure¬
ments made with several species of algae other than Chlorella.
The Anabaena circinalis for the run of August 4 (Fig. 9)
was gathered in Mann Lake, using a tow net just below the sur¬
face. The material gathered in this way was suspended in Mann
Lake water and used without filtering or centrifuging. A con¬
siderable amount of Chroococcus and Microcystis was also pres¬
ent, but the Anabaena probably constituted 80 per cent or more
of the plant material. Mann Lake contains about the same total
carbon dioxide concentration as does Trout Lake, though the pH
is usually somewhat higher.
The Anabaena limnetica for the run of August 17 was sim¬
ply dipped from near the surface of Little Star Lake and used
undiluted. The material was abundant enough to form an almost
Manning , Juday & Wolf — Photosynthesis 395
opaque layer at and immediately below the surface, and ap¬
peared to be a nearly pure culture of Anabaena. The total car¬
bon dioxide concentration of Little Star Lake water is about
34 mg/1.
MULTIPLY by I04
Fig. 8. Relative reaction rates for long and short runs. To get actual
values of AO* m mg/1 for points marked with solid circles, multiply ordinates
by 1.03/4.00. Tne curve is drawn to fit the four-hour points. August 19,
time 1.03 and 4.00 hours, cell concentration 718,000 per cc.
The Cladophora (Fig. 10) was found growing on rocks in a
few centimeters of water near the north shore of Arbor Vitae
Lake, where it is often exposed to intense sunlight. The fila¬
ments were cut in lengths of about two centimeters, suspended
in filtered Trout Lake water and constantly stirred in order to
maintain an even distribution while the bottles were being filled.
The surface intensity for this run was high, yet the maximum
rate occurred at the surface. Apparently the Cladophora was
adapted to the high surface intensity. The curve shows an
apparent tendency for a secondary maximum at four meters.
The species of Cladophora used in this experiment was not de¬
termined.
In Figure 11 are given the results for two undetermined
species of Spirogyra. In both cases the material was gathered
near the surface, that for the run of July 29 coming from Little
Arbor Vitae Lake and that for the run of August 6 coming from
a swampy stream near High Lake. The material was prepared
for use in the same manner as was the Cladophora. The shapes
of the two curves are quite different, though both show inhibi¬
tion at the surface, with maxima at 0.5 and 2.0 m., and a later
396 Wisconsin Academy of Sciences , Arts , and Letters
break in the downward progress of the curves. There is a
sharper break in curve A than in B.
Fig. 9. Photosynthetic rate (A(h, mg/1) as a function of light intensity
(ergs/cm2/sec) for two species of Anabaena. A, A. circinalis} August 4,
3.17 hours, cell concentration 142,000 per cc. ; B, A. limnetica, August 17,
3.07 hours, cell concentration 151,000 per cc.
Figure 12 shows the combined results for two runs with
Spiro gyra crassa. This is a very large celled species and was
found growing in the bottom of Muskellunge Lake at a depth of
about 6 m. The run for August 22 extended only from the sur¬
face to a depth of 5 m. The run for August 24, carried out on a
cloudy day, extended from 2 to 11 m. Points for both runs indi¬
cate a second maximum in the rate curve at about 3 X 104 ergs
/cm2/sec. This is approximately the intensity for sunlight at
the depth in Muskellunge Lake at which this species was ob¬
tained. The figures for a02 at the various depths in the August
24 run were multiplied by 1.95/3.79 in order to make the height
correspond to that of August 22 run. The ratio 1.95/3.79 was
chosen because it brought the two-meter point of the August 24
run on the curve of the August 22 run.
For all of the runs with algae, concentrations were deter¬
mined in terms of cells per cubic centimeter of suspension, but
no effort has been made to correlate absolute reaction rates for
other species with those for Chlorella pyrenoidosa, since not only
the cell size, but type and distribution of chloroplasts, must in¬
fluence the rate per cell for different species. However, to per¬
mit comparison with the results of Curtis and Juday (1937),
Table V gives the rates, expressed in milligrams of oxygen per
106 cells per hour, for all of the runs in which algae were used.
Manning , Juday & Wolf — Photosynthesis
397
Fig. 10. Photosynthetic rate (A02, mg/1) as a function of light inten¬
sity (ergs/cm2/sec) for a species of Cladophora. July 31, 3.25 hours, cell
concentration 851 per cc.
Experiments with Angiosperms
Figure 13 shows the results of the rate measurements with
three species of aquatic angiosperms. In all three cases, the
material used in the experiments was taken from actively grow¬
ing portions of the plants.
The Potamogeton was collected in Trout Lake at a depth of
about 1 m. Leaves of approximately equal size were removed
from the region just back of the tip. The leaves were sus¬
pended in filtered lake water and then four leaves were chosen
at random and placed in each bottle. In addition, a portion of
stem bearing four leaves was placed in each of eight bottles
(four light, four dark). These last eight bottles were all sus¬
pended at 2 m. during the run and the mean rate value for them
is given by the higher two-meter point in Figure 13. Appar¬
ently the power of the leaf to photosynthesize is only slightly
impaired by removal from the stem. Checks for duplicate bottles
were only fair, probably because of failure to obtain equal
amounts of material in each bottle.
The Sagittaria was collected from a depth of about one-third
to one-half meter in Trout Lake. Six leaves chosen at random
were placed in each bottle. The Vallisneria was collected from
about the same depth, but from Arbor Vitae Lake rather than
Trout Lake. For this plant, ten centimeter strips, each includ¬
ing the growing tip, were cut from each leaf. Three such strips
were placed in each bottle.
No definite evidence for adaptation to high light intensities
can be obtained from the curves for the three species shown in
398 Wisconsin Academy of Sciences , Arts , and Letters
Figure 13. It would be desirable to have observations on repre¬
sentatives of these species which had grown at greater depths.
Ruttner (1926a) made such observations on Elodea canadensis
and found definite evidence of light adaptation for this species.
Fig. 11. /Photosynthetic rate (A02, mg/1) as a function of light in¬
tensity (ergs/cm2/sec) for two species of Spirogyra. To obtain observed
values of A02, the points in curve B must be multplied by 2. Curve A,
August 6, 3.00 hours, cell concentration 4,700 per cc.; B, July 29, 3.00 hours,
cell concentration 1,370 per cc.
Fig. 12. Photosynthetic rate (A02, mg/1) as a function of light inten¬
sity (ergs/cm2/sec) for Spirogyra crassa. To obtain observed values of
A02, the points represented by open circles must be multiplied by 379/195;
the solid circles represent observed values. Solid black circles, August 22,
3.00 hours, cell concentration 630 per cc.; open circles, August 24, 4.50
hours, cell concentration 338 per cc.
Manning, Juday & Wolf — Photosynthesis
399
Table V
Quantity of oxygen produced per million algal cells per hour at different
depths in Trout Lake. Incident energy values are for visible light only.
400 Wisconsin Academy of Sciences , Arts, and Letters
Table V — Continued
Manning , Juday & Wolf — Photosynthesis 401
Discussion
Quantum efficiency. It may be seen from Figure 5 (lower
curve) that the value found for the quantum efficiency of photo¬
synthesis at low light intensities, namely, 0.05, is far below the
value of 0.25 obtained by Warburg and Negelein (1923) for
Chlorella. They used monochromatic light at intensities near
103 ergs/cm2/sec.
Even if the 14 m. point in Figure 5 be ignored, and the curve
continued as a straight line through the 7 and 10 m. points, it
still intersects the 103 intensity ordinate at a value about one-
third that found by Warburg and Negelein. Moreover, there
seems no reason to doubt the validity of the 14 m. point, within
the wide range of experimental error. The value given on the
curve, 0.041, is the mean of the values 0.033 and 0.049, obtained
for the two duplicate bottles. The temperature at 14 m. for this
run was 10.8° C., as compared with 16.2° at 10 m., 21.3° at 7 m.,
23.5° at 5 m. and 24.5° at the surface; but the low temperature
at 14 m. should have had no appreciable effect on the rate, since
experiments with Chlorella by Warburg (1919) showed that the
temperature coefficient had dropped to unity at intensities as
high as 6 or 7 X 103 ergs/cm2/sec. (Warburg's data permit
only an approximate estimate of the absolute values of the light
intensities used in his work.) This is in accord with the results
for Hormidium obtained by Van den Honert (1930) and for
Elodea obtained by Ruttner (1926b).
The photosynthetic rate at 14 m. shown in Figures 3 and 4
is about twice the rate of respiration at the same depth. (See
Table V) . In the quantum efficiency measurements of Warburg
and Negelein (1923), the respiration rate exceeded the photo¬
synthetic rate, doubtless because their suspensions were so con¬
centrated that a large fraction of the cells received almost no
light.
It is difficult to see how the use of monochromatic light could
lead to efficiencies very much greater than those obtained for
polychromatic light, unless the observation by Warburg and
Negelein (1923) that quantum efficiency varies only slightly with
wave-length be disregarded.
Experiments in the laboratory at the University of Wiscon¬
sin, in which monochromatic light was used, indicate a quantum
efficiency for Chlorella that is much lower than the values ob-
402 Wisconsin Academy of Sciences , Arts , and Letters
Fig. 13. Photosynthetic rate (A02, mg/1) as a function of light inten¬
sity (ergs/cm2/sec) for three species of aquatic angiosperms. Curve A,
Potamogeton richardsonii, August 3, 2.05 hours; B, Vallisneria spiralis,
August 10, 3.00 hours; C, Sagittaria graminea, August 10, 2.25 hours.
tained by Warburg and Negelein and in approximate agreement
with our values. (These results will appear in a paper soon to
be published.) i
Actually, kinetic considerations make it probable that the
quantum efficiency should possess a maximum value at a fairly
low light intensity, and fall off again at still lower intensities.
Apparently three or more quanta of visible light must be suc¬
cessively absorbed by chlorophyll in order to reduce a single
molecule of carbon dioxide to formaldehyde. If any of the inter¬
mediate photochemical steps are appreciably reversible, either
thermally or by fluorescence, then the rate should fall off more
rapidly than the first power of the light intensity, as the inten¬
sity is itself diminished. How low the intensity would have to
be to make this effect appreciable would depend on the rate of
the reverse steps. If such a loss in efficiency were due to fluor¬
escence, then the temperature should have little effect on the
Manning , Juday & Wolf — Photosynthesis 403
value of the light intensity for which the loss is just noticeable.
If, as would be more probable, the decrease in efficiency were
due to thermal reversal of one or more of the photochemical
steps, then the effect should be observable for a higher light in¬
tensity at a temperature of 25° or 30° C. than at 5° or 10°.
In a recent review, Emerson (1936) observed that if there
were appreciable decomposition of intermediates in the photo¬
chemical steps of photosynthesis, then the quantum efficiency
should decrease at low light intensities. Emerson cited the high
efficiencies found at low intensities by Warburg and Negelein
(1923) as an indication that intermediate compounds do not
appreciably decompose in the intervals between absorption of
light quanta.
However, the experiments of Warburg and Negelein were
carried out a temperature of 10° C., where thermal decomposi¬
tion of intermediates might be much slower than at higher tem¬
peratures. At 10° C., light intensites still lower than those used
by Warburg and Negelein might show a lower efficiency.
If a reversal of one or more of the photochemical steps can
occur, then curves of the type shown in Figure 3 should be con¬
cave upward for sufficiently low intensities. So far as we know,
such a form of curve has not been reported, but since little work
has been done at very low intensities, this type of behavior may
easily have escaped observation. In future work, we plan to
study more fully the rate and efficiency of photosynthesis at
very low light intensities.
Limiting factors at high intensities. In the region of maxi¬
mum reaction rate, as for two, three and four meters in Figure
3, it would be desirable to know what factor or factors are re¬
sponsible for limiting the rate, and consequently for reducing
the efficiency. It is probable that in our work both temperature
and carbon dioxide concentration are factors influencing the
maximum rate. In the preliminary run of July 11 (Table V),
about 3 mg/1 of free carbon dioxide was added to the water used
for the algal suspension. The average reaction rate for the four
concentrations, 0.0041 mg. 02/106 cells/hour, is distinctly high¬
er than the maximum for the other Chlorella runs, but since the
concentrations were very low, total reaction was small and the
results are subject to rather large error. The amount of added
404 Wisconsin Academy of Sciences , Arts , and Letters
carbon dioxide is difficult to control, since much of it is likely to
re-escape during the operations incident to starting a run. There
are no data in our work to show the influence of temperature
variations. The temperatures at two, three and four meters for
the rates shown in Figure 3 were 24.1°, 23.9° and 23.7° C., re¬
spectively. Most investigations of the effect of temperature on
photosynthetic rate indicate only a small increase in rate as the
temperature increases above 24°, at least for measurements ex¬
tending over a period of several hours. It therefore seems prob¬
able that the maximum rates in the present work are somewhat
more sensitive to changes in the carbon dioxide concentration
than to changes in temperature.
Inhibition of photosynthesis at the lake surface . As was
pointed out in the presentation of results (Fig. 7), the effect of
high light intensity alone is not sufficient to explain the rate
^eduction observed at the surface for Chlorella. A comparison
of color distribution at the surface and at 2 m. in Trout Lake
(Table IV — 2 m. value obtained by interpolation) shows that the
percentage of red is smaller at 2 m. (about 14 per cent as against
22 per cent at the surface) . The percentages for violet and blue
are only slightly reduced and those for orange, yellow and green
are increased at 2 m. In addition to the visible radiation at the
surface, about as much more energy is contributed by infra-red
radiation. Practically all of this is absorbed in the first meter
of water. There is also a small amount of ultra-violet radiation
at the surface, although it is necessary to consider only the por¬
tion of wave-length longer than about 3300 a, since the bottles
used do not transmit shorter wave-lengths. The small amount
of organic stain in Trout Lake water is enough to cut out ultra¬
violet fairly rapidly, especially that of short wave-length. There¬
fore, it is probable that for high intensities of visible light
(around 1.5 X 105 ergs/cm^/sec. or higher), one or more of the
following changes will cause a reduction in rate: (1) Increase
of the percentage of red in the incident light, with the total in¬
tensity remaining constant; (2) addition of a large amount of
infra-red (probably short wave-length infra-red) to the incident
light; (3) addition of a moderate amount of long wave-length
ultra-violet to the incident light,
Manning , Juday & Wolf — Photosynthesis 405
From Figure 7 it seems probable that the inhibiting effect
becomes less as the intensity of visible light is decreased. On
very cloudy days, the maximum reaction occurs at the surface,
as in the run of August 14. (See also Schomer and Juday
1935). However, on such a day, the infra-red intensity is also
very greatly reduced.
The observed facts can be explained by assuming that at
high light intensities, a photochemical oxidation of chlorophyll
can take place. Evidence indicates that the effective light ab¬
sorbing agent for photosynthesis is not free chlorophyll, but a
chlorophyll-C02 or a chlor ophy 11-H2 C 03 complex. The evidence
for the complex as the active agent has been reviewed by Burk
and Lineweaver ( 1935) . In the dark or at low light intensities,
it is probable that nearly all of the active chlorophyll is com¬
bined with carbon dioxide. However, as the light intensity be¬
comes higher and the rate of photosynthesis increases, the rate
of formation of free chlorophyll (free as a result of the comple¬
tion of a cycle of carbon dioxide reduction) will also increase.
The rate of combination of chlorophyll with carbon dioxide will
probably depend on the temperature as well as on concentration
of carbon dioxide and chlorophyll. Other variables being kept
constant, the equilibrium concentration of free chlorophyll will
increase as the light intensity increases. This means that as the
intensity increases, a larger and larger fraction of the energy
will be absorbed by free chlorophyll and hence be unable to pro¬
duce photosynthesis. It is, however, reasonable to suppose that
this activated free chlorophyll molecule may be oxidized on col¬
lision with an oxygen molecule. Since oxygen is a product of
photosynthesis, its concentration in the illuminated chlor oplast
is high.
The mechanism just postulated can also explain the observed
facts if it be modified to permit photo-oxidation of the chloro¬
phyll whether or not it is in combination with carbon dioxide.
According to this modification, the rate of oxidation would be
proportional to light intensity, total chlorophyll concentration
and to oxygen concentration. The quantum efficiency for oxida¬
tion would probably increase slightly with increasing light inten¬
sity as long as the net photosynthetic rate were to increase, be¬
cause the concentration of oxygen in the chloroplast would also
increase. Above this intensity, the efficiency of photo-oxidation
406 Wisconsin Academy of Sciences , Arts , and Letters
would decrease, but the rate of oxidation would continue to in¬
crease even at much higher intensities.
A choice between these alternative mechanisms might be
made on the basis of experiments with Chlorella at high light
intensity and high carbon dioxide concentration, but at a low
temperature, such as 5° or 10° C. Under these conditions the
maximum rate of photosynthesis should be limited by the rate
of the so-called Blackman reaction (or reactions) which prob¬
ably follows the photochemical reaction in photosynthesis
(Burk and Lineweaver 1935). The concentration of free chlor¬
ophyll should then be much lower than under our experimental
conditions, and if the first mechanism is the correct one, fur¬
ther increase in intensity should result in much less oxidation
than would be necessary to explain the inhibition we observe at
higher temperatures. On the other hand, if the second alterna¬
tive is correct, then the relative inhibition at high intensities
should be somewhat increased or at least not decreased, since
the rate maximum would be lower and would be reached at
lower intensities than for experiments at higher temperature.
It is also possible that the actual mechanism may be inter¬
mediate between these two alternatives, that is, that either
chlorophyll-C02 or free chlorophyll can be oxidized, but that
oxidation proceeds more easily for the free chlorophyll.
The consumption of oxygen in this photo-oxidation would in
itself tend to cause an apparent reduction in the rate of photo¬
synthesis, since the rate is measured in terms of oxygen pro¬
duced. Complete oxidation of a chlorophyll molecule would re¬
quire about 70 oxygen molecules, but probably only one, or at
most a few, of the oxygen molecules would be involved in photo¬
oxidation of chlorophyll. It is possible that the first stages of
oxidation would not entirely destroy the optical properties of
the chlorophyll. Chlorophyll b may be the first stage in the oxi¬
dation of chlorophyll a (-CH3 group replaced by -CHO group) .
Moreover, evidence indicates that in the photochemical forma¬
tion of chlorophyll in the etiolated leaf of Zea mays (Inman
1937), chlorophyll a is the first to be formed. It is possible that
the reaction forming chlorophyll results exclusively in produc¬
tion of chlorophyll a and that chlorophyll b results from a sec¬
ondary oxidation reaction. Further oxidation of the product or
products might proceed thermally and thus cause an apparent
Manning, Juday & Wolf — Photosynthesis 407
increase in respiration after a period of rapid photosynthesis,
such as noted by Van der Paauw (1932). In Van der Paauw’s
measurements of stimulation of respiration in light at exceed¬
ingly low carbon dioxide concentration, he may have been meas¬
uring, at least in part, a rate of photo-oxidation of chlorophyll.
Oxidation of any considerable fraction of the chlorophyll in a
cell would also lower the rate of photosynthesis by reducing the
amount of effective light absorption. Over periods of an hour or
longer, such as in our measurements, this reduction of light
absorption would probably contribute more to the inhibiting
effect than would the actual process of oxidation, unless the oxi¬
dation were to approach completion for most of the chlorophyll
destroyed.
The rate at which photosynthesis increases with increasing
light intensity becomes less as the intensity becomes greater,
whereas the postulated retarding effect, due to photo-oxidation
of the chlorophyll, would become greater as the intensity in¬
creases. The maximum observed rate would occur at an inten¬
sity where the opposing effects become equal to each other. The
picture is complicated by the fact that chlorophyll concentration
must also be treated as a variable. This leads to the conclusion
that at a given light intensity, an observed inhibiting effect will
become greater as the time of measurement is increased. To
complete the picture, it is necessary to consider the fact that
another photo-reaction, resulting in chlorophyll formation, is
also taking place. Moreover, for long periods of time, increase
in plant material, or growth, must be considered.
So far, only the effect of increasing the light intensity, with¬
out altering the wave-length distribution, has been considered.
If, to a given intensity of light, there be added light of a wave¬
length nearly or entirely incapable of producing photosynthesis,
but still able to cause oxidation of free chlorophyll, then inhibi¬
tion would be more marked than in the case of an equal inten¬
sity change without change in wave-length distribution. Thus,
from the results observed at the surface in Trout Lake, light of
wave-length between 3300 and 4000a, or of wave-length longer
than 6900 or 7000 A, may be suspected of causing relatively more
oxidation than photosynthesis, at least for high total intensities.
If the process postulated in the foregoing discussion is cor¬
rect, then for a simple alga like Chlorella, rate of photosynthesis
408 Wisconsin Academy of Sciences , Arts, and Letters
would tend to reach an equilibrium value for constant values of
temperature, light intensity, wave-length distribution and car¬
bon dioxide concentration. At equilibrium a balance would
exist between the rates of three photo-reactions, namely, photo¬
synthesis, chlorophyll formation and chlorophyll destruction. A
change in either light intensity or carbon dioxide concentration
would result in a gradual shifting in rate of photosynthesis until
a new equilibrium concentration of chlorophyll could be at¬
tained.
For some plant species, studies have been made on chloro¬
phyll concentration and rate of photosynthesis as a function of
the intensity of light in which the plants were grown (Spoehr
and Smith, 1936; Harder, 1933). The results are in agreement
with the process just discussed.
A quantitative discussion of the process postulated above,
with applications to specific examples, will appear in a later
paper.
The observed behavior for Anabaena (Fig. 9) appears quite
similar to that for the Chlorella. This is perhaps not surprising
since both are plankton algae, and hence unable to adopt a fixed
position for growth.
The types of light adaptation shown by Cladophora (Fig.
10) and Spirogyra crassa (Fig. 12) cannot easily be explained
in terms of the picture just presented, unless there be intro¬
duced additional postulates, such as different behavior for chlor¬
ophylls a and b, or differences in mechanism for light absorbed
in the various regions of the chlorophyll absorption spectra.
In any event, a more complex behavior would be expected
for plants with large and complex cells, where distribution of
chloroplasts and lengths of diffusion paths must influence the
process of photosynthesis.
Summary
Rates of photosynthesis at various depths in Trout Lake were
measured for several species of algae and for three species of
higher plants.
Experiments with Chlorella pyrenoidosa, made at a series of
algal cell concentrations, show that the carbon dioxide supply
in Trout Lake water is adequately maintained during the course
Manning , Juday & Wolf — Photosynthesis 409
of a three hour experiment, unless the concentration of material
is very high.
From experiments with Chlorella, the quantum efficiency of
carbon dioxide assimilation has been calculated for light inten¬
sities varying from approximately 1 X 103 to 5 X 105 ergs/cm2/
sec. The value approached for low light intensities is approxi¬
mately 0.05, much lower than the value found by Warburg and
Negelein (1923).
Experiments with Chlorella show that the failure to main¬
tain a rate of photosynthesis at the surface as high as that for
depths of 2 or 3 m. below the surface, is due not only to exces¬
sive intensities of light, but also to a difference in wave-length
distribution at the surface.
For two species of Anabaena, variations in rate at different
depths were similar to those for Chlorella.
A species of Cladophora, found growing where often sub¬
jected to intense illumination, was the only alga studied in 1936
to show a maximum rate of photosynthesis at the surface of the
lake on a clear day.
Striking evidence for light adaptation was found for Spiro-
gyra crassa, a very large-celled species collected at a depth of
6 m. in Muskellunge Lake. It showed maximum rates of photo¬
synthesis for two light intensities, one just below the surface of
the water, the other at a depth where the intensity was approxi¬
mately 3 X 104 ergs/cmVsec., which is about the amount for
sunlight at 6 m. in Muskellunge Lake.
Two other species of Spirogyra, found growing in shallow
water, showed only a single rate maximum.
Experiments with Potamogeton, Vallisneria and Sagittaria
showed no definite evidence of light adaptation. None of these
plants showed diminished rates at the surface. A different type
of rate-intensity curve was obtained for each species.
The diminished rates at the surface, observed for most of
the runs with algae, may be explained on the basis of a photo¬
oxidation, to which the cholorphyll may be susceptible at high
light intensities. A process of this type can explain the ability
of plants to adapt themselves to low or high light intensities.
Literature
Barker, H. A. 1935. Photosynthesis in diatoms. Archiv. f. Mikrobiologie 6:
141-156,
410 Wisconsin Academy of Sciences , Arts , and Letters
Birge, E. A. and Juday, C. 1931. A third report on solar radiation and
inland lakes. Trans. Wis. Acad. Sci., Arts and Let. 26 : 383-425. Also
27: 523-562. (1932).
Burk, D. and Lineweaver, H. 1935. The kinetic mechanism of photosyn¬
thesis. Cold Spring Harbor Symposia 3: 165-179.
Curtis, J. T. and Juday, C. 19317. Photosynthesis of algae in Wisconsin
Lakes. III. Observations of 1935. Internat. Rev. ges. Hydrobiol. u.
Hydrog. 35: 122-133.
Ehrke, G. 1932. Ueber die Assimilation komplementar gefarbter Meeres-
algen im Lichte von verschiedenen Wellenlangen. Planta Arch. Wiss.
Bot. 17: 650-665.
Emerson, R. 1936. Review of investigations in photosynthesis. Ergebnisse
d. Enzymforschung 5 : 305.
Emerson, R. and Green, L. 1934. Manometric measurements of photosyn¬
thesis in the marine alga Gigartina. Jour. Gen. Physiol. 17: 817-842.
Harder, R. 1933. Ueber die Assimilation der Kohlensaure bei konstanten
Aussenbedingungen II. Planta Arch. Wiss. Bot. 20: 699-733.
Inman, O. L. 1937. The ratios of chlorophyll a and b and the mechanism
of photosynthesis. Science 85: 52.
Marshall, S. M. and Orr, A. P. 1928. The photosynthesis of diatom cultures
in the sea. Jour. Mar. Biol. Assoc. United Kingdom, N. S. 15: 321-360.
Ruttner, F. 1926a. Ueber die Kohlensaureassimilation einiger Wasser-
pflanzen in verschiedenen Tiefen des Lunzer Untersees. Internat. Rev.
ges. Hydrobiol. u. Hydrog. 15: 1-30.
Ruttner, F. 1926b. Ueber den Gaswechsel von Elodeasprossen verschiedener
Tiefenstandorte unter den Lichtbedingungen grosserer Seetiefen. Planta
Arch. Wiss. Bot. 2: 588-599.
Schomer, H. A. 1934. Photosynthesis of water plants at various depths in
the lakes of northeastern Wisconsin. Ecology 15: 217-218.
Schomer, H. A. and Juday, C. 1935. Photosynthesis of algae at different
depths in some lakes of northeastern Wisconsin. Trans. Wis. Acad.
Sci., Arts and Let. 29: 173-193.
Spoehr, H. A. and Smith, J. H. C. 1936. The light factor in photosynthesis.
Paper XXXI (p. 1015-1058) in “Biological Effects of Radiation”, B. M.
Duggar, Editor. McGraw-Hill. 1936.
Van den Honert, T. H. 1930. Carbon dioxide assimilation and limiting
factors. Rec. Trav. Bot. Neerland. 27: 149-284.
Van der Paauw, F. 1932. The indirect action of external factors on photo¬
synthesis. Rec. Trav. Bot. Neerland. 29: 500-620.
Warburg, O. 1919. Ueber die Geschwindigkeit der photochemischen Kohlen-
saurezersetzung in lebenden Zellen. Biochem. Zeitschr. 100: 230-262.
Warburg, O. and Negelein, E. 1923. Ueber den Einfluss der Wellenlange
auf den Energieumsatz bei der Kohlensaureassimilation. Zeitschr. f.
Physik. Chem. 106: 191-216.
Zscheile, F. P., Jr. 1934. An improved method for the purification of
chlorophylls a and b; quantitative measurement of their absorption
spectra; evidence for the existence of a third component of chloro¬
phyll. Bot. Gaz. 95 : 529-561,
AMOUNT AND DISTRIBUTION OF THE CHLOROPHYLL
IN SOME LAKES OF NORTHEASTERN WISCONSIN
Zygmunt Kozminski
Wigry Hydvobiological Station, Poland
From the Limnological Laboratory of the Wisconsin Geological
and Natural History Survey. Notes and reports No. 85.
The greatest obstacle in investigations concerning lake pro¬
ductivity has been the difficulty of quantitatively determining the
life activity of the lake inhabitants. Neither the most accurate
knowledge of the quantitative occurrence and distribution of
every species in a lake, nor data concerning the total organic
content of a lake can be directly used as a measure of the rate of
production and consumption of organic matter. The chief reason
for this is the physiological heterogeneity of the material investi¬
gated.
However, modern limnology appears to have found an effec¬
tive approach to a study of lake metabolism in the measurement
of photosynthetic and respiratory activity of water organisms in
closed glass vessels, which are exposed at different depths in
lakes (see, for instance, Schomer and Juday 1935; Curtis and
Juday 1937; Manning, Juday and Wolf 1938). Research in this
direction, when considered in connection with advancing knowl¬
edge of light conditions in inland lakes, is extraordinarily prom¬
ising, especially when the experiments are made with natural
populations of water organisms.
The investigations described in this paper represent an en¬
deavor to approach the same problem from another direction.
The experimental measurement of photosynthetic and respira¬
tory activity establishes more or less artificial life conditions
which may be compared with natural conditions only with cau¬
tion. Consequently, it seemed to be desirable also to approach the
problem by measuring the quantity and vertical distribution in
lakes of a substance which has an important share in the photo¬
synthetic process, namely, chlorophyll.
The investigations of the author concerning the quantity and
vertical distribution of chlorophyll occurring in the phytoplank¬
ton of different lake types began in 1935 at the Wigry Hydrobio-
411
412 Wisconsin Academy of Sciences, Arts, and Letters
logical Station in Poland. When in 1937 the author received the
opportunity to work on the lakes of Northeastern Wisconsin, he
decided to continue there this chlorophyll investigation. A large
amount of general limnological data has been obtained for the
lakes of this very interesting region. This, together with the ex¬
cellent organization and technical equipment of the Trout Lake
Limnological Laboratory, enabled the author to collect during
a short time (July- August, 1937) the material for the present
paper.
Acknowledgements
My thanks are due Professor C. Juday, Director, for the
opportunity to carry on this investigation at the Trout Lake
Limnological Laboratory, as well as for the privilege of working
in his laboratory at the University of Wisconsin. Both published
and unpublished data regarding the various lakes have been gen¬
erously placed at my disposal, as well as the apparatus of the
Trout Lake Laboratory. The lively interest of Professor Juday
and Dr. E. A. Birge in these chlorophyll studies is greatly appre¬
ciated. I am also indebted to several members of the staff of the
Trout Lake Laboratory for assistance in collecting material and
for many other courtesies.
I am especially indebted to Dr. W. M. Manning for co-operat¬
ing with me in overcoming experimental difficulties and for con¬
tributing some helpful suggestions during my field and labora¬
tory work. He has also undertaken the language revision of this
paper.
My thankful acknowledgement is due to Prof. dr. J. Wlodek,
Cracow, Poland, who suggested to me the use of a photometric
method for chlorophyll determination and gave me other useful
technical advice.
A grant from the Polish National Culture Fund made possible
my investigations in the United States.
Method
Acetone extracts, which were made of plankton samples,
taken from different depths in the lakes of Northeastern Wiscon¬
sin, varied widely in quality of color as well as in intensity. For
example, in samples from lakes having an abundant plankton
crop, the color of the acetone extract of a surface sample was
Kozminski— Chlorophyll in Lakes
413
usually yellow or orange-yellow; in most eases extracts from
the plankton of deeper layers were successively yellow-green,
green, brown-green and finally— from near the bottom— brown.
For lakes having a low plankton production, extracts of surface
samples were yellow, and of deep-water samples yellow-green
or yellow-brown.
This color variation is an indication that the relative
bution of chlorophyll and the other acetone-soluble plai
ments (animal as well as plant) changes with changes
factors in the plankton environment. Because of this
it was not possible to determine chlorophyll quantitati
method, such as that described by Harvey (1934), in w
comparisons are made visually with prepared standarc
The rather difficult and time-consuming chemical s
of the chlorophyll pigments from the other acetone-so
ments was impractical because of the large number of
ations necessary in a limnological investigation of 1 ? f - i
Consequently, a photometric method was used, in which
of the light was confined to the wave-length 6200-6800
are absorbed by chlorophyll but which probably are n
ciably absorbed by any other pigments that may have been pres¬
ent. Similar methods have been used by other investigators
(Fleischer 1935; Godnew and Kalischewicz 1936).
There follows a description of the method used in this in¬
vestigation.
Water samples of 5 to 15 liters (18 liters in the case of very
low plankton content) were taken from different depths by using
a hand operated vacuum pump (actually a modified tire pump)
and a rubber hose, except for the surface samples which were
dipped up. The water was centrifuged in a Sharpies supercentri¬
fuge (a continuous flow type), which operated at about 25,006
revolutions per minute. Centrifuging a 10 liter sample required
about 30 minutes. The residue, consisting of phytoplankton and
other particulate matter, was carefully washed out of the cen¬
trifuge bowl with 98% acetone into a 250 ml. casserole. The
small water content of the residue did not reduce the acetone
concentration below 90%. A little CaCCh was added in order
to neutralize the organic acids which might otherwise accelerate
decomposition of the chlorophyll. The residue-acetone mixture
was carefully ground with a pestle in order to hasten complete
414 Wisconsin Academy of Sciences , Arts , and Letters
extraction of the chlorophyll. Complete extraction is not easy
in some cases and requires large amounts of acetone. It would
probably be better to grind the residue with addition of some
diatomaceous earth. The acetone extract was then filtered
through filter paper into a 25 or 50 ml. flask, the residue being
thoroughly washed with pure acetone. After diluting with ace¬
tone to exactly 25 or 50 ml., the light absorption of the extract
was measured. The entire procedure was always carried out
during a single day, care being taken to protect the chlorophyll
solutions from very intense light.
In preliminary work at the Wigry Hydrobiological Station,
ght absorption measurements were made with a Zeiss Pulfrich
hotometer using the accompanying red filter No. S 66.6/3.5.
the Trout Lake Laboratory a Cenco Photelometer (Central
itific Co., Chicago) was used for light absorption measure-
nts (see Sanford, Sheard and Osterberg, 1933). This instru¬
ment is equipped with a photronic cell, a diaphragm to regulate
the incident light intensity and a microammeter for recording
the current through the photronic cell. The microammeter
scale is arbitrarily divided into 100 units. For a light source,
this instrument employs a 50 c.p., 6 volt tungsten filament
lamp, the current being supplied from a constant voltage trans¬
former connected to the usual 110 volt A.C. circuit. Two red
filters transmitting only between 6200 A and 6800 A (Corning
Signal Red No. 243) were placed in series immediately in front
of the photronic cell. The absorption vessels supplied with the
photelometer were too thin (approximately 1 cm. optical path)
to give satisfactory deflections on the microammeter at the low
chlorophyll concentrations often encountered in this work. To
avoid the necessity of collecting and centrifuging unduly large
samples, a cylindrical glass cell was constructed with an optical
path of about 4 cm. and a volume of about 12 ml. A special
mounting was also constructed to permit the use of this cell in
the photelometer. When this cell was used instead of the stand¬
ard type, the sensitivity was increased by a factor of nearly 4.
In making a measurement with the photelometer, the light in¬
tensity was adjusted, with the help of the diaphram, to give a
microammeter deflection of 100 units when the absorption cell
was filled with pure acetone. The acetone was then replaced with
the chlorophyll solution and the resulting deflection noted.
Kozminski — Chlorophyll in Lakes
415
A calibration curve was prepared from absorption measure¬
ments made with a large number of known concentrations of
ethyl chlorophyllide. This compound was obtained from the
firm of R. Sandoz, Basel, Switzerland. With this curve, mi¬
croammeter readings could be expressed directly as milligrams of
ethyl chlorophyllide. Readings obtained with plankton extracts
were used to compute the ethyl chlorophyllide content in one
cubic meter of lake water. The ethyl chlorophyllide molecule is
smaller than that of phytyl chlorophyllide and, for a given weight
it is, therefore, more intense in color. The ratio of color intensity
for the two compounds is, according to Willstatter and Stoll
(1928, p. 74) , 50 :38. This ratio was used in the present paper in
computing the phytyl chlorophyllide content of the lakes. There¬
fore, all ethyl chlorophyllide values were multiplied by the coeffi¬
cient 1.316 (50/38) in order to obtain the corresponding phytyl
chlorophyllide values.
The shape of the calibration curve indicated that chlorophyll
concentractions giving deflections of between 60 and 90 on the
microammeter scale could be determined more accurately than
lower or higher concentrations. Consequently, 95% of the 142
chlorophyll determinations were made with concentrations in
this range. The remaining 5% fell within the range 58-60 or
90-92. In cases where higher concentrations were encountered,
the sample was diluted.
The sensitivity of the method just described is such that one
may determine chlorophyll amounts as small as 0.005 mg. Sev¬
eral experiments were carried out in order to determine the
magnitude of the experimental error. The error of the whole
procedure of centrifuging, washing, extracting and filtering
amounted to not more than 5 % for a chlorophyll content of about
0.01 mg. in the sample; it fell, moreover, to about 1% for a chlo¬
rophyll content of 0.1 mg. A larger error may occur in the pho-
telometer reading ; for a single reading the error may reach one
division of the microammeter scale but the accuracy may be con¬
siderably improved by repeating the determination several times
and using the average value thus obtained.
The total error of the method just described may vary from
10% at a chlorophyll concentration in lake water of about 1
mg/m3 to 4% at chlorophyll concentrations of 10 mg/m3 and
more.
416 Wisconsin Academy of Sciences, Arts , and Letters
The tables accompanying this report contain data not only on
chlorophyll concentration but also on several other important
limnological characteristics of the lakes investigated. These data
were collected by the staff of the Trout Lake Limnological Lab¬
oratory, using well known methods which need not be cited here.
It need only be said that the data concerning the organic content
in the centrifuge seston correspond to the difference in weight
between the dry and the ignited residue obtained by centrifuging
one liter of water. The relative chlorophyll content in the seston
was expressed in per cent of this loss on ignition. The possible
error of these relative values is considerably larger than the
error of the absolute chlorophyll values, because of the additional
uncertainty in determinations of loss on ignition.
Results
The results of these investigations are presented in tables
at the end of the paper. During the time from July 19 till August
27, 19 vertical series of determinations and 2 single determina¬
tions of chlorophyll concentration were carried out. These in¬
vestigations were accomplished on 17 different lakes situated in
the neighborhood of the Trout Lake Laboratory. Nearly all water
samples were taken in the morning, near the region of maximum
lake depth. The only exceptions were series 1 and 19 (Table I)
from Trout Lake which were taken at points where the lake was
only 20 m. deep.
O Mg/M3 50 100 150 200
Fig. 1. Examples of different types of vertical distribution of chlorophyll.
Lakes : C — Crystal, H — Helmet, Ma — Mary, Mu — Muskellunge,
Sc — Scaffold. The ordinate represents depth in meters.
Kozminski — Chlorophyll in Lakes
417
The smallest chlorophyll content, 1.0 mg/m3, was found at
1 m. in Trout Lake on July 21 ; the largest content, 386.2 mg/m,3
was found at 4 m. in Cardinal Bog on August 4.
As could be foreseen, there was in a large majority of eases a
distinct vertical differentiation in the chlorophyll distribution.
The results show five types of chlorophyll stratification in the
lakes that were studied (Fig. 1 and Table I).
I. The chlorophyll content at all depths is very low, with the
epilimnion extremely poor. There is a small increase in chloro¬
phyll content in the hypolimnion but no distinct maximum layer
—Crystal Lake.
II. The chlorophyll content at all depths, while somewhat
higher than for type I, is still relatively low. In the thermo-
cline there is a distinct though not very large maximum layer.
The epilimnion is richer in chlorophyll than the hypolimnion;
there is in the latter a gradual decrease of chlorophyll content
with increasing depth — Trout Lake (Fig. 2).
III. The epilimnion is rather poor in chlorophyll, with the
thermocline layer containing a somewhat higher concentration.
A further gradual increase in the hypolimnion is terminated with
a rapid increase in the deepest layer, forming a very distinct
chlorophyll maximum at or just above the bottom — Weber,
Silver, Little Rock, Muskellunge and Nebish lakes.
IV. Chlorophyll content in the epilimnion is rather high; it
decreases gradually with increasing depth, with a distinct chloro¬
phyll maximum at or near the surface of the lake — Big, Helmet
and Ike Walton lakes.
V. There is a relatively high chlorophyll content at all depths,
the epilimnion usually containing less than the other layers.
There is a very strong chlorophyll maximum in or just below the
thermocline. Below this maximum there is a layer of water
with distinctly less chlorophyll, while just above the bottom there
usually occurs a second chlorophyll maximum which may be
higher or lower than the first one. The vertical differentiation in
this type is very marked, with the contrast between adjacent
levels frequently large — Mary, Scaffold and Wildcat lakes.
418 Wisconsin Academy of Sciences , Arts, and Letters
Og/Mo/L 0 2 4 6 8
Fig. 2. Vertical stratification of chlorophyll (Ch), temperature (T) and
oxygen (02) in Trout Lake. Curve J represents chlorophyll data
taken July 26-27, 1937 ; curve A data taken August 23, 1837. The
ordinate represents depth in meters.
The chlorophyll series from Midge Lake and Cardinal Bog
can hardly be classed among these types. The chlorophyll curve
for Midge Lake is irregular with no distinct maximum (Table I) .
Cardinal Bog has an enormous chlorophyll content, far exceed¬
ing the content of the other lakes for which data were obtained.
However, this small lakelet seems to receive from outside a rela¬
tively large amount of fresh plant material (chiefly fragments
of Sphagnum mosses). Therefore, its chlorophyll content prob¬
ably cannot be directly compared with that of other lakes.
Kozminski—Chlorophyll in Lakes
419
There are, of course, some lakes which show an intermediate
type of stratification. Weber Lake, for instance, combines to a
certain degree the characteristics of types I and III ; Nebish and
Wildcat lakes could probably be classed among the lakes of either
type III or type V. However, the occurrence of these intermedi¬
ate lakes does not obliterate the general diversity in type of ver¬
tical chlorophyll distribution in the lakes investigated.
All the series were taken in mid-summer during a period of
generally warm and sunny weather. It is known, however, that
the phytoplankton may sometimes, within a few days, reach a
high degree of development or, on the contrary, it may show a
considerable decrease within a short time. It is not certain then
that the data in this paper give in all cases the most characteris¬
tic distribution of chlorophyll for a given lake. Only for Trout
Lake are the data sufficient to give some idea of the variation in
distribution during the summer.
A comparison of series 1 and 2 (Table I) shows that during a
few days in July the chlorophyll content in the epilimnion of
Trout Lake nearly doubled. One month later (series 19, Table I;
also Fig. 2) a further slight increase was observed. The chloro¬
phyll content in the hypolimnion, however, showed only a very
slight change during this month. The general type of distribu¬
tion did not change.
The problem of horizontal differentiation in chlorophyll dis¬
tribution was studied only on Trout Lake. Series 19 (Table I)
was taken at a place where the lake approaches its greatest
depth, namely 35 m. ; series 20, however, was taken the follow¬
ing day at a place where the lake is only 20 m. deep. The agree¬
ment between the two series is fairly good except for the deep¬
est point in series 20, which contains 5 mg/m3 of chlorophyll,
while at the same depth in series 19 there were only about 2
mg/m3 ; the deepest layer in series 19 contained only 1.4 mg/m3
(Table I).
This observation suggested that there may be a distinct dif¬
ferentiation in the chlorophyll content of the deepest water layer
of the lake. Consequently, on the following day further samples
were taken from the layer adjacent to the bottom at different
depths of Trout Lake. The results are shown in Table II and in
Fig. 3.
420 Wisconsin Academy of Sciences, Arts, and Letters
0 Chlo.,Mg/M* 2 3 4 5
Fig. 3. Chlorophyll content of the deepest water layer (1 m. above the
bottom) at different places in Trout Lake, August 23-25, 1937.
Ther., region of thermocline. The dotted line (Trp.) indicates the
depth of the Secchi disc reading. The ordinate represents depth
in meters.
In portions of the lake, where the bottom is above the thermo¬
cline, the chlorophyll content of the lowest layer of water is
low and in general corresponds to the content at the same depth
in places where the lake is deeper. This is doubtless due to the
mixing action of the wind which extends to the bottom in shallow
regions. Below the epilimnion, however, the chlorophyll content
of the bottom water appears to differ significantly from the con¬
tent of the same layer in deeper regions of the lake. As far as
one can judge on the basis of the available data, there is in Trout
Lake a distinct zone of higher chlorophyll content extending
along the bottom from the depth of the lower part of the thermo¬
cline to a depth of about 20 m. Below this zone the chlorophyll
content of the bottom layer is lower than that of adjacent layers
nearer the surface.
Kozminski — Chlorophyll in Lakes
421
Table I lists not only the absolute but also the relative chlo¬
rophyll content, computed as per cent of the organic content (loss
on ignition) of the seston of the same water layer. The variation
in these relative values is, of course, considerably smaller, for
the lakes investigated, than the variation in absolute values. The
extreme range is from 0.08% (Trout, 1 m., July 21) to 3.85%
(Mary, 4 m., August 13). However, the general character of
stratification in the lakes is similar when expressed either in
terms of absolute or relative chlorophyll content, except that
there are more irregularities in the figures for relative chloro¬
phyll content, probably due in part to the larger experimental
error.
Discussion
Since the present paper is believed to be the first report con¬
cerning the chlorophyll content of lakes and its vertical distribu¬
tion, the principal purpose has been to determine the general ap¬
pearance of these features and to create in this way a basis for
further, more detailed and more fundamental studies in this field
of limnological research. The data presented have been obtained
from lakes of very different and relatively very well known lim¬
nological characters ; except for very deep lakes and those with a
high calcium content, there are among the lakes investigated
many combinations of various limnological factors. Table III
shows some important characteristics of the 17 lakes that have
been investigated. Much of this information is from unpublished
records.
Light is certainly a very important factor governing the
chlorophyll occurrence and distribution in a lake. Thanks to
the comprehensive research of Birge and Juday (1929-1932) and
of Whitney (1938) the light conditions in the Northeastern Wis¬
consin lakes are relatively well-known.
Investigations concerning the rate of photosynthesis in lake
waters indicate that the maximum rate for this process (at
least in full sunlight) usually occurs not in the surface layer but
somewhere below the surface, the exact depth depending prin¬
cipally on the light transmission of the given water. On the
other hand it is fairly well established that plants living under
conditions of moderately low light intensity usually have a rela¬
tively higher chlorophyll content than plants living in full sun¬
light.
422 Wisconsin Academy of Sciences , Arts , and Letters
On the basis of these two facts alone it might be expected
that the chlorophyll content of the surface water of a lake would
be lower than at depths below the surface, and that the chloro¬
phyll maximum would occur in a layer somewhere between the
depth of maximum photosynthetic activity and the depth at
which the average photosynthetic rate is equivalent to the aver¬
age respiratory rate of the phytoplankton (compensation point).
Below the maximum layer a decrease in chlorophyll content
would be expected with complete disappearance below the com¬
pensation point. It would, of course, be expected that the addi¬
tion of various other factors actually encountered in a lake, such
as temperature and food conditions (external factors) and the
varying ecological requirements and responses of different spe¬
cies (internal factors), would greatly modify the results of this
simple hypothesis.
The chlorophyll stratification found in the lakes investigated
conforms only partially to these theoretical expectations. Par¬
ticularly striking is the occurrence of relatively large amounts
of chlorophyll in the deepest water layers; this is especially
marked for lakes of types III and V described above.
The position of the compensation point in the lakes investi¬
gated cannot be determined exactly. Previously published inves¬
tigations in this direction (Schomer and Juday, 1935; Curtis and
Juday, 1937) are based not on experiments with natural plankton
populations occurring in a given lake at a given depth, but on
experiments with algal cultures. Moreover, most of these mea¬
surements were made only during the period of maximum illum¬
ination, rather than on a 24-hour basis. Assuming, however,
that the position of the compensation point as found by the cited
writers corresponds approximately to its position under natural
conditions, it may be concluded that probably in all the lakes in¬
vestigated (except for four shallow lakes: Little Rock, Mann,
Starrett and Cardinal Bog) this compensation point is situated
above the greatest depth of the lake, usually far higher than this
depth. Consequently, chlorophyll sometimes occurs in water
layers which must be far below the compensation point ; in some
lakes it even reaches its maximum value in these deep layers.
Near the bottom of even the deepest lake investigated, 33 m. in
Trout Lake, there is still an appreciable chlorophyll content.
Kozminski — Chlorophyll in Lakes
423
This depth is probably situated about 20 m. below the compensa¬
tion point (compare Schomer and Juday, 1935, p. 183).
One possible explanation for this deep-water chlorophyll is
that it may have been produced for the most part in the upper
layers of water and that its presence in the deepest layers is due
to a continual sinking of dead, senescent or resting plant cells.
Decomposition of part of this chlorophyll would probably soon
commence, but with the low temperature and low oxygen con¬
tent usually occurring in these deep waters, the rate of decompo¬
sition may be low. Thus there may occur in these depths a partial
equilibrium between the arrival of chlorophyll material from
above and its decomposition.
This hypothesis gives rise to many questions which are at
present unanswerable. What is the rate at which phytoplankton
cells sink into the bottom layer ? What is the chlorophyll content
of these cells before and after sinking? With given thermal,
light, oxygen and pressure conditions, how rapidly does chloro¬
phyll decompose in the lower water layers? What effect have
the fauna, which may feed on the sinking plankton, and the
bacterial flora, which may attack the senescent and dead organ¬
isms, on the rate of disappearance of chlorophyll in the lake
depths ?
For several of the lakes investigated, the chlorophyll in
the lowest water layers must be at least temporarily inactive;
that is, not carrying on appreciable photosynthesis. These lakes
are the following : Trout, Btig, Ike Walton, Mary, Midge, Helmet
and probably also Wildcat and Scaffold.
The first of the lakes mentioned, in spite of its rather trans¬
parent water, is deep enough so that the light intensity in its
deepest layers would not be sufficient for appreciable photo¬
synthesis. At a depth of 26 m. Whitney (1938) found less than
0.001% of incident radiation; at a depth of 33 m., assuming an
unchanged transmission in the deeper water, the light cannot
amount to more than a few millionths of 1% of the surface in¬
tensity. The general character of the chlorophyll distribution in
Trout Lake (Table I and Fig. 2), especially the gradual decrease
below the thermocline and the low chlorophyll content of the bot¬
tom, support the supposition that this chlorophyll owes its pres¬
ence there to a slowly sinking “rain” of senescent or otherwise
photosynthetically inactive cells of phytoplankton.
424 Wisconsin Academy of Sciences, Arts, and Letters
In the remaining lakes mentioned above, the chlorophyll con¬
centration found in one or more of the upper layers is very con¬
siderable. Thus, the upper layers may easily serve as abundant
sources of chlorophyll for the lower layers. On the other hand,
the light intensity in the deepest water of these lakes is undoubt¬
edly extremely low. The disc readings in these lakes varied
from 0.75 m. to 3.8 m. (Table III). In Midge, the most transpar¬
ent of these lakes, Birge and Juday (1932, p. 550) found at a
depth of 9 m. only 0.0035% of the incident radiation. In Mary
there was only 0.035% at 5 m., and in Helmet only 0.045% at
3 m. below the surface (computed for zenith sun) . At the bot¬
toms of these lakes (Tables I, III, IV) there can be only traces
©f light.
Two experiments made in cooperation with Dr. W. M. Man¬
ning gave some information concerning the photosynthetic capac¬
ity of this deep-water phytoplankton. Photosynthetic and respi¬
ratory activity were measured for the plankton contained in
water samples from two lakes with a very sharp chlorophyll
stratification; i.e., from Wildcat and Scaffold lakes (Table I).
The samples were exposed (for method, see Manning, 1938) in
transparent and in black bottles at different depths of Trout
Lake. The surface plankton from Wildcat showed a higher pho¬
tosynthetic rate than the plankton from a depth of 7 m., though
the latter contained about 10 times as much chlorophyll as the
former. This result would indicate a much lower photosynthetic
capacity for the deep-water phytoplankton. On the other hand,
the plankton from 8.7 m. in Scaffold Lake showed a photosyn¬
thetic rate twice as large as for the surface water, while the
chlorophyll content at 8.7 meters in Scaffold was about 6 times
as large as in the surface water. It is certain that the light in¬
tensity at a depth of 8.7 m. in Scaffold was much smaller than
the intensity at 7 m. in Wildcat; the disc reading at the time
of taking the samples was 2.7 m. in Wildcat and only 0.75 m. in
Scaffold. Thus, in spite of the low light intensity, the deep¬
water plankton of Scaffold Lake had not entirely lost its photo¬
synthetic capacity.
The lakes which show chlorophyll stratifications of types I
and III are rather transparent and have clear, low-colored water,
at least in the upper layers. Except for Little Rock Lake which
is only 6 m. deep and showed a Secchi disc reading of 4.0-4.1 rn.,
Kozminski — Chlorophyll in Lakes
425
the disc readings for these lakes varied from 5.2 to 9.5 m. They
contained small amounts of chlorophyll in the upper strata, but
some of them showed a considerable increase in chlorophyll in
the lowest stratum of water. The evident lack of a source in the
upper strata from which the deep water layers could draw these
comparatively large amounts of chlorophyll and also the more
favorable light conditions in the depths of these lakes combine
to suggest that the greater part of this deep water chlorophyll
is probably autochthonic. This implies that a limited photosyn¬
thetic activity is carried on at the bottoms of all these lakes,
except perhaps Nebish Lake.
It is true that the compensation point in these lakes (except
for the shallow Little Rock) probably lies above their greatest
depths. In Crystal Lake it was found (under maximum light
conditions) by Schomer and Juday (1935, p. 179) at a depth of
17 m. ; in the remaining less transparent lakes the compensation
point should be higher. However, this does not exclude the possi¬
bility of a limited amount of photosynthesis regularly occurring
at greater depths (compare, for instance, Schomer and Juday,
1935, p. 179-180), provided that plants growing at these depths
can obtain part of their food saprophytically, that is, from dis¬
solved organic matter. The total effect of this photosynthetic
activity cannot be very large and probably is practically limited
to a few hours on sunny days. Low light intensity would prob¬
ably cause the plants living in these layers to develop high chloro¬
phyll concentrations, which would be favorable for absorbing
most of the light reaching them.
The light intensity in the lowest water layers of Crystal and
Weber lakes may sometimes exceed 1% (for zenith sun) of the
incident radiation (Birge and Juday, 1932, p. 539 and 545). In
both lakes three species of Bryophytes grow abundantly even in
the deepest places (Juday 1934). It is probable therefore that
phytoplankton organisms are also carrying on some photosynthe¬
sis there.
The Secchi disc reading was 6.2 m. in Nebish Lake, which in¬
dicated that the water in the epilimnion was quite clear and
transparent. In and below the thermocline, i.e., from 6-7 m.
depth, the light transmission decreased rapidly because of the
sudden increase in color and turbidity of the water (Whitney
1938a). For Nebish Lake the chlorophyll maximum occurred not
426 Wisconsin Academy of Sciences, Arts, and Letters
in the deepest water layers but at a depth of 10 m. (Table I).
This water layer may correspond to a stratum of minimum light
intensity below which there may be only photosynthetically in¬
active chlorophyll. Therefore, Nebish Lake was considered as
an example of a transitional lake between type III and type V
of chlorophyll distribution.
Muskellunge and Silver lakes are about 20 m. deep. Although
their upper water layers are also rather clear and transparent,
light conditions in the deepest layers are probably not very fav¬
orable. However, light intensity at the bottom of both lakes was
higher than for the lakes with chlorophyll stratification of types
II and V. At a depth of 17.5 m. in Muskellunge, Whitney (1938)
found 0.01% of the incident radiation while at a depth of 15 m.
in Silver there was 0.02% of the radiation computed for zenith
sun (Birge and Juday, 1932, p. 547). Other investigators have
recorded photosynthesis at intensities lower than these (Spoehr
and Smith; in Duggar, 1936, p. 1037-1038). Therefore, the light
intensities at the bottoms of Muskellunge and Silver Lakes are
probably sufficient to permit appreciable photosynthesis.
Despite the possibility of some photosynthesis in regions of
low light intensity, it is surprising that the chlorophyll maxima
in these lakes fall so often in the deepest water layers and not
somewhat higher where light conditions are certainly better.
Part of the explanation may lie in the sinking process postulated
earlier in the discussion, but another contributing factor may be
the better food conditions in the lowest water layers. It is prob¬
able that in water layers within one meter of the bottom the
content of dissolved inorganic salts and of carbon dioxide as
well as the content of organic matter, is considerably higher than
in layers situated nearer the surface; such an increase in these
substances in the bottom layers has been recorded frequently in
limnological investigatons. All these substances are probably
necessary for plant organisms which live both autotrophically
and saprophytically in these water layers.
Attention may again be called to the chlorophyll distribution
in the lowest water layers of Trout Lake (Fig. 3 and Table II).
In the regions of this lake where the depth varies from 10 to
20 m., the chlorophyll maximum occurs in the bottom layer, in a
manner similar to that found for lakes with a chlorophyll strati¬
fication of type III. Here again a sinking process may at least
Kozminski — Chlorophyll in Lakes
427
partially account for the observed behavior, but it is also possi¬
ble that in all these lakes this water layer furnishes optimal food
and minimal light conditions for photosynthetic activity. Hence,
lakes of type III should show maximum chlorophyll at the bot¬
tom because they are not deep enough to have water layers in
which light conditions fall below this minimum. According to
this viewpoint, the bottom layers below 20 m. in Trout Lake con¬
tain less chlorophyll than bottom layers between 10 and 20 m.
(possibly in spite of food conditions as favorable as in shallower
bottom layers) because of sub-minimal light intensities.
In considering Trout Lake, attention should be directed to
the simultaneous occurrence of the chlorophyll and oxygen
maxima in the thermocline. This coincidence appears to be a
constant and probably not accidental characteristic of Trout
Lake. However, in the other lakes which have an oxygen max¬
imum in the thermocline (e.g., in Silver Lake, compare Tables I
and III) there is not such a distinct coincidence as in Trout Lake.
Some conclusions of limnological value may probably be
drawn from these considerations. The limit between the tropho-
genic and tropholytic water layer is not sharp ; some production
of organic from inorganic matter is carried on below the compen¬
sation point. The effectiveness of this deep-water activity can¬
not be determined at the present time, but it should not be dis¬
regarded in future investigations.
On the other hand the presence of physiologically inactive
chlorophyll in some of the lakes investigated serves to warn
against attributing too much significance to single chlorophyll
determinations. Such determinations cannot give any exact mea¬
sure either of the mass of phytoplankton or of the intensity of
photosynthesis occurring at a given depth in a lake.
More direct significance can probably be attached to calcu¬
lated values of the mean chlorophyll concentration in a lake.
Table V shows for each lake the mean chlorophyll concentration
in a water column extending from the surface to the bottom at
the point of maximum depth. These mean values, expressed in
mg/m3 and in per cent of loss on ignition, were computed from
the vertical distribution data.
The increase in mean chlorophyll value in different lakes
seems to parallel approximately the increase in productivity of
these lakes. The first of the lakes, Crystal, is known to be the
428 Wisconsin Academy of Sciences, Arts, and Letters
most oligotrophic lake of the region (Table IV) and it has also
the lowest mean chlorophyll content. Trout Lake is considered
to be less oligotrophic, perhaps intermediate between the oligo¬
trophic and eutrophic lakes (Juday and Birge, 1932, p. 449) ;
its greater depth, however, causes large parts of the hypolimnion
to be unfavorable for organic production ; this fact is reflected in
a low mean chlorophyll concentration. Weber Lake was some
years ago nearly as oligotrophic as Crystal Lake ; fertilization ex¬
periments in recent years have materially increased its produc¬
tivity, and it is therefore not surprising that the mean chloro¬
phyll content of Weber is nearly twice as large as that for
Crystal Lake.
For the lakes with a higher mean chlorophyll content, it is not
possible to draw a simple parallel between this factor and the
other limnological characteristics shown in Table IV. It is prob¬
able, however, that such a parallel may be brought out in the
future. The highest mean chlorophyll content is shown by Card¬
inal Bog which — as mentioned above — probably receives from
outside relatively large amounts of fresh plant material. Next
highest is Scaffold Lake, the only lake investigated in which
there was a distinct and strong water-bloom.
The mean chlorophyll values for different lakes would cer¬
tainly be more valuable and better comparable if they could be
computed only on the basis of chlorophyll which is actually ac¬
tive in photosynthesis. The rate of disappearance in deep water
of inactive chlorophyll is probably very different in different
lakes; it must depend on many external and internal factors
whose effects cannot easily be quantitatively estimated. There¬
fore, the actual amounts of this inactive chlorophyll in different
lakes may be often not comparable.
On the other hand, we shall probably soon be able to deter¬
mine — at least approximately — the depth to which photosynthetic
activity of the phytoplankton extends. Many advances have
been made recently in our knowledge concerning light conditions
in lakes, and light efficiency for photosynthesis of different kinds
of water plants in their natural environment.
The mean chlorophyll content found above the depth of this
photosynthesis limit will probably represent a more characteris¬
tic and significant measure than any other used up to the present
time in calculations of lake productivity. This will be especially
Kozminski — Chlorophyll in Lakes
429
true, when chlorophyll investigations will be carried on during
the whole year, permitting a knowledge of chlorophyll production
of a lake per unit of time and per unit of volume.
Summary
1. A photometric method is described for the determination of
chlorophyll in centrifuge plankton.
2. The vertical distribution of chlorophyll concentration in 17
lakes of Northeastern Wisconsin was investigated during
July and August, 1937.
3. The lowest observed chlorophyll concentration amounted to
1.0 mg/m3, the highest to 386.2 mg/m3.
4. A large variation was found in the vertical distribution of
chlorophyll in different lakes. Five different types of stratifi¬
cation may be distinguished.
5. Some data are given for Trout Lake concerning changes in
chlorophyll content during the summer season of 1937.
6. The horizontal distribution of chlorophyll in Trout Lake
was investigated. Distinct concentration differences were
found in the water layers situated immediately above the
bottom, depending on the depth of the lake at the point in¬
vestigated.
7. The influence of light conditions on the vertical chlorophyll
distribution is discussed.
8. It is concluded that the relatively large amounts of chloro¬
phyll, found in the deepest water layers of some lakes, must
be inactive photosynthetically.
9. Possible causes are considered for the different types of
chlorophyll stratification.
10. The average chlorophyll concentration has been calculated
for different lakes. The limnological application of these val¬
ues is discussed.
Literature
Birge, E. A. and C. Juday. 1929-32. Transmission of solar radiation by
the waters of inland lakes. Trans. Wis. Acad. Sci., Arts & Let. 24;
509-580. Second report. 1930. Ibid. 25: 285-335. Third report.
1931. Ibid. 26: 383-425. Fourth report. 1932. Ibid. 27: 523-562.
Curtis, J. T. and C. Juday. 1937. Photosynthesis of algae in Wisconsin lakes.
III. Observations of 1935. Internal. Rev. Hydrobiol. u. Hydrog. 35:
122-133.
430 Wisconsin Academy of Sciences, Arts, and Letters
Duggar, B. M. 1936. Biological effects of radiation. Vol. II. Me Graw-Hill
New York.
Fleischer, W. E. 1935. The relation between chlorophyll content and rate
of photosynthesis. Jour. Gen. Physiol. 18 (4): 573-597.
Godnew, T. N. and S. W. Kalischewicz. 1936. Die quantitative Bestimmung
des Chlorophylls vermittels des lichtelektrischen Kolorimeters von
Lange. Planta 25: 194-196.
Harvey, H. W. 1933-34. Measurement of phytoplankton population. Jour.
Mar. Biol. Assoc. U. K. 19: 761-773.
Juday, C. and E. A. Birge. 1932. Dissolved oxygen and oxygen consumed
in the lake waters of northeastern Wisconsin. Trans. Wis. Acad. Sci.,
Arts & Let. 27: 415-486.
Juday, C. 1934. The depth distribution of some aquatic plants. Ecology
15: 325.
Manning, W. M., C. Juday and M. Wolf. 1938. Photosynthesis of aquatic
plants at different depths in Trout Lake, Wisconsin. Trans. Wis. Acad.
Sci., Arts & Let. 31 : 377-410.
Sanford, A. H. C. Sheard and A. E. Osterberg. 1933. The photelometer and
its use in the clinical laboratory. Amer. Journ. Clin. Pathol. 3: 405-420.
Schomer, H. A. and C. Juday. 1935. Photosynthesis of algae at different
depths in some lakes of northeastern Wisconsin. I. Observations of
1933. Trans. Wis. Acad. Sci., Arts & Let. 29: 173-193.
Whitney L V. 1938. Transmission of solar energy and the scattering pro¬
duced by suspensoids in lake waters. Trans. Wis. Acad. Sci., Arts & Let.
31: 201-221.
Whitney, L. V. 1938a. Microstratification of inland lakes. Trans. Wis. Acad.
Sci., Arts & Let. 31 : 155-173.
Willstatter, R. and A. Stoll. 1928. Investigations on chlorophyll. English
Translation. Science Press, Lancaster.
Kozminski — Chlorophyll in Lakes
431
TABLE I
The amount and vertical distribution of the chlorophyll in 17 lakes.
The deepest sample for each lake was taken about 1 m. above the bottom ,
except for series 1 , U and 5. The phybyl chlorophyllide values in mg /m3
are computed from the values for ethyl chlorophyllide by multiplication of
the latter by the coefficient 1.316. The relative values for phytyl chlorophyl¬
lide are computed from absolute values as the per cent of loss on ignition.
432 Wisconsin Academy of Sciences , Arts, and Letters
Kozminski — Chlorophyll in Lakes
433
TABLE II
Chlorophyll content of Trout Lake water one meter above the bottom
at different places in the lake.
434 Wisconsin Academy of Sciences , Arts , and Letters
TABLE III
Some limnological characteristics of the lakes investigated (data taken
at the same time as the data of Table I). Transparency means the Secchi
disc readings; color is expressed in the units of the platinum-cobalt scale;
it was determined on centrifuged water.
Kozminski — Chlorophyll in Lakes
435
436 Wisconsin Academy of Sciences , Arts , and Letters
TABLE IV
Some morphometrical, physical and chemical data concerning the 17 lakes investigated. When two numbers are given they represent the
ge of variation observed either at different times or at different depths or both.
438 Wisconsin Academy of Sciences , Arts , and Letters
TABLE V
Mean chlorophyll concentration in a water column from the surface
to the bottom in different lakes.
A STUDY OF THE FISH PARASITE RELATIONSHIPS
IN THE TROUT LAKE REGION OF WISCONSIN*
Samuel X. Cross
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 76.
Introduction
It is a commonly accepted fact that large numbers of para¬
sites of any kind are harmful, in one way or another, to their
host. It seems to make little difference what type of parasite
one may consider or to what division of the animal kingdom the
host may belong, the result of large numbers of parasites is
usually a very obvious handicap to the animal harboring them.
The results of heavy parasitism are most clearly seen in young,
growing animals.
While considerable work has been done on the effects of para¬
sites on the growth and development of various animals, very
little work of this character has been done on fishes. Quantita¬
tive studies on this subject are particularly lacking for fish.
Various phases of the problem of fish parasites have been
studied by a number of workers in this country and elsewhere,
but little has been said in any of the literature regarding the
relationship of the parasite infections to the growth of the fish.
Ward (1911), speaking of fish, stated that “there are no gen¬
eral diseases produced by internal parasitic worms unless gen¬
eral weakness, loss of flesh, and power of growth and reproduc¬
tion be considered as such.” Hubbs (1927) has given figures
to show that minnows ( Platygobio gracilis) which were heavily
parasitized in their first year retained many of their infantile
features and were, on the average, shorter than their non-para-
sitized companions of the same age. He was unable to show any
differences between normal and heavily infested fish after the
first year. Reighard (1928) suggested that the variation in
weight of bass of the same length and age might be related to
* This investigation was supported by a grant from the Wisconsin Alumni Research Foundation.
439
440 Wisconsin Academy of Sciences, Arts, and Letters
the varying degrees of parasitic infection occurring in these
fish.
The fish parasite study presented here was carried on during
the summers of 1931 and 1932 as part of the research program
of the Trout Lake Limnological Laboratory, which was estab¬
lished in 1925 for the purpose of studying the physics, the chem¬
istry and the biology of the lakes of that region. A general study
of the fish life of certain types of these lakes was begun in 1928.
The chief problem in these studies at that time was the rate of
growth of the fish and the factors which affect it. This natu¬
rally raised the question of the number and kinds of internal
parasites harbored by the fish and the effect of these parasites
on their growth; the effect of heavy and light infestation was
determined for the perch in three different lakes and these re¬
sults have been published (Cross 1935).
In the summer of 1930, Dr. C. A. Herrick of the University
of Wisconsin made a preliminary study of the fish parasite con¬
ditions in several lakes of the district. His results showed that
there was a wide variation in the extent of the parasitism in the
different lakes, with a very low incidence in some and a rather
high incidence in others. In this preliminary investigation, no
attempt was made to correlate the number and kinds of para¬
sites with the rate of growth of the fish.
With this as a background, the present study was begun in
the summer of 1931 and was continued in 1932. It was made as
nearly quantitative as possible in view of the large amount of
material which was collected. Each host was considered sep¬
arately with regard to the number of each species of parasite
harbored. Such parasites as the larvae of the cestode Triaeno-
phorus nodulosus and the myxosporidia form cysts of varying
sizes in the viscera and other parts of the body; in these cases
note was taken of the relative mass of the parasite tissue pres¬
ent.
Material and Methods
During the two summers more than 4400 fish from 14 differ¬
ent lakes were examined for parasites. Internal, flesh and ex¬
ternal parasites were counted when possible and specimens of
the various species were preserved for further study. Anglers
are particularly interested in the flesh parasites of fish; nearly
Cross — Fish Parasites
441
all of these forms are encysted larval trematodes which pass
their adult stage in fish eating birds and a study of the para¬
sites of 64 of these birds was made in 1932. The purpose of
this investigation was to determine the importance of these birds
as disseminators of the flesh parasites which often make the fish
undesirable for human food.
The fish used in this investigation were taken with gill nets
and by hook and line. The latter method was used chiefly in
lakes that were not fished with nets. The fish were measured,
weighed and scale samples were taken for growth studies before
they were examined for parasites. After the ages were deter¬
mined, the relative abundance of parasites could be correlated
with rate of growth.
The exterior of the fish, the mouth, the gills, the skin, the
flesh and the entire digestive tract were examined for parasites.
All parasites retained for future study were either preserved
directly with 70 per cent alcohol, or they were placed in a killing
solution and then washed and preserved in 70 per cent alcohol.
Mercuric chloride containing 5 per cent glacial acetic acid was
found to be the best killing agent for cestodes and trematodes,
while a modified form of Hetherington’s solution proved best
for nematodes and acanthocephala.
Distribution of Parasites
The lakes of this district vary widely in physical, chemical
and biological characteristics and these factors appear to have
some effect upon the distribution of the parasites. These bodies
of water belong to two general types, namely seepage lakes or
those which have neither an inlet nor an outlet, and drainage
lakes or those which have an outlet. The former usually have
very soft, acid waters which constitute a rather unfavorable
environment for the aquatic Mollusca ; as a result the mollusk
population of the seepage lakes is generally sparse and this tends
to restrict the parasites that are dependent upon them as hosts.
Most of the drainage lakes, on the other hand, have moderate
amounts of carbonates and neutral or alkaline waters, which
afford a more favorable environment for the mollusks. Table I
is a record of the various fish parasites found in the investiga¬
tion, together with the hosts and the lakes in which they were
found.
442 Wisconsin Academy of Sciences , Arts, and Letters
Table I
Host catalogue of the fish parasites of the Trout Lake region.
ACANTHOCEPHALA
Cross-Fish Parasites
443
Perch. No adult cestodes of any kind were taken from yel¬
low perch ( Perea flavescens) . Cestode cysts in the liver or other
parts of the viscera were not uncommon. The majority of the
perch from Silver and Trout lakes harbored cysts of the two
cestodes Triaenophorus nodulosus and T. robustus in large num¬
bers, as shown in Table II.
The most common parasites found in the perch, in addition
to cestode parasites in the flesh, were larval trematodes encysted
throughout the body and adult nematodes in the intestine. Some
specimens harbored so many encysted strigeids in the skin and
flesh that, over large areas, the black pigmented cysts often
touched each other. Many of these same specimens had large
numbers of unencysted strigeid larvae in the vitreous humor of
the eyes.
Adult intestinal trematodes were rather uncommon in perch
in all of the lakes. In 1931, 12 per cent of the perch of Nebish
Lake harbored varying numbers of a trematode, Cryptogonimus
sp., a member of the family Heterophyidae. The only other
adult trematodes in perch were found in two specimens from
Muskellunge Lake and in those from Lake Vieux Desert. These
trematodes belonged to the species Azygia angusticauda (Staff).
This seems to be a new host record for this species which wa3
reported in 1926 for Wisconsin lakes from Micropterus dolo-
mieu, Esox Indus and Amia calva.
A few perch in Muskellunge Lake, in 1931 and 1932, were
infested with the acanthocephalan E chinorhynchus thecatus,
Table II
Percentage of parasite infestation in perch.
444 Wisconsin Academy of Sciences , Arts , and Letters
which also was found in this lake in large numbers in small¬
mouthed black bass and in rock bass.
Rock bass. The rock bass ( Ambloplites rupestris) from Ne-
bish, Muskellunge and Silver lakes, and from all other sources
were heavily infested with encysted strigeid larvae in the flesh
and skin (Table III). Most of these fish harbored some Neascus
van-cleavei encysted in the liver, heart and kidney tissues. Some
of the rock bass from Muskellunge and Nebish lakes were so
heavily infested with the cysts of this species that the liver,
heart and kidney tissues were largely supplanted by cysts.
The most numerous parasites in the intestines of the rock
bass were the Heterophid trematodes of the genus Crypto goni-
mus, similar to those found in perch. These small trematodes
were often so numerous that they gave the intestinal contents
the appearance of being made up of about 50 per cent black pep¬
per.
Table III
Percentage of parasite infestation in rock bass.
Most of the nematodes from the intestines of rock bass were
large ascarids, probably Ascaris lucii (Pearse). It was impos¬
sible to be certain of the identity of this parasite from the origi¬
nal description. These hardy worms were often found alive in
fish that had been kept in the refrigerator for two days after
their death. Frequently these worms migrated from the intes¬
tine into the stomach or mouth of the host if the latter had been
dead only a few hours.
Cross — Fish Parasites
445
Cysts of the cestode Triaenophorus robustus were found in
the wall of the stomachs of a few rock bass from most of the
lakes that were studied. As in the perch, no adult cestodes were
found in the rock bass, with the exception of one specimen that
contained a small Proteocephalus. This specimen also contained
fish remains and the tapeworm, which was found dead in the
stomach, was probably taken with the fish that was eaten.
About 98 per cent of the rock bass and a few of the perch
of Nebish Lake were afflicted with a fungous disease.
Bluegills. In the bulegills ( Helioperca macrochira ), speci¬
mens were examined only from Muskellunge Lake. The most
striking infestation was that of Neascus van-cleavei. Many
specimens were found in this lake in which it was difficult to
see how it was possible for the heart and kidneys to function
when so much of the host tissue had been replaced by parasite
cysts. Many of the hearts and kidneys, at first glance, appeared
to be tightly packed masses of cysts with no original host tissue
remaining.
Small-mouthed black bass. Table IV shows that 76 to 100
per cent of the small-mouthed black bass ( Micropterus dolo-
mieu) from Nebish, Muskellunge and Silver lakes were infested
with trematode cysts. All of these infestations were strigeid
cysts, with the exception of two specimens from Muskellunge
Lake that harbored, in addition, one Clinostomum marginatum
cyst each.
The trematodes found in the intestines of these bass all be¬
longed to the genus Crypto gonimus, family Heterophyidae,
Table IV
Percentage of parasite infestation in the small-mouthed black bass.
446 Wisconsin Academy of Sciences , Arts , and Letters
which also occurred in the perch and rock bass. These trem-
atodes often occurred in very large numbers.
In 1932 the small-mouthed bass in Muskellunge Lake showed
a rather high incidence of Triaenophorus noduiosus, but no in¬
festation of this kind was found in 1931. One bass with such
an infestation was taken in Silver Lake in 1931.
Adult cestodes ( Proteocephalus ambloplites) were very com¬
mon in the intestine of the small-mouth both in 1931 and 1932,
but they were never as numerous as the tapeworms in the intes¬
tines of the wall-eyed pike and the ciscoes. Few specimens were
seen in which mechanical injury had been done to the viscera, as
reported by other authors. A few specimens from Muskellunge
Lake showed numerous adhesions, but none was found with the
parasite active in the viscera.
Nematodes do not play an important role among the para¬
sites of the small-mouthed black bass in the lakes that were
studied. Few specimens were found and these were very small.
Acanthocephala (E chinorhynchus thecatus) were very com¬
mon in the small-mouth and were often found in very large
numbers in the anterior part of the intestine and in the gastric
caeca.
Wall-eyed pike. Wall-eyed pike ( Stizostedion vitreum) from
all sources were heavily infested with tapeworms ( Abothrium
eras sum) . In many specimens these parasites nearly occluded
the lumen of the intestine and crowded into the gastric caeca in
large numbers. Frequently where the infections were light, the
distal ends of the caeca were distended with putrid food mate¬
rial that seemed to have been held there by the parasites. (See
Table V).
Table V
Percentage of parasite infestation in wall-eyed pike.
Cross — Fish Parasites
447
Acanthocephala (Neoechinorhynchus tenellus) were found in
one specimen of wall-eyed pike from Trout Lake in 1931 and in
one specimen from the Manitowish River, also taken in 1931.
The strigeid cysts that were found in such abundance in some
species of fish in this region were rather scarce in the wall-eyed
pike, with the exception of those taken in Clear Lake in 1932.
The infected fish from Clear Lake harbored a large number of
these cysts, which raises the question as to why the wall-eyes
were not more frequently infected in lakes where nearly 100 per
cent of other species of fish were infested with this parasite.
Ciscoes. Table VI shows that adult cestodes were the prin¬
cipal parasites in the intestine of the cisco ( Leucichthys artedi) .
With the exception of the Clear Lake ciscoes which were virtu¬
ally uninfested, 63 to 100 per cent of the other specimens car¬
ried cestodes. Proteocephalus exiguus was the most common ces-
tode in the ciscoes of all lakes that were investigated.
Acanthocephala were very common in the 1931 ciscoes of
Silver Lake and some specimens from Trout Lake also harbored
large numbers of them. No acanthocephala were found in the
Trout Lake ciscoes in 1932 and no specimens from Silver Lake
were examined in 1932. None of the ciscoes harbored strigeid
cysts in the flesh or skin.
Table VI
Percentage of parasite infestation in ciscoes
Parasitic copepods ( Ergasilus confusus) were common para¬
sites on the gills of Trout Lake ciscoes and a few were found on
the gills of Silver Lake ciscoes also.
Common sucker. The common suckers ( Catostomus commer-
sonnii) from Trout and Muskellunge lakes were very heavily
448 Wisconsin Academy of Sciences, Arts, and Letters
infested with the cysts of Triaenophorus nodulosus in the vis¬
cera. The infestations in many specimens were so heavy that
the entire visceral mass was tightly bound together by adhesions
around these cysts. A number of suckers from Lost Lake did
not have any of these cysts.
Acanthocephala ( Pomphorhynchus bulbocolli, Neoechinorhy-
chus crassus and Octospinifer macilentus) were very numerous
in suckers taken from Trout Lake. The relative numbers of
these parasites in an individual host were generally in the order
named ; usually P . bulbocolli outnumbered the other two species
combined. (See Table VII) .
Several very large specimens of Ligula intestinalis were ob¬
tained from Muskellunge Lake suckers in 1932. The largest of
these specimens was 52 cm. in length when alive. Only five
suckers contained these large Ligula and these were the only
specimens of this parasite taken from suckers in either 1931 or
1932.
Filometra were fairly common parasites of suckers in Trout
Lake in 1932 but no specimens of it were found in 1931.
With the exception of the ciscoes, the suckers of the region
had a smaller number of strigeid cysts than any other species of
fish examined in these studies.
Table VII
Percentage of parasite infestation in suckers.
The suckers were by far the most heavily infested with the
parasitic copepod Argulus of all the species of fish in the Trout
Lake region. These so-called fish lice were very common on the
suckers of Muskellunge Lake.
Cross — Fish Parasites
449
Discussion
Strigeid larvae were very common parasites in the eyes of
perch and rock bass of this region and were found in all species
of fish that were examined. These larval trematodes were often
so numerous that they could be seen in large numbers before the
eye was opened. The vision of the fish must be considerably
impaired by the movements of 50 to several hundred of these
active little worms in the vitreous humor of the eyes.
Weber Lake was unique in that practically no parasites were
found in the perch and small-mouthed black bass taken in it. It
is a seepage lake with very soft, transparent water ; its area is
15.6 ha. (38 a.) and its maximum depth 13.5 m. (44 ft.). The
perch of this lake reached a larger size and a greater age than
most of those taken from neighboring lakes.
Clear Lake is also a seepage lake with soft, transparent wa¬
ter; it has an area of 349 ha. (863 a.) and a maximum depth of
29 m. (95 ft.). The ciscoes of this lake were virtually unpara¬
sitized. A few specimens carried very light infestations of M'yxo-
sporidia. One cisco taken in this lake in 1931 contained a Tria -
enophorus nodulosus cyst and one taken in 1932 contained one
specimen of acanthocephala. The ciscoes obtained from Clear
Lake were very much larger than those taken from any other
lake of the region. The freedom of the ciscoes of Clear Lake
from internal parasites was particularly interesting because the
wall-eyed pike and the rock bass of this lake were as heavily in¬
fested with them as in any of the other lakes that were investi¬
gated.
Another interesting series of non-parasitized fish was noted
in 1931. Some 60 specimens of the common sucker were ob¬
tained from Lost Lake in that year; they proved to be entirely
negative for all parasites except a few Arguli. No suckers were
obtained from this lake in 1932, but other species of fish were
taken and every specimen, regardless of species, harbored para¬
sites of some sort. These two examples of non-parasitized fish,
the ciscoes of Clear Lake and the suckers of Lost Lake, are par¬
ticularly interesting because the specimens of both of these spe¬
cies of fish taken from all other lakes in the region were very
heavily infested with parasites. In both lakes also, all of the
other species of fish were infested with various kinds of para¬
sites,
450 Wisconsin Academy of Sciences , Arts, and Letters
Effect of Parasitism on Growth
An interesting and important relation between parasitism
and the rate of growth of the perch was found in these studies.
The results obtained in three lakes were embodied in a recent re¬
port (Cross 1935) and the data on which the report is based are
given in Table VIII. The problem of determining the relative
intensity of parasite infestation proved to be a difficult one be¬
cause several species of parasites which differ greatly in size
and in relative abundance were involved. A rather arbitrary
system was finally adopted which was based on the general size
and relative abundance of the different kinds of parasites pres¬
ent, as well as on the evident pathological conditions produced
by them as seen in a macroscopic examination. In preparing
the data for Table VIII therefore, an attempt was made to group
Table VIII
Size variation in perch with different parasite burdens.
Silver Lake
Muskellugne Lake
Nebish Lake
Cross — Fish Parasites
451
those perch together that were seemingly about equally handi¬
capped by the parasites they harbored when examined.
The perch were divided into three groups. Those with few
or no parasites were classed as “light infestations” ; those in¬
cluded in “heavy infestations” had large numbers of parasites
either in the body cavity, the intestine or the flesh, or in all three
of these locations combined, while those with intermediate num¬
bers were classed as “medium infestations”. The results showed
that the effect of the parasites on the growth of the perch was
independent of the location of the infestation. Trematode or
cestode cysts in the flesh, when numerous, had as much effect as
adult parasites in the intestine.
The figures given for the ages of the perch in Table VIII
represent the number of annular rings on the scales; thus age
4 represents four of these rings. It means that the perch had
lived through four winters and was in its fifth summer when it
was caught.
Regarding the three lakes represented in the table, the fishing
records indicated that the fish population was most crowded,
both for food and space, in Muskellunge Lake and least in Ne-
bish Lake, with approximately the same population density in
Silver Lake as in Muskellunge Lake.
The data for Nebish Lake, the most lightly populated of the
three lakes, show two important facts, namely (1) the perch
were larger for their age in Nebish than those in the other two
lakes and (2) generally there was less difference between the
heavily and lightly infested perch in this lake than in either of
the other two. The infestations were just as heavy in Nebish as
in the other two lakes, but the differences were probably due to
the relatively smaller population and greater abundance of food
in the former.
The last column in Table VIII shows the percentage of dif¬
ference in length and in weight of the average perch with light
infestations over the average heavily infested ones. The great¬
est difference in length and weight was found in the four year
old perch from Nebish Lake, where the lightly infested fish were
27 per cent longer and 120 per cent heavier than their heavily
infested fellows of the same age. The smallest difference was
also noted in Nebish Lake; in the three year old group for 1931,
the difference in length was only 11 per cent and in weight 39
452 Wisconsin Academy of Sciences, Arts, and Letters
per cent. In the three lakes represented in the table, the perch
with light parasite infestations were more than 10 per cent
longer and more than one-third heavier than those with heavy
infestations. With the exception of the Nebish Lake perch taken
in 1931, the older specimens showed more difference in size be¬
tween heavy and light infestations than those one year younger.
In the four year old perch taken in Silver Lake in 1931 for ex¬
ample, there was a difference of 12 per cent in length and 43 per
cent in weight, while the five year olds showed differences of 18
per cent in length and 94 per cent in weight.
An explanation of this increasing difference between lightly
and heavily parasitized perch with increasing age is not evident
from these data. Slow growing individuals are found in the
various species of fish and the present results suggest the pos¬
sibility that fish that become heavily infested with internal para¬
sites in early life are more or less handicapped throughout their
lives, and thus show a slower growth. In the older fish on the
other hand, no differences of this character that could be corre¬
lated with parasitic infestations were found in these studies.
Most of the perch that were 7 years of age or older had very
slight infestations, if any at all. This suggests that the light
infestations of the older perch may be due to an age immunity.
Changes in food habits with advancing age may also be involved
in the problem.
Immunity between Parasites
Several investigators have noted that the presence of large
numbers of one species of parasites has a tendency to limit in¬
festation by other species. Some results bearing on this prob¬
lem were obtained from 92 ciscoes taken in Silver Lake in 1933.
In this group of ciscoes, no parasites were found other than
adult cestodes of the species Proteocephalus exiguus and acantho-
cephala of the genus Neoicanthorinchus. In the series of 92 cis¬
coes, the average number of tapeworms harbored was 16.9 and
the average number of acanthocephala was 8.3 ; 84 of them had
tapeworms, 73 harbored acanthocephala and all of them had one
or both parasites. (Cross 1934) .
The degree of infestation is shown in Table IX. Those with
15 or more acanthocephala had a very limited infestation of
tapeworms, and conversely those with more than 25 tapeworms
Cross— Fish Parasites
453
had few or no acanthocephala. On the other hand, the ciscoes
with fewer parasites often had them quite evenly divided be¬
tween the two types, as shown in items 4 and 5 in the table.
The first item in the table shows a high tapeworm infestation
with a low acanthocephala incidence, while item 2 shows the
converse, namely a high acanthocephala infestation with a low
tapeworm incidence.
Table IX
Average of parasite infestations in arbitrary groups of Silver Lake ciscoes .
That this phenomenon is not due to crowding of the para¬
sites in the digestive tract is shown by the fact that the two
forms occupy different parts of the tract ; the tapeworms prefer
the gastric caeca as points of attachment while the acanthoce¬
phala attach themselves near the middle of the intestine. This
seems to indicate that there is a non-specific immunity which
limits either tapeworms or acanthocephala when one of these
parasites is present in large numbers.
Fish Eating Birds
In 1932, 64 fish eating birds were collected in order to de¬
termine their part in the fish parasite problem.
Of the 13 Merganser ducks examined, only one was negative
for fish parasites and this bird was taken from Nebish Lake in
October. Twelve of these ducks contained adult Ligula in the
intestine and three were heavily infested with adult strigeids.
Ligula intestinalis was usually found in young ducks, those that
were not yet fully feathered. Two adults contained three Ligula
each. In the young birds the infestation was often heavy enough
to distend the intestine; one young specimen contained 108
adult worms.
454 Wisconsin Academy of Sciences, Arts, and Letters
One young osprey was taken, but it had no parasites.
Two loons were examined and one of them contained about
200 adult strigeids.
Ten belted kingfishers were taken and all of them were heav¬
ily infested with adult strigeids; one specimen harbored 250 of
these worms. Opisthorcis was a common liver parasite of these
birds; this form may or may not be an important fish para¬
site, but no cysts were seen in any of the fish examined.
Seven great blue herons were examined, all of which were
heavily parasitized with adults of both Clinostomum margina¬
tum and strigeids.
Thirty-one crows were studied but no parasites traceable to
fish were found. These birds were examined because they were
seen feeding on the carcasses of dead fish that had drifted
ashore.
These data seem to show that most of the fish eating birds
harbor strigeids of one or another species and are, therefore,
potential spreaders of these small parasites that heavily infest
many of the fish. The Ligula found in the Merganser ducks has
little direct importance for the game fish of this region; their
larvae were confined to the small minnows with the exception
of 5 suckers from Muskellunge Lake. The Ligula infestation of
the Notropis of Trout Lake may be a very definite check on the
productivity of these minnows which are a welcome addition to
the diet of the game fish.
The great blue heron was the only fish eating bird found to
harbor the adults of Clinostomum marginatum. There may be
other definitive hosts among the herons or bitterns that have not
been recorded.
There has been some agitation for an open hunting season on
the great blue heron for the purpose of checking the fish para¬
sites that they carry. The value of the reduction of the heron
population for this purpose is very doubtful. The problem of
the infestation of game fish and pan fish with C. marginatum ,
which is carried only by the heron, is a very small one in com¬
parison with that of the strigeid parasites which are carried by
all of the fish eating birds represented in this study, except the
young osprey that was negative for all kinds of parasites.
Cross — Fish Parasites
455
Summary
The degree of parasitism varies a great deal in the various
lakes as well as in the various species of fish in any one lake.
There is also an annual variation.
Fish that are very heavily parasitized are retarded in growth.
This retardation of growth seems to be greater when the weight
of the fish is considered than when the length alone is used as a
measure of growth. This retardation is so great in some cases
that the parasites present at the time the fish is examined can
hardly account for the difference unless it is assumed that the
individual has been heavily parasitized during most of its life.
Fish that harbor large numbers of parasites of one kind have
a tendency toward a light infestation, or none, with other para¬
sites.
Most of the fish eating birds are carriers of one or more
kinds of parasites that enter fish during part of their life cycle.
Merganser ducks and kingfishers are perhaps the most impor¬
tant birds in this role.
Literature
Bere, Ruby. 1931. Copepods parasitic on the fish of the Trout Lake region,
with descriptions of two new species. Trans. Wis. Acad. Sci., Arts
& Let. 26:427-436.
Bere, Ruby. 1935. Further notes on the occurrence of parasitic copepods on
fish of the Trout Lake region, with a description of the male of
Argulus biramosus. Trans. Wis. Acad. Sci., Arts & Let. 29:83-88.
Cooper, A. R. 1918. North American Pseudophyllidean cestodes from fish.
Ill. Biol. Monog. 4:288-542.
Cross, S. X. 1934. A probable case of non-specific immunity between two
parasites of ciscoes of the Trout Lake region of northern Wisconsin.
Jour. Parasitol. 20 : 244-245.
Cross, S. X. 1935. The effect of parasitism on growth of perch in the Trout
Lake region. Jour. Parasitol. 21:267-273.
Essex, H. E. 1927. The structure and development of Corallobothrium.
Ill Biol. Monog. 11:1-75.
Hubbs, C. L. 1927. The related effects of a parasite on a fish. Jour. Para¬
sitol. 14:75-84.
Hughes, R. C. 1928. Studies on the t remat ode family Strigeidae. No. IX.
Neascus van-cleavei (Agersborg). Trans. Amer. Micro. Soc. 47:320-
340.
Hughes, R. C. 1928. Studies on the trematode family Strigeidae (Holos-
tomidae). Pa. Mich. Acad. Sci., Arts & Let. 10:483-500.
456 Wisconsin Academy of Sciences, Arts, and Letters
Hunter, G. W. Ill 1927. Studies on the Caryophyllaeidae of North America.
Ill. Biol. Monog. 11:1-186.
LaRue, G. R. 1914. A revision of the cestode family Proteocephalidae. Ill.
Biol. Monog. 1:1-850.
Manter, H. W. 1925. Some North American fish trematodes. Ill. Biol.
Monog. 10:127-264.
Miueller, J. F., and H. J. Van Cleave. 1932w Parasites of Oneida Lake
fishes. Roosevelt Wild Life Annals. 3:79-154.
Pearse, A. S. 1924. The parasites of lake fishes. Trans. Wis. Acad. Sci.,
Arts & Let. 21:161-194.
Pearse, A. S. 1924. Observations on parasitic worms from Wisconsin
fishes. Trans. Wis. Acad. Sci., Arts & Let. 21:147-160.
Reighard, J. 1928. A biological examination of Loon Lake, Gogebic County,
Michigan, with suggestions for increasing its yield of small mouth bass
(Micropterus dolomieu). Pa. Mich. Acad. Sci., Arts & Let. 10:589-612.
Van Cleave, H. J. 1919. Acanthocephala from the Illinois River, with de¬
scriptions of species and a synopsis of the family Neoechinorhynchidae.
Bull. Ill. State Lab. Nat. Hist. 13:225-257.
Ward, H. B. 1911. The distribution and frequence of animal parasites and
parasitic diseases in North American fresh-water fish. Trans. Amer.
Fish. Soc. 1912:207-241.
Ward, H. B., and G. C. Whipple. 1918. Fresh-Water Biology. John Wiley
and Sons, New York.
AGE AND GROWTH OF THE SUCKER, Catostomus commer-
sonnii (Lacepede) , IN MUSKELLUNGE LAKE,
VILAS COUNTY, WISCONSIN
William A. Spoor
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey.* Notes and reports No. 75.
Introduction
Since the establishment of the Trout Lake Limnological Lab¬
oratory in 1925 by the Wisconsin Geological and Natural His¬
tory Survey, investigations have been in progress on a number
of lakes in northeastern Wisconsin. The lakes under observa¬
tion differ from each other in a number of respects, and offer a
wide variety of habitats for aquatic life. This variety of habi¬
tats, and the accumulation of information concerning the physi¬
cal, chemical and biological conditions in these lakes present an
excellent opportunity to make an analysis of the influence of
various factors in the environment upon their fish populations.
Accordingly, in cooperation with the United States Bureau of
Fisheries and the Wisconsin Conservation Commission, the Sur¬
vey has included studies of the fishes of this region in its pro¬
gram since 1927.
The investigation has included studies of the growth, para¬
sites, food, distribution, and population density of perch, rock
bass, ciscoes, bluegills, small-mouthed black bass and suckers.
Data on muskellunge and wall-eyed pike have also been gathered
when available. The results of some of these studies have been
published by Wright (1929) in a preliminary report on the
growth of the rock bass, Juday and Schneberger (1930, 1933),
and Juday and Bennett (1935) in reports on the growth of game
fish, Couey (1935) in a study of the food habits, Cross (1934,
1935) in studies of the parasites, Schneberger (1935) in an
analysis of the growth of perch, and Hile (1936) in a study of
the growth of the ciscoes of the region. At present, studies
* This investigation was made in co-operation with the U. S. Bureau of Fisheries and the
results are published with the permission of the Commissioner of Fisheries.
457
458 Wisconsin Academy of Sciences, Arts, and Letters
dealing with the growth of muskellunge and wall-eyed pike, and
with the growth and population of blue-gills, rock bass, perch,
small-mouth black bass and suckers are in progress.
In addition to contributing to our knowledge of the biology
of the fishes mentioned, these studies are of interest in compar¬
ing the conditions of one lake with those of another, or in com¬
paring the conditions within a lake for one year with those of
another year. Further, when the ecological position of a species
of fish is considered, a study of its growth rate is of interest in
connection with the relative abundance or availability of the
organisms upon which that species feeds.
With these considerations in mind, a study of the growth of
the sucker in certain type lakes being examined by the staff of
the Laboratory was undertaken. A review of the literature re¬
vealed, however, that while the food and habits of the sucker
have received considerable attention (see Adams and Hankin-
son, 1928), but little is known about its growth rate. Since the
scale method is of great importance in making a detailed study
of the age and growth of the sucker, and since no papers dealing
with this subject in any detail have appeared, it was decided
that before the growth of the suckers in various lakes could be
compared there must be a reliable means for measuring growth.
It is also necessary that we have growth data for fish from a
single lake in numbers sufficient to make a reliable basis for
comparison. This paper is an attempt, therefore, to apply the
scale method to the study of the growth of the suckers of Mus¬
kellunge Lake, to establish growth curves which will serve as a
basis for later work from the comparative standpoint, and to
make such further contributions to our knowledge of the life
history of the sucker as the data at hand permit.
I wish to express my thanks to Dr. Chancey Juday for pre¬
senting the opportunity to conduct this study, for providing
much of the apparatus, and for advice and criticism throughout
the course of the investigation. My thanks are also due Dr.
Edward Schneberger, Mr. G. W. Bennett, Mr. D. G. Frey, and
Mr. C. L. Schloemer for their willing cooperation in the field.
Muskellunge Lake has been studied extensively, and the data
at hand are sufficient to present a fairly clear picture of its
physical, chemical and biological organization. Therefore, the
suckers of this lake have been chosen to form the basis of this
Spoor — Growth of Sucker
459
paper. Muskellunge Lake is a eutrophic lake situated in Town¬
ship 41 N., Kange VII E., Vilas County, Wisconsin. It is 3.8
kilometers (2 miles) long by 1.18 kilometers (0.7 mile) wide,
and has a surface area of 372 hectares (920 acres). Its maxi¬
mum depth is 21 meters, but the greater part of the lake is less
than 10 meters deep. The average Secchi disc transparency is
4.0 meters, bound C02 is 10.0 milligrams per liter, and organic
matter of plankton is 1.16 milligrams per liter. The color on
the platinum cobalt scale is 4, pH is 7.3, and the conductivity is
40 reciprocal megohms. There is no inlet, but there is a tem¬
porary outlet which is used only during periods of high water.
The ice disappears about the first of May, and by the middle of
May the water temperature is about 7.5°C. By July it is 20 to
25°C. at the surface, and about 8°C. at the bottom. The thermo-
cline lies between 8 and 12 meters during the summer. The
lake starts cooling in late August, and by the end of October the
temperature is down to 10° C. The ice forms again in Novem¬
ber.
Materials
This study is based upon the examination of 3,697 specimens
of the common sucker {Catostomus commersonnii (Lacepede) )
taken from Muskellunge Lake, chiefly during the summers of
the period 1927 to 1935. Hubbs (1926) lists the species as hav¬
ing fewer than 80 scales along the lateral line ; these specimens
agree with this description, the scale count ranging between 59
and 72, averaging around 64. The common name and the spell¬
ing of the specific name are those used by Hubbs (1926) and
Greene (1935).
Of the 3,697 specimens employed, 1068 were males, 1065
were females, and 1564 were not sexed. The numbers captured
in each year are : 1927, 47 not sexed ; 1928, 49 not sexed ; 1930,
5 males, 4 females, and 89 not sexed ; 1931, 442 not sexed ; 1932,
458 males, 466 females, and 1 not sexed; 1933, 125 males, 140
females, and 186 not sexed ; 1934, 182 males, 193 females, and
622 not sexed ; 1935, 298 males, 262 females, and 128 not sexed.
No fishing was done in 1929. As fish between the lengths of
60 and 110 millimeters could not be captured in Muskellunge
Lake, a few specimens of this size from other localities were
examined ; such fish have been so indicated when used.
460 Wisconsin Academy of Sciences, Arts, and Letters
Methods
The majority of the specimens used in this study were caught
either in gill nets or in fyke nets. During the period 1927 to
1932, gill nets were used exclusively ; starting in 1933, fyke nets
as well as gill nets were used. Specimens less than 60 milli¬
meters in length were captured by means of a minnow seine;
these specimens are listed as “not sexed” for 1933 and 1934.
Gill Nets
Seven sizes of gill nets were available ; the mesh sizes were
as follows: %, %, %> 1, 1%, 1%, and IV2 inch bar measure.
Each net was 150 feet long and 6 feet deep ; the floats and leads
were arranged so that the net rested upright along the bottom of
the lake when set. Usually seven nets were set in a gang, one of
each mesh size being used. During the period 1933 to 1935, how¬
ever, the % inch net was usually omitted from the set. The nets
were left in the water for varying lengths of time, and at vari¬
ous depths, according to the nature of the data desired. They
were never left for more than 24 hours before being examined,
however, and they were usually set in water ranging from six to
ten meters in depth.
Fyke Nets
In connection with a series of marking experiments, the use
of fyke nets was begun in 1933 ; their use was continued during
the summers of 1934 and 1935. These nets were four feet in
diameter at the mouth, and were 18 feet long; they were equip¬
ped with wings, leads, and two funnels. The mesh was % inch
bar measure, and was protected by a heavy coating of tar. As a
rule they were set in from 4 to 8 meters of water, being anchored
in the desired position. They were usually left in one location
for a week or so, at the end of which period they were returned
to the laboratory and dried thoroughly. While in the water they
were examined every day or so.
Handling of the fish
The routine practice in handling the fish was to empty the
nets early in the morning and return the fish to the laboratory,
where standard length, weight, sex, and the condition of the
Spoor — Growth of Sucker
461
gonads were noted. A scale sample was taken from each fish
and placed in a standard envelope upon which the date, locality,
gear, and the data mentioned above were recorded.
During 1927, length only was recorded; during 1928, 1930,
and 1931, weight and length were recorded. During the period
1932 to 1935, all of the above data were recorded for all suckers
returned to the laboratory with the exception of a few which had
decomposed to such an extent that the sex could not be deter¬
mined.
The scales were removed from the left side of the fish, about
midway between the base of the dorsal fin and the lateral line.
These scales were later mounted in white Karo syrup on glass
slides, and were examined with the aid of a Promar projection
apparatus. Most scales were not difficult to read and measure;
scales which could not be read after several examinations were
discarded.
To measure the length, the fish were placed upon a board and
straightened out with the tip of the snout touching a fixed peg, a
dissecting needle was then stuck into the board at the end of the
last scale on the lateral line, the fish was removed, and the dis¬
tance between the peg and the needle was measured to the near¬
est millimeter on a steel tape. Weights were estimated to the
nearest gram on a Chatillon spring platform balance of 500
gram capacity.
Sex was determined by examination of the gonads. Nota¬
tions regarding the condition of the gonads were based primarily
on the thickness and degree of coiling in the case of the testes,
and upon thickness, the presence or absence of ova, and the size
of the ova in the case of the ovaries. Further details will be
considered in another section.
During 1933 and 1934 some of the fish used in this study
were tagged and liberated. Weight and sex were not determined,
but a scale sample was taken from each fish and standard length,
tag number, locality, and date were recorded on the scale en¬
velope. These fish were measured in a measuring trough cali¬
brated in millimeters, but, as a live sucker is quite slippery and
very active, it was difficult to obtain accurate measurements in
some cases. To test the extent of the error, 60 fish were meas¬
ured alive in the trough in the field, returned to the laboratory,
and measured with the tape. Since the average difference was
462 Wisconsin Academy of Sciences, Arts, and Letters
but 0.5 millimeter no distinction was made between tape-meas¬
ured and trough-measured fish in the length calculations.
During 1933, 1934 and 1935, several hundred specimens
were preserved after length, weight, and other data had been
taken. These fish were used in determining the ratio between
girth and length.
Data other than those which have been mentioned were taken
from some of the fish. These data, together with certain varia¬
tions in the technique of capturing and handling specimens, will
be discussed in the sections with which they are associated.
Determination of Age and Growth
Inasmuch as the scale method has not proved to be satisfac¬
tory for studies of the age and growth of certain species of fish,
it is generally agreed that its validity should be tested for each
species of fish to which it is applied. With the exception of a
few age determinations recorded by Stewart (1926), the scale
method has not been applied to a study of the life history of the
sucker. It is necessary, therefore, that the applicability of the
method be tested.
Description of the scales
The scales of the sucker are of the cycloid type (Figs. 1-8),
and cover the entire body with the exception of the head and
fins. When first formed the typical scale is circular, but when
about two years old it becomes more or less quadrilateral, with
the dorsal and ventral margins straight and the anterior and
posterior margins rounded. (Stewart 1926). During the first
year the focus is nearer the anterior than the posterior margin,
but, due to the differential growth of the anterior and posterior
quadrants, it later comes to lie near the center, or, if excentric,
toward the posterior margin.
Radii. The radii are confined chiefly to the anterior and
posterior quadrants, but a few may occur at the anterior and
posterior ends of the dorsal and ventral quadrants. In the pos¬
terior quadrant, most of the radii originate at the focus ; in the
anterior quadrant, only the radii nearest the midline originate
at the focus, the rest take their origins more or less at random,
Spoor — Growth of Sucker
463
frequently in the vicinity of annuli. As the scale increases in
size, the radii increase in number.
Circuli. The circuli are most numerous and most closely
spaced in the anterior, dorsal, and ventral quadrants of the
scale ; relatively few are found in the posterior quadrant. They
parallel the edge of the scale in the anterior and posterior quad¬
rants; in the dorsal and ventral quadrants, however, they do
not quite parallel the edge, but seem to spread outward as they
extend back from the anterior quadrant, so that many of them
appear to be cut off by the edge of the scale as the posterior
quadrant is approached. As the scale grows, however, many of
these circuli are continued across the posterior quadrant; some
fail to do so and become encircled by succeeding circuli. This
condition is of value as an indicator of the annulus.
Annuli . As is the case with the scales of many other fishes,
the circuli of the sucker scale are arranged in zones of widely
spaced circuli alternating with zones of closely spaced circuli.
About half of the circuli in the widely spaced zones cross the
posterior quadrant of the scale, while the majority of those in
the closely spaced zones are cut off either shortly before or im¬
mediately after entering the posterior field. These circuli seem
to be cut off by the inmost circuli of the succeeding widely spaced
zone, thus giving the “cutting over” effect mentioned by Hile
(1931) in connection with the identification of the annulus. The
annulus of the common sucker, then, appears to be marked by a
crowding of the circuli and cutting over of these crowded circuli
at or near the point where they enter the posterior quadrant of
the scale. In the anterior quadrant, the annulus is frequently
associated with irregularities in the structure and course of the
circuli, in addition to their crowded condition.
The application of the method 1
As Van Oosten (1929) pointed out, the applicability of the
scale method to the determination of the length of a fish at the
end of any year of its life, and the amount of growth per year,
depends upon the following propositions :
1 Lack of space prohibits the inclusion of an extended discussion of the applicability of the
scale method to this study. The data upon which these conclusions are based are on file in the
library of the University of Wisconsin.
464 Wisconsin Academy of Sciences , Arts, and Letters
“1. That the scales remain constant in number and (retain their)
identity throughout the life of the fish.
2,. That the annual increment in length (or some other dimension
which must then be used) of the scale maintains, throughout the
life of the fish, a constant ratio with the annual increment in
length.
3. That the annuli are formed yearly and at the same time each
year (or that some other discoverable relation exists between their
formation and increment of time).”
The scales of the suckers used in this study were analyzed with
these requirements in mind; in general, the methods described
by Van Oosten were employed.
The constancy in number and identity of the scales through¬
out life was tested by scale counts, the nature of the central area
of the scale, regenerated scales, and by a comparison of the
growth of the scale with that of the fish. The average number
of scales along the lateral line for each length class of 10 milli¬
meters is shown in Table I. The counts ranged from 59 to 72,
chiefly between 60 and 69 ; they averaged 63.7. Since the devia¬
tions from this average are slight, it was concluded that the
scales remain constant in number throughout life. The high
degree of correlation between scale size and fish size (see p. 468)
is further evidence that the scales increase in size rather than in
number as the fish grows. That the scale retains its identity
throughout the life of the fish is shown in Figures 1-8 in which
the characteristic structure imparted to the scale by the rela¬
tively slow growth of the anterior quadrant during the first
summer (Fig. 3) can be seen to remain in the central area of
older and larger scales (Fig. 6). Further, if a scale is lost, it is
replaced by a typical regenerated scale in which the character¬
istic structure of the central area of a normal scale is missing.
That the annuli increase in number as the fish increase in
size is shown in Table II, in which the average standard lengths
of male and female suckers are arranged according to the number
of annuli on the corresponding scales. This is also shown in
Figures 1-7, in which it is clear that the scales of the larger fish,
in addition to being larger, show more annuli than do the scales
of smaller fish.
Evidence to show that one annulus is formed each year, and
that it is formed in winter, is presented in Figures 1-8. Figure
Plate I
Spoor — Growth of Sucker
465
Fig. 1. Scale from a Ill-group sucker, 144 mm. long. Muskellunge Lake,
August 23, 1935. Two annuli; X26.
Fig. 2, Scale from a II-group sucker, 114 mm. long. Muskellunge
Lake, August 16, 1931. One annulus; x2:6.
Fig. 3. Scale from a I-group sucker, 62, mm. long. Trout Lake, August
7, 1935. No annuli; X26.
Fig. 4. Scale from a IV-group sucker, 186 mm. long. Muskellunge
Lake, July 22, 1933. Three annuli; X26.
Fig. 5. Scale from a Vll-group sucker, 215 mm. long. Muskellunge
Lake, August 4, 1933. Six annuli; X26.
Fig. 6. Scale from a V-group sucker, 200 mm. long. Muskellunge
Lake, September 4, 1933. Four annuli; X26.
Fig. 7. Scale from an VUI-group sucker, 254 mm. long. Muskellunge
Lake, August 20, 1934. Seven annuli; x26.
Fig. 8. Scale from a IV-group sucker, 154 mm. long. Muskellunge
Lake, May 12, 1935. Third annulus just formed; X40.
466 Wisconsin Academy of Sciences, Arts , and Letters
Table I
Average scale measurements , average ratio between length (L) and
girth (G), average ratio between length and head length (H), and aver¬
age number of scales along the lateral line per length class of 10 milli¬
meters. Muskellunge Lake suckers.
(Number of specimens in parentheses)
8 shows the scale of a fish taken from Muskellunge Lake on May
12, 1936. Since the water temperature at this time was about
7.5° C. throughout the lake, we may safely assume that the grow¬
ing season was not very far advanced. This scale shows two
annuli which are certainly complete, and one at the edge which
appears to be well nigh, if not entirely, completed. Evidence
Spoor — Growth of Sucker
467
Table II
Average actual standard lengths per age-group of male and female
suckers taken from Muskellunge Lake , September , 1985.
Collections combined.
(Number of specimens in parentheses)
that it has been completed, or is very near to completion, is to
be found along the left margin, where it will be noted that the
zone of closely spaced circuli is just beginning to yield its place
at the edge of the scale to the succeeding zone of widely spaced
circuli. At the left side of the posterior quadrant, the first signs
of cutting over are appearing, and the new season's growth is
apparently just beginning.
Further evidence that one annulus is formed each year is
presented in Tables XIX to XXII. These tables are based upon
measurements of the amount of scale laid down outside the outer¬
most annulus as the season progresses. From these tables we may
conclude that the season's growth has not progressed very far
by the middle of May, and that as the season advances the width
of the band widely spaced circuli outside the outermost annulus
increases, rapidly at first, then more slowly as the succeeding
winter is approached.
It has been noted by several authors that accessory annuli
frequently cause difficulty in determining the age and growth of
various species of fish. Structures which might be considered
accessory annuli were rare among the scales used in this study.
If a structure had all the characteristics of an annulus it was
considered as such, regardless of its position on the scale. It is
felt that this procedure was necessary because the data pre¬
sented no reason for believing that unusual spacing of annuli
meant anything other than unusual growth for the periods rep¬
resented by those annuli.
Calculation of length
Although various areas of the scale have been measured for
use in length calculations, it is usually reasoned that, since the
468 Wisconsin Academy of Sciences , Arts , and Letters
length of the fish is being calculated, the scales should be meas¬
ured along the antero-posterior axis to keep the method consis¬
tent. Accordingly, the majority of investigators have measured
along the anterior radius of the scale. Van Oosten (1929) and
Hile (1936) modified this procedure by measuring the antero¬
posterior diameter. Schultz (1933) measured along the antero¬
lateral radius.
The most satisfactory axis for measuring the scales of the
sucker proved to be the dorso- ventral diameter. Since the annulus
is not distinct across the posterior quadrant, Van Oosten’s meth¬
od cannot be employed. The dorso-ventral diameter is more sat¬
isfactory than the anterior radius for two reasons: (1) The
annuli are more distinct in the dorsal and ventral quadrants,
and (2) the growth of the scale along the dorso-ventral diam¬
eter is more closely correlated with growth in body length than
is the growth of the scale along the anterior radius. The latter
difference is due to the fact that the anterior quadrant grows
more slowly than the rest of the scale during the first summer
of life. The measurements of a series of corresponding scales
are averaged for each length-class of 10 mm. in Table I. The
coefficient of correlation between body length and the dorso-ven¬
tral diameter of the scale is .985 plus or minus .001; between
body length and the length of the anterior radius it is .944 plus
or minus .006. Since the relative girth of the suckers from
Muskellunge Lake remains practically constant (Table I), it has
no influence upon length calculations based upon measurements
of the dorso-ventral diameter of the scale.1
Formula
Length calculations were based upon the standard formula:
S*
s
L«
L
in which S represents the total dorso-ventral
diameter of the scale, S« the dorso-ventral diameter included
within the annulus being considered, L the standard length of
the fish at the time the scale is removed, and La the length of the
fish at the time the annulus was formed.
1 Observations made upon a female cisco captured during the spawning season are of interest
in this, connection. The body cavity of this fish was literally filled with ripe eggs; the resulting
distension caused the horizontal scale rows to be stretched so far apart that they hardly overlapped;
normally, the scales of the cisco show considerable overlapping) It was further noted that the
scales of the horizontal rows ventrad the lateral line were the only ones involved in this spreading;
those dorsad the lateral line did not seem to be affected.
Spoor — Growth of Sucker
469
The scales of the sucker first appear when the fish is about
25 millimeters long, but no correction factor was introduced to
compensate for late scale formation. As Van Oosten found for
the lake herring, the scales of the sucker grow at an accelerated
rate for a short period after their formation, so that by the time
the body length reaches 30 to 35 millimeters the scales overlap
in a manner similar to that found in the adult. Lee’s “phenome¬
non” of apparent change in growth rate appears among the
calculated lengths (Tables IV to XI), but seems to be due, not
to late scale formation, but to the selective action of the gear.
The use of a factor to compensate for late scale formation had
no effect upon the appearance of Lee’s “phenomenon”.
Comparison of actual and calculated lengths
In a later section, data are presented which indicate that the
growing season of the suckers of Muskellunge Lake is practi¬
cally over by September. Therefore, it is possible to guage the
accuracy of the scale method by comparing calculated lengths
with the corresponding actual lengths of specimens captured in
September. Such a comparison is shown in Table III.
Table III
Average calculated lengths to outermost annuli compared with the
corresponding average actual lengths of Musekllunge Lake suckers cap¬
tured in September of the years 1927 , 1930 , 1931 } 1932, 1933, 193 4, and 1935.
Collections combined.
(Number of specimens in parentheses)
Length in millimeters at end of growing season
To construct this table, all September specimens were segre¬
gated into their age-groups according to the number of annuli
on their scales ; the average length of each age-group1 was then
placed below the average calculated lengths to the outermost
annulus on the scales of fish in the next older age-group. The
lengths were arranged in this manner to reduce the discrepan-
1 The term “age-group” is used to indicate the age of the fish in summers.
470 Wisconsin Academy of Sciences, Arts, and Letters
cies due to Lee's ‘'phenomenon". As is apparent in the first two
columns, these discrepancies could not be avoided entirely among
the younger age-groups. From the third column on, however,
the calculated and corresponding actual lengths agree fairly
well. An analysis of the data upon which these figures are based
showed that the discrepancies which do occur are due chiefly to
sampling, and not to deficiencies in the scale method.
The influence upon length calculations of such factors as the
ratio between head length and body length (Table I), and vari¬
ation in the body area from which the scales were taken was
investigated, but proved to be negligible.
Summary of calculated lengths
Tables IV to XI present a summary of the calculated lengths
in millimeters for all fish for which age and growth determina¬
tions were made. Each table summarizes the collections for one
year; these collections have been divided into age-groups ac¬
cording to the number of annuli on the scales. All lengths shown
are calculated lengths ; since the fish were caught over a period
of several months during the summer of each year, the actual
lengths are not comparable when the year's collections are com¬
bined.
The Selective Action of the Gear
The frequency column at the right of each table indicates
that the sampling may be the cause of the occurrence of Lee's
Average calculated lengths in millimeters to the end of each year of life
for each age- group of Muskellunge Lake suckers captured in 1927.
Spoor — Growth of Sucker
471
T able V
Average calculated lengths in millimeters to the end of each year of
life for each age-group of Muskellunge Lake suckers captured in 1928.
Table VI
Average calculated lengths in millimeters to the end of each year of
life for each age-group of Muskellunge Lake suckers captured in 1930.
Table VII
Average calculated lengths in millimeters to the end of each year of
life for each age-group of Muskellunge Lake suckers captured in 1931.
472 Wisconsin Academy of Sciences, Arts, and Letters
Table VIII
Average calculated lengths in millimeters to the end of each year of
life for each age-group of Muskellunge Lake suckers collected in 1932.
Table IX
Average calculated lengths in millimeters to the end of each year of
life for each age-group of Muskellunge Lake suckers collected in 1933.
Table X
Average calculated length in millimeters to the end of each year of
life for each age-group of Muskellunge Lake suckers collected in 193U.
Spoor — Growth of Sucker
473
Table XI
Average calculated length in millimeters to the end of each year of
life for each age-group of Muskellunge Lake suckers collected in 1935.
Table XII
Number of Muskellunge Lake suckers in each age- group in the col¬
lections of each year.
phenomenon among the calculated lengths. These frequencies
are brought together in Table 12, in which the fish are arranged
according to age-group and the year collected.
Obviously, the younger age-groups are not well represented.
No. I-group fish, and very few II-group fish were collected. If
the sampling had been perfect, that is, if each year class were
represented in the catch in proportion to its abundance in the
lake, we might reasonably expect to find the I-group to contain
the greatest number of individuals in the collection, the II-
group next, and so on. This expectation would be based upon
the assumption that the relative abundance of each age-group
would decrease as predation, disease, accident, and perhaps old
474 Wisconsin Academy of Sciences , Arts , and Letters
age, took their toll. It is quite evident that the sampling was
far from perfect ; this can be traced directly to the gear used in
fishing.
As mentioned previously, all fish were caught in gill nets and
fyke nets except several hundred young suckers which were col¬
lected in shallow water with a minnow seine. It has been found
that the gill nets have a pronounced selective action among the
younger age-groups of ciscoes (Hile 1936) and perch (Schne-
berger 1935). This also appears to be the case among the suck¬
ers.
During 1930, 1931, and 1932, the mesh size of the gill net in
which each fish was caught was recorded on the scale envelope.
A summary of these records is presented in Figure 9, in which
the fish caught in each size of mesh have been arranged accord¬
ing to frequency per length class of three millimeters. It is
readily seen that the gill nets are quite selective; except for a
Fig. 9. Number of fish per length-class of three millimeters caught
in each mesh-size of gill nets; Muskellunge Lake suckers, 1930, 1931 and
1932 collections.
Spoor — Growth of Sucker
475
few “accidental” captures, each net appears to have selected fish
over a rather limited range of length. The % inch net took fish
over a length range of 116 to 133 millimeters, the % inch net
from 123 to 159 millimeters, the % inch net from 145 to 181
millimeters, the 1 inch net from 171 to 227 millimeters, the 1%
inch net from 186 to 249 millimeters, the 1% inch net from 205
to 277 millimeters, and the l1/^ inch net from 241 to 321 milli¬
meters. The overlapping is sufficiently great that if all nets are
used they can be expected to catch suckers between the lengths
of 116 and 321 millimeters.
The suckers are cylindrical fishes, and since they have no
spines projecting from their bodies, and since their mouths and
heads present no projections which might become entangled in
the meshes of the net, the only way they can become caught in a
gill net is by entering a mesh which has a circumference measur¬
ing less than the greatest girth of the fish. When suckers are
caught in gill nets, they are usually held somewhere in the region
between the eyes and the beginning of the dorsal fin. If the
girth in this region is less than the circumference of the mesh,
they can pass through the net. The circumference of a mesh %
inch bar measure is 21/2 inches, or 63.5 millimeters. By using the
average value for the ratio between girth and length (Table I),
we can calculate the length of a fish which has this girth, and
which, presumably, would just be caught by a % inch mesh.
This length is 113 millimeters. Therefore, we may conclude that
any average sucker less than 113 millimeters in length would be
able to pass through the meshes of the net. For this reason,
none of the first year fish would be taken, and since the average
calculated length at the end of the second winter is around 100
millimeters, the nets would catch only those II-group fish which
were above the average in length and girth. Similarly, if the %
inch net were omitted from the set, fish smaller than the smallest
caught by the % inch net (126 mm.) would escape, and so on.
Since the nets with the smaller meshes were frequently omitted
from the sets, a large number of the smaller fish escaped. The
use of fyke nets during 1933, 1934 and 1935 did little to remedy
the situation because, since the mesh used was % inch bar meas¬
ure, they too were selective and did not retain the smaller fish.
As a result of this selective action of the gear, only the larger
and more rapidly growing suckers of the first few age groups
476 Wisconsin Academy of Sciences , Arts, and Letters
were captured; consequently the averages of the calculated
lengths of these age groups are too high. Since the older age
groups were better represented, the slowly growing as well as
the rapidly growing fish would be included in the averages of
the calculated lengths; as a result, these averages tend to be
relatively lower than the average calculated lengths of the se¬
lected samples of the younger age groups. In Tables IV to XI it
will be noted that the major indications of Lee’s phenomenon
occur among the younger fish. Therefore, it is concluded that
the appearance of Lee’s phenomenon among the calculated
lengths of the specimens used in this study is due chiefly to the
method of sampling. It has been suggested by various investi¬
gators (Schneberger 1935; Van Oosten 1929; Hile 1936) that
Lee’s phenomenon can be accounted for in part by assuming that
the more slowly growing fish reach a greater age than the more
rapidly growing members of the same year class. This sugges¬
tion, if valid, may be considered as an additional factor contri¬
buting to the appearance of the phenomenon among the calcu¬
lated lengths of these specimens.
Hile (1936) points out that since the gear selects only the
larger individuals of the younger age-groups these age-groups
should be excluded from the data upon which general growth
curves are based. In this study the II-group seems to be the
worst offender in this respect, but since it is represented by a
very small number of specimens this group has not been elimi¬
nated from the data.
GROWTH OF THE SUCKERS OF MUSKELLUNGE LAKE
I. Growth in Length
The analysis of the growth in length of the suckers of Mus-
kellunge Lake is based upon age and growth determinations
from the scales of 2,990 of the specimens collected during the
period 1927 to 1935. Most of the tables and figures are based
upon the calculated lengths ; since the specimens employed were
captured at different times during the growing season of each
year, the actual lengths are not suitable for the construction of
growth curves because they do not represent the lengths reached
by the end of each year of life.
Spoor — Growth of Sucker
477
As was mentioned in the introduction, the growth of the
sucker has been studied by several investigators. Crawford
(1923) studied the growth of young suckers in Maine, and found
that during the period May 15 to June 22 the average length
increased from 16 to 29 millimeters in a hatchery pond. Hubbs
and Greaser (1924) found that the young suckers of Douglas
Lake, Michigan, increased from an average length of 16 milli¬
meters in early June to an average length of 51 millimeters by
mid-August. These figures are in agreement with those given
by Bigelow (1923), which show that the average length of young
suckers in Lake Nipigon increased from 19 millimeters in the
middle of June to 40 millimeters by the middle of August. Em¬
body (1915) records the growth as follows: Five months, 50 mil¬
limeters; one year, 75 to 100 millimeters; two years, 150 to 175
millimeters. Stewart (1926) states that the suckers in the vi¬
cinity of Ithaca, New York, reach a length of 38 millimeters by
the end of their first season, and gives the following figures for
the lengths reached at the end of each subsequent year of life:
Second, 75 millimeters; third, 100 millimeters; fourth, 112 to
150 millimeters ; fifth, 150 to 200 millimeters ; sixth, 200 to 343
millimeters ; seventh, 343 to 394 millimeters. These figures are
averages of the actual lengths of about a dozen fish of each age-
group ; age was determined from the scales. Adams and Hank-
inson (1928) state that the usual length reached by Oneida Lake
suckers at the end of their first season is between 75 and 100
millimeters.
The suckers of Muskellunge Lake apparently do not grow at
the same rate as those studied by Adams and Hankinson, Em¬
body, and Stewart. In Tables IV to XI are summarized the age
and growth determinations of all specimens with readable scales.
Each table includes the average calculated lengths to the end of
each year of life for each age-group of the suckers captured in
one year. The grand averages of the calculated lengths for each
year of life are brought together in Table 13, in which the grand
average calculated lengths and grand average annual increments
for all fish are also shown.
As has been shown (p. 469), the calculated lengths agree fair¬
ly well with the corresponding actual lengths for the older age-
groups. We need, however, a check upon the accuracy of the
calculated lengths for the first season's growth. The calculated
478 Wisconsin Academy of Sciences , Arts, and Letters
length to the first annulus of Muskellunge Lake suckers averages
61 millimeters. This figure is not in agreement with that given
by Stewart (38 millimeters), and is considerably smaller than
those given by Embody and Adams, and Hankinson (75 to 100
millimeters) for the first year’s growth. It is compatible, how¬
ever, with those given by Bigelow (40 millimeters) and Hubbs
and Creaser (51 millimeters) since these figures represent the
average lengths in mid- August, before the close of the growing
season. Bigelow (1923) records the capture of three young
suckers in Lake Nipigon on June 6, 1922; these specimens aver¬
aged 55 millimeters in length. Sixty specimens taken from that
lake on June 14 ranged between 17 and 21 millimeters, and
thirty-four specimens taken on August 12 ranged between 32
and 49 millimeters. It seems, therefore, that the 55 millimeter
fish captured on June 6 were not in their first season, but that
they were captured at the beginning of their second season (see
p. 481). The August 16 collection shown among the data of
Hubbs and Creaser contains suckers ranging from 41 to 62 milli¬
meters in length. On August 11, 1934, two young suckers col¬
lected in Muskelunge Lake had standard lengths of 48 milli¬
meters; a young sucker captured in Trout Lake (in the vicinity
of Muskellunge Lake) on August 7, 1935, was 51 millimeters
long. These fish showed no annuli on their scales, and were in
their first season. The fish from which the I-group scale shown
in Figure 3 was taken had reached a length of 62 millimeters
by August 7, when it was captured. On July 12, 1935, a young
sucker measuring 71 millimeters in length and showing one
annulus and marginal growth on its scales was taken from a
creek in the vicinity of Muskellunge Lake. It appears, there¬
fore, that a calculated length of 61 millimeters at the end of the
first winter is in harmony with the actual lengths of fish cap¬
tured before and after the first winter of life.1
Direct evidence of the validity of the calculated length to the
second annulus is meagre. The II-group fishes captured in Mus¬
kellunge Lake in September of 1931 average 123 millimeters in
standard length. Because of the selective action of the gill nets
in which they were caught, however, this average represents
1 Since the I-group hatches about the if iddle of June in Muskellunge Lake, the fish have not
lived for an entire year by the time the first annulus is formed, but since they develop and hatch
toward the beginning of the growing season, and since it is the growth during this season with
which we are most concerned, 61 millimeters can be considered as the growth of the first year
of life.
Spoor — Growth of Sucker
479
the lengths of only the largest of the II-group fish, and cannot
be accepted as the true average length at the end of the second
growing season. Another line of evidence is of assistance in
determining approximately what this average should be.
In studies of the growth rates of the sunfish (Greaser 1926),
lake herring (Van Oosten 1929), cisco (Hile 1936), perch
(Schneberger 1935), bay smelt (Schultz 1933), jack-smelt
(Clark 1929), and many other fishes, it has been found that the
increase in length during the second year of life is usually less
than that during the first year, and either equals or exceeds
that of the third year of life. We are not justified in concluding
that the same is true of the sucker on this basis, but since the
condition appears to be rather general among fishes, we may
reasonably expect to find it true of the sucker. The average
length increments for each year of life shown in Table XIII are
plotted against age in Figure 10. The average increment of the
calculated growth of the second season is 39 millimeters, making
the average length at the end of the second year 100 millimeters.
Since the increment for the third year is 38 millimeters, and
Fig. 10. Growth in length of Muskellunge Lake suckers; all collections
combined.
480 Wisconsin Academy of Sciences , Arts , emd Letters
that for the first year is 61 millimeters, an increment of 39 milli¬
meters is acceptable as a fair measure of the growth of the Mus-
kellunge Lake sucker during its second year of life.
In addition to the average annual increments, the grand av¬
erages of the calculated lengths to the end of each year of life
for all fish are shown plotted against age in Figure 10. The
Table XIII
Grand averages of the calculated lengths to the end of each year of
life of Muskellunge Lake suckers collected during the period 1927 to 1935;
sexes combined.
(Number of specimens in parentheses)
Grand Average Calculated Length to End of Year
* One very doubtful specimen.
oldest fish of the 2990 specimens with readable scales had ten
annuli on its scales, and therefore was in its eleventh summer
of life. The curve is in dashed lines beyond the eighth annulus
because the numbers of specimens used to establish the calcu¬
lated lengths for the nine- and ten-annulus fish were small. To
sum up the information given in Table XIII and Figure 10, the
increase in length is most rapid during the first few years ; the
annual increment decreases gradually with each succeeding year
Spoor — Growth of Sucker
481
of life. The average calculated lengths of Muskollunge Lake
suckers at the end of each year of life are as follows : First year,
61 millimeters; second year, 100 millimeters; third year, 188
millimeters ; fourth year, 173 millimeters ; fifth year, 197 milli¬
meters; sixth year, 222 millimeters; seventh year, 245 milli¬
meters ; eighth year, 262 millimeters. The lengths at the end of
the ninth and tenth years are based upon such a small number
of specimens that they can not be considered representative of
the population as a whole. They indicate, however, that the fish
continue to grow after the eighth winter.
During the years 1932, 1933, 1934 and 1935, the sex of each
specimen returned to the laboratory was determined whenever
possible. The sexed fish for which age and growth were de¬
termined were arranged in groups according to sex and year of
capture ; the average calculated lengths to the end of each year
of life for each group are shown in Table XIV,, in which the grand
averages of the calculated lengths of each sex for all years are
also included. The numbers of specimens employed in the de¬
termination of each average is included in parenthesis. The
grand average calculated lengths are plotted against age in Fig-
Fig. 11. Growth in length of male and female Muskellunge Lake
suckers.
482 Wisconsin Academy of Sciences, Arts, and Letters
ure 11. According” to Figure 11, the males and females increase
in length at about the same rate for the first four or five years
of life. From the fifth year on, however, the females increase
more rapidly than the males in length ; the difference is 4 milli¬
meters at the end of the sixth year, 8 millimeters at the end of
the seventh year, and 11 millimeters by the end of the eighth
year. Since there were but two males and five females available
for the determination of the average lengths at the end of the
ninth year of life, but little significance can be attached to the
curves beyond the eighth year. Apparently, however, the fe¬
males continue to grow more rapidly than the males.
That this apparent difference between the growth rates of
males and females is not due to a mere accident in sampling is
shown in Table XIV, in which it is evident that after the fifth
year of life the calculated lengths of the females are consistently
greater than those of the males in the collections of four different
Table XIV
Comparison between the calculated lengths to the end of each year of
life of male and female suckers collected in Muskellunge Lake during the
period 1932 to 1935.
(Number of specimens in parentheses)
Spoor — Growth of Sucker
483
years. That it is not due to an error in the application of the scale
method is shown in Table XV, wherein it is evident that after
the fifth year the average actual lengths of the females exceed
those of the males of the same age and collection. It appears,
therefore, that the difference between the growth rates of the
males and females is real.
Table XV
Average actual lengths of females compared with average actual lengths
of males of same age and collection. Muskellunge Lake Suckers.
II. Growth in Weight
During the period 1928 to 1935, the weight in grams of each
sucker returned to the laboratory was recorded on the scale
envelope. No weights were taken during 1927, and fish which
were tagged in the field during 1933, 1934 and 1935 were not
weighed. The I-group fish captured with a minnow seine were
not weighed at Trout Lake, but were preserved in alcohol and
returned to the Zoological Laboratory of the University of Wis¬
consin, where the weights of 266 specimens were measured. A
freshly killed young sucker sinks in distilled water, while one
which has been preserved in alcohol will float. To compensate
somewhat for the error introduced into the weight measurement
by the lighter alcohol, each fish was left in distilled water for
from 12 to 24 hours, or until it sank, before it was weighed.
484 Wisconsin Academy of Sciences, Arts, and Letters
Excess water was removed by placing the fish between moist
blotters, and weights were determined to the nearest milligram
on a balance. The weights which were determined in this man¬
ner are probably not strictly accurate, but are sufficiently exact
for use in this section.
The average weight in grams per length-class of 10 milli¬
meters was determined for all weighed specimens with the ex¬
ception of those collected in May, 1935; the weights of speci¬
mens in this collection were not included in the averages because
the fish appeared to be in slightly poorer condition than those
collected later in the season. These averages are tabulated in
Table XVI, which also includes the average weights of males and
females considered separately. The unsmoothed averages of the
combined collections are plotted against length in Figure 12.
Since the data on fish between the lengths of 50 and 110 milli¬
meters were lacking, this section of the curve is indicated by a
dashed line. Similar curves for males and females treated sep¬
arately are shown in Figure 13.
Fig. 12. Growth in weight of Muskellunge Lake suckers, all collections
coinbined. Age indicated, by Roman numerals.
Spoor — Growth of Sucker
485
Figure 12 shows that while the average annual increment of
length decreases, the average annual increment of weight tends
to increase with each successive year of life. In Figure 13, it is
shown that the males and females increase in weight at about
the same rate per unit of length. Since, however, the females
increase in length at a rate which exceeds that of the males, they
are correspondingly heavier than the males at the end of each
year of life after the fifth.
Since the collections were made during the growing season
of each year, it is not feasible to determine the average weight
attained by the end of each year of life by averaging the weights
of all fish of an age-group captured in the course of the season.
It is possible, however, to determine from the graph the weight
which corresponds to the average calculated length at the end of
each year of life, and thus arrive at an approximation of the
average annual gain in weight. Figure 12 shows that the aver¬
age weights of Muskellunge Lake suckers at the end of each
year of life after the second are as follows: Third year, 47.5
50 100 150 200 250 300 350 400 450 500 550
WEIGHT IN GRAMS
Fig. 13. Growth in weight of male and female Miuskellunge Lake
suckers.
486 Wisconsin Academy of Sciences, Arts, and Letters
grams; fourth year, 77.5 grams; fifth year, 130 grams; sixth
year, 177.5 grams; seventh year, 244 grams; eighth year, 320
grams ; ninth year, 401 grams. Data for the direct determina¬
tion of the average weights at lengths corresponding to those
calculated for the first and second years of life are not available,
but by projecting the curve across the gap between the actual
Table XVI
Average weight in grams per length class of 10 millimeters. Muskel-
lunge Lake suckers; collections of 1928, 1930, 1931, 1932 , 1933j 193U, and
1935.
Spoor — Growth of Sucker
487
measurements at 48 and 115 millimeters, we are enabled to find
the approximate weights at those lengths. These approximate
weights are 4 grams for the first year, and 18.5 grams for the
second.
Relationship between Length and Weight
It has been found by many investigators that the length and
weight of fishes are closely correlated. Crozier and Hecht (1914)
found that for the squeteague (Cy noscion regalis) the coefficient
of correlation was 0.952, that the weight varied as the third
power of the length, and that this relationship could be ex¬
pressed, therefore, by an equation in the form of y = a x3; in
which y represents weight, x length, and a is a constant, the
value of which depends upon the units employed. Hecht (1916)
extended these observations to various other species of fish. Keys
(1928) found that the cube formula was not a perfect expres¬
sion of the relation between length and* weight in the sardine,
herring and killifish, and that the relationship was much more
accurately expressed by the more general equation W = a(L)n,
in which W is the weight, L the length, and a is a constant de¬
pending upon the form of the species and the units employed.
The exponent “n” expresses the rate of change in weight with
length. The exponent “n” was found to be 3.1 for the sardine,
3.5 for the herring, and 3.7 for the killifish. This formula has
been applied by subsequent investigators to the sardine (Clark
1928), the whitefish (Hart 1931), smallmouth black bass (Tes¬
ter 1932), baysmelt (Schultz 1933), perch (Schneberger 1935),
and many others. It has usually been found that the weight
increases at a rate slightly greater than the cube of the length,
and that no single value of n holds for all lengths of a species.
Over the length range 115 to 300 millimeters the weight of the
suckers of Muskellunge Lake increases as the cube of the length.
Coefficient of Condition
Hecht (1916) pointed out that the factor “a” of his formula
y — a xs could be used as an index to the “condition”, or relative
fatness, of the fish at different seasons of the year. This formula
is usually expressed as W = fL3, in which W is the weight, L the
length, and f the equivalent of Hecht’s coefficient (a). This co¬
efficient has been given various names, all denoting the condition
488 Wisconsin Academy of Sciences , Arts , and Letters
of the fish, but is most frequently referred to as the “coefficient
of condition”. It is designated by various symbols, but most
frequently by the letter K. When length is expressed in milli¬
meters, and weight in grams, the first significant figure of this
cofficient usually falls far to the right of the decimal. To avoid
the use of such small decimals, Hile (1936) defined the coefficient
of condition (K) as: K = f (L) X 105. This definition has been
accepted for use in this study. If the relative weight increases
or decreases with respect to the length, the value of K will in¬
crease or decrease accordingly. Since the chief factors usually
considered to influence the weight of fishes are fatness and the
degree of maturity of the gonads, many authors have used the
fluctuations of the value of K as an indicator of the relative
fatness, the time of spawning, or both.
Hile (1936) calls attention to the fact that numerous investi¬
gators have confused the coefficient of the formula expressing
the length-weight relationship (W = a(L)n) with the coefficient
of the cube formula, and shows that when “condition” is being
analyzed it is important that the exponent be kept constant so
that the relative weights of fish of different lengths and ages
will remain comparable. Since the exponent remains 3 in the
formula expressing the relationship between the length and
weight of Muskellunge Lake suckers, it is not necessary to dis¬
tinguish between these two coefficients in this particular case.
The coefficient of condition (K) was calculated for each spec¬
imen of Muskellunge Lake sucker for which weight was re¬
corded. Since no collections were made during the period
November to April, and since the months of May, June, and
October are represented by relatively few specimens, the data
at hand are not suitable for an analysis of the changes in condi¬
tion throughout the year. Inasmuch as data more suitable than
the coefficients of condition are available for a study of the time
of spawning, condition is not considered in this connection. A
comparison of the coefficients of specimens captured during
May, 1935, with those of specimens captured during July,, Au¬
gust and September of the same year affords some information
regarding the- variation in condition as the season advances, but
the coefficients of condition are included here primarily to fur¬
nish a basis for comparing the condition of the suckers of Mus-
Spoor — Growth of Sucker
489
kellunge Lake with that of suckers from other lakes to be con¬
sidered in subsequent investigations.
The 1985 collections were made during the period May to
September ; while the collection made in May is not large,
enough specimens were taken to make possible a comparison of
the average coefficients of condition at different times during
the growing season. These averages are tabulated in Table XVII
according to the age and sex of the specimens employed. The
season was divided into three periods, early, middle and late;
fish captured in May represent the early period, those captured
during the period July 25 to August 10 represent the middle
period, and those captured on dates between August 23 and Sep¬
tember 6 represent the late period.
Although the numbers of specimens entering into the calcula¬
tion of each of the averages shown in Table XVII are hardly large
enough to render these averages entirely reliable, there seems
to be a general tendency for the sucker to improve in condition
between May and the end of July. This is in agreement with
evidence to be presented in another section which shows that
a large percentage of the season's growth is completed by the
end of July. The slightly lower values of K for the May fish are
not associated with a change in relative weight due to spawning
because but one of the females of the collection had ripe eggs,
the rest were immature, and while the gonads of the mature
Table XVII
Average values of K for each age-group and sex of Muskellunge Lake
suckers captured at different times during the period May to September ,
19SS- ik i iM
(Number of specimens in parentheses)
490 Wisconsin Academy of Sciences , Arts , and Letters
males appeared to be ripe, there were no indications that spawn¬
ing had occurred.
The average values of K for each age-group of Muskellunge
Lake suckers collected during the period July to September of
the years 1928 to 1935 inclusive are tabulated in Table XVIII.
Collections of 1932 to 1935 are further grouped according to
sex. Although there are some indications of fluctuations in the
average values of K, there are no pronunced trends which would
indicate significant variation with the calender year or sex.
There are indications that there is a tendency for fish to become
poorer in condition in the ninth and tenth years of life, but the
numbers of specimens representing these age-groups are so small
that it is not advisable to consider these averages as representa¬
tive of the population as a whole. The slightly higher values
Table XVIII
Average values of K for each age-group of Muskellunge Lake suckers
collected during the period July to September, 1928 , 1930, 1931, 1932 ,
1933, 193b and 1935.
(Number of specimens in parentheses)
*1932 male, age-group XI, 1.6S.
Spoor — Growth of Sucker
491
for certain of the II- and Ill-group fish may be attributed to the
selective action of the nets.
III. The Growing Season
During the years 1927 to 1933, collections were made only
during July, August and September. In 1934, collections were
also made in June and October, and in 1935, in May and June as
well as during July, August and September. The dates on which
collections were made outside of the regular season are June 18
to 28 and October 27, 1934, and May 12 and June 29, 1935, so
that the collections of 1934 and 1935 represent a period extend¬
ing from the middle of May to the end of October. The ice
usually leaves the lakes of this region around the first of May
(Juday and Birge 1930), by the middle of May the temperature
of Muskellunge Lake is about 7° C., and by July it is 20 to 25°
at the surface. The lake starts to cool in late August, and by
the end of October is about 10° C. throughout. The collections
of 1934 and 1935 may therefore be considered as representative
of the early, middle, and late growing season. By analyzing the
marginal growth for each month, that is, the amount of growth
occurring since the formation of the annulus of the preceding
winter, it should be possible to determine the duration of the
growing season and the period of most rapid growth.
The averages of the marginal growth for each age-group of
sucker represented in the collections of 1934 and 1935 are tabu¬
lated in Tables XIX and XX. To obtain a measure of the growth
of each age-group in an entire year, the average growth incre¬
ment of the preceding year of each age-group of 1935 fish was
derived from the calculated lengths. These averages are arranged
in the last column of Table XX, so that the growth of each age-
group in the current season is compared with the growth of the
next older age-group in the preceding season. For example,
since the IV-group fish of the 1935 collections were the Ill-group
fish of the 1934 collections, the growth of the Ill-group fish dur¬
ing the 1935 season is compared with the 1934 increment of the
IV-group fish of 1935 ; the latter figure gives a measure of the
annual growth increment of Ill-group fish. The annual incre¬
ments of each age-group were employed instead of the grand
average increments of the population as a whole (shown in Table
492 Wisconsin Academy of Sciences , Arts, and Letters
XIII and Figure 10) to avoid the discrepancies due to Lee’s phe¬
nomenon.
Since the collections of each year contained representatives
of each age-group, it is possible to compare the marginal growth
of the fish taken during 1934 with the total growth increment
for that year as calculated from the scales of fish of the same
year class (fish hatched in the same calendar year) but captured
during the 1935 season, when they were a year older. For ex¬
ample, since fish hatched in 1932 would be in their third year of
life in 1934, and in their fourth year of life in 1935, the annual
increment of the third year of life of the IV-group fish of 1935
represents the increment which would have been gained by the
Ill-group fish of the 1934 collection had they been allowed to
Table XIX
Average marginal growth of each year-class of Muskellunge Lake
suckers reached during each month of the period June to October, 1934 ,
compared with the corresponding increment gained during the entire 1934
season by the same year-class. Sexes combined.
(Number of specimens in parentheses)
Table XX
Average marginal growth of each age-group of Muskellunge Lake
suckers reached during each month of the period May to September , 1935,
compared with increment gained by the same age-group during the entire
1934 season; sexes combined.
(Number of specimens in parentheses)
Spoor — Growth of Sucker
493
Table XXI
Average marginal growth of each age-group of Muskellunge Lake
suckers reached during each month compared with the corresponding
annual increment; collections of 1934 and 1935 combined , sexes combined.
(Number of specimens in parentheses)
Table XXII
Percentage of the total season’s growth shown by the collections
of each month. Muskellunge Lake suckers , collections of 1934- and 1935
combined , sexes combined.
Per cent of annual increment
^Percentages omitted from averages; see text.
remain in the lake. In the last column of Table XIX, the 1934
increment of each year-class represented in the 1935 collection
is compared with the marginal growth of fish of the same year-
class captured in 1934. When the number of specimens upon
which each average is based is taken into consideration, it is
evident that the two methods of comparing total annual growth
with marginal growth gives approximately the same result.
To present a general summary of the growth during the
growing season, and to increase the number of specimens used
in the determination of each average, the data of Tables XIX
and XX have been combined in Table XXI. The averages shown
in the annual increment column of this table were derived by
the method employed for the corresponding column of Table
XX; the average 1933 growth of each age-group of the 1934
494 Wisconsin Academy of Sciences, Arts, and Letters
collection was combined with the average 1934 growth of the
same age-group of the 1935 collection, and compared with the
average marginal growth of corresponding age-groups of the
combined 1934 and 1935 collections. These tables show that
there is a progressive increase in the length of each age-group
from month to month, and that the value of the marginal growth
tends to approach that representing the full year’s growth as
the season advances. The few discrepancies which occur are
probably due to the small number of specimens upon which the
particular aberrant average is based. The data shown in Table
XXI are expressed as percentages of the corresponding annual
increments in Table XXII. Since the averages of the X-group
were quite irregular, and since the available data did not contain
information relative to the annual increment of the tenth year
of life, the X-group was omitted from Table XXII. The irregu¬
larities in the X-group fish are probably caused partly by the
small number of specimens, and partly by the difficulty in mak¬
ing accurate growth and age determinations from the scales
peripheral to the eighth or ninth annuli. In order to present a
summary of the seasonal growth of the population as a whole
(as represented by these age-groups), the percentages of each
month were averaged. These grand average percentages are
shown along the bottom line of the table. The IX-group per¬
centages were not included in the grand averages because of the
pronounced discrepancies occurring in this group. These dis¬
crepancies are due to the unusually low average annual incre¬
ment ; since this average increment is based upon but six speci¬
mens, it was eliminated in order to avoid giving undue signifi¬
cance to the percentages calculated from it. The grand averages
of the percentages of the annual increment gained through each
month of the growing season are based, therefore, upon data
taken from age-groups III to VIII.
As noted previously, the May collection was made on the
12th, the June collections from the 18th to 29th, and the October
collection on the 27th ; during the regular season, collections were
made throughout the months July and August, and during the
first two weeks of September. Accordingly, the percentages
represent the season’s growth to the middle of May, the latter
part of June, mid- July, mid- August, early September, and the
end of October. On the basis of the data given above, it may be
Spoor — Growth of Sucker
495
concluded that the growth during the period May to October is
completed as follows: 24% by the middle of May, 48% by the
latter part of June, 79% by mid- July, 92% by mid- August, and
97% by early September. Growth appears to be completed by
the end of October, but the full season’s growth is probably at¬
tained by the middle of September. Hile (1936) concluded that
the ciscoes of Muskellunge Lake completed their season’s growth
by the middle of July; they appear to grow most rapidly during
the first half of the season, namely, May to mid- July. Because
of the difficulty involved in determining the edge of the outer¬
most annulus on the scales of some of the specimens in the May
collection, there is reason to question the accuracy of the average
percentage for May. The October averages, although repre¬
sented by but few specimens, are probably acceptable since they
do not depart from the general trend. The remaining averages
appear to be well established.
Time of Spawning
The suckers spawn in the spring, but the dates appear to vary
somewhat with the locality. Reighard (1920) found that spawn¬
ing occurred in southern Michigan during April and early May.
Stewart (1926) notes that suckers in the region of Ithaca, New
York spawn from the end of April until the end of May. Adams
and Hankinson (1928) report that spawning occurs in the vicin¬
ity of Oneida Lake in early spring, soon after the ice leaves.
The suckers spawn in the latter part of April and early May in
southern Wisconsin. Since the ice does not leave the lakes of
northeastern Wisconsin until about the first of May, the season
seems to be somewhat later in this region. Apparently the
suckers of Muskellunge Lake spawn later than those previously
noted. Since this lake has no inlet, and no outlet except during
unusually high water, the suckers spawn in the lake; conse¬
quently, they could not be seen in the act of spawning. However,
it is possible to determine the approximate time of spawning by
observing the condition of the gonads of the fish in early season
collections, and, since the incubation period is known, by comput¬
ing back from the date on which young suckers are first found.
Suckers were seen spawning in a stream near Muskellunge
Lake on May 20, 1933. As noted earlier, suckers were captured
in Muskellunge Lake on May 12, 1935. Only one of the 23 fe-
496 Wisconsin Academy of Sciences, Arts, and Letters
males was mature. The eggs of this specimen appeared to be
fully developed, but they would not flow when pressure was
exerted; it was concluded, therefore, that the fish was not yet
ready to spawn. Most of the males collected were fully mature ;
the milt flowed freely and the pearl organs were well developed.
In the collections of June 18 to 28, 1984, some fish were ripe;
the sperm and ova flowed freely while the fish were being han¬
dled. Examination of the gonads showed that the majority of
the specimens in the collections were spent. It was further noted
that the genital apertures of mature specimens were inflamed;
at no other time of the year was this condition observed. A
number of females collected during the last week of June, 1934
and 1935, apparently had spawned, but the presence of minute
ova in the ovaries indicated that the spawning period had been
completed some time earlier. The spawning period, then, ap¬
pears to occur in June, apparently during the first part of the
month. Judging from the state of the ovaries of the mature
female captured on May 12, it seems probable that spawning
commences toward the end of May.
Stewart (1926) states that the incubation period of the
sucker lasts for 18 to 20 days ; the fish are about 8 millimeters
long when hatched. His records show that a length of 12 milli¬
meters is reached by the ninth day, 14 millimeters by the elev¬
enth day, and 16 to 18 millimeters by the twentieth day after
hatching. Therefore, the parents of young fish between the
lengths of 16 to 18 millimeters would have spawned 38 to 40 days
earlier than the date on which the young reached this length.
Young suckers were first seen in Muskellunge Lake on June 18,
1934, over a sand and pebble beach in water about 8 inches deep.
They were in large schools, .and were not over 10 or 12 milli¬
meters long. On June 26, 119 specimens of one school were
captured and measured. The lengths ranged between 12 and 21
millimeters, and averaged 16.5 millimeters. According to the
findings of Stewart (1926), the average age of these fish would
be 20 days, which would set the date of spawning 38 to 40 days
previous to this. The estimated date at which they were spawned,
therefore, is May 18 to 20. These dates agree fairly well with
those found by examination of adult specimens, and it may be
concluded that spawning occurs in Muskellunge Lake during the
latter part of May and early June.
Spoor — Growth of Sucker
497
Age at Maturity
Several investigators have recorded the size at which the
sucker reaches sexual maturity. Fowler (1912) found speci¬
mens three inches long with well developed milt and roe. Reig-
hard (1920) records that both males and females spawn when
about six inches long. Stewart (1926) agrees with Reighard’s
conclusion, and adds that the fish are four or five years old at
maturity.
In an effort to learn at what age Muskellunge Lake suckers
reach sexual maturity, the condition of the gonads was deter¬
mined for as many as possible of the sexed specimens. The
gonads are paired, and extend the length of the coelom laterad
the air-bladder. Those of the mature males are white, from two
to ten millimeters in thickness, convoluted, and of a cheesy con¬
sistency. Those of immature males are similar to those of the
adult in color and consistency, but are about a millimeter in
thickness, and instead of being convoluted they are straight. It
was noted that when the males were in spawning condition the
testes were strongly convoluted. After the spawning season,
these convolutions remained, but the testes were usually smaller
in diameter. Many of the smaller males had testes which were
but little over a millimeter in thickness, but which showed the
pronounced convolutions of the adult. It was impossible to de¬
termine whether these fish had spawned, but since the testes of
known adults were' always convoluted, this condition was con¬
sidered indicative of the approach of sexual maturity. There¬
fore, those males which had convoluted testes were classed as
mature, regardless of thickness; those which had straight and
small testes were classified as immature. In no case was a thick
testis found coupled with a straight condition.
The position of the ovaries of the female corresponds to that
of the testes of the male. The ovaries of mature females are
yellow, from three to fifteen millimeters in thickness, and almost
invariably contain ova, the size of which depends upon the sea¬
son. The gonads of very young females are similar to those of
the immature males in size and shape, but are semitransparent,
yellow to orange in color, and have the consistency of a gel. Fish
with such ovaries were termed “immature”. As the females
grow older, the ovaries become from three to five millimeters
thick, but remain semi-transparent and contain no ova; they
498 Wisconsin Academy of Sciences, Arts, and Letters
were classified as “clear”. Still later, the ovaries become opaque
and finely granular; these granules are ova. At this stage the
fish were classified under the heading VSE (very small eggs).
Those females whose ovaries were in the “immature” or “clear”
condition were considered immature, those with ova in any stage
of development were considered mature. That the appearance
of the clear condition in summer did not indicate that maturity
would be reached by the following spring is shown by the fact
that most of the 22 immature females captured on May 12, 1935,
shortly before the spawning season, had clear ovaries. The class¬
ification of both males and females was based almost entirely
upon macroscopic examination.
The males and females of the combined collections of 1932 to
1935 are arranged according to age and degree of maturity in
Table XXIII. If the age at which 50 per cent of the specimens
reach maturity is considered as the average age at sexual ma¬
turity, it can be seen that in general the males mature one year
earlier than the females, during the fifth and sixth years, re¬
spectively. 75 per cent of the males become mature during the
fifth year, but not until the seventh year does this percentage of
the females reach maturity. Since the Ill-group contains 7 per
cent mature females, and no mature males, it appears that cer¬
tain of the females mature before the males. Since the age-
groups after the eighth are not well represented, percentages
shown for the ninth, tenth and eleventh years do not have great
significance. Judging from the trends of the percentages, it
appears that about 10 per cent of the population fails to become
Table XXIII
Number of immature and mature males and females, and per cent .
mature, in each age-group of Muskellunge Lake suckers collected during the
period 1932 to 1935.
Spoor— Groivth of Sucker
499
sexually mature by the end of the eighth year of life. Inasmuch
as the relative abundance of the older age-groups in the collec¬
tions indicates that the normal span of life is not more than eight
or nine years, this is of considerable interest. It may be, how¬
ever, that all suckers do not spawn each year after sexual ma¬
turity has been reached.
Sex Ratio
No observations regarding the sex ratio of the suckers have
been found in the available literature. Since information bear¬
ing upon this subject is to be found among the records of the
collections made in Muskellunge Lake from 1932 to 1935, an
analysis of the sex ratio has been attempted.
It will be noted (section on “materials”) that for the sexed
specimens the males and females were taken in about equal
numbers each year, and that the total number of males (1068)
is practically identical with the total number of females (1065).
This implies that the males and females are present in equal
numbers in the population as a whole, and that the sex ratio,
therefore, is 1 :1.
Since some of the scales could not be read, it was impossible
to determine the ages of some of the specimens. Of the 1058
males, the ages were determined for 967, and of the 1065 fe¬
males, the ages of 972 were determined. The ratio between all
males and females of which age was determined remains about
1 :1. This ratio is not constant at all ages, however. The speci¬
mens are arranged according to sex, age-group, and year of cap¬
ture in Table XXIV in which the number of specimens of each
sex in each age-group in the combined collections is also shown.
The relative number of females in each age-group is expressed
in the form of percentage. In the per cent column of the com¬
bined collections, it is evident that the females are more abun¬
dant than the males (58 per cent for females and 42 per cent for
males) during the third and fourth years of life. During the
fifth, sixth, and seventh years the condition is reversed; the
males are more numerous than the females. In the eighth year
the females appear to be more abundant than the males, and
while the IX and X groups are not well represented, there seems
to be a tendency for the females to maintain their numerical
superiority in the later years of life.
500 Wisconsin Academy of Sciences, Arts, and Letters
Since the differences between the numbers of males and fe¬
males are not great, it is possible that the apparent changes in
relative abundance are due to chance sampling. That this is
not the case is indicated by the fact that the same general trend
is shown when each of the collections is considered alone. The
females predominate among the younger age-groups and again
among the older age-groups in each collection. This does not
rule out entirely the factor of chance sampling, of course, but it
does indicate that the percentages express the actual condition.
Table XXIV
Number of male and female suckers and per cent female in each age-
group in the collections of 1932, 1933, 193J, \, 1935.
It is to be noted that when the age-groups of each sex are
combined so that the entire population is considered, the ratio
remains approximately 1 : 1 from year to year. The factors
underlying the changes in the relative abundance of males and
females as age increases are not clear, nor do the available data
furnish a probable explanation. Hile (1936) found that a simi¬
lar change occurs among the ciscoes of Muskellunge Lake and
other lakes in the vicinity, and presented evidence to show that
it is caused by a differential mortality of the two sexes.
Spoor — Growth of Sucker
501
Summary
1. This study of the age and growth of the suckers of Muskel-
lunge Lake is based on data taken from 3,697 specimens collected
during the period 1927 to 1935. Age and growth determinations
were made from the scales of 2,990 of these specimens.
2. The scale method was found to be applicable to a study of
the age and growth of the sucker because (a) the scales remain
constant in number and persist throughout life, (b) the scales
increase in size as the fish increases in length and (c) a dis¬
tinct annulus is formed each winter.
3. The annulus is characterized by a crowding of the circuli
and by “cutting over” as they approach the posterior quadrant
of the scale. It is formed between October and May.
4. The scale of the sucker can be measured either along the
anterior radius or along the dorso-ventral axis. Since the annu¬
lus is more distinct in the dorsal and ventral quadrants, and
since the length of the fish appears to be more closely correlated
with scale width than with the length of the anterior radius,
measurements were made along the dorso-ventral axis of the
scale.
5. The scales of the sucker form when the fish is from 23 to
25 millimeters long. For a short period after they are formed
the scales grow at an accelerated rate, so that by the time the
body length reaches 30 to 35 millimeters they overlap in a man¬
ner similar to that found in the adult. Thenceforth they increase
in size at about the same rate at which the fish increases in
length. Because of this acceleration of growth until the scales
overlap, no correction was added to the scale formula to com¬
pensate for late scale formation.
6. Lee’s phenomenon of apparent change in growth rate is
quite pronounced among the calculated lengths of the younger
age-groups. This phenomenon appears to be due chiefly to the
selective action of the gear used in making the collections.
7. The gill nets and fyke nets have a pronounced selective
action. Fish of less than 113 millimeters standard length could
not be captured in the gill nets, and the fyke nets allowed speci-
502 Wisconsin Academy of Sciences , Arts , and Letters
mens less than 150 millimeters in length to escape. For this
reason no I-group fish, and only the largest of the Il-group fish
are represented in the collections made with these nets.
8. The average calculated lengths of fish three or more years
old are in agreement with the corresponding average actual
lengths.
9. Scales from all parts of the body show the same number
of annuli and growth calculations from these scales are practi¬
cally identical.
10. The growth in length of Muskellunge Lake suckers, based
upon the averages of the calculated lengths for each year of life
of 2,990 specimens, is as follows: First year, 61 millimeters;
second year, 100 millimeters; third year, 138 millimeters;
fourth year, 173 millimeters; fifth year, 197 millimeters; sixth
year, 222 millimeters; seventh year, 245 millimeters; eighth
year, 262 millimeters ; ninth year, 292 millimeters. The annual
increment tends to decrease as the fish grow older.
11. The females grow in length more rapidly than the males
after the fifth winter. The average difference for the sixth
year is 4 millimeters, 8 millimeters for the seventh year, and 11
millimeters for the eighth year. The females seem to maintain
a higher rate of growth beyond the eighth year, but specimens
with more than eight annuli were so few that the averages can¬
not be considered significant.
12. The estimated average weights at the end of each year of
life are: First year, 4 grams; second year, 18.5 grams; third
year, 47.5 grams; fourth year, 77.5 grams; fifth year, 130
grams; sixth year, 177.5 grams; seventh year, 244 grams;
eighth year, 320 grams; ninth year, 401 grams.
13. The males and females increase in weight at the same
rate per unit of length, but since the females outgrow the males
in length after the fifth winter the females are correspondingly
heavier than males of the same age at the end of each year after
the fifth.
14. The weight of the suckers of Muskellunge Lake increases
as the cube of the length.
Spoor — Growth of Sucker
503
15. There is some evidence that the suckers are in better
condition in the middle of the growing season than they are
early in the season.
16. The season's growth appears to begin early in May, and
is completed as follows : 24 per cent by the middle of May, 48
per cent by the latter part of June, 79 per cent by mid- July, 92
per cent by mid-August, and 97 per cent by early September.
Growth appears to be completed by the end of October, but the
full season's growth is probably attained by the middle of Sep¬
tember.
17. Growth appears to proceed most rapidly during the first
half of the growing season.
18. The suckers spawn in Muskellunge Lake during late May
and early June.
19. The average age at sexual maturity is five years for the
males and six years for the females. Apparently about 10 per
cent of the population fails to reach sexual maturity by the end
of the eighth year of life.
20. Males and females seem to be present in the population
as a whole in equal numbers, but the sexes are not equally repre¬
sented in each age-group. The females are most numerous dur¬
ing the third and fourth years; the males predominate during
the fifth, sixth and seventh years, and from the eighth year on
the females are again more abundant than the males.
Literature
Adams, Charles C. and T. L. Hankinson. 1928. The ecology and economics
of Oneida Lake fish. N. Y. State Col. Forestry. Roosevelt Wild Life
Annals 1:241-548.
Bigelow, N. K. 1923. The food of young suckers (Catostomus commersonii)
in Lake Nipigon. Univ. Toronto Stud., Biol. Ser. 24. Pub. Ontario
Fish. Res. Lab. 21:81-115.
Clark, F. N. 1928. The weight-length relationship of the California sar¬
dine (Sardina caerulea) at San Pedro. Div. Fish and Game of Cal¬
ifornia, Fish Bull. 12:1-59.
Clark, F. N. 1929. The life history of the California jack smelt, Atherinop-
sis calif omiensis. Div. Fish and Game of California, Fish Bull.
16:1-23.
504 Wisconsin Academy of Sciences, Arts, and Letters
Couey, F. M. 1935. Fish food studies of a number of northeastern Wisconsin
lakes. Trans. Wis. Acad. Sci., Arts and Let. 29:131-172.
Crawford, D. R. 1923. The significance of food supply in the larval de¬
velopment of fishes. Ecology, 4:147-153.
Creaser, C. W. 1926. The structure and growth of the scales of fishes in
relation to the interpretation of their life history, with special refer¬
ence to the sunfish, Eupomotis gibbosus. Univ. Mich., Mus. Zool.,
Misc. Pub. 17:1-82.
Cross, S. X. 1934. A probable case of non-specific immunity between two
parasites of ciscoes of the Trout Lake region of northern Wisconsin.
Jour. Parasitol. 20 (4) :244-245.
Cross, S. X. 1935. The effect of parasitism on growth of perch in the Trout
Lake region. Jour. Parasitol. 21 (4) : 267-273.
Crozier, W. J. and S. Hecht. 1914. Correlations of weight, length, and
other body measurements in the weakfish, Cynoscion regalis. Bull. U.
S. Bur. Fish. 33:139-147.
Embody, G. C. 1915. The farm fishpond. Cornell Reading Courses. Country
Life Ser. 3:213-252.
Fowler, H. W. 1912. Some features of ornamentation in freshwater fishes.
Amer. Nat. 46:470-476.
Greene, C. W. 1935. The distribution of Wisconsin fishes. Wis. Cons.
Comm. 235 pp.
Hart, J. L. 1931. The growth of the whitefish, Coregonus clupeaf ormis
(Mitchill). Contrib. Canadian Biol, and Fish., N. S. (20) : 429-444.
Hecht, S. 1916. Form and growth in fishes. Jour. Morph. 27(2) :379-400.
Hile, Ralph. 1931. The rate of growth of fishes of Indiana. Investigations
of Indiana lakes, No. 2. Dept. Cons., Div. Fish and Game, Indiana.
Pub. 107:9-55.
Hile, Ralph. 1936. Age and growth of the cisco, Leucichthys artedi (Le
Sueur), in the lakes of the northeastern highlands, Wisconsin. Bull.
U. S. Bur. Fish. 48(19) :211-317.
Hubbs, C. L. 1926. A checklist of the fishes of the Great Lakes and tribu¬
tary waters, with nomenclatorial notes and analytical keys. Univ. Mich.,
Mus. Zool., Misc. Pub. 15:1-77.
Hubbs, C. L. and C. W. Creaser. 1924. On the growth of young suckers and
the propagation of trout. Ecology, 5:372-378.
Juday, C. and G. W. Bennett. 1935. Growth of game fish in Wisconsin
waters — third report. Notes from the Limnological Lab. Wis. Geol.
and Nat. Hist. Survey. (Mimeographed).
Juday, C. and E. A, Birge. 1930. The highland lake district of north¬
eastern Wisconsin and the Trout Lake Limnological Laboratory. Trans.
Wis. Acad. Sci., Arts and Let. 25:337-352.
Juday, C. and E. Schneberger. 1930. Growth studies of game fish in Wis¬
consin lakes. Notes from the Biol. Lab. Wis. Geol, and Nat. Hist, Sur¬
vey (Mimeographed ) .
Spoor — Growth of Sucker
505
Juday, C. and E. Schneberger. 1933. Growth studies of game fish in Wis¬
consin waters — second report. Notes from the Limnological Lab. Wis.
Geol. and Nat. Hist. Survey. (Mimeographed).
Keys, A. B. 1928. The length-weight relation in fishes. Proc. Nat. Acad.
Sci. 14:922-925.
Lee, R. M. (Mlrs. T. L. Williams) 1920. A review of the methods of age and
growth determination in fishes by means of scales. Min. Agric. and
Fish., Fish. Invest., Ser. II, 4(2) : 1-32.
Reighard, J. 1920. The breeding behavior of the suckers and minnows.
Biol. Bull. 38:1-32.
Schneberger, E. 193-5. Growth of the yellow perch (Perea flavescens
Mitchill) in Nebish, Silver and Weber Lakes, Vilas County, Wisconsin.
Trans. Wis. Acad. Sci., Arts and Let. 29 :103-130.
Schultz, L. P. 1933. The age and growth of Atherinops affinis oregonia
Jordan and Snyder and of other subspecies of baysmelt along the
Pacific coast of the United States. Univ. Wash. Pub. in Biol. 2(3) :45-
102.
Stewart, N. H. 1926. Development, growth, and food habits of the white
sucker, Catostomus commersonii Le Sueur. Bull. U. S. Bur. Fish.
42:147-184.
Tester, A. L. 1932. Rate of growth of the small-mouthed black bass ( Microp -
terns dolomieu) in some Ontario waters. Univ. Toronto Stud., Pub.
Ontario Fish. Res. Lab. 47:207-222.
Van Oosten, J. 1929. Life history of the lake herring (Leucichthys artedi
Le Sueur) of Lake Huron as revealed by its scales, with a critique of
the scale method. Bull. U. S. Bur. Fish. 44:265-428.
Wright, S. 1929. A preliminary report on the growth of the rock bass,
Ambloplites rupestris (Rafinesque), in two lakes of northern Wisconsin.
Trans. Wis. Acad. Sci., Arts and Let. 24:581-595.
A SECOND REPORT ON THE GROWTH OF THE MUSKEL-
LUNGE, Esox masquinongy immaculatus (Garrard),
IN WISCONSIN WATERS
Clarence L. Schloemer
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey, Notes and reports No. 77.
INTRODUCTION
For the past nine years much of the data on the growth of
Wisconsin’s important species of game fish have come from
fishermen interested in learning the ages of the fish they caught
in the state’s inland waters. These data include such informa¬
tion as length, weight, sex of the fish, as well as, scales and the
date and locality from which the fish were taken. With these
facts fishery biologists are able to make quite an accurate esti¬
mate of the age of a given fish and are able to formulate corre¬
lations between length and weight and between age and length
which are valuable to an understanding of some of the life his¬
tory features of any species considered. Furthermore, fishery
investigators have found that the length of a fish during any
year of its past life can be determined from measurements of a
scale and a few simple calculations, provided the scale used is
not regenerated or injured in any way and the growth of the
scale is fairly proportional to the growth of the body in length.
This scale method, as it is called, has proven to be a most valu¬
able aid to scientific investigators interested in understanding
the growth and abundance of species of commercial importance.
Only relatively recently however, have fishery biologists begun
to utilize it as a tool in the “sport fishing” conservation pro¬
gram. By its use it is possible to establish the rate of growth
of a species of game fish in a given body of water ; consequently
lakes can be compared with each other as regards their ability
to support a population of game fish. Eschmeyer (1936) has
summarized this newer approach in lake management in his
statement that “since the rate of growth of fishes is dependent,
within limits, upon the amount of food available for the fish,
507
508 Wisconsin Academy of Sciences , Arts, and Letters
stocking intensity may be determined on the basis of growth
without an intensive study of the food or of the concentration of
fish in a lake. Based on a periodic study of fishing intensity
and on a continuous study of growth, stocking may eventually
be placed on a scientific foundation.”
Growth of Muskellunge
It was with this idea in mind that work was begun on a study
of the growth of the muskellunge two years ago. The main body
of the results of this study appeared in published form last year
(Schloemer 1936). The study showed that the muskellunge
grows most rapidly during the first three years of life, after
which time its rate of growth gradually decreases. The growth
calculations based on 351 specimens showed that the average
length of the muskellunge is 7.8 inches at the end of the first
year of life, 16.0 inches at the end of the second year and 22.8
inches at the end of the third year. They also showed that the
average muskellunge attains the legal length of 30 inches during
its fifth summer of life. After the muskellunge is eleven years
of age its increase in length is about 1 inch per year for the next
several years. The study indicated that the muskellunge actu¬
ally never stops growing in length although at advanced ages
the annual increase is very slight.
At the time legal length is attained the muskellunge weighs
from 6 to 9 pounds. During the next several years the increase
in weight of the average muskellunge runs parallel to the in¬
crease in length. The data show, for instance, that the average
weight of a 7 year old muskellunge, measuring 38 inches from
tip to tip, is approximately 15 pounds and that with an addi¬
tional year's growth 2 inches are added to its length and 2 pounds
to its weight. After the 10th summer, however, this parallelism
between growth in length and weight changes. For the next
few years the muskellunge increases over twice as much in
weight as it does in length, while in the later years of life the
annual increase in weight is altogether disproportionate to the
yearly increment in length. An extremely old muskellunge may
grow but a fraction of an inch in a growing season, but it may
add as much as 3 pounds to its weight in this time.
Schloemer— Growth of Muskellunge
509
Growth in Different Drainage Areas
In the previous study a comparison was made between the
growth of muskellunge in 12 Wisconsin lakes grouped into three
drainage areas ; namely, the Chippewa, Court Oreilles and Man -
itowish. These lakes together with 3 additional lakes from the
Manitowish drainage area and also Pelican and Big Arbor Vitae
lakes and the Wisconsin River are presented in Table I where
they are arranged according to the time when muskellunge taken
from them reached the legal length of 30 inches. The “average”
indicated in the chart refers to the average time when legal
length was attained by the 351 muskellunge mentioned before.
It should be understood that the number of specimens upon
which these growth curves for a given body of water are based
are too few to make this classification of much practical value
at the present time ; nevertheless, the table shows the extent of
variation that exists beween muskellunge from different lakes
as concerns the average time when they reach their legal size.
The rates of muskellunge growth in the 3 additional lakes in
the Manitowish drainage, as well as the average growth of 5
individuals from the Wisconsin River are presented in Figure 1.
Table I
Classification of some lakes in Sawyer and Vilas counties based on the
time when muskellunge taken from these lakes attain legal length.
510 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 1. Comparative growth rates of muskellunge in 3 lakes of the
Manitowish drainage area and in the Wisconsin River.
As was indicated previously in the table, the Wisconsin River
specimens show a faster rate of growth than do muskellunge
from all lakes with the exception of Grindstone. A detailed com¬
parison of the growth of the Wisconsin ‘River specimens with
the growth of those from Lake Muskellunge shows that the indi¬
viduals from the former maintain a fairly consistent higher
rate of growth than do those of the latter. The greatest differ¬
ence in growth between specimens from these two bodies of wa¬
ter occured in the 1st and 2nd years and amounted to almost
3 inches. This would indicate that the conditions for growth of
Schloemer — Growth of Muskellunge
511
young muskellunge are more favorable in the Wisconsin River
than they are in Lake Muskellunge. Generally, the specimens
from Lake Muskellunge show better growth than do specimens
from any other lake considered in the Manitowish drainage area.
In the previous report on the growth of muskellunge in lakes
belonging to this drainage system, the specimens from High
Lake were shown to have a better rate of growth than the speci¬
mens from Clear, Spider and Island lakes. In comparing the
rate of growth of specimens from High Lake with specimens
from Big Lake no great difference is noticeable until after the
6th year of life when the individuals from Big Lake begin to
grow much more rapidly than those from High Lake. If the
growth of these Big Lake specimens is compared with the
growth of the muskellunge from Allequash and Muskellunge
lakes, however, we find that the average growth of individuals
from these other lakes of the Manitowish drainage area is much
greater than the growth of the Big Lake specimens, with the
exception of the Allequash Lake muskellunge after the 6th year
of life. The greatest difference in growth between individuals
from Allequash and Muskellunge lakes occurred during the 7th
year and amounted to 3 inches. The average time when the
specimens from all of the lakes reach legal length is indicated in
Figure 1.
Since Big Arbor Vitae and Pelican lakes are drained by rivers
other than the Manitowish, they cannot be included in the Mani¬
towish group. The growth of muskellunge from these two lakes
has been studied and the study shows that both of them are
capable of supporting a fast growing stock of muskellunge. Peli¬
can Lake is of especial interest for from it was taken a 19 year
old specimen, the oldest one on record. The heaviest and longest
muskellunge on record was taken from White Sand Lake in the
Lac du Flambeau region. It measured 57 inches from tip to tip
and weighed 52 pounds. It was in its sixteenth summer.
This study of muskellunge growth shows the kind of informa¬
tion that can be derived from data sent in by fishermen. With in¬
creased cooperation more data will accumulate and consequently
the growth curves of muskellunge from various lakes through¬
out the state will become more reliable because they will repre¬
sent a better sample of the muskellunge population in a given
body of water. Fishermen can also furnish information con-
512 Wisconsin Academy of Sciences , Arts , and Letters
cerning some of the food habits of the muskellunge which in¬
vestigators would find difficult to obtain. Some of the most inter¬
esting facts regarding the food of muskellunge have come from
this source; for example, one fisherman reported catching a
muskellunge with a full grown muskrat in its stomach. In
another instance, a 37 inch muskellunge was caught having a
21 inch muskellunge in its alimentary tract. It is observations
such as these that will complete the picture of the kinds of food
utilized by the “tiger” muskellunge.
Bibliography
Eschmeyer, R. W. 1936. Essential Considerations for Fish Management in
Lakes. Proc. of N. Amer. Wildlife Conference, 1936:332-339.
Schloemer, C. L. 1936. The Growth of the Muskellunge, Esox masquinongy
immaculatus (Garrard), in Various Lakes and Drainage Areas of
Northern Wisconsin. Copeia, 4, 1936: 185-193.
GROWTH OF THE BUFFALO IN WISCONSIN LAKES
AND STREAMS
David G. Frey
University of Wisconsin
and
Hubert Pedracine
Works Progress Administration
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 78.
Introduction
In October, 1935, the Wisconsin Geological and Natural His¬
tory Survey undertook the study of the growth of rough fish in
Wisconsin lakes and streams as W.P.A. project No. 5590. Aided
by the Wisconsin Conservation Commission, four W.P.A. work¬
ers visited regularly the lakes and streams being seined by com¬
mercial fishermen and by state seining crews, and obtained scales
and measurements of the various species of rough fish hauled in.
The region covered by these operations extended from Lake Lit¬
tle Butte des Morts on the north to Lake Koshkonong and the
Crawfish River on the south. Time of collecting included the
fall of 1935 and the spring and fall of 1936.
Three scales were taken from the left side of the body from a
region even with the anterior edge of the dorsal fin and two or
three rows above the lateral line. Also from each fish there was
obtained the standard length in millimeters from the tip of the
nose to the base of the tail, and in the fall of 1935 the weight in
grams. These measurements were recorded on the envelopes
in which the scales from the fish were placed.
Assuming, as has been shown for other species of fish, that
the length of the scale increases in proportion to the increase
in body length, calculations were made of the length of the fish
at the end of each growing season. These calculations form the
basis of the present report.
513
514 Wisconsin Academy of Sciences , Arts, and Letters
Scales and measurements were obtained from carp, buffalo,
sheepshead, suckers, and several other species of rough fish of
lesser importance and represented by only a few specimens in
the collection at the present time. The present paper deals with
the buffalo.
Growth of the Buffalo
Not until the project was well under way was it realized that
the commercial term “buffalo” included three species of rough
fish — Megastomatobus cyyrinella (Cuvier and Valenciennes),
the big-mouthed buffalo; Ictiobus niger (Rafinesque) , the black
buffalo; and Ictiobus bubalus (Rafinesque), the small-mouthed
buffalo. An examination of the scales from the different species
showed that the growth rate was nearly the same and, more¬
over, that the scale structure was almost identical. Rather than
discard this first material which includes the bulk of the collec¬
tion, all three species have been lumped together in this report
and will henceforth be designated by the rather indefinite name
“buffalo”. It was not possible to determine the numbers of the
three species of buffalo in the collection. Megastomatobus cy->
prinella occurs in small numbers because of its more southern
distribution, while Ictiobus niger and Ictiobus bubalus , probably
occurring in nearly equal numbers, make up the bulk of the col¬
lection.
The collection includes 1364 buffalo from 4 lakes and 6 rivers.
All fish were taken from the hauls of the commercial fishermen
Table I
Distribution of the buffalo in the collection as regards locality and
year of collecting.
Frey & Pedracine — Growth of Buffalo Fish 515
with the exception of the buffalo from Bass Lake; they were
fish which had been killed by the cold winter of 1935-1936 and
cast up on shore. 1130 of the fish have been used in this report.
151 buffalo taken from Lake Koshkonong in the fall of 1936 at
the end of their fourth season of growth were not measured be¬
cause their growth was found to be the same as that of the other
141 four year olds. Only 83 specimens or 6 per cent of the fish
could not be used because of regeneration injury or uncertainty
in markings. Table I shows the number of specimens and the
streams and lakes from which they were taken.
All the data for both years when averaged and plotted give
a very uniform growth curve (Figure 1). It is seen from the
graph that the growths during the first two years are nearly the
Fig. 1. Growth curve and growth increment curve for 1130 buffalo
caught during 1935 and 1936.
516 Wisconsin Academy of Sciences , Arts, and Letters
same, the second year's growth being slightly the larger by
4 mm. After the second year of growth the increment for each
following year as far as was accurately ascertainable became
progressively and regularly smaller. This is best shown by the
increment curve in Figure 1. Each point on the curve up to and
including the fourth year is calculated from more than 300 speci¬
mens.
Table II
Calculated lengths at the end of each year of growth for 1130 buffalo
caught during the years 1935 and 1936. All lengths are standard lengths
in millimeters.
Table II gives a summary of the data from which the growth
curve and the growth increment curve were plotted. If age
group VII is left out of consideration because of being repre¬
sented by only 2 specimens, the calculated lengths for the various
age groups of buffalo at the end of any growing season do not
differ very much. The two year old fish show the most rapid
rate of growth ; they are well above the average of the older fish
for both the first and second years. It is interesting that the
five year old fish show the next most rapid rate of growth. This
was found to be the case not only in the grand average but also
in the averages for 1935 and 1936. Judging from this growth
chart there is apparently no marked decrease in calculated
length as the fish become older.
Figure 2 represents the growth of the buffalo caught during
the fall of 1935. Several points of interest are evident from the
graph. In the first place the growth rate of the buffalo for any
lake or river is markedly constant for the first two years ; none
of the lines cross during the second season of growth except
Frey & P e dr acine— Growth of Buffalo Fish
517
Fig. 2. Growth curves for 459 buffalo caught during the fall of 1935.
those of Lake La Belle and Crawfish River. The buffalo from
three localities, Lake Monona, Lake Koshkonong, and the Craw¬
fish River, continue to increase during the third growing season
at only a slightly decreased rate; whereas, fish from the other
three localities, Horicon Marsh, Sugar River, and Lake La Belle,
all show a marked decrease in growth rate. In the fourth grow¬
ing season Lake Koshonong and Crawfish River, which grow at
nearly the same rate from the end of the first year, show a strong
decline in growih rate, with a sudden spurt during the fifth
growing season.
The growth curves for the 1936 buffalo, Figure 3, appear to
be more irregular than those of 1935, but most of this irregular¬
ity can be attributed to the presence of older fish from more
lakes in the 1936 collection, and also to the fact that the 1936 col¬
lection is composed of fish from more localities.
Several interesting facts arise from these curves. Buffalo
from four bodies of water — Lake Koshkonong and the Crawfish,
Wisconsin, and Rock rivers — all have a very similar growth dur¬
ing the first year. Lake Koshkonong and the Crawfish River
518 Wisconsin Academy of Sciences , Arts, and Letters
buffalo have an identical growth through the third year, but
during the fourth and fifth growing seasons the Crawfish River
buffalo slacken considerably in their rate of growth. One thing
that is very apparent is that fish from Bass Lake consistently
show a smaller length at the end of any year than fish from the
other bodies of water.
Fig. 3. Growth curves for 671 buffalo caught during the spring and
fall of 1936. Only the Bass Lake and Rockdale Creek specimens were taken
in the spring.
Comparing the curves for the two years one sees that the
range in length at the end of the first and second seasons of
growth is almost identical for the 1935 and 1936 buffalo. At the
end of the third year the range for the 1936 fish is much less
than that for the 1935 fish. This is brought about by the sudden
decrease in growth rate of the Lake La Belle and Rockdale Creek
buffalo, which during the first two years exhibit the same growth
as the buffalo from Lake Monona and Horicon Marsh, but dur¬
ing the third year have a much slower growth rate.
Undoubtedly some of these curves are complicated to a cer¬
tain extent by the few old fish from each lake which often show
Frey & Pedracine — Groivth of Buffalo Fish
519
Fig. 4. Growth curves for 264 four year old buffalo caught during 1936.
an abnormally small or abnormally large growth as calculated
from the annuli present on the scale. A much better picture of
the growth rate than that given by the curves for all the 1935 or
1936 buffalo can be obtained from the growth curves of the four
year old buffalo caught during 1936. (Figure 4). The four
year olds include one half of the buffalo caught that year. Their
growth rates are very similar— a fact unusual not only because
one would expect noticeably different growth rates in the various
lakes and streams due to different environmental conditions, but
also because the word “buffalo”, as used in this report, includes
three different species of fish. The only bodies of water showing
any departure from this uniformity are Wisconsin River during
the first year, Lake La Belle during the first and second years,
and Bass Lake during the third and fourth.
Table III shows the calculated lengths of these fish. At the
end of the first year the buffalo from Lake La Belle were unusu-
520 Wisconsin Academy of Sciences , Arts, and Letters
Table III
Calculated growths for the four year old buffalo caught during 1986.
ally large and those from Wisconsin River unusually small, but
the rest of them showed a variation in length of only 3 mm. At
the end of the second year, excluding Lake La Belle, the varia¬
tion was 20 mm. At the end of the third and fourth years, omit¬
ting Bass Lake in both instances, the variations were 17 and 18
mm. respectively. These fish were all about 385 mm. (15 inches)
in length. It is remarkable that the variation in length among
these fish from five different bodies of water should be only 18
mm. (% inch) or 5 per cent of the length. It is also apparent
that the average calculated length for these four year old buffalo
corresponds very closely to those for the entire buffalo popula¬
tion. (See Tables II and III.) The growth of the four year
olds is the same as the general average at the end of the first
year, 3 mm. below the average at the end of the second, 8 mm.
above the average at the end of the third, and 2 mm. above the
general average at the end of the fourth year.
Three bodies of water — lakes Koshkonong, La Belle, and the
Crawfish River — are represented by specimens both from 1935
and 1936. A very significant fact is best brought out by Table
IV. It is evident that in the 1935 collections most of the fish
were in their third year of growth, and that in 1936, one year
later, the modal class comprised fish in their fourth year. The
surprising thing is that although these collections were made as
randomly as possible from the fishermen's catches, there are
such low frequencies in the adjacent age groups. Thus, in 1935,
192 fish from Lake Koshkonong were three years old, in 1936
there was not one three year old among the 146 fish included in
Frey & Pedracine — Growth of Buffalo Fish 521
this report. In 1935 only three year olds were taken from Lake
La Belle ; in 1936 there were 3 three year olds, but there were
also 52 four year olds. The comparison in the Crawfish River
collections is not so striking because so few buffalo were col¬
lected in 1936 ; however the distribution gives the same picture.
Table IV
Age-frequenoy distribution of buffalo in three localities for 1935 and 1936.
Data from the other lakes are of much the same nature. Thus
in 1935 collections, every lake had its greatest frequency in the
three year old group, and three localities — lakes Monona, La
Blelle, and Horicon Marsh — are represented solely by three year
old fish. In the 1936 collections Bass Lake and Rock River, in
addition to lakes Koshkonong, La Belle, and the Crawfish River,
have the greatest frequency in the four year old group ; whereas
Rockdale consists almost entirely of two year olds and the Wis¬
consin River of two and three year olds.
There is apparently no selective factor of importance to be
considered here in the sampling. For any one lake the hauls
were made each year in the fall by the same commercial fisher¬
men using the same equipment. There seems to be fairly good
evidence then that the buffalo probably run in cycles. Appar¬
ently at quite regular intervals an extra large number of fish
hatch out and dominate the population until the hatching of the
next large crop. There is nothing to indicate how far apart
these large hatches are, except that the large growths in age
groups II and V of Table II may indicate that they come every
third year. Clarification or negation of this idea must await
data from several more years of collecting.
There is apparently no great difference in growth rate be¬
tween river buffalo and lake buffalo. From Figure 5 it can be
seen that the growth curves of the two groups of fish represent-
522 Wisconsin Academy of Sciences, Arts, and Letters
Fig. 5. A comparison of the growth curves for the lake and river buffalo
caught during 1935 and 1936, and the Crawfish and Sugar river buffalo
which were seined from running water.
in g the combined 1935 and 1936 material cross each other quite
often and at no time are widely separated. No one type of lo¬
cality as judged from these curves can be said to be most favor¬
able for the growth of the fish. There is to be considered the
fact, however, that most of the buffalo listed as coming from
rivers were actually seined from millponds or embayments. In
only two cases were the buffalo seined from running water.
Crawfish River and Sugar River. Figure 5 shows that the growth
rate of the buffalo from these two localities is almost consist¬
ently slower than the growth rates of the buffalo in the lakes or
in all the streams. This may possibly indicate that the growth
rate of buffalo in running water is slower than that for buffalo
in standing water.
Data which might be used to determine the constancy of the
calculated growth as the fish increase in age is rather conflicting.
Again the best evidence is found in the modal classes from lakes
La Belle, Koshkonong, and the Crawfish River. (Table V).
Lake Koshkonong buffalo have a smaller calculated growth in
1936 than in 1935; this might be indicative of Lee’s phenome¬
non among the buffalo were it not that the Crawfish River and
Lake La Belle fish both show an increase in the calculated length
Frey & Pedracine — Growth of Buffalo Fish
523
Table V
Comparison of calculated growth rates of buffalo caught in 1935 and 1936.
cf 1936 over that of 1935. There have been found no reasons to
explain the great discrepancies in the growth calculations of the
Lake La Belle buffalo for 1935 and 1936.
The relationship between length and weight in the buffalo is
shown by Figure 6, which is the length-weight curve for the
1935 buffalo on which has been superposed the average lengths
of all the buffalo at the end of the various years of growth (Table
II), and the weights which fish of these lengths are likely to
have. This curve shows that the ratio of length to weight in the
buffalo is not constant but is continually becoming smaller as
Fig. 6. Length-weight curve for 459 buffalo caught during the fall
of 1935.
524 Wisconsin Academy of Sciences , Arts, and Letters
the fish grow older. An increase in length of 10 mm. in the
fourth year is accompanied by a much greater change in weight
than an increase in length of 10 mm. during the first year. Hence
it is that a length-weight ratio of 1:2 at the end of the second
year drops to 1 :4 by the end of the fourth year.
Fig. 7. Length-weight curves for the river and lake buffalo caught
during the fall of 1935.
A comparison of the length-weight curves for buffalo col¬
lected in the fall of 1935 (Figure 7) shows that there is no
marked difference in the length-weight ratio between buffalo
caught in rivers or lakes. The lake buffalo were heavier than
the river buffalo during the latter part of the second and most
of the third years, but the scarcity of old fish in the 1935 collec¬
tion makes accurate comparisons beyond this impossible. The
point in the upper right hand corner of the graph is based on
only 4 river buffalo. While this point itself may not be accu¬
rate with respect to the entire population, yet it helps to indicate
that for a given length the river buffalo are lighter than the lake
buffalo. These differences are more significant than differences
in the age-growth curves because all except 29 of the 1935 river
buffalo are from the Crawfish and Sugar rivers ; the remaining
29 buffalo from Horicon Marsh while not seined from running
water nevertheless show the same length-weight relationships
as the other river buffalo.
Frey & Pedraeine — Growth of Buffalo Fish
525
Summary
1. This report deals with three species of fish considered to¬
gether under the term “buffalo”.
2. The growth of the buffalo during the first two years is nearly
the same, 116 and 120 mm. respectively, and after that de¬
clines gradually and regularly.
3. The growth and growth rates of buffalo from different bodies
of water are remarkably similar.
4. There is some indication that buffalo living in running water
grow more slowly in length and in weight with relation to
length than buffalo in standing water.
5. There seems to be fairly good evidence that the buffalo may
possibly run in cycles, with the good seasons coming every
third year.
A CENSUS OF THE FISH CAUGHT BY ANGLERS
IN LAKE KEGONSA
Chancey Juday
University of Wisconsin
and
Lawrence E. Vike
Works Progress Administration
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and reports No. 79.
Introduction
During the summer and autumn of 1936, Mr. Lawrence E.
Yike was employed by the Works Progress Administration as a
part of Work Project No. 5590 in obtaining a record of the num¬
ber and kinds of fish caught by anglers in Lake Kegonsa, which
is one of the Four Lakes located in the Madison area. This in¬
vestigation covered a period of 20 weeks and extended from
June 14 to October 31, 1936; it was made in cooperation with
the Wisconsin Conservation Department and the University of
Wisconsin.
Cordial thanks are due the owners and operators of the vari¬
ous boat liveries located on the lake for the excellent cooperation
which they gave in obtaining catch records from their patrons
and many of the anglers as well rendered valuable assistance in
the investigation.
Lake Kegonsa lies in the basin drained by the Yahara River
and it has this stream for an inlet and an outlet; it also has
several other inlets. The lake is rather shallow, with a maxi¬
mum depth of 31 feet and a mean depth of 15 feet. The volume
of the lake is 77,250,000 cubic yards when the water is at its
normal level. The water is fairly hard since it contains con¬
siderable amounts of calcium and magnesium. Large aquatic
plants grow abundantly in the shallower parts of the lake and
fish food organisms are generally abundant also.
527
528 Wisconsin Academy of Sciences, Arts, and Letters
Data
A daily record of the catches of fish was obtained for the 20
weeks already indicated. These records were summarized by
weeks and the results are given in Table I ; also the results for
the four leading species of fish are shown graphically in Figure
1. In addition the table shows the total number of the different
kinds of fish caught during the 20 weeks. The open season for
certain species began on May 15, so that those taken before
June 14 are not represented in the census.
Table I
Number of fish caught by anglers in Lake Kegonsa between June 14
and October 31, 1936.
Bluegills. The table shows that the anglers caught more blue-
gills than any other species of fish and most of them were taken
Juday & Vike — Fish Census
529
between September 1 and October 10. The number taken in
June was rather small ; this was followed by a rather large catch
in the latter part of July and then a maximum in late September
and in October. Very few were caught in August and none after
October 11. The bluegill curve in Figure 1 shows four peaks;
a small one represents the week of June 21-27, a larger one in the
last week of July, a maximum peak during September 20-26 and
a secondary one during October 4-10.
White Bass. The white bass ranked second in the total num¬
ber caught during the season. The largest number was taken
during the week of October 4-10. The minimum came the last
week in August and the first week in September. More than
100 specimens were taken each week during 12 of the 20 weeks
included in these observations; while the number taken each
week showed wide variations, yet they furnished the steadiest
yield for the entire season. The wide variation in weekly catches
is well shown by the curve for white bass in Figure 1; it con¬
sists of a series of peaks and hollows throughout the entire 20
weeks.
Wall-eyed Pike. The wall-eyed pike ranked third in total
number for the season. The best catches of wall-eyes were ob¬
tained between the middle of June and the middle of July, and
again in late September and early October. Only a few were
taken between August 1 and September 12. The wall-eye curve
in Figure 1 shows that the June maximum was considerably
larger than that of early October ; it also brings out clearly the
fact that they bit most vigorously in the early and the late part
of the season, with very few taken between August 1 and Sep¬
tember 15.
Perch. The perch were fourth in total number for the sea¬
son. The curve for perch in Figure 1 shows that only a small
number was caught between June 14 and August 29 ; this was
followed by a maximum catch during the first week in September
and thereafter the number steadily declined to 39 during the
week of October 4-10. Only 2 were taken after the latter date.
Other Species. All of the other species of fish were repre¬
sented by comparatively small numbers. The catch of large¬
mouthed black bass amounted to 161 specimens for the 20 weeks,
530 Wisconsin Academy of Sciences , Arts , cmcZ Letters
Fig. 1. Number and kinds of fish caught by anglers each week in Lake
Kegonsa during 1936.
as indicated in Table I; small-mouthed black bass reached 35,
pickerel or northern pike 67, crappies 106 and rock bass 19.
Miscellaneous. The column marked “miscellaneous” includes
several species that were represented by only a few specimens
each ; more bullheads than any other single species are found in
this column.
Juday & Vike — Fish Census 531
Weekly Catches
The weekly totals given in the last column of Table I show
that the best fishing in Lake Kegonsa was had during June and
July, and again in September and early October. The catches
fell to a minimum in August, if the last week in October is ex¬
cluded; in late October practically all fishing was done from a
bridge across the outlet of the lake and not from the open lake.
The smallest number (84) was taken during the week of August
23-29 and August 9-15 was second with 86. The number rose
to more than a thousand during the week of September 6-12,
while the following week was second with 765.
With respect to the total number of fish caught by anglers
during the season, it may be said that many summer residents
occupy cottages on the shores of Lake Kegonsa during the
months of July and August and it was impracticable to secure
complete records from all of them. It is estimated, however,
that their unrecorded catches amounted to about 25 per cent of
the total number included in the table up to September 1. The
record shows a total of 3042 fish taken between June 14 and
September 1 and 25 per cent of this number is 760; adding the
latter to the former number gives a total of 3802 specimens for
this period. This total added to the number taken after Sep¬
tember 1 (4704) gives a grand total of 8506 fish caught by
anglers during the 20 weeks covered by this investigation.
Weights
Average weights of the various species of fish were obtained
so that the weight of the total catch can be estimated; the re¬
sults are indicated in the last line of Table I. The total weight
of the fish included in the table is approximately 8800 pounds.
This amount should be increased by about 800 pounds to cover
the weight of the fish caught by summer residents up to Sep¬
tember 1 and which are not included in the table. This gives a
total of 9600 pounds of fish caught by anglers during that part
of the season covered by these studies. The area of Lake Ke¬
gonsa is 3145 acres, so that the 9600 pounds of game and pan
fish represents an average yield of a little more than three
pounds per acre of lake. This is a very small yield in compari¬
son with that of carp ; in the 1934 rough fish seining operations,
532 Wisconsin Academy of Sciences, Arts, and Letters
229 pounds of carp per acre were removed from Lake Kegonsa,
while the catch amounted to 125 pounds per acre in 1935.
Fishing Records
The lake is so large that it proved impractical to obtain a de¬
tailed record of the number of anglers and the amount of time
spent in fishing. Some observations were made on the number
of boats and the number of anglers in each boat for a period of
3 weeks in August. This happened to be the poorest part of the
fishing season, so that the yield per fisherman was small. Dur¬
ing the week of August 2-8, 35 boats were counted on the 3 days
on which observations were made; they contained 62 anglers
who caught 142 fish on these 3 days. During 4 days of the fol¬
lowing week, 33 fishing boats with 64 anglers were noted and
their total catch was 65 fish. In 3 days of the week August 16-
22, there were 24 boats and 49 anglers; their total catch
amounted to 82 fish. Thus the average catch of fish was 1.6 per
angler for the 10 days on which observations were made.
FISH RECORDS FOR LAKE WINGRA
Chancey Juday
From the Limnological Laboratory of the Wisconsin Geological and
Natural History Survey. Notes and Reports No. 80.
Introduction
In recent years the Wisconsin Conservation Department has
been operating seines in the various waters of southern Wis¬
consin for the removal of the carp. In connection with these
operations, two seine hauls were made in Lake Wingra, which
is situated at the southern edge of the city of Madison, on No¬
vember 15 and 18, 1936, for the purpose of removing carp from
this body of water. This gave an opportunity for collecting
some data regarding the fish population of the lake. Data re¬
garding the total number of fish caught and the weight of the
carp were obtained by the seining crew of the Conservation De¬
partment. Three assistants employed by the Works Progress
Administration obtained material for a determination of the
weights of the other species of fish, as well as scales for a study
of the rate of growth of some of the species.
Lake Wingra is roughly rectangular in shape and has an
area of 200 acres. It is a shallow body of water, with a maxi¬
mum depth of 14 feet ; only a very small area, however, is more
than 10 feet deep.
A small-meshed seine was used in making the hauls ; it was
long enough to reach entirely across the lake in its widest part
and deep enough to cover the entire depth of water. As the hauls
extended the entire length of the lake, the two catches give a fair
idea of the fish population, except those specimens that were
small enough to escape through the meshes of the net.
Data
Table I gives a summary of the data obtained. The weight
of all species of fish, except the carp, was determined by using
the average weight of random samples taken at the time the
seine hauls were made. The sunfish group consisted chiefly of
533
534 Wisconsin Academy of Sciences, Arts, and Letters
bluegills. Both large-mouth and small-mouthed black bass were
caught in the hauls, but they are combined in the table; the
former were a little more abundant than the latter.
Table I
Number and weight of fish caught in Lake Wingra on November 15
and 18, 1936. Weights are indicated in pounds.
Kinds of fish Number caught Total weight Pounds per acre
Carp . 6,000 41,850 209.0
Buffalo fish . 652 1,300 6.5
Gar fish . 2.500 3,500 17.5
Black bass . 1,100 1,600 8.0
Wall-eyed pike . . . 1,000 1,500 7.5
Sunfish . 20,000 6,600 33.0
Crappies . 40,000 13,300 66.0
White bass . 1,500 1,900 9.5
Totals . 72,752 71,550 357.0
On the weight basis, the table shows the great dominance of
the carp in the fish life of Lake Wingra; by weight the carp
makes up a little more than 58 per cent of the fish crop, and the
three species of rough fish (carp, buffalo and gar fish) consti¬
tute 65 per cent of the catch. By weight the pan fish amount to
30 per cent of the two catches and the game fish a little less than
5 per cent.
The last column in Table I shows the weight in pounds per
acre of lake surface; the total catch was 357 pounds per acre,
of which 124 pounds consisted of game and pan fish. The small¬
est fish taken in the seine hauls were in their second year of age,
so that a few pounds per acre must be added for those that were
small enough to escape. The entire fish crop at the time of sein¬
ing can be conservatively estimated at about 365 pounds per
acre.
PROFESSOR C. DWIGHT MARSH AND HIS
INVESTIGATIONS OF LAKES
Mrs. Florence W. Marsh
Dr. Charles Dwight Marsh was born at Hadley, Mass., De¬
cember 20, 1855; he graduated from Amherst College in 1877,
and received from the same college the degree of A.M. in 1881.
After graduating, he taught for four years in secondary schools,
and in 1883 came to Ripon College, Wisconsin, as Professor of
Natural Science. During the earlier years in this position his
energy was fully claimed by the duties of organizing the college
work and the teaching of his classes. But from boyhood he had
interested in the minute forms of fresh-water life, and he
recognized at once the opportunities which Green Lake offered
for such studies. He spent the summers of 1885 and 1887 at
the marine laboratories at Annisquam and Juniper Point, work¬
ing at the microscopic forms of marine life. In 1885 he made his
first collections in Green Lake ; and in the summer of 1886 he col¬
lected from Green, Winnebago, and Puckaway lakes.
In the following years Green Lake became the center of num¬
erous investigations which grew in extent and importance. He
kept a boat on the Lake, which he had fitted with a sail so that he
could collect at all points of the Lake and at all depths. Collec¬
tions were made throughout the year, from the boat in the open
season and in the winter through the ice. From the first he was
attracted by the unique opportunities offered by the great depth
of the Lake. His earlier studies extended to the depth of nearly
200 feet, which was greater than any which had been explored in
an inland lake of the United States. His dredgings revealed an
“abyssal” fauna quite like that of Lake Michigan, as explored by
Dr. Wheeler and Dr. Hoy in 1879, almost 20 years earlier. These
studies of Dr. Marsh were the first of the kind to be done in this
country on a small inland lake.
This early work had a decisive influence on all of the later
studies of Dr. Marsh. The Copepoda, and especially the Diapto-
midae, are the most numerous and most conspicuous members of
the micro-fauna of these deep waters of lakes; and they be-
535
536 Wisconsin Academy of Sciences, Arts, and Letters
came the special objects of his later studies. The choice was a
natural one, indeed, an inevitable one; and perhaps the limita¬
tion of his special studies to these groups was the more natural,
since Dr. Birge had already done much work on the other great
group of freshwater micro-crustacea, the Cladocera. Dr. Marsh
became a member of the Wisconsin Academy of Sciences, Arts,
and Letters in 1888. In 1891 he presented his first papers on
Green Lake ; one related to its deep water life and another gave
its surface and bottom temperatures. He found a minimum
temperature of 5.28 °C. at 58 meters depth. He noted that this
was substantially the same as the temperatures of 5. 0-5.5 found
by Dr. and Mrs. Peckham in the lakes of southeastern Wisconsin.
The observations of the Peckhams were made in 1879 ; they were
the first of the kind to be made in Wisconsin, and those of Dr.
Marsh were the second.
Dr. Marsh continued his work on Green Lake through the
’90’s; he constructed a dredge which could be closed at any
depth ; with this he made elaborate studies of the vertical distri¬
bution of the plankton Crustacea, using five meter intervals. Thus
he ascertained the diurnal variation of distribution, making
twelve hauls in twenty-four hours in order to determine the
effect of light and darkness. He also worked out the seasonal ver¬
tical distribution of the Crustacea, through similar methods, dur¬
ing the period Sept., 1894-Dec., 1896. He reported the results in
Vol. 11 of the Transactions of the Wisconsin Academy, having
determined them for eight genera of Crustacea. It is worth re¬
cording that this work coincided in date with similar study of
Dr. Birge on the limnetic Crustacea of Lake Mendota, which ran
from July, 1894-Dec., 1896. Thus there is a comparable record
for these two Wisconsin lakes, not far separated in distance, but
very different in their ecological character and in their crustacean
population.
The Wisconsin Geological and Natural History Survey was
established by the State in 1897, and after this date much of the
limnological work of Dr. Marsh was done in cooperation with
the Survey and with its aid. In 1897-98 he directed and took part
in a careful hydrographic survey of Green Lake, which resulted
in a map on the scale of three inches to the mile and with con¬
tours of twenty feet. This was published by the Survey in 1899.
The soundings were later converted into the metric scale and
Marsh — Work of C. Dwight Marsh
537
were employed for the map of Green Lake which appears in Bul¬
letin XXVII of the Survey, by Dr. C. Juday, on the hydrography
and morphometry of the inland lakes ; this was published in 1914.
The last important work on lakes done by Dr. Marsh in Wis¬
consin was a careful comparative study of the total plankton col¬
lected by the net in Green Lake and Lake Winnebago. This was
carried on from July 1899 to 1902 by Dr. Marsh and assistants in
the two lakes. A summer laboratory was set up on Lake Winne¬
bago, near Oshkosh, and work was done continuously there. Dr.
Marsh also visited some thirty lakes of Wisconsin, scattered all
over the State, and determined the quantity of their total plank¬
ton.
The results of this study were published in 1903, in Bulletin
XII of the Survey, entitled The Plankton of Lake Winnebago and
Green Lake. It is a thorough examination of the quantity of
plankton found in these two large lakes, with especial attention
to the micro-crustacea, which constitute much of the food for
the fish of the lakes. The comparative study has a peculiar value
since the two lakes are as different, ecologically, as it is possible
for lakes to be.
Dr. Marsh was President of the Wisconsin Academy, 1897-
1899, and on retiring from that office gave an address on The
Plankton of Fresh-Water Lakes. This is a general account of the
work done in this field of study, both in Europe and America,
with a statement of the factors controlling the quantity and types
of plankton, so far as these were known. It is published in the
Transactions of the Academy, Vol. 13.
Dr. Marsh left Ripon in 1903, having leave of absence from
the College for the year 1903-1904. He employed this leisure in
study at the University of Chicago, from which he received the
degree of Ph.D. at the close of the year. It may be added here
that his alma mater, Amherst College, gave him the well-earned
honorary degree, Sc.D., in 1927, on the fiftieth anniversary of
his graduation from the College.
In March, 1905, Dr. Marsh was asked by Dr. Rodney H. True
to undertake work with the Drug Plant Division of the United
States Department of Agriculture. This was an experimental
study of domestic animals poisoned by forage plants ; it led to a
permanent connection with the Department and to extensive in¬
vestigations of poisonous forage plants and their effects. These
538 Wisconsin Academy of Sciences , Arts, and Letters
continued for twenty-five years, and in their prosecution Dr.
Marsh became the first authority on the subject.
The work took him into the new country of the far West in
pioneer days. Transportation was primitive and slow ; conditions
must be provided in field camps, not only for living, but also for
scientific studies; under such surroundings stations must be set
up and operated for experiments on animals. He engaged in and
directed such work on the high plains of Colorado, 1905-09; in
South Central Montana, 1912-15; in Southern Utah, in the Fish-
lake Forest, 1915-30. The work increased in proportions and in
value, and in nearly every state of the Union were found cases of
poisonous forage plants affecting horses, cattle or sheep. All
came to Dr. Marsh and they resulted in innumerable experiments,
conferences and addresses. There remain as witness to this work,
besides many temporary publications, a list of 82 reports and
papers, mostly published by the Department.
But through these twenty-five years Dr. Marsh maintained
his interest in the fresh-water Crustacea; the group of animals
which had kindled in him the love of research and had trained
him in its meaning and methods. He received and studied collec¬
tions of Copepoda from all parts of the United States and from
foreign countries, and the results are published in twenty papers
which are listed in the appendix.
After retiring from the work on forage plants Dr. Marsh was
at Put-in-Bay with Dr. Stillman Wright, working on the Plank¬
ton of Lake Erie. Later through the kindness of Dr. Schmitt,
curator of the Marine Invertebrate Division of the United States
National Museum, he was allotted a table for his private study
of Crustacea. He received an honorary title from the authorities
of the National Museum, that of Curator of the Collections of
Freshwater Copepoda. In the Laboratories of the Museum he
continued the studies which were his first love and interest until
his death. This came on April 3, 1932, at Washington, D. C.
LIST OF PAPERS ON LAKES AND ON CRUSTACEA
1891. On the deep water Crustacea of Green Lake. Trans. Wis¬
consin Acad. Sci., Arts, and Letters, vol. 8, pp. 211-213..
Preliminary list of deep water Crustacea in Green Lake,
Wis., U.S.A. Zool. Anz., vol. 14, no. 370, pp. 275-276.
Marsh — Work of C. Dwight Marsh
539
1892. Notes on Depth and Temperature of Green Lake. Trans.
Wisconsin Acad. Sci., Arts, and Letters, vol. 8, pp. 214-
218.
1893. Notes on the Copepoda of Wisconsin. Science, vol. 22, no.
544, pp. 3-4. On the Cyclopidae and Calanidae of central
Wisconsin. Transactions Wisconsin Acad. Sci., Arts,
and Letters, vol. 9, pp. 189-224, pis. 3-6.
1894. On the vertical distribution of pelagic Crustacea in Green
Lake, Wis., Amer. Nat., vol. 28, no. 333, pp. 807-809.
On two new species of Diaptomus. Trans. Wisconsin
Acad. Sci., Arts, and Letters, vol. 9, pp. 15-18, 1 pi.
1895. On the Cyclopidae and Calanidae of Lake St. Clair, Lake
Michigan and certain of the inland lakes of Michigan.
Michigan Fish Comm. Bull. 5, pp. 1-24, 9 pis.
1897. On the Limnetic Crustacea of Green Lake. Trans. Wis¬
consin Ac. Sci., Arts, and Letters, vol. 11, pp. 179-224,
pis. 5-14.
1899. Methods of making microscopic preparations of Copepoda.
Journ. Applied Microscopy, vol. 2, no. 3, pp. 295-296.
The Plankton of fresh-water lakes. Trans. Wisconsin
Acad. Sci., Arts, and Letters, vol. 13, pt. 1, pp. 163-187.
(Published also in Science, new ser., vol. 11, no. 271,
pp. 374-389, 1900.)
Hydrographic Map of Green Lake. Wisconsin Geol. Sur¬
vey. Map No. 7.
1900. On some points in the structure of the larva of Epischura
lacustris Forbes. Trans. Wisconsin Acad. Sci., Arts,
and Letters, vol. 12, pp. 544-549, pis. 12, 13.
1901. The Plankton of Fresh-Water Lakes. Trans. Wisconsin
Acad. Sci., Arts, and Letters, vol. XIII, pp. 163-187.
1903. On a new species of Canthocamptus from Idaho. Trans.
Wisconsin Acad. Sci., Arts, and Letters, vol. 14, pt. 1,
pp. 112-116, 10 figs.
The Plankton of Lake Winnebago and Green Lake. Bull.
Wisconsin Geol-Survey, No. XII, pp. i-vi, 1-94, 22 pis.
1904. Report on the Copepoda. In Henry Baldwin Ward's “A
biological reconnaissance of some elevated lakes in the
Sierras and Rockies." Studies Zool. Lab. Univ. Ne¬
braska, no. 60, pp. 146-149, pis. 30, 31.
540 Wisconsin Academy of Sciences , Arts , and Letters
1905. The groups and distribution of the North American spe¬
cies of Diaptomus. An address before the A.A.A.S. Ab¬
stract published in Science, new. ser., vol. 21, pp. 270-
271.
1906. Copepodes. In Maurice Neveu-Lemaire’s “Les lacs des
hauts plateaux de l’Amerique du Sud”, pp. 175-188, pis.
17, 18. Paris.
1907. A revision of the North American species of Diaptomus.
Trans. Wisconsin Acad. Sci., Arts, and Letters, vol. 15,
pt. 2, pp. 381-516, pis. 15-28.
1910. A revision of the North American species of Cyclops.
Trans. Wisconsin Acad. Sci. Arts, and Letters, vol. 16,
pt. 2, pp. 1067-1135, pis. 72-81.
1911. On a new species of Diaptomus from Colorado. Trans.
Wisconsin Acad. Sci., Arts, and Letters, vol. 17, pt. 1,
pp. 197-199, pi. 10.
Structural abnormalities in Copepoda. Trans. Wisconsin
Acad. Sci., Arts, and Letters, vol. 17, pt. 1, pp. 195-196,
9 figs.
1912. Notes on fresh-water Copepoda in the United States Na¬
tional Museum. Proc. U. S. Nat. Mus., vol. 42, pp. 245-
255, 14 figs.
1913. Report on fresh-water Copepoda from Panama, with de¬
scriptions of new species. Smithsonian Misc., Coll., vol.
61, no. 3, pp. 1-30, 5 pis.
1915. A new crustacean, Diaptomus virginiensis, and a descrip¬
tion of Diaptomus tyrelli Poppe. Proc. U. S. Nat. Mus.,
vol. 49, pp. 457-462, 7 figs.
1918. Copepoda. In Henry Baldwin Ward and George Chandler
Whipple's “Fresh-Water Biology", pp. 741-789, figs.
1171-1243. New York and London.
1919. Report on a collection of Copepoda made in Honduras by
F. J. Dyer, Proc. U. S. Nat. Mus., vol. 55, pp. 545-548,
pi. 49
1920 The fresh-water Copepoda of the Canadian Arctic Expe¬
dition, 1913-18. Rep. Canadian Arctic Exped. 1913-18
(Southern Party, 1913-16), vol. 7: Crustacea, pt. J:
Copepoda, 25 pp., 5 pis.
1924. A new locality for a species of Diaptomus. Science, new
ser., vol. 59, no. 1535, pp. 485-486.
Marsh — Work of C. Dwight Marsh
541
1926. Crustaces copepodes recoltes par M. Henri Gadeau de
Kerville pendant son voyage zoologique en Syrie (Avril-
Juin 1908).
In Gadeau de Kerville’s "Voyage zoologique d’Henri Gad¬
eau de Kerville en Syrie’", vol. 1, pp. 171-185, pis. 23-26.
Paris.
On a collection of Copepoda from Florida, with a descrip¬
tion of Diaptomus floridanus, new species. Proc. U. S.
Nat. Mus., vol. 70, art. 10, pp. 1-4, 6 figs.
1929. Distribution and key of the North America copepods of
the genus Diaptomus, with the description of a new
species. Proc. U. S. Nat. Mus., vol. 75, art. 14, pp, 1-27,
16 figs.
1931. On a collection of Copepoda made in El Salvador by Sam¬
uel F. Hildebrand and Fred. J. Foster of the U. S. Bu¬
reau of Fisheries. Journ. Washington Acad. Sci., vol.
21, no. 16, pp. 397-405, 3 figs.
1932. A new species of Cyclops from the Phillipine Islands.
Journ. Washington Acad. Sci., vol. 22, no. 7, pp. 182-
184, 8 figs.
1933. Synopsis of the calanoid crustaceans, exclusive of the
Diaptomidae, found in fresh and brackish waters,
chiefly of North America. Proc. U.S. Nat. Mus., vol.
82, art. 18, pp. 1-58, 24 pis.
The Crustacea of the plankton of western Lake Erie. (To
be published in the Bulletin of the U.S. Bureau of Fish¬
eries as a chapter in a general report on the lake.)
Note
The active scientific life of Dr. Marsh covered nearly a half cen¬
tury; it may be dated approximately from 1883, when he joined the
faculty of Ripon College, to his death in 1932. This period is divided
into two parts, not very unequal. The first is twenty-one years long,
from 1883 to 1904; this period was spent at Ripon and its scientific
activities concerned limnological subjects, and especially the fresh¬
water Copepoda. In 1905 he became a member of the United States
Department of Agriculture ; in that capacity he gave most of his atten¬
tion to poisonous forage plants, and their effects; and he became the
leading authority in that field.
542 Wisconsin Academy of Sciences , Arts, and Letters
This long and active service in an important field has tended to
obscure his services in another and far distant field of science, which
were contributed during his life in Wisconsin. He was among the
earliest and the most important students of fresh-water life and of
fresh-water ecology in the United States.
Ripon College is seven miles from Green Lake, the deepest inland
lake between the Finger Lakes of New York and the mountain lakes
of the Rocky Mountains. Dr. Marsh fully appreciated the situation and
he began his work on Green Lake early in his life at Ripon. We should
not fail to note the initiative and independence of Dr. Marsh in this
early limnological work. Green Lake is seven miles from Ripon, and
this was by no means a negligible distance in the days of the horse and
buggy. Every series of readings in the lake called for much time in
coming and going. So, also, limnological study was not suggested to
Dr. Marsh by other work of the sort in Wisconsin, for there was then
no such work in progress. I had been interested in fresh- water life in
the 70’s; but my time from 1881 to 1891 was more than filled by the
teaching of pre-medical students ; and the fire of 1884, which destroyed
Science Hall, also burned all of my notes on Cladocera and discouraged
such incidental study as I might otherwise have given to our lakes. A
short paper on the Cladocera of Madison, presented at the meeting of
the Wisconsin Academy in 1891, marked my return to fresh-water stud¬
ies. At the same meeting Dr. Marsh read his first papers to the
Academy, on the deep water Crustacea of Green Lake and on the depth
and temperature of the Lake.
In 1888, Dr. Marsh became a member of the Wisconsin Academy of
Sciences, Arts, and Letters. He was an active member, constant in at¬
tendance at its meetings, and contributing papers to its Transactions.
He held various offices, including that of President, 1897-99. In 1903
he left Ripon College and in 1905 he went to the United States Depart¬
ment of Agriculture, engaging in the study of poisonous forage plants.
To this work he gave the following twenty-five years, retiring in 1930.
This period constitutes the most important side of the scientific work
of Dr. Marsh, in which it developed a national scope and a national
value. No adequate account of it can be given in these Transactions,
which are concerned with that part of his life and work which is more
intimately related to Wisconsin and to this Academy. The larger story
must be left to the Department to which he gave such distinguished
service during a quarter-century. This sketch of his work in Wisconsin
Marsh — Work of C. Dwight Marsh
543
and on the life found in its fresh-waters has been prepared at my re¬
quest by Mrs. Florence W. Marsh.
E. A. Birge, Biologist,
Wisconsin Geological and Natural
History Survey.
A NEW SPECIES OF RECEPTACULITES (R. PEDUNCU-
LATUS) FROM THE SILURIAN STRATA
OF EASTERN WISCONSIN
W. H. Twenhofel
According to Shrock1, the Silurian formations of eastern
Wisconsin in ascending order consist of Mayville, Byron, Wau¬
kesha, Coral Beds, Racine-Guelph and Waubakee. The Wauke¬
sha has been postulated to be the southern equivalent of the Coral
Beds. The rocks, with minor exception, are dolomites or dolo-
mitic limestones. The fossil described in this paper was col¬
lected in a quarry near the town of Cedarburg, Wisconsin, about
sixteen miles north and a little west of Milwaukee, from strata
which seem to be either in the upper part of the Coral Beds or
the lower part of the Racine.
The specimen is excellently preserved with respect to show¬
ing the general form and size, but the material of which it is
composed has been entirely recrystallized from the original con¬
dition so that detail cannot be made out to perfection. The
specimen is cone-shaped, but the enlargement from the base is
not the same throughout so that the surface below the middle
is concave outward. It seems to be open at the top and thus to
have the shape of a glass goblet. The maximum diameter at the
top is 38 mm. and it is about 50 mm. high. The base is slender
and cylindrical, forming a stem about 12 mm. long, thence up¬
ward the increase in diameter is rapid. The head plates are
arranged in the form of two spirals, trending in opposite direc¬
tions and making nearly one turn in ascending from the base to
the summit. Each spiral makes an angle of about 221/6° with
the vertical axis. Plates also fall into an arrangement perpen¬
dicular to the axis, but it is not sure that this arrangement is
annulate and not in the form of a low spiral.
The preservation is such that on a part of the specimen the
interior impression of the skeletal structure is shown and on
another part of the surface there is shown an impression of the
1 R. R. Shrock — Silurian Geology of Wisconsin, Proc. Geol. Soc. Am. for 1934: 107-108,
(1935).
545
546 Wisconsin Academy of Sciences, Arts, and Letters
outer surface of the skeletal matter. The greatest number of
plates observed in any one of the spirals is thirty-nine. Each
plate shows a small basal depression which in many cases has
an opening at the base. In the parts showing something of the
skeletal structure there are two depressions in each plate ; one
is at the base and the other is at the upper corner. The latter
seems to communicate with the lower end of the plate just above.
There are four plates in 10 mm. along the line of the spiral in
the upper part of the specimen ; five in 10 mm. near the middle,
and at least eight in 10 mm. on the pedunculate end.
The plates are rhomboidal in shape with the longer diameter
of the rhomboid in vertical position and the shorter in horizon¬
tal. There are five plates in 10 mm. in the horizontal position
near the top where there are four plates in 10 mm. in the spiral.
Two other species of Receptaculites have been described from
the Silurian of Wisconsin. R. hemisphericus Hall has a hemi¬
spherical shape with a truncated base. This was collected from
the Racine. R. tesselatus Winchell and Marcy has about the same
dimension as this new species but is widest a little above mid¬
length and narrows toward the top. It has a moderately sized
base of attachment that is not pedunculate. The plates at the
top of the specimen are stated to be wider than high, whereas
in this new form they are higher than wide.
The characters of the new form seem to be distinct from
that of any Silurian species already described with the differ¬
ence chiefly in the form of the pedunculate base. It is proposed
to designate the new species under the name R. pedunculatus.
The holotype is in the collection of W. H. Twenhofel.
Figure 1. Skeleton in vertical position. About % natural size.
Figure 2. A part of the surface near the top as shown in figure 1, x F/2
TOPOGRAPHY OF ABANDONED BEACH RIDGES
AT ELLISON BAY, DOOR COUNTY, WISCONSIN
0. L. Kowalke and E. F. Kowalke
The abandoned shore lines and beach ridges near the village
of Ellison Bay, Wisconsin indicate that the waters of Green Bay
and of Lake Michigan were once joined by the route of Rogers
Lake, locally called the Mink River. The Mink River (Rogers
Lake) shown on the map in Fig. I, is a very sluggish stream
with little fall. In fact its head waters near the junctions of
sections 11, 12, 13 and 14 are only three feet above the level of
Green Bay. At its mouth the solid rock sloping gently eastward
is found at the water level. In Ellison B,ay in contrast the
ground slopes quite abruptly into the water ; the slope is precipi¬
tous off the bluff in sections 15 and 16; it is likewise precipi¬
tous about 150 feet off shore in sections 10 and 2.
It is easy to see how ice on Ellison Bay driven by wind might
push rock and sand from the beach inland and how waves would
do likewise and thus fill this water way. As the water level fell
the solids so transported became exposed to the air and thus
formed the land on which the village of Ellison Bay now stands.
The storms that raged threw up some interesting beach ridges
now to be described.
A topographic survey to show the relationship of these ridges
had long been under consideration, but it was not until the sum¬
mer of 1935 that both of us were free to collaborate and do the
job. The measurements in this report were made with an engi¬
neer's transit equipped with stadia wires and level attachment,
a steel tape, and a level rod. The starting point of the survey
was the joint corner of sections 10, 11, 14 and 15 and the county
surveyor's location of the line east and west between sections 10
and 15. All elevations were based on the U. S. Coast and Geode¬
tic Survey Bench Mark No. U 21, elevation 589.06 feet above sea
level. The maximum deviation in levels closure was 0.15 feet
and that of distances was 2.0 feet. The data gathered on the
survey were combined in the map shown in Fig. II. It is re-
547
548 Wisconsin Academy of Sciences, Arts, and Letters
Fig. I.
Kowalke & Kowalke— Abandoned Beach Ridges 549
550 Wisconsin Academy of Sciences, Arts, and Letters
grettable that the necessary reduction in size for the lantern
slide and the Transactions makes some details obscure.
There are three distinct groups of beach ridges whose eleva¬
tions above sea level are respectively 640 feet, 603 feet and 588
feet. They are all composed of well rounded limestone frag¬
ments ranging in size from one inch to eight inches diameter
and are devoid of sand and soil material except for small
amounts of humus originating from decayed leaves and shrubs.
The vegetation growing on them, consists of balsam firs, white
cedar, white and Norway pine, hard maple, white birch, poplars,
and shrubs.
The beach ridge at the 640 feet level shown in Fig. II at J-J
is somewhat oval in shape, about 220 feet by 500 feet, and is
located at the end of a solid rock projection. The waves caused
by storms coming out of the northwest rolled the beach pebbles
southward along a small shelf in the cliff in sections 2 and 10
and then deposited them at J-J. At the 640 feet above sea level
stage of water the cliff in sections 10 and 2 rose only a short
distance out of the water. Indeed at this stage most of the land
north and east of the line from Ellison Bay through the Mink
River (Rogers Lake) to Rowleys Bay was submerged. There
are no other evidences in the immediate vicinity of beach ridges
at the 640 feet stage of the water.
The ridges of the 603 feet level marked K-K and L-L in Fig.
II are on opposite sides of Ellison Bay waters, and their crests
are not hachured but are left in white for the sake of clearness.
A noteworthy feature of them is that they were apparently
formed by one great storm. Note that the landward slopes are
not indented or lobed. The ridge K-K is about 2400 feet long;
its crest is not flat but slightly rounded, and a profile at the
section B-B shows that a succession of storms added deposits of
pebbles, but that the first crest was never topped. At the eastern
end of the ridge the crest is in the form of a flat shelf against the
hill at the back and it is also deflected toward the south because
of the solid rock humps in sections 10 and 14 respectively. The
forest covers consists of balsam fir. white cedar, some pine,
poplar, and dogwood.
The ridge L-L is more certainly the work of one big storm ;
both seaward and the landward slopes are regular and not lobed.
The crest at the section G-G is only 6 feet above the surface at
Kowalke & Kowalke — Abandoned Beach Ridges
551
the landward side and is continuous with the shore line at the
603 feet level of water where it begins on the west at the State
Highway 42 and ends 2100 feet to the eastward near the Rowleys
Biay Road. It has no forest cover. Another notable feature of
this ridge is that it closes the north end of the gulley M-M ; and
this dyke has never been broken. The water that runs off the
high hill to the south and to the west is received by this gulley
and is then drained rapidly to the shore of Ellison Bay under¬
ground through the voids of the very coarse gravel. At the point
(N) about 500 feet north from the gulley the drill record of a
well showed that solid rock lay over 90 feet beneath the surface
and that the intervening material was coarse gravel.
In contrast with the ridges just discussed, the ridge R-R at
the 588 feet level is the work of a succession of storms. From a
little north of the section line between sections 10 and 15 it runs
in a southeasterly direction to State Highway 42. Notice the
lobed form of the contour at the foot of the ridge on the land¬
ward side ; notice particularly the long finger lobes at the north
end near the section line ; and finally how at the head of the bay
the pebble ridge has been extended farther toward the south and
the west leaving a circular area (S) which is a marsh. The
ridge R-R east and north of the dock is grown over thickly with
white cedar together with a few white pine ; south and west of
the dock there is no vegetation at present.
The saddle or col between Ellison Bay and the Mink River is
on or near the profile section F-F about 400 feet east of the
boundary line between sections 14 and 15 and its elevation is
596 feet. The slopes of the ground from the col to the east and
to the west to water are gentle as shown in the profile F-F in
Fig. III. The top soil is fine beach sand and is now mostly under
cultivation.
For contrast and for comparison of the ridges it seemed de¬
sirable to construct a series of profiles across the area surveyed.
The sectioning planes are shown on Fig. II by the lines A-A,
B-B _ _ GG, and the respective profiles are shown in Fig. III.
If the ridges were eliminated the pitch from land to water is
nearly the same in profiles B-B, C-C, D-D, and E-E. The pitch
in profiles F-F and G-G at the head of Ellison Bay is gentle due
to the breaking of the waves. Note the ridge at the 604 feet
level in profile G-G how steep the landward slope is which would
552 Wisconsin Academy of Sciences , Arts , and Letters
seem to predicate that the frontal thrusts of the waves piled up
the ridge. Note in profile D-D the succession of ridges formed
when the water was at the 588 feet stage.
The origin of the pebbles forming the ridges J-J, K-K, and
R-R was the cliff that extends along almost the entire eastern
shore of Ellison Bay. The weathering and disintegration of
this cliff is fairly rapid owing to the large fluctuations in tem¬
perature, to freezing and thawing, and to its favorable exposure
to the sun. Note on profile A-A in Fig. Ill, that there is an
escarpment about 12 feet under water nearly 120 feet from the
shore line, and this extends almost the entire length of the cliff
toward the north. So as the cliff disintegrated and fell the frag¬
ments were rounded by wave actions and on the occasion of big
storms were rolled into the ridges.
The contrast between the cliff in sections 15 and 16 on the
south shore and that in sections 10, 2, and 35 on the east shore
is striking. Because of the north exposure, the cliff in sections
15 and 16 has disintegrated but slowly. White cedars are now
growing luxuriantly on every little shelf so that the face of the
cliff is partially hidden ; no such growth occurs on the east shore
in sections 10 and 2 except on the benches in section 35. Fur¬
thermore along the shore in sections 15 and 16 and on some of
the higher benches of old shore lines one may find glacial boul¬
ders in abundance; they are not in evidence on the east shore.
Some interesting correlations were made possible by combin¬
ing our data with those on elevations and grade for the Wiscon¬
sin State Highway 42 from Sister Bay through Ellison Bay to
Gills Rock. For example, the pebble ridge (SB) in Fig. I north
of Sister Bay village has the same elevation, 633 feet above sea
level, as the old shore line marked (OS) in Section 15 south of
Ellison Bay village. The highest elevation above sea level on
Highway 42 between Sister Bay and Ellison Bay is 767 ft. and
is shown at (HE) in Fig. I. Between Ellison Bay and Gills
Rock the highest elevation is 672 feet; so that at the 640 feet
stage of the water much of that land was submerged. Further
interest in the 640 feet stage centers about a cave in the cliff
near the place where the north and south quarter line of Section
16 intersects the shore. The height above water of this cave
was measured by Mr. Hilder Erickson of Ellison Bay to be about
65 feet which corresponds to 643 feet above sea level.
Kowalke & Kowalke • — Abandoned Beach Ridges
553
Fig. III.
Grateful acknowledgments are here made to Mr. Frank
Cnare of the Wisconsin Highway Commission for the data on
State Highway 42 and to Mr. H. Erickson for his assistance.
University of Wisconsin
Madison
and
Milwaukee, Wisconsin
A NORTH AMERICAN RECORD FOR JUNCUS
CAPITATUS WEIG
S. C. Wadmond
Included with other Junci sent for determination in Febru¬
ary 1936 from the Herbarium of the University of California, at
Berkeley, was one sheet which looked quite unlike any Juncus
listed in Abrams' Ill. Flora of the Pacific States; in fact, it
failed to fit the description of any North American species, its
nearest congener apparently being Juncus bulbosus L., a rather
common European species, known in the western hemisphere
from Labrador, Newfoundland and Nova Scotia.
Through the kindly offices of the Gray Herbarium of Har¬
vard University, the plant was identified as J. capitatus Weig., a
well-known rush of northern Africa, the islands of the Medi¬
terranean and southern and central Europe, coming up as far
as south Sweden and central Russia. Here was indeed an inter¬
esting and unorthodox find, much like the occurrence of J. bulb¬
osus on Vancouver Island which was reported a year or so ago.
Dr. Herbert L. Mason, Associate Curator of the University of
California Herbarium, was immediately apprised of the “find"
and asked if he had other sheets of this species from California.
He promptly forwarded another sheet on suspicion, and sure
enough, it was another J. capitatus, collected only a few miles
from the first! Both of the collections are from Sacramento
County, California. The earlier collection, Ethel K. Crum No.
1599, April 2, 1934, is from Keithling Ranch, Rio Linda, “in
field pastured but never cultivated." The other was taken at
Winding Way, 2.5 miles east of Fair Oaks, altitude 300 ft., “in an
uncultivated field with Briza, Hemizonia and Eryngium" Ann-
etta M. Carter No. 813, May 12, 1935. The two localities are
apparently not over ten miles apart. Miss Carter, a botanist on
the Herbarium staff, made a second trip to the Winding Way
station on May 30, 1936, and secured more material. She re¬
ported that the plants were taken in an open, uncultivated field,
the larger plants growing around the margin of a drying vernal
555
556 Wisconsin Academy of Sciences , Arts, and Letters
pool ; the depauperate ones from the drier part of the field. The
most luxuriant plants from either of the two stations are not
over 5 cm. tall, ranging from that down to tiny depauperate
specimens scarcely 1 cm. high, and bearing a single flower only !
The European manuals describe it as from 5 to 8 cm. tall.
A first question which naturally arises is whether the species
is native or introduced. Is it perhaps a native species with
something of the same strikingly disjunct distribution as J.
sphaerocarpus Nees, a well-known plant of central and southern
Europe and Asia, which makes a jump of several thousand miles
to appear in Idaho, Washington, Oregon, California and Ari¬
zona? Or has it been introduced in some manner ? My own
thought is that the latter hypothesis is the more correct one.
We know J. Dudleyi has been accidentally introduced into Ger¬
many, and the plants of this species reported from Scotland may
have had a similar origin. J. effusus compactus is believed to
have been introduced on Vancouver Island with cranberry plants
and moss from Nova Scotia. J. bulbosus which turned up a
year or so ago on Vancouver Island may have come there in a
similar manner. J. compressus, plants of which Fassett found
at Galetta, Ontario,1 the most westerly North American station
so far reported, is annotated in Gray’s Manual 7th ed. as “Pos¬
sibly naturalized from Europe.” Dr. Fassett believes this to be
the case without doubt. In Rider’s Hand Book of California
Fair Oaks is described as located in the lowest foothills of the
Sacramento Valley, near the American River, on the route of the
historical Pony Express. It is likewise on the trail which the
Argonauts of 1849 took as they swarmed into the Sacramento
Valley on their trek towards near-by Coloma, the historic spot
where James W. Marshall discovered gold in 1848. Is it con¬
ceivable that J. capitatus may have come in during this period
in some manner, and finding conditions congenial, has persisted ?
It is an interesting speculation, even if a fanciful one. Nearer
the truth perhaps, it may have been brought in in a much more
prosaic fashion with other plants introduced from Europe for
economic or cultural purposes.
Dr. Mason writes he has seen no record of the occurrence of
J. capitatus in California. The late Dr. F. V. Coville reported
there was no collection of this species in the National Herbarium
1 Rhodora XXXV. 391 (1933)
Wadmond — American Record for Juncus 557
from North America, nor is it represented in the Gray Herbari¬
um from this continent, according to Associate Curator Weath-
erby. Bluchenau2 cites a collection of this species from New¬
foundland (de la Pylaie, 1826; hb. Candoll.) . Concerning this
citation Mr. Weatherby writes : “The Newfoundland plant which
Buchenau called J. capitatus was probably J. bulbosus, which in
habit somewhat resembles J. capitatus. So far as Prof. Fernald
knows, there is no authentic record of J. capitatus from North
America. If La Pylaie’s plant was really that, there must, he
thinks, have been a mixture of labels.” The writer is quite in¬
clined to agree with this conclusion because Prof. Fernald has
collected extensively in Newfoundland, and would hardly have
failed to locate it if occurring there. It looks, then, as though this
were the first authentic record of the collection of this species in
North America.
3 Engler, Pflanzenr. IV. 36. 256 (1906)
LANDLOCKED SALMON IN WISCONSIN
By T. E. B. Pope, Curator of Lower Zoology, Public Museum
of Milwaukee
On November 12th, 1937, Mr. Charles D. McCommons, of
Delavan, Wisconsin, brought to the Public Museum of Mil¬
waukee two large fine fish for identification. Mr. McCommons,
who is something of an amateur fisherman-ichthyologist, sus¬
pected that they were not what they were being called by
sportsmen, taxidermists and local fishermen at Lake Geneva,
Walworth County, Wisconsin, as “Brown Trout”. A careful
examination of them and consultation of such literature as was
available caused him to believe that they were instead Land¬
locked Salmon.
The specimens are now in the official possession of the Mil¬
waukee Museum where they have been photographed by both
ordinary and color processes, measured, plaster molds made,
original skins being mounted by taxidermists, the gillrakers
and a sample of the eggs preserved, etc. The specimens are now
designated as Cat. No. 5103 (<*) and No. 5104 (?). The writer
confirmed Mr. McCommon’s suspicions and definitely declares
that these fishes are true specimens of Landlocked Salmon (Salmo
salar sebago) .
Before specifying the facts, measurements, etc., upon which
this identification has been based, it may be of interest to note
that according to Mr. McCommons this species of fish has been
observed in the lake for at least fifteen years and specimens
have frequently been taken of varying weights up to fifteen
pounds. The two specimens, shown by the accompanying photo¬
graph, represents a male of 24 inches in length with a weight
of 5 lbs. 8 oz. and a female 26 inches long with a weight of
7 lbs. 4 oz. This particular pair of beautiful fish, in their tem¬
porary dark-pink coloration, were taken by Mr. L. Coventry, a
garage man of Fontana, from the Fontana Creek, a tributary of
Lake Geneva, by netting. Mr. McCommons, in his letter to the
writer dated November 14, 1937, says, “They are very easy to
559
560 Wisconsin Academy of Sciences , Arts, and Letters
catch as they run into the shallows as far as they can go,
usually two or three males to one female during their spawning
activities. They are about thru spawning now.” The first fish
that Mr. McCommons examined was a female of about four
pounds which, as he says, was much more ‘silvery' in color,
had finished spawning, and was taken from the Williams Bay
Creek, another Lake Geneva tributary some two miles distant
from Fontana Creek. That fish was caught by a Mr. Joseph
Ambrose. I quote further from this letter of Mr. McCommons,
“I have had many reports of immense schools of ‘large
trout' being seen in the lake at different times — ‘big trout
with hooked jaws'. The first part of one I ever saw was the
head of a male of 9 pounds shown to me by the caretaker of
the T. J. Lefens estate which is located near the so-called
narrows on the south shore. This man had the head nailed
up on the wall of his tool shed. He said he had caught two
years previous ‘‘when they are up here around the pier in
schools jumping all over the place”. He went into the house
and brought out a photograph of the fish to prove he had
caught it — -if that proves anything. Said it was the “darnd-
est thing I ever got ahold of, and it took me a half hour to
land him”. Later, in my calls around the lake, I stopped again
at Lefen’s estate, and this time he had one of five pounds, all
dressed and on ice. He stated that they were running then
and . . . that there were some that would go ten to fifteen
pounds. Small boys and some not so small have taken, for
over fifteen years, large numbers of them each year during
November, mostly at night by walking along the banks of
the creeks with flashlights and using a club or spear or
anything. Every once in a while one has been taken thru
the ice in the winter, and off and on during the summer
months while deep fishing for northern pike and bass.”
The two specimens of Salmonidae under consideration and
designated in this article as Cat. No. 5103 and No. 5104 or the
“Lake Geneva” specimens, were identified by the writer, as
stated above on November 12, 1937. Two days later, on Novem¬
ber 14, there appeared in the Sunday issue of the Milwaukee
Journal a fairly long account of the discovery of this, fish in
Lake Geneva and containing significant information that the
writer was entirely unaware of when he made his official iden¬
tification. This significant paragraph is as follows :
Pope — Landlocked Salmon
561
“Mr. McCommons has found out that 40 years ago the
Fairbanks estate on Lake Geneva had a fish hatchery. Work¬
ing at the hatchery was the father of Bert Welsher of
Delavan. Welsher told McCommons that Fairbanks imported
landlocked salmon eggs and that his father hatched and
planted them in 147-foot-deep Lake Geneva. In addition, the
United States bureau of fisheries has planted these fish in
all of the Great Lakes and other waters.”
In a letter of November 16, 1937 addressed to the writer, U. S.
Fish Commissioner, Frank T. Bell, states,
“In response to your letter of November 13, I would ad¬
vise that our records show no plants of landlocked salmon
(Salmo salar sebago) in Lake Geneva. Our records do show,
however, that in the fiscal year 1911 8,000 fingerling landlock¬
ed salmon were planted in Pleasant Lake, the point of deposit
being Coloma, Wisconsin. In the very early days eggs were
shipped around rather indiscriminately and it might be that
the State Fish and Game Department obtained some of them.
However, we have no definite record of this.
I do not know whether the foregoing will help to solve
your mystery, but it does indicate that landlocked salmon
have been introduced in the waters of the State of Wiscon¬
sin.”
S. B. Locke, in Bureau of Fisheries Document No. 1062, pp. 184-
185, states that Salmo sebago has been introduced into
“Payette Lake and several lakes in the Sawtooth Mountains
and in the latter locality have given encouraging results.
They have also been planted with some success in Fish Lake,
Utah.” \ , *
Thus it appears that Landlocked Salmon has been planted in
the West, in Wisconsin, and that private individuals had planted
them in Lake Geneva and showing that a valid and reasonable
ground exists for the presence of this species of fish in the
waters of Lake Geneva. This is especially so when specimens
like those recently received by the Milwaukee Museum are found
to conform so closely to the published type of Landlocked
Salmon.
The identification of the Lake Geneva specimens has been
based principally and primarily upon the recognized permanent
physical structure and only secondarily upon coloration fea¬
tures. For instance,
562
Wisconsin Academy of Sciences , Arts , and Letters
1. Scale Count. Not over 120 on the later line but a true
count which excludes the small scales at the base of the
caudal fin will be about 116 which is the number found
by Dr. W. C. Kendall in his “Connecticut Lakes” paper
(Bur. Fish. Doc. No. 633) for three specimens of this
species. On page 48 of this same paper Kendall cites 140
scales for the Brown Trout.
2. Dorsal Fin. Kendall reports 12 rays for two of the three
specimens described on pages 42-44 and 11 for the third.
For Brown Trout he cites 10. Both of our Lake Geneva
specimens have at least 11 articulated rays.
3. Gillrakers. Kendall gives an average for his three speci¬
mens of Landlocked Salmon of not less than 7 plus 12,
while for Brown Trout he mentions 5 plus 11. The
Lake Geneva specimens have the higher count.
4. Proportional Measurements. (These can be verified from
the accompanying photograph).
The head of the Landlocked Salmon in proportion to its
body length ranges from 3% to nearly 5 times. The
head of the Brown Trout is generally about % of its
body length. The Lake Geneva specimens have heads
4.40 and 4.50 respectively.
The eye of the Landlocked Salmon, according to Kendall
for his three specimens ranges from 6.82 to 9.12 with
an average of 7.89. For Brown Trout he gives 5. Our
Lake Geneva specimens will average 8.
The snout of the Landlocked Salmon, according to Ken¬
dall for his three specimens will average about 3.
For Brown Trout he gives 3.7. The Lake Geneva speci¬
mens have 3 for the male and 3.40 for the female.
In the above it is seen that the Landlocked Salmon has a
much lower scale count, more gillrakers, a smaller eye, a
smaller head and a longer snout.
5. Coloration. All marks on the bodies of the Lake Geneva
specimens are relatively small, closely-crowded, distinctly
cross-like, and blackish as opposed to the large, circular,
brightly-colored and widely-separated spots of the Brown
Trout. The Lake Geneva specimens have numerous large
spots on sides of head, color on the adipose fin, and no
white margins to the ventral and anal fins. In the male
the tip of the lower jaw was black and with black streaks
on chin. In fact, they very closely correspond to Ken¬
dall's description of the three specimens in the Connec-
Pope — Landlocked Salmon
563
ticut Lakes paper. Under the Brown Trout Kendall states
adipose fin plain, throat and under part of lower jaw
white and shows spots on sides of head quite few.
Both of these specimens (Lake Geneva) had bodies
flushed with dark-pink spawning coloration. (This color
has been recorded by a museum color-photograph). From
the body of the female a pint, simply a sample, of eggs
were taken that were of a pure golden color and that
have since in preservative been slowly changing to pink.
The male is in the hands of the taxidermist and the
skin of the female with its head and fins attached is in a
10% solution of formaldehyde. At this writing, the skin
of the female has lost all trace of the original temporary
dark-pink coloration and has reverted to the probable
normal color which is gray and covered with a silvery
sheen. The side of the head is also silvery.
Correspondence and opinions received to date from a num¬
ber of the leading ichthyologists have not revealed any definite
denial that the specimens under consideration are Landlocked
salmon and yet, on the other hand, not one reply has been re¬
ceived confirming the name of Brown trout. Some of these
authorities say openly that they do not know and others sug¬
gest the possibility of the specimens being Loch Leven trout. The
situation, therefor, by reason of this confusion and the lack of
denial, entitles the writer to declare them Landlocked salmon
by reason of the evidence shown above. Smith and Kendall in
Bureau of Fisheries Document No. 904 on the “Fishes of the
Yellowstone National Park”, recognize, comment upon and figure
as different and distinct species Loch Leven trout (Salmo leven-
ensis) and Brown trout (Salmo fario). While some authorities
may claim close relationship between these species and even that
they might be one and the same, the writer has not, as yet, seen
or received any recognized publication showing that they are
the same. If it can be shown that Loch Leven trout has the basic
physical structure that the Lake Geneva specimens possess and
that, therefor, these specimens are Loch Leven trout, then the
situation has not changed in any respect from that claimed above
except in the name that the writer has declared. Thus it appears
that the specimens in question are not Brown trout. Again, in
the latest published list of the state’s fish-fauna by the Con¬
servation Commission of Wisconsin entitled, “The Distribution
of Wisconsin Fishes”, by C. Willard Greene, 1935, there is no
564 Wisconsin Academy of Sciences, Arts, and Letters
mention of either Landlocked salmon or Loch Leven trout. Con¬
sequently in these Lake Geneva specimens we have a new and
important addition to the fish-fauna of the state.
One of the principal questions, and perhaps the foremost,
that will arise by virtue of this discovery of numerous adult
and spawning Landlocked salmon in Lake Geneva, will prob¬
ably be as to why they had not been recognized as salmon by
some scientist throughout all these past years. The answer can
easily be given. These fish were in the deeper parts of the lake
when any investigation of the fish-fauna of the region was being
conducted, if any. All specimens that were captured at any
season of the year apparently only fell into the hands of natives,
sportsmen and others who were not familiar with salmon recog¬
nition, that were not observant enough or curious enough to in¬
vestigate if they had suspicions. No true salmon having been
before included in the list of the state's fishes, not suspecting that
such might be in the state, not being technically trained to rec¬
ognize identifying characteristics of fishes and not, as stated
above, being curious enough to investigate, all such men were
simply content to continue calling them ‘big Brown Trout'.
Since the announcement in the Milwaukee newspaper that
Landlocked salmon were present in Lake Geneva in considerable
numbers much harm threatens these fine fish. The writer is in
receipt of a letter dated November 29, from a prominent citi¬
zen who is a naturalist and conservationist of that region read¬
ing as follows:
“I enclose a little clip from our local paper. I do not know
how true it is but I am told that the fishermen in that area
are awfully sorry that all this publicity developed just at this
time; they say that just as soon as it was definitely learned
that the species in question was not a brown trout but a land¬
locked salmon, and therefore not covered by existing fish and
game laws of Wis., a small army set out after them and took
great quantities, including many spawning females which had
retreated up into the shallow spring-runs and in some cases
could almost be taken out by hand. I hope they wont deci¬
mate the species completely before some restrictive measures
can be put into effect.''
The above letter speaks for itself. It should prove to be an
incentive, as well as a duty, for all true sportsmen and fish-lovers
to urge and see that the State Conservation Department ade-
Plate II.
Pope — Landlocked Salmon
565
quately protects these fishes whatever they may be and how¬
ever they may be called. Furthermore, it would seem advisable
for both State and Federal governments to immediately take
steps for the artificial reproduction of these large and beautiful
game fishes for the benefit of the State of Wisconsin.
PROCEEDINGS OF THE ACADEMY
Sixty-seventh Annual Meeting
The sixty-seventh annual meeting of the Wisconsin Academy of
Sciences, Arts and Letters was held conjointly with the Wisconsin Archeo¬
logical Society and the Midwest Museums Conference at the Milwaukee
Public Museum, April 9 and 10, 1937. The following program of papers,
special lectures and other activities was presented.
Friday morning , 9 :00-10 :00 o’clock. — Registration.
Friday morning , 10:00-10:15 o’clock. — Opening session. Mr. George A.
West, President of the Board of Trustees of the Milwaukee Public Museum
delivered a short address of welcome to members and guests of the three
societies. There followed several announcements, after which those in at¬
tendance separated into two sections to hear the various papers.
Friday morning , Section A, 10 : 15-12: 30 o’clock. — Herbert W. Kuhn,
Indian uses of shells; Zida C. Ivey, The Fort Atkinson Museum; Charles
G. Schoewe, Wooden vessels of the Wisconsin Indians; Louis S. Buttles,
Middle and Lower Mississippi Valley pottery; John B. McHarg, Miniature
slides for schools and museums; Alonzo W. Pond, Wisconsin joins the ranks
of earliest inhabited areas in America.
Friday morning, Section B , 10 : 15-12: 30 o’clock. — James R. Neidhoefer,
Carterius tenosperma Potts, a fresh-water sponge new to Wisconsin;
Paul L. Carroll, Blood supply of the tympanic membrane of the Frog;
Hilary J. Deason, The distribution of sculpins (Cottidae) in Lake Michigan
(By title) ; H. A. Schuette, Honey in the “Primitive Physic” of John Wesley;
Carl A. Bays, The relation of Mlohawkian facies to the Wisconsin Arch;
W. H. Twenhofel, A new species of Receptaculites from the Silurian of
Wisconsin; Rufus M* Bagg, Conquering the Frozen North, or the Romance
of mining gold near the Arctic Circle.
Friday afternoon , Section A, 2:00 — 4:15 o’clock. — Hjalmar R. Holand,
A fourteenth century battleaxe unearthed in upper Michigan; Alton K.
Fisher, The archeology of Washington Island; A. P. Kannenberg, The
Winnebago culture focus; A. P. Kannenberg, A French Traders burial plot;
Ralph Buckstaff, A cache of Ohio blue flint blades; Louise P. Kellogg,
Proposed removal of French inhabitants from Wisconsin, 1816-1820 (By
title).
Friday afternoon, Section B , 2:00-4:15 o’clock. — Rufus M. Bagg, A
new Wisconsin meteorite; Laurence F. Dake and Carl A. Bays, Insoluble
residues of the Mohawkian series; W. H. Twenhofel, The bottom sediments
567
568 Wisconsin Academy of Sciences , Arts , Letters
of Lake Monona; Ira A. Edwards, Recent drainage changes in Jackson
County; S. C. Wadmond, A Juncus new to North America; W. M. Manning,
C. Juday, and M. Wolf, Photosynthesis of aquatic plants at different depths
in Trout Lake, Wisconsin; George W. Bennett, The growth of the large¬
mouthed black bass in the waters of Wisconsin; V. C. Finch, Observations
on the distribution of cultural features on the Mississippi delta fringe.
Friday afternoon, Business Meeting, 4:30-5:30 o’clock. — The minutes of
this meeting are discussed below.
Friday afternoon , Tour of the Museum , 4:30-5:30 o’clock.
Friday evening, Academy Dinner, 6:00 o’clock. — The Academy dinner
was held in the Hotel Schroeder with 54 in attendance. Immdeiately after
the dinner members and guests returned to the Lecture Hall of the Mil¬
waukee Public museum to hear the evening address by Drs. Stebbins and
Whitford.
Friday evening, Evening Address, 8:00 o’clock. — ‘'‘Gadgets and Gal¬
axies”, by Dr. Joel Stebbins and Dr. A. E. Whitford, Washburn Observa¬
tory. About 150 members and guests attend this address.
Saturday morning, Section A, 10:00-12:30 o’clock. — John G. Gregory,
Electoral suffrage in Wisconsin; Rudolf B. Gottfried, Spenser as an his¬
torian in prose; Ruth J. Shuttleworth, University of Wisconsin genesis;
Victor S. Taylor, Jefferson County cemetery lore (by title) ; Dorothy M.
(Miller) Brown, Wisconsin circus lore; Victor S. Craun, Cave legends;
Charles E. Brown, The Wisconsin Guide, Federal Writers Project; George
Overton, The last French Traders of Butte des Morts; Paul H. Nesbitt,
Recently discovered Mpgollon culture in southwest New Mexico (By title) ;
Wilton E. Erdman, Indian mounds at Horicon and vicinity; Albert 0.
Barton, Types of Pioneer stories and songs; C. M. Partridge, The work
of the Wisconsin Federal Arts Project; J. E. Boell, Wisconsin Historical
Records Survey.
Saturday morning , Section B, 10:00-12:30 o’clock. — Edward A. Halbach
(Introduced by S. A. Barrett), Co-ordinating meteor observations by radio;
Lynn Mathias (Introduced by S. A. Barrett), Celestial photometry with
small cameras; M. J. W. Phillips (Introduced by S. A. Barrett), Amateur
telescope making by high school students; H. A. Schuette, The American
elm as a source of capric acid; Raymond Zehnpfennig (Introduced by H. A.
Schuette), The characteristics and composition of the seed oil of the hack-
berry; H. N. Calderwood, Seasonal variations in the needle oil of the White
Spruce, Picea glauca (Moench) Voss. [ Picea canadensis (Miller), P. alba
(Link)]; Frank W. Gould, The present status of the Dane County prairie
flora; Charles F. McGraw, The root systems of some Wisconsin prairie
plants — preliminary report; J. Walter Thomson, Dynamics of some prairie
plants in Juneau County, Wisconsin; Loyal Durand, Jr., Baranowice, an
estate in Polish Silesia; J. J. Yen and V. W. Meloche, Analysis by a new
micro-centrifuge tube (By title).
Proceedings of Academy
569
At the close of this sectional meeting the Secretary read a vote of
thanks to the Director of the Milwaukee Public Museum and his associates
for their splendid cooperation before and during the meeting, and to Mtr.
H. W. Cornell, President of the Milwaukee Astronomical Society, for his
valuable assistance in arranging the program.
The annual business meeting of the Academy was held immediately
following the sectional meetings on Friday afternoon. The Secretary pre¬
sented the following report on membership as of April 8, 1937: Honorary
members, 3; Life members, 14; Corresponding members, 15; Active mem¬
bers, 330; total, 362. Membership losses during the year: deceased, 6;
resigned, 14; dropped for non-payment of dues or because of loss of ad¬
dress, 20; total, 40. Applications for membership from twenty-three in¬
dividuals being presented, the Secretary was unanimously instructed to cast
the ballot of the Academy in their favor. The list follows: Paul Alcorn,
Storrs, Connecticut; Kenneth Bertrand, Madison; John Bruggink, Mil¬
waukee; Joseph J. Chopp, Milwaukee; Ithel B. Davies, Delavan; Gilbert
H. Doane, Madison ; Frank W. Gould, Madison; Hjalmar B. Holand,
Ephraim; V. E. Huntzicker, Milwaukee; Paul Icke, Madison; Paul J.
Jannke, Madison; Charles H. Krohn, Milwaukee; Eugene S. McDonough,
Milwaukee; Benjamin W. Meek, Mladison ; John P. Eetzer, Whitefish Bay;
Mrs. Jeanette Robinson, Milwaukee; Adolph G. Schwerfel, Milwaukee;
Theodore M. Sperry, Madison; Halvor O. Teisberg, Madison; J. W. Thom¬
son, Madison; Stanley A. Tyler, Madison; Ray Wilcox, Madison; and Claude
P. Zens, St. Francis.
The executive committee accepted the resignation of the Librarian,
Walter ML Smith and the Secretary-Treasurer, Robert R. Shrock, and
elected Gilbert H. Doane, Librarian, and Loyal Durand, Jr., Secretary-
Treasurer.
The Secretary-Treasurer reported informally on the present condition
of the Academy’s finances and the President then appointed H. A. Schuette
and Lowell E. Noland to audit the Treasurer’s books.
Receipts
Balance transferred to R. R. Shrock by H. A. Schuette, November
2, 1936 . .$1,204.94
Securities matured . . . . . 600.00
Securities sold . 500.00
Interest on investments . 104.72
Dues received from members . 435.50
Academy publications sold . 109.29
Milwaukee Astronomical Society . 2.00
Total
$2,956.45
570 Wisconsin Academy of Sciences , Arts, and Letters
Disbursements
Secretary’s allowance . .$ 100.00
Purchase of securities . 1,031.62
Printing of Vol. 30 of Transactions . 1,152.00
Printing of programs . 25.00
Postage . 22.00
Safe Deposit box . 3.30
Stationery . 5.60
Ledger sheets, rubber stamps, filing boxes and miscellaneous - 8.95
Total . $2,348.47
Balance in treasury, April 8, 1937 . $ 607.98
Endowment Fund
U. S. Treasury Bond . $ 500.00
Chapman Block bonds . 400.00
Home Owners Loan Corporation bond . . . 50.00
Home Owners Loan Corporation bond . . . 1,000.00
Capitol Square Realty Company bond . . . 100.00
Capitol Square Realty Company bond . 100.00
U. S. Treasury bond . . 1,000.00
Total . $3,150.00
The auditing committee, consisting of L. E. Noland and H. A. Schuette,
reported that it had examined the accounts of the Treasury and had found
them correct.
The following members of the Academy died during the past year :
Mrs. Ellen F. Butterfield (date of death unknown)
J. J. Davis, February 26, 1937
Paul B. Jenkins, August 4, 1936
J. B. Overton, March 18, 1937
Fred G. Russell, December 6, 1936
R. R. Shrock, Secretary-Treasurer
June 4, 1937.
We have examined the books of the treasurer, checked his accounts,
and examined the securities on deposit, and find them to be correct as
stated in his report for the fiscal year ending April 8, 1937.
Signed Lowell E. Noland
Signed H, A. Schuette
_