/ \
c^V ^
^ ::
ilipa
* ” ^ \
d^agmS
P' A <j
^ -V*
^ A ,::i!
a _ <►*«■
A
ifi:
%.
AP V
rA
A A
<0 . >>
% J9
"V v •■*-
a
x> 4/,
# %
r*k
j>
A"
% $
% JP
% P ,,
J*\% KgF*/ \
P % '^^^.P A
A - A A M§*^ '
*' ^ ».-. *.*,
i;.;
\ ^ w <*^®ss*
^ A
A \
A
./ %
S>
A A jpsp
^ %, fe#
Ax
V
Y> ^
o
Ax
V,
% 0^
X^3
il,ir '%
pt %.
AA %
«jF „ vqn
5>
A <r 0gpm, >
/ VfcP/’
, mm 4 VT#-
ww*- W«*
‘H.
-■3 is^q__
IS
___\
“ \ <?
> 'w
>N /a
\ p
A
-r
/ %
A~r <A
A
'a
V v fw/ ^ AtkVi
^ ftfe ^ <W
0> 4< to, ™ ,
A ' / a>
A </# o>
>0 A •-?„ A
°* A *>■ * ,;i
Z5 \ <p* .yji
<, -vni.y
A„ A
^ \ ::;;;
<$P
vV
ra
% ^
^ % t
% #
:!, li
A
v - “V
A ■=
.# A. f%
aV a, 4if.
A
Quaest
lones
entomologicae
f/i >
//!*
f $
i
II rn
1*. CP
W
co
vO
tc
Cj?
o:i
-Hjj
Co’/
o/.
,.£f
^a.//
A periodical record of entomological investigations,
published at the Department of Entomology, Uni-
versity of Alberta, Edmonton, Canada
VOLUME I
1965
CONTENTS
Editorial- Words, words, words i
Khan - Behaviour of Aedes mosquitoes in relation
to repellents 1
Book review 36
Editorial- Beastly teachers 39
Pucat - The functional morphology of the mouthparts
of some mosquito larvae 41
Freitag - A revision of the North American species of the
Cicindela maritima group with a study of hybridization
between Cicindela duodecimguttata and oregona 87
Guest editorial- Two cultures and the information explosion . . * 171
Wellington - An approach to a problem in
population dynamics 175
Wada - Population studies on Edmonton mosquitoes 187
Wada - Effect of larval density on the development
of Aedes aegypti (L. ) and the size of adults 223
Announcement 250
Corrigenda 250
INDEX
Aedes , 1,41, 187,223
aegypti , 1, 46, 64, 71
campestris > 198
canadensis , 58, 61, 69> 78, 81,
198
cantator » 29
cataphylla ) 5, 69, 71, 195
communis , 69, 73, 188, 195
dorsalis } 198
excrucians f 61,71,78,195,197
fasciata , 51
fitchi , 46,50,54,61,63,76,
78, 195
hexodontus , 59, 61, 78, 188, 195
impiger , 61
implicatus , 61, 195
increpitus , 61, 195, 197
intrudens 5, 188, 195
niphadopsis , 189
pionips , 61, 78
pullatus , 189
punctor , 5,61,71, 188, 195
riparius , 61,69, 195, 197
sollicitans » 29
Stic tic us t 61,69, 195, 197
stimulans » 61,78, 195, 197
vexons* 49,61,78
Amoore, J.E., 4,31
Anabaena , 7 6
Anderson, E. , 89, 167
Andrewartha, H. G. , 201,221
Ankistrodesmus , 76
Anopheles , 44, 72, 75, 81, 194
earlei , 72, 81
fasciatus , 44
gambiae » 223
maculipennist 24, 29, 42, 62, 80
messeae » 79
quadrimaculatus » 42, 49, 60, 190
rossi > 44
Anophelinae, 43
anopheline larvae, 80, 249
Anthon, H. , 49, 82
Anscombe, F. J. , 201,221
Apoidea, 163
Applegarth, A. G. , 56, 82
Apterobittacus , 56, 58
Asaphidion , 36
attractant, 4,21,28
Atwood, C.E., 3,34
autogeny, 223
Baker, F. C. , 219, 221
Ball, G. E. , 38
Banks, C.S., 13,31
Bates, M. , 5, 31
barrier, communication, 172
geographic, 115, 133, 165, 166
Bar-Zeev, M. , 18,31
Beckel, W. E. , 188, 221
bees, dancing, 180
honey, 120, 164
behavior , blood feeding, 1,2
feeding, 41, 72, 73
group, 66
individual, 178
mosquito, 1
orthokinetic, 72
variations, 179
Bekker, E. , 42, 82
Bembidion , 36
graciliforme > 37
humboldtienset 37
'Kl
immaturum » J '
mcrematum » J 1
nigrum f 37
Bibio » 56, 58
binomial distribution, 192,201,
208
Bishop, A. , 2, 31
Blackwelder, E. , 133, 167
Blatchley, W. S. , 103,167
Bliss, C.I. , 201,221
Bock, J. W. , 163, 167
Bowman, M.C., 28,32
Brown, A. W. A. , 4, 31
Brown, W. L. , 90, 167
Browne, B. L. , 15, 31
browser, 42,58,65,74,80,81
Burgess, L. , 28,31
Butt, F. H. , 49, 82
Cain, A. J. , 160, 167
Calliphora erythrocephalat 80
cannibalism, 73, 74
Carabidae, 36, 120
Carabus , 120
Carmichael, A. G. , 29,31
Carpenter, S. J. , 64, 82
Carr, F.S. , 152, 168
Casey, T. L. , 36
Chadwick, L. E. , 27,32
Chaoboridae, 43
Chaoborus , 43, 73, 75, 80
americanus , 41,63,73,75,78
chemor eceptor s, 1, 3, 8, 10, 15
Chir onomidae, 49
Chironomus , 46, 51, 60
hyperboreus , 80
Chitty, D., 177,185
Chlamydomonas , 77
Christophers, S.R., 3,31,42
Cicindela ,audax , 111
bellissima , 87, 94, 101
bucolica , 102
californica , 91
columbica , 87, 94, 101
depressula , 87, 92, 101, 138
duodecimguttata , 87, 91> 101,144
guttifera » 111, 126
hirticollis » 87 > 94, 101, 161
hudsonicat 102
limbata . 87,94,101,161
oregona > 87,92,101,111,144
ovalipennis ,111
praetextata ,91
prove ns is , 111
quadripennis , 1 1 1
repanda , 94, 103
s cute llaris » 111,126
sterope , 1 1 1
tranquebarica , 126
theatina » 87 ,97, 102
Cladosporium , 76
Clements, A.N., 42,82
corrigenda, 250
Coggeshall, A. S. , 80
Cohn, G. , 27, 31
color, elytra, 122, 139, 140,
pattern, 88,113,120,122,144
Compositae , 77
Contia tenuis , 143
Cook, E. F. , 42, 82
Corvus corone , 159
Cox, E. L. , 209, 222
Cryophila lapponica , 74
Cu/e*, 41, 44, 58, 74, 80, 194
annulatus , 44
Culex( cont. )
atratus , 44
fatigans , 44
molestus , 24,49,223
nemorosus , 44
peccator , 44
pipiens , 29,44,61,223
tarsalis , 6 1
territans , 51,54,58,61,72,75
Culicoides circumscriptus , 80
Culicidae, 58
Culicinae, 43
Culiseta f 41, 44, 58, 74, 80, 194
impatiens , 58, 61, 69, 78, 81
incidens , 55, 60, 78
inomata , 46,55,58, 63,66,71 ,
73,75,77,79,80,81
morsitans , 51, 54, 58, 61, 63, 66,
71, 73, 75, 77, 79, 187
current feeding, 41,68,79
Cyclops , 68, 77
Daphnia , 68
Das, G. M. , 56, 82
Davidson, R . H. , 4,32
Davies, J. T. , 4, 31
DeLong,D.M., 2,31
Dethier, V. G. , 2,32
diapause, 219
Dicaelus , 12 0
Diptera, 56, 58, 62
Dobzhansky, T. , 158, 168
DuPorte, E.M., 50,82
Dyar, H. G. , 42, 83
Dyson, G.M., 27,32
ecophenotypes , 126
elytral pattern, 88, 103, 105, 109,
112, 120, 146, 160
emergence, 218
Eucorethra , 43, 81
underwoodi , 7 3
Euglena , 77, 80
Evans, D.R., 11,32
evolution, 43, 80, 109, 133,163
Ferris, G.F. , 49,83
filter feeders, 42,58,72,74, 79
Findley, J.S., 143,168
Fisher, R. A. , 204, 221
flight, 220
food, 190, 192
shortage of, 224, 228
Foskett, D. J. , 172,174
Fowler, H. W. , 219, 221
Fraenkel, G.S., 72,83
Fragilaria , 7 6
Fr eitag, R . , 87
Frings, H. , 3, 32
Frisch, K. von, 180,185
Geminella , 7 6
genes, 158
infiltration, 150
pleiotropic, 146
geneticist, 173
genitalia, 88, 91, 102, 161
geologist, 172
Gilchrist, B.M., 2,31
Gillies, M. T. , 223, 249
Gleocapsa , 76
Goeldi, E.A., 5,32
Gomphonema , 7 6
Gordon, R.M., 5,32
Gouck, H.K. , 28,32
Gouin, F. J. , 60, 83
Graves, R.C., 103,168
Gressitt, J.S., 165,168
Gryllus luctuosus , 49
Gunn, D. L. , 72, 83
Gunther, A., ii
Haddow, A. J. , 5, 32
Hagen, H.} ii
Hamilton, C.C., 87,168
Hammond, A.R., 60,84
Hamrum, C. L. , 3, 32
Hanson, N.R . , 183, 185
Harrison, G.A., 160,167
Hatch, M. H. , 111, 168
Haufe, W.O. , 190,221
Hayward,R., 36
Henry, L. M. , 49, 83
Hinton, E. H. , 56,83
Hocking, B. , ii, 19, 40, 32
Hodgson, E.S., 29,32
homodynamy, 163
Hooke, R . , 41,83
Horn, W. , i
Horsfall, W.R. , 190,221
Howard, L. O. , 42, 83
Howland, L. J. , 80, 83
Howlett, F.M., 4,33
Hoyle, F., 172,174
Hubbell, T.H., 120,168
Hubbs, C. L. , 133, 168
hybrid index, 87, 89, 105, 144* 146
zone, 89, 90, 144, 160
hybridization, 87, 144, 150, 152,
158, 163, 166
Imms, A.D., 56,84
Inger, R . F. , 90, 169
intergradation, 87, 90, 103, 134,
144, 148, 159
introgression, 126, 152, 162
isolation, 159, 166
differentiation, 143
geographical, 164
spatial, 134
James, H. G. , 74> 84
Johannsen, O.A. , 42,84
Johnston, J. W. , 4, 31
Jones, F. N. , 4,33
Jones, J. C . , 60, 84
Kalmus, H. , 1,33
Kellogg, F.E., 5,33
Kemper, H. , 28, 33
Kendrew, W. G. , 125,169
Kennedy, J.S., 24,33
key, 101
Khan, A. A. , 1
Khelevin, N. V. , 219,222
King, P. B. , 164, 169
Klomp, H. , 224, 249
Knab, F. , 42,83
Knight, K. L. , 189, 222
Krishnamurthy, B . S. , 223,249
Kupka, H. , 28, 34
LaCasse, W. J. , 64, 82
larvae, mosquito
active, 180
browsing, 41,51,62,70,74,77
density, 223
development, 223
filter feeding, 41,42,63,77
labium, 59
labrum, 50
mortality, 223
non- predatory, 41
predatory, 41, 43, 63, 73, 74, 77
overcrowding, 223, 249
sluggish, 180
Laven, H. , 223,249
Leng, C. W. , 87, 169
Lepidoptera, 62, 163
Lindroth, C.H. , 36
Linsley, E. G. , 163, 169
Lotmar , R . , 8,35
Lumsden, W.H.R., 5,32
Lulzia » 43
halifaxi > 50
Macfie, J. W.S., 10,33
Malacosoma pluviale , 175
Manton, S.M., 46
Martin, P. S. , 164
mating, 13,25,27,157
Mayr, E. , 90, 169
MacGinitie, H. D. , 164, 169
McGregor, D. , 80, 84
McLintock, J. , 46, 84
Mecham, J.S., 158,169
mechanor eceptor s, 1,3,29
Mecoptera, 46, 56
Meinert, F. , 42, 84
Melanoplus puer , 120
Mellon, De F. , 29, 32
Menees, J.H., 46,84
Mengel, R.M. , 164, 169
Miall, L. C. , 42, 84
Microspora , 7 6
migration, 220
Miller, R.R. , 133, 168
Miocene, 164
Mitchell, E., 42,84
Mochlonyx , 43, 75, 80, 81
culiciformis , 74
'velulinus > 41,63,73,74,78
Montchadsky, A.S., 42,84
Morita, H. , 30, 33
morphology, 91
Morris, R.F. , 204,222
mortality, 187,189,221,223,
225,235
mosquitoes, black-legged, 187
control of, 187,217
Edmonton, 187
mouthparts, 41,42,64,81
mutation, 109, 163
Navicula , 76
Nearctic, 165
Nematocera, 46, 56, 58, 62
Nuttall, G.H.F., 42,85
Ochlerotatus , 187, 196, 197
Olbiogaster , 49
olfaction, 4, 27
Omus californicus
Oncopeltus fasciatus , 49
Opifex fuscus , 80
overwintering, 220
oviposition, 15, 17, 25, 27, 30,
190, 194, 219, 221
Owen, A. R . G. , 202,221
Palearctic, 165
Panorpa , 56, 58, 62
Panorpoidea, 58
Papp, H. , 91, 169
Peffly, R. L. , 4, 32
Peters, W. , 8, 33
Peterson, A. , 43, 85
Phacus , 77
phenology, 139
Phormia regina , 11,15,29
Phrypeus > 36
phylogeny, 160
P innularia , 7 6
Pinus , 77
Platt, J.R., 177
Pleistocene, 133, 144, 158, 163
Pliocene, 165, 166
Poisson distribution, 200, 203
population, allopatric, 158
alpine, 122, 125
boreal, 122
density, 190,192,201,210, 216
desert, 125
dynamic s, 175,221,223
ecology, 177, 184
literature, 176
primitive, 166
samples, 89, 121
studies, 187
theory, 177, 179, 184
world, 171,172
Populus , 77
Potter, E. , 56, 85
predators, 43,73,74,81
Provost, M. W. , 220, 222
Pucat, A.M., 41
pupation, 224, 228
Puri, I. M. , 42, 85
Putnam, P. , 14, 34
Quate, L. W. , 163, 169
Quiscalus quiscula 159
Rahm, U. , 3, 33
R ana aurora * ^43
Rao, T.R. , 24, 34
Raschke, E. W. , 42, 85
Rausch, R . L. , 120,169
Reaumur, M. , 41,85
receptors, 1
olfactory, 4, 16, 30
R eed, W. , 1
Rees, B. E. , 56, 83
Rempel, J. G. , 64,85
Renn, C. E. , 42,85
repellents, 1
R euter , J. , 4, 34
Richards, D. W. , 14,34
Ridgway, R., 88,169
Rivalier, E. , 91, 169
Roeder, K.D. , 29, 33
R oss, R . , 1, 34
Roth, L.M. , 3, 34
Rubin, M . , 4,31
Rumpp, N. L. , 91,170
Sabrosky, C. W. , 163,170
Salem, H. H. , 51, 85
Saltatoria, 164
Sass, J. E. , 75,85
Scenedesmus, 76
Schenkling, K. , i
Schremmer, F. , 43, 85
Sekhon, S.S., 3,34
sex hybrid, 90
Shaerffenber g, B. , 28, 34
Shalaby, A.M., 46,85
Shannon, R . C. , 14, 34
Shelford, V. E. , 122,170
Shipley, A.E., 42,85
Short, L. L. , 89, 170
Shute, G. T. , 223, 249
Sibley, C. G. , 89, 170
Simulium , 7 9
Slifer, E.H. , 3, 34
Smith, C.N., 29,35
Snodgrass, R.E., 42,86
Snow, C . P. , 172
Sorex vagrans , 143
Southwood, T.R.E., 220,222
Spielman, A., 223,249
Spiro gyra , 76, 77
Stace-Smith, G. , 148
Stagmomantis Carolina , 49
Stahler , N. , 223,249
Stauroneis f 76
Stebbins, R.C., 143,170
Steward, C. C. , 3,34
Sturckow, B. , 30, 34
Sturtevant, A. H. , 163,170
subspecies, 125,434,139,143
sugar feeding, 10,25,27, 30
Sullivan, C.R., 177,185
Surtees, G. , 42, 86
Sylvester-Bradley, P.C.,163,170
Sylvester, E. S. , 209,222
synonymy, 103, 112, 139
Systematic Zoology, 90
taxonomist, 173
taxonomy, 97, 125
teachers, 39
Telford, A. D. , 219, 222
temperature, 190, 192, 219, 233
Tertiary, 163,164
Theobaldia incidensJS5 , 60
thermoreceptors, 1,2,30
Thiel, Van, P.H. , 4, 34
Tipula , 56, 58
Travis, B. V. , 29, 35
Trechus , 36
Trembley, H. L. , 46, 86
Ulmus , 7 1
variation, color , 105, 106,112,141
geographic, 89, 103, 112, 134
interspecific, 88, 91, 97
intraspecific, 88
population, 109, 112
Venard, C.E., 4,32
Vimmer, A. , 49, 86
Vockeroth, J.R., 188,222
Wada, Y., 187,223
Wallis, J. B. , 4, 35
Waters, W.E., 201,222
Weismann, R., 8,35
Wellington, W. G. , 175,185
Wesenberg- Lund, C.N., 42,86
Wheeler, W.M., 49,86
Williams, T.R., 80,86
Willis, E. R. , 3, 35
Wilson, E. O. , 90, 170
Winteringham, F.P.W., 174
Wright, R. H. , 5, 35
Xiphidium ensiferum , 49
Yost, M. T. , 8, 32
Zeuner, F. E. , 163,170
zoogeography, 160, 163
Quaestiones
entomologicae
A periodical record of entomological investigations,
published at the Department of Entomology, Uni-
versity of Alberta, Edmonton, Canada.
VOLUME I
NUMBER 1
JANUARY 1965
QUAESTIONES ENTOMOLOGICAE
A periodical record of entomological investigations, published
at the Department of Entomology, University of Alberta, Edmonton,
Alberta.
Volume 1 Number 1 2 January 1965
CONTENTS
Editorial i
Khan - Behaviour of Aedes mosquitoes in relation
to repellents 1
Book review 36
Editorial - Words, words, words
The first edition of the World List of Scientific Periodicals,
published in 1921, listed 25,000 titles. The second edition in 1934
listed more than 36,000; the third edition in 1952 listed more than
50, 000. The fourth edition now appearing lists over 60, 000, despite
the fact that "some 10, 000 titles included in the third edition have been
left out as being of social or commercial rather than scientific inter est".
Most periodicals have recently waxed fat, so that one may estimate 25
years as the time in which the flow of scientific literature doubles itself.
By comparison with science as a whole, the growth of entomo-
logical literature seems somewhat pedestrian; the Insecta portion of
the Zoological Record listed 1970 titles of papers in 1921 and 4024 in
1953. The applied literature, as represented by the Review of Applied
Entomology has been, somewhat surprisingly, growing more slowly
than this, so that one may estimate 35 years for the entomological lit-
erature to double itself. Even so the 25,229 entomological articles
listed in Horn and Schenkling as published from the beginning of history
until the end of 1863, at current rates would be produced in about four
years, and the total number of scientific papers now published in the
field of entomology must exceed a quarter of a million. One may sus-
pect, however, a shrinkage in the mean length of papers under the
joint influence of mounting page charges and the philosophy of "publish
or perish" coupled with the waning ability of administrators to judge
publications by anything beyond their number.
Some may say that in this situation a new periodical should be
offered with an apology - if at all. But if we would slow down the march
of science, we must stop research before it has begun, not lose the
results of it when it is all but finished. Certainly we must see to it
11
that we do not produce new facts faster than we can assimilate them
into generalizations, although this process calls for that very breadth of
outlook which the literature flood makes it difficult for us to achieve.
If we can no longer achieve individual breadth, we must provide for
composite breadth by facilitating diversity of training and the unusual
combination of subjects.
If we stagger under the impact of a swelling literature, before
we call for a slow down in research we should remind ourselves that a
quarter of a million entomological papers only represents less than one
per described species of beetle, and that more than half the species of
insects remain to be found and described.
If then, this growth of the literature must go on, what can we do
to keep abreast of it? A great many things: fight the trend to shorter
papers, which has now reached the ridiculous stage when an index card
for a paper may be larger than the content of the paper itself. It costs
more in time, money, and effort, to produce, file, store, retrieve, and
read ten one page papers than one ten page paper. Publish in the most
appropriate periodical from the subject viewpoint; publish promptly;
index and abstract everything diversely; and make full use of modern
techniques such as microforms, punch cards, and even computers. It
may seem redundant to say that material should be published once only,
yet how often do we find it difficult to avoid duplicate publication of
material from the proceedings of a meeting, and how often is this due
to inappropriate publication in the first place? A marriage between
microcards and punch cards is long overdue; if sufficiently prolific, the
hybrid offspring would be of inestimable value to the bibliogr? pher .
There are signs that things are beginning to move in this direc-
tion; perhaps this periodical is one of them. But one may question
whether the move is fast enough to get us out of chaos: movable type ,
despite its name, is conservative stuff.
Despite our concern for the future, we should be both remiss
and churlish to enter 1965 without a backward glance to 1865 and the
beginning of the Zoological R ecord. Let us pay both dollars and respect
to our venerable abstracting and indexing services - in no other field
of endeavour is continuity more important. I wonder whether any other
branch of science is as fortunate as entomology with its Hagen, Horn
and Schenkling, and Zoological Record. Many complain of the increasing
delay in publication of successive volumes of Zoological Record, but
how many of the complainants have ever attempted a similar task?
And whose fault is this? As Glinther pointed out in his preface to vol-
ume one in August 1865, many journals of learned societies which would
carry the date 1864 on their title pages, had still not appeared; but here
we are treading on dangerous ground. We regard it as a most fortunate
and propitious honour, to commence publication in the year in which the
Zoological Record celebrates its centenary.
Brian Hocking
EFFECTS OF REPELLENTS ON MOSQUITO BEHAVIOR
Department o t Entomology Qvaestiones entomologicae
University of Alberta i.i-jo. ivo d
The behavior of Aedes aegypti L- and other species of Aedes in relation to
repellent chemicals was studied. The repellents used were dimethyl phthalate, ethyl hexaned-
iol, N, N-diethyl metatoluamide and indalone. The effect of these repellents on the behaviour
of mosquitoes was studied firstly by placing the repellents on selected parts of the environment
and secondly by painting them on parts of mosquitoes themselves where chemoreceptors are
known to occur, such as the antennae, labium, and tarsi. The aspects of behavior studied were:
feeding on blood and on sugars, mating, oviposition, the reactions to wind, geotaxis and orient-
ation to centrifugal force, and the visual response to black stripes. All these aspects of behav-
ior are affected significantly by repellents. Dimethyl phthalate has the greatest effect of the
four repellents on blood feeding behavior when they are painted on the tarsal receptors and the
smallest effect when they are painted on the receptors of the antennae and the labium.
The experiments provided some understanding of the mode of action of insect repel-
lents. They suggest that repellents interfere with normal behavior perhaps by blocking the
olfactory receptors mediating attraction to food and the contact chemoreceptors invoking feeding
on blood and those used in the selection of oviposition sites. The experiments show that mech-
anoreceptors effecting orientation to gravity and air flow and visual receptors effecting orientat-
ion to black stripes are also interfered with by repellents. There is also some evidence that
repellents block the thermoreceptors which may mediate piercing for feeding on blood and perhaps
auditory organs involved in mating. The only receptors which the repellents do not appear to
interfere with seem to be those of the common chemical sense.
INTRODUCTION
The discovery of the transmis sion of malarial parasites by Ross
(1898) and the discovery by Walter Reed and his collaborators that
yellow fever was transmitted by Aedes aegypti led to the realization of the
importance of mosquitoes as carrier of disease. Repellents being a
cheap and efficient means of individual protection, many worker s studied
their effects mainly against the blood feeding behavior of insects.
Kalmus Hocking (I960), however, studied some other aspects of
behavior as well. I have studied the behavior of Aedes aegypti in the
presence of repellents not only in relation to blood feeding but also in
relation to sugar feeding, mating, oviposition, geotaxis, wind direction
and speeds, and visual responses to black stripes. The repellents used
were: dimethyl phthalate, indalone, diethyl toluamide, and Rutger's
612. The first two are esters, the third an amide and the last named
an alcohol. These are compounds of low volatility and moderate
molecular weight. They are insoluble or only very slightly soluble in
water but are miscible with alcohol and ether. Their physical and
chemical properties are listed in table 1.
z
Repellent Effects
TABLE 1 - Chemical and physical properties of the repellents used in
the study of behavior oi Aedes .
Blood feeding behavior was studied by applying the repellents
in various ways in the environment and on different chemosensory fields
of female Aedes aegypti. Some general observations were also made on the
blood feeding behavior of Aedes spp. mosquitoes in the field and Aedes aegypti
in the laboratory. A preliminary test was made of the effect of washing
chemosensory areas with lipoid solvents on blood feeding by Aedes aegypti.
REVIEW
Sense Organs of Aedes aegypti L.
An Aedes aegypti female is attracted to its host in part through the
chemoreceptors located on head appendages, mainly the antennae.
Bishop and Gilchrist (1946) showed that in Aedes aegypti eyes are not
essential for feeding on blood. Roth (1951, p. 60) also reported that
eyes are not necessary in locating the host in a small cage.
DeLong (1946) considered the anteranae and the palps as the chief
organs for locating the host and stimulating probing. According to him
the antennae may perform both functions but the palps can receive
stimuli only when the insect is directly on the skin. Roth (1951)
considered that the antennae function as directional thermoreceptor s and
probably chemoreceptors as well. R oth ( 1951 ) also reported temperature
receptors on the palps of A. aegypti. Dethier (1952) considered that
Khan
3
different receptor fields function at different levels of sensitivity. The
antennae according to him are the most sensitive and the various mouth-
parts less so. Rahm (1958) showed by antennal amputation that these
organs are essential for host finding and attraction from a distance.
He also reported that antenna-less mosquitoes can probe and suck if the
palps remain intact.
Antennae
The antennae in the male and female consist of a basal ring-like
scape, aglobular pedicel, and a long flagellum of thirteen articles. The
pedicel inboth sexes contains Johnston’s organ, which is more developed
in the male.
Roth and Willis (1952) reported that many thin walled trichoid
sensilla are present on each of the thirteen flagellar articles of the
female A. aegypti and on the two terminal flagellar articles of the male.
They concluded on experimental evidence that they serve as hygro-
receptor s .
Christophers (I960, p. 663) described the trichoid sensilla as
. .40-50 p, in length, thin walled and without articulated base, arising
from thin membrane over a pore canal surrounded distally by a semi-
circular ridge in the article. ”
Steward and Atwood (1963) identified five structural types of
sensilla on the antenna of the female A. aegypti. Three of these types they
found thin walled and classified them as Al, A2 and A3. According to
them a typical Al sensillum is 0. 06 mm long, curved and tapering to a
sharp point. Type A2 is shorter, 0. 04 mm long and with a blunt tip.
Both are about the same diameter. The innervation of the two types is
essentially the same. Steward and Atwood described type A3 as a short,
curved, thin-walled peg organ which is innervated by a group of nerve
cells. Sensilla of type Al and A2 are more numerous on the distal
articles while sensilla of type A3 are found to be located chiefly on the
proximal articles of the antennal flagellum. They concluded from
experimental evidence that type Al and perhaps A3 play a major role in
mediating attraction while type A2 are responsible for mediating
repulsion.
Slifer and Sekhon (1962) studied the structur e of the sense organs
in the flagellum of A. aegypti. The heavy walled hairs according to them
ar e mechanor eceptor s . The thin walled hair s with sharp tips they thought
to be chemor eceptor s . The thin- walled hairs with blunt tips they
supposed to be olfactory in function.
Palpi
Roth and Willis (1952) described the palps of female Aedes aegypti
as abundantly supplied with thin- walled club-shaped sensilla on the
terminal segment. Pointed trichoid sensilla are also present. There
is also a central short sclerotized peg at the tip of the palp.
Labium
Frings and Hamrum (1950) noted four kinds of hairs on both
sexes of A. aegypti. Of these, hairs about 40 \l long and lying at the tip of
4
Repellent Effects
the labella are considered to be tactile in function while curved hairs
about 20 pin length at the tip and on the ventral surface are believed to
be chemor eceptor s .
Tarsi
On the tarsi of the fore and mid legs of A. aegypti are many
slightly curved hairs probably tactile in function (Frings and Hamrum,
1950). Wallis (1954) found that in A . aegypti all tarsal segments were
provided with thin-walled curved spines. Slifer (1962) described the
hairs on the tarsias approximately 100 in number in the female. These
hairs stain at the tip when dye is applied to the external surface of the
insect. She concluded: "Little doubt now remains that the hairs with
stainable tips are the tarsal gustatory receptors of the mosquito. "
Mode of Action of Olfactory Receptors
Several theories have been advanced to explain the mode of action
of olfactory receptors. Jones and Jones (1953) reviewed the modern
theories on olfaction and classified them as; mechanical, chemical,
steric, radiation and vibration theories.
Davies (1962) proposed that the mechanism of olfaction is the
penetration and dislocation of a small region of the wall of an olfactory
nerve cell. This dislocation allows the potassium and sodium ions to
move across the membrane, so initiating the nerve impulse.
Amoore (1963), and Amoore, Johnston and Rubin (1964) favor
the stereochemical theory of olfaction. According to them the odor of
a chemical is determined by the structure of the molecule, in particular
by its size and shape. If a chemical is volatile, and its molecules have
the appropriate configurations to fit closely into the receptor site, then
a nerve impulse will be initiated, possibly through a mechanism involving
disorientation and hence depolarization of the receptor cell membrane.
Factors Attracting Mosquitoes to the Host
The mode of action of repellents cannot be fully studied without
an understanding of the factors that attract the insect to the host.
Contradictory views can be found in the literature on this point; all
workers accept temperature and humidity, as attr actant factors; others
consider factors like carbon dioxide and host odor, or only carbon
dioxide to be also important in attracting the mosquito to its host.
Howlett (1910) believed temperature to be the chief attr actant
and said that the smell of sweat or of blood was not attractive. Reuter
(1936) showed that moisture was distinctly attractive to A. aegypti. Van
Thiel (1937) assigned the role of attr action chiefly to the physical factors
of temperature and humidity and the chemical factor, carbon dioxide.
Later Van Thiel (1953) considered that the scent of the host plays an
important part in the orientation of the mosquito toward it.
DeLong, Davidson, Peffly and Venard (1945) found moistened
warm air more attractive to A. aegypti than warm air. Most of their tests
were conducted with olfactometers or inanimate objects. Brown (1958)
recognized six factors which guide female mosquitoes to their animal
hosts, three of these being air-borne (water vapor, carbon dioxide, and
Khan
5
convective heat) and three visual (movement, contour, and r eflectivity) .
Kellogg and Wright ( 1 957 ) and Wright ( 1 96Z ) considered moisture
and carbon dioxide to be the main attractant factor s . Christophers (I960,
p. 535) remarked: "The evidence that smell is an important stimulus in the
attraction of A. aegypti to feed is not very strong. "
On the other hand, many have said that body odor plays an important
role in the attraction of mosquitoes . Goeldi (1905) reported per spiration to
be the agent attracting mosquitoes to man. Haddow ( 1 942 ) r eported that an
unwashed African child attracts more Anopheles spp. than a clean child.
Willis ( 1 947 ) r eported that females of A. aegypti and Anopheles quadrimaculatus
Say were attracted by the odor of the human arm. He also found CO^ in con-
centrations of 1, 10, or 50 per cent in the air not attractive to females of A.
aegypti or Anopheles quadrimaculatus when tested in an olfactometer. Bates
(1949) thought smell to be the primary stimulus in guiding the mosquito in
its sear ch for food. R ahm ( 1 956 ) r eported that CO2 emitted by the skin did
not determine attractiveness and remarked ( 1 957) that human odor and
sweat may play a part in the attraction of mosquitoes to the human hand.
Again in 1 957 he r eported that per spiration did not seem to attract mosq -
uitoes but the odors given out by the host did. Rahm ( 1958) further remark -
ed that the olfactory substances of man were found to be alone responsible
for greater activity offemale A. aegypti. Dethier (1957) wrote: "Host finding
and discrimination, trail following, orientation to odor s by flying insects
and courtship are shown to depend largely on the chemical stimuli. "
EXPERIMENTAL - BEHAVIOUR
Blood Feeding in Relation to Repellents
Christophers (I960, p. 486) remarked on blood feeding by a. aegypti
in the following words: "Another striking feature of feeding is that the ins-
ect once it has begun to suck blood, appears to become oblivious to all dan-
ger and considerable physical force is required to make it give up its hold."
Thi s featureis referredtoby Gordon and Lumsden (1939 )who wr ote that they
were only able to get A. aegypti tofeed on the frog's foot by employing mos -
quitoes which had been allowed to start feeding on the human arm. When
nearing repletion, however , the insect usually leaves readily if disturbed.
Kalmus and Hocking (I960) observed the effect of painting repellent
with a fine camel hair brush on the backs of feeding mosquitoes . A lead was
taken from this study andmore observations were made on the effect of re-
pellents on other species of Aedes in the field and Aedes aegypti in the labor-
atory.
Observations on Aedes spp. in the Field
For studies on the species of Aedes in the field a thicket of poplar
trees was selected. The four repellents , dimethylphthalate, ethyl hexan-
ediol, indalone, andN-N-diethylmetatoluamide were used. The mosqui-
toes reacted to all four repellents in the same way. The species of Aedes
studied were A. punctor Kirby, A. cataphylla Dyar, and A. intrudens Dyar.
The time to take a complete blood meal, from the insertion of
Repellent Effects
6
the proboscis to its retraction after complete engorgement ranged from
two to four minutes. (Mean = 2 min 31 sec with standard deviation 41
sec). It was observed that the mosquitoes could be very easily disturbed
in the early stages of their blood meal. If a clean brush were brought
near them soon after the insertion of the proboscis, they could be seen
retracting it. If a repellent or olive oil were placed near the antennae
or painted on the mesonotum, the mosquitoes invariably flew away. As
reported byKalmus and Hocking ( 1 96 0, p. 7 ) "A contact between r epellent
chemicals as liquids and substantial areas of the proboscis, tarsi and
tibiae, mesonotum or the wings leads to the interruption of biting, and
in mosquitoes not engaged in biting to the retraction of the touched limb
or limbs or to take off. ^ But the behavior of mosquitoes was found
quite different in relation to repellents and other stimuli if they had been
feeding for a minute or more, i. e. roughly in the middle of their meal;
e. g. :
(i) The mesonotum was rubbed with a dry brush, painted with
repellents or olive oil until the whole mesonotum was covered with
liquid, but the mosquito never flew away, instead it completed its blood
meal, continuing to feed for another 45 seconds tov one minute.
(ii) The antennae were painted with repellents, were in fact
soaked in repellent, but the mosquitoes continued to feed.
(iii) A drop of repellent was made to flow near the tarsi, there
was no reaction until it made contact with them. As soon as contact
was made the tarsus was lifted. The same reaction was observed with
olive oil. However, the mosquitoes continued to feed even when the
tarsi of all the six legs were lifted. The mosquito then came to rest on
its abdomen. When the repellent was presented on a brush near the
lifted tarsi, they sometimes rested the tarsi on the repellent soaked
brush, without showing any other abnormal behavior, and continued to
feed.
(iv) Similar behavior was observed inmosquitoes feeding on the
foot through socks. Mosquitoes coming to feed landed only on clean
areas of the sock and avoided areas where repellent had been placed.
However, mosquitoes which had been feeding through the sock for some
time were not affected if a repellent was placed on the sock underneath
them, and they continued to feed to completion although they lifted the
abdomen.
(v) Chloroform or ether was brought near the abdomen of a
feeding mosquito. It always flew away, even when it had been feeding
for a minute or more.
( vi ) A hot spatula was brought near the mosquito (about 1 mm).
The spatula was heated for two minutes in a flame of a spirit lamp.
Eighty per cent of the mosquitoes took off in 5 to 10 seconds. When the
spatula heated for the same time was kept at the same distance from
the mercury bulb of a Fahrenheit thermometer, the thermometer
registered a rise of 4-6 degrees.
(vii) Repellent was painted on the wing of a feeding mosquito.
The mosquito always flew away but when the wing was rubbed with a dry
brush or painted with olive oil it continued to feed.
(viii) Physical injury was inflicted on the mosquito to the extent
Khan
7
that all the six legs were clipped off at the femoro - tibial joint, but it
continued to feed and did not fly away.
The observations were made at a temperature of 65°F and R.H.
of 57%.
Observations on .Aedes aegypti
In the laboratory the same behavior was studied in Aedes aegypti.
A one cubic foot cage made of steel wire and covered with nylon net was
fitted with a sleeve on each of two adjacent walls, i. e.at right angles to
one another. Mosquitoes were allowed to feed on a hand inserted through
one sleeve while the other hand was introduced through the other sleeve
to apply the repellent.
As observed in the other species of Aedes, Aedes aegypti couldalso
be easily disturbed in the initial stages of blood feeding, but after one
minute of feeding they could not be disturbed so easily:
(i) When the mesonotumwas rubbed with a dry brush or painted
with olive oil or any of the four repellents under study.
(ii) When their wings were painted with repellents. This was
contrary to the behavior observed in the field spe'cies which invariably
flew away whenever repellents were painted on the wings.
(iii) They continued to feed even when they were made to rest
their tarsi on the repellent soaked brush.
(iv) Being small in size, it was not possible to paint their antennae
with repellent while they were feeding, but when a drop of repellent was
placed very close to the proboscis they continued to feed.
(v) Almost every mosquito continued to feed when the tarsi of
its hind legs were clipped off, but some flew away when the tarsi of
their other legs were clipped.
(vi) When a heated spatula was brought near them they always
flew away even when the spatula was as far as 1-2 cm away. It had to
be brought much near er to mosquitoes in the field to elicit this response.
When the spatula heated for the same time was kept at the same distance
from the mercury bulb of a Fahrenheit thermometer this registered a
rise of 1. 5 to 2 degrees.
Experiments were conducted by applying the repellent on differ ent
chemosensory fields of female A. aegypti and observing the behavior and
recording the number feeding on an untreated human arm. As the
repellent was not applied on the skin, there was no interaction between the
skin and the repellent or the chemical stimuli emanating from the skin and
the repellent on the surface of the skin. The experiments provided some
under standing of the site of action of different repellents as well as
providing a quantitative basis for comparing the repellents with each
other. The experiments also provided a quantitative basis for evaluating
the function and efficiency with which the different chemosensory fields
play their role in the act of feeding as well as some grounds for accepting
the role of smell in attracting mosquitoes to feed and the function of
the repellent when applied on the skin in offsetting this role.
8
Repellent Effects
10- 12 female mosquitoes, 7-8 days old, previously fed on raisins and
sugar solution only, in a sucking tube and then chilling them for 1. 5 min
at 15°F, in order to immobilize them. Their probosicides , either one
or both antennae, or all the tarsi, were then painted with repellents with
a fine brush in separate sets of experiments. This operation was
performed over a cold petri dish covered with a filter paper and placed
under a binocular microscope. A radius was drawn in ink on the filter
paper and mosquitoes were treated one by one, starting on one side of
the radius until all of them were treated. They were then sucked back
into the sucking tube and released in a paper lined petri dish to revive
in a one cubic foot cage of steel wire covered with nylon net. The mos -
quitoes recovered from the chill in 2-3 minutes. The behavior and the
number that fed on blood on introducing the arm into the cage through a
sleeve were noted, firstly ten minutes after the treatment and then at
greater intervals from the treatment until the number fed in a given time
approached the number fed in controls. Two controls were run with
each set of experiments, one a plain control when the receptor field
that was intended to be treated was rubbed with a vdry brush only, and
another when it was painted with olive oil. The palps could not be treated
separately without running some repellent on the proboscis and the
antennae, because of their close proximity to these structures.
Results - The figures given in table 2 give the cumulative mean
percentages of mosquitoes feeding on blood after different chemor eceptor
sites were painted with repellents . The standard error of themean was
used to find statistical significance between the means.
The results show that Rutger's 612, diethyl toluamide, and
indalone reduce the number of mosquitoes feeding on blood more than
dimethyl phthalate after the first ten minutes when the proboscis was
painted, and the effect lasted longer. Indalone remained significantly
more effective as compared to Rutger's 612 and diethyl toluamide after
two hours when painted on the proboscis.
When painted on both the antennae, diethyl toluamide, Rutger's
612 and indalone again reduced the number of mosquitoes feeding more
than dimethyl phthalate. The effect of dimethyl phthalate was found to
have been lost within one hour but the effect of the other three repellents
lasted more than six hours.
When painted on one antenna, the same significant differences
were found between the repellents as when both the antennae were
painted, i. e. , diethyl toluamide, Rutger's 612 and indalone were
significantly more effective than dimethyl phthalate.
The results obtained on painting all the tarsi with repellents
were, however, different. Dimethyl phthalate was found to reduce
feeding more effectively when painted on tarsi than when painted on
both the antennae or on the proboscis, and to maintain this effect at
least as long as the other three materials.
There is evidence that many repellents work byway of specialized
chemor eceptor s (Weismann and Lotmar, 1949; Dethier and Yost,
1952; Peters, 1956; Dethier, 1956 a). Peters (1956) reported that
Calliphora erythrocephala could detect dimethyl benzamide with the tarsal
Khan
9
a)
rd •
U CO
4-> U
d £
lj fn
S <0
MH "d
id u
^ as
n 'g
0 5
+J TO
MH +->
d to
XS +1
^ 0
aj d
- 3
Cfl Ph
0 0
d ^
s s
® o
’S ^
& 0
^ CO
cn d
. d
^ 0
0 co
OjO +*
d S
4J 0
fl H
<1> 0
9 f1*
H 0
•5 ^
£ 'S
g a
H -H
2 d
U A
d
O
•rH
4->
d
>
0
co
£
o •.
a "d
0 0
U d
CO 00 CM
+! +1 +1
ro CO ^ ro
+1 +1 +1 +1
o- ro
. o .
CM r-H (M
+1 +1 +1
CM CO uD
ID CM
+! +1
10
Repellent Effects
receptors only, while other materials like indalone and dimethyl carbate
could be detected with the tarsal receptors, labella, and antennae.
The significant difference in the number of mosquitoes landing
on the hand after treatment of chemor eceptor s on different head
appendages and on the tarsi can be explained on the basis of the population
of chemor eceptor s getting such treatment. As most of the chemo-
receptors are situated on the antennae, their treatment with repellents
would inhibit the landing of mosquitoes on the hand more than the treat-
ment of other head appendages. The ineffectiveness of the painting of
one antennae only in keeping the mosquitoes from a blood meal for two
hours can be explained by the same argument, i. e. , a large population
of chemor ec eptor s r emained functioning effectively when only one antenna
was painted. The painting of any one of these chemor eceptor sites with
repellent must be affecting the mosquito in two ways, affecting the
chemor eceptor s of the chemosensory area painted in liquid form and
also affecting the adjacent chemosensory sensilla in vapor form. The
greater the area painted, the greater the number of sensilla affected,
resulting in inhibition of feeding for a longer period.
Painting Repellent on Mosquito Antennae and Host Skin
3y the procedure described above one antenna of each of about 10
A. aegypti females was painted with diethyl toluamide. An arm also
treated with diethyl toluamide was then introduced into the cage and the
behavior of the mosquitoes was studied. A little more flight activity
and some searching on the wing was observed in these mosquitoes as
compared to those in the control where no repellent was applied on the
mosquitoes themselves but only on the hand. A similar behavior was
observed in experiments with the other three repellents as well. In
controls, mosquitoes were seen mostly sitting on the walls of the cage.
There was little or no flight activity.
When both the antennae of mosquitoes were treated with diethyl
toluamide, indalone or Rutger’s 612, and the same repellent was applied
on the hand introduced into the cage, the mosquitoes could be seen
searching on the wing. Many landed on the repellent coated surface of
the hand, walked about and even probed but did not take a blood meal.
The behavior was observed for ten minutes every hour for four hours
but no mosquito bit. In similar experiments with dimethyl phthalate,
however, no landings on the hand were observed though the mosquitoes
came quite close to it and sometimes even touched the skin.
Sugar Feeding
The principal food of female Aedes aegypti is blood from a human
host though they can exist for long periods on food other than blood .
Male Aedes aegypti do not take blood at all but feed entirely on sugary
materials. Goeldi (1905) kept females alive for 31 to 102 days on honey
alone. Macfie (1915) observed that the females feed on honey for the
first couple of days but the males feed only on honey at anytime. Gordon
(1922 b) observed both males and females of Aedes aegypti sucking nectar
from flowers. Many observers have noted that sugary fluids, raisins,
bananas, and other fruits are sucked by both sexes.
Khan
11
Many workers have devoted much time to studies of the effect of
repellents on blood feeding of mosquitoes but their effect on sugar feeding
has not attracted much attention. Evans (1961) has studied the effects
by the blowfly Phormia regina Meigen. Experiments were conducted to
study the effect of repellents on the feeding of Aedes aegypti on raisins.
Kalmus and Hocking (I960) conducted some tests on blood
feeding in relation to repellents with Aedes aegypti by keeping a 10 cm
length of 3mm outside diameter glass tubing which was clamped in a
vertical position so that the lower end was about 1 cm above the middle
of a 6 cm bare circle on the back of a gloved hand. A few drops of
repellent were placed in the lower end of the tube. In this way a circle
of skin about 1. 5 cm diameter was kept free of bites. A lead was taken
from this experiment in exploring the effect of repellents on the feeding
of Aedes aegypti on raisins.
Experiment 1
About 100 male and 100 female mosquitoes were taken in a cubic
foot cage of steel wire covered with nylon net. Thqage of the mosquitoes
was 2-4 days and they were not fed anything for six hours prior to
experiments. Ten raisins were fixed with 1 cm clear space between
each on a horizontal steel wire hanging 4 inches below the top of the
cage. The wire was hung by bending its ends and hooking them on top
of the side walls of the cage. A 2 cm wide strip of paper was fixed
above the raisins, running parallel to them at a distance of 1.5 cm.
Half of this paper strip (covering 5 raisins) was painted with repellent
and the other half (covering the other 5 raisins) was kept as a control.
Observations were made on the number of mosquitoes settling on either
side at intervals of 5 minutes. After each observation the cage was
shaken and another observation recorded after five minutes. In this
way five replicates were taken for each repellent. Separate batches of
mosquitoes were taken in separate cages for experiments with different
repellents .
The observations are recorded in table 3. The vapor of repel-
lents significantly reduced the number of mosquitoes feeding on raisins .
The standard error of the mean was used as statistical test for
significance.
Experiment 2
Ten raisins were fixed on the wire lying as close to each other
as possible without touching. Five alternate raisins were then painted
with repellent leaving the other five as controls. The numbers of
mosquitoes that settled on the treated and untreated raisins are recorded
in table 4, column 1 to 3. The figures are means of 5 replicates.
Observations were recorded every five minutes as in the previous
experiment. The total number of mosquitoes in the cage for each
experiment was 200.
In these experiments mosquitoes were seen coming close to the
raisins to land but they usually flew away without landing. No significant
difference was found between the number of mosquitoes settling on treated
Repellent Effects
12
and untreated raisins in the control with olive oil. The results show
that the repellent on the treated raisins kept the mosquitoes away from
the untreated raisins as well. Kalmus and Hocking (I960, p. 23) obtained
bites up to almost a mosquito half- width (about 2. 3 mm) from a repellent
painted circle on the back of the hand. In these experiments the mean
width of untreated raisin separating the two treated ones with repellent
was 10 ± 0. 3 mm. This greater distance was perhaps due to the factors
of heat, CC>2 and probably skin odor, which were missing as attractant
factors in the raisins.
TABLE 3 - Numbers oLA. aegypti settling on raisins separated by 1 cm,
under the plain and repellent coated halves of a paper strip.
Means of five replicates ± standard errors.
Experiment 3
Raisins were kept 1 cm clear apart from each other and alternate
raisins were painted with repellent. Other factors were the same as in
the previous experiments.
The mean numbers of mosquitoes that landed on the treated and
untreated raisins are given in table 4, columns 4 and 5. The numbers
of mosquitoes feeding or settling on the untreated raisins were still very
low and no significant differ ence was found in the number of mosquitoes
feeding on untreated raisins in this experiment as compared to the number
of mosquitoes feeding on untreated raisins in the previous experiment.
Experiment 4
Only 5 raisins were taken and were placed 1 cm clear apart and
the portion of wire between them was painted with repellent. Since the
raisins were not painted with repellent in this experiment their number
was reduced to five so that the number of mosquitoes landing on them
could be compared with the number of mosquitoes landing on the untreated
raisins in previous experiments.
The results are given in table 4, column 6. The comparison of
results in table 4 shows that significantly more mosquitoes settled on
raisins in this experiment than in experiments where- treated and
untreated raisins were placed close to each other. This is perhaps due
to the small surface area of wire between the raisins as compared to
the much greater area of the raisins in the previous experiments. This
Khan
13
would result in a much slower production of repellent vapor.
TABLE 4 - Numbers of A. aegypti settling on raisins in the presence of
repellents. Means of five replicates ± standard errors.
Mating
In Aedes aegypti "The stimulus which induces the male to copulate
is the sound produced by the female during flight. " . odor plays no
part in the sexual behavior of aegypti ..." (Roth, 1948, pp. 284, 282) .
Roth also observed that in Aedes aegypti the male is the aggressor and is
attracted by the female in flight and that the female is passive and does
not show any mating behavior similar to that of the male, "...never
in our observations was a male seen to initiate copulation with a resting
female" (Roth, 1948, p. 276). Banks (1908, p. 246) on the contrary
stated that specimens of aegypti confined in small jars "...have been
seen to copulate while the female hangs from the gauze covering the
vessel, the male always approaching her from the ventral surface. "
Christophers (I960, p. 502) observed that copulation takes place quite
commonly with the female at rest.
During the course of this work it was observed that a female
Aedes aegypti is not entirely passive and that copulation does take place
when a female is at rest. It was observed that when a flying male came
close to a sitting female, the female would take flight and the male would
grasp her for copulation. Many times females were seen taking flight
spontaneously and males were seen getting hold of them in mid-air.
The males were also observed coming to land sideways with a female,
14
Repellent Effects
then trying to take a ventral position and many a time they succeeded.
At other times because of his efforts to gain a ventral position to the
resting female the male roused the female to fly and copulation took
place on the wing or the two could be seen falling to the floor copulating.
But mostly copulation took place with a female in flight.
Roth (1948) also observed that the male would copulate repeatedly
with the same or different females. After repeated matings, females
become more and more reluctant to fly and would resist the attempts of
the males to copulate. Richards (1927) suggested that repeated
copulations exhaust the individuals. Shannon and Putnam (1934) in their
laboratory study of A. aegypti observed that the average pupal period of
females was 14 hours longer than that of males. Roth (1948, p. 308)
observed that by the time the female begins to fly and becomes ‘attractive1
the male’s antennae have reached a state where the sound stimulus can
be perceived and his genitalia have rotated sufficiently so that copulation
can be successful (usually about 15 to 24 hours after emergence). In
view of these observations it was neces sary in this work to separate the
sexes before they started mating and to keep the observation time
reasonably short. To forestall fatigue in the females due to repeated
copulations, the males were separated from the females 14 hours after
emergence.
Ten females 2-4 days old and 10 males 5-6 days old were used
for each experiment. The females were chilled in a sucking tube for
1. 5 minutes at 15 F and then all their tarsi were painted with repellent
with a fine brush while on a cold petri dish under a binocular microscope.
Since Aedes aegypti mate venter to venter and the female does not clasp the
male to her, her legs remaining out- stretched and serving as structures
to which the male clings (Roth, 1948, pp. 27 0, 301), it was decided to
paint the tarsi of the female mosquito with repellent. After the tarsi
were painted the females were released in a one cubic foot cage and
allowed to recover from chill. They recovered in 3-4 minutes. Ten
minutes after the treatment 10 males were released in the cage. After
application of the repellent on the tarsi of the female Aedes aegypti few
flew spontaneously. Most females sat quietly on the walls of the cage.
Males hardly ever succeededin persuading the female at rest to copulate.
It was also observed, though no quantitative basis could be laid down
for this, that the efforts of the male to copulate with the resting female,
as well as with the female in flight, were less persistent and quite often
they were seen releasing the female soon after coming in contact. The
cage was therefore shaken every minute to make the females fly and the
number of matings in a period of 30 minutes was recorded. Each
experiment was performed with a new batch of mosquitoes.
The results are recorded in table 5. The standard error of the
mean was used as a test of significance. The highly significant reduction
in the number of matings in A. aegypti in association with repellents can
be explained as a result of two factors: 1) a decrease in the flying
activity of the females and 2) less persistent efforts by males and
premature release of the female.
Though the cage was shaken every minute in experiments with
repellents as well as in the control it was observed that the females in
Khan
15
the controls continued to fly for a much longer time after shaking than
in experiments with repellents. With repellents, most of the time the
females could be seen coming to rest on the wall immediately after
shaking the cage, and many a time on shaking they would fly only from
one wall of the cage to another. There was also a lack of spontaneous
flight activity on the part of the females.
TABLE 5 - Numbers of matings in a 30 minute period in a population
of 1 0 male and 1 0 female _Aedes aegypti with repellents applied
to the tarsi of the females. Means of four replicates ±
standard errors.
Oviposition
Wallis (1954) in his studies on the oviposition activity of
mosquitoes, including A. aegypti , found that the female could detect an
objectionable amount of salt even when the movements of the abdomen
were restricted. Likewise surgical removal of the palpi, proboscis,
and antennae from the head did not result in loss of sensitivity. Surgical
removal or wax coating of various combinations of legs and leg articles
resulted in the demonstration that sensitivity was localized in the tarsal
articles of all the species of mosquitoes studied by him. His investi-
gations also showed that the sensitivity was present in all the tarsal
articles of Aedes aegypti. The thin walled chemor eceptor s of the tarsi
enabled the mosquitoes to detect differences in saline concentrations as
slight as 0. 02 M.
Browne (I960) studied the role of olfaction in the stimulation of
oviposition in the blowfly Phormia regina Meigen. He found that the odor of
a liquid medium containing powdered milk and yeast stimulated the blow-
fly to oviposit. He also provided evidence for olfactory perception by
the ovipositor of the blowfly.
In this study oviposition in Aedes aegypti was obs er ved by as s ociating
potential oviposition sites with repellent vapors as well as by applying
repellents on the tarsal chemor eceptor s.
Experiment 1
Five, 7-8 day old blood fed females in a one cubic foot cage were
taken for each experiment. The cage was provided with a rectangular
platform, 7 " x 4 " made of a steel wir e frame (diameter of wir e 2 . 5 mm).
The platform was covered with nylon net on one side and with two paper
towel strips pasted on the other except in the center where a gap of 1
16
Repellent Effects
cm was left in between the strips, see figure 1.
The platform was placed in the cage, nylon net side upwards,
the ends resting on two glass bottles filled with water. On the nylon net
was spread a piece of cheese cloth, the two ends of which remained
dipped in the water in the glass bottles. The cheesecloth was kept wet
by capillary action by the water in the two bottles. One of the two paper
strips was painted with repellent while the other was left untreated.
Thus an oviposition platform for the mosquitoes was provided, one half
of which had repellent vapor coming from underneath through the nylon
screen, while the other half served as control. The nylon net under-
neath the cheese cloth served as a support for it and did not allow it to
come in contact with the repellent on the paper strip below but allowed
the repellent vapors to pass through. Most of the eggs were found to
be laid on the cheese cloth but some were laid on the paper strip. They
were counted separately 72 hours after the blood meal, and the results
are recorded in table 6. Four experiments were run with each repellent.
The behavior of Aedes aegypti during egg laying is described in
detail by Wallis (1954). During the experiments it was observed that a
female mosquito could sample the oviposition sites while on the wing
and would land on the control half rather than on the repellent treated
half of the oviposition platform. At other times when she landed on the
repellent half she walked for a few seconds and then flew away and
landed on the control side. This behavior demonstrates the function of
olfactory receptors in the selection of an oviposition site when repellent
vapors are associated with it. The complete absence of egg laying on
the repellent coated as well as olive oil coated paper towels on the lower
side of the platform seems to be the result of tarsal chemor eceptor s
which select the suitability of the egg laying medium on contact. The
significantly small numbers of eggs laid on chees e cloth on the repellent
side as compared to the number of eggs laid on the control side show
that Aedes aegypti rejects oviposition sites when these are as sociated with
repellent vapor.
Experiment 2
Experiments were also conducted by painting the tarsi with
repellent by the same technique as described in previous experiments
and recording the number of eggs laid in 24 hours. Christophers (I960,
p. 507) records that egg laying in Aedes aegypti usually begins on the after-
noon of the third day from blood feeding, counting the day of feed as
zero. Female mosquitoes 6-7 days old were fed on blood and left in a
cage with raisins for three days. On the fourth day their tarsi were
painted with repellent and the mosquitoes were placed singly in separate
vials with water soaked cottonwool in the bottom and a nylon net cap on
the top on which was placed a raisin. Eggs laid in a 24 hour period
were then counted. Four replicates were run for each experiment.
The mean numbers of eggs laid are recorded in table 7.
The difference in the number of eggs laid in the control and
those laid by repellent treated mosquitoes is not significant, using
standard error of the mean as a test of significance. This is perhaps
17
Oviposition platform
Oviposition p latform - frame
of steel wire (2.5mmqauqe)
b.
Cheese cloth
Nylon net pasted on
oviposition platform
Water bottle
Paper towel strips pasted on
oviposition platform-frame
Figure 1 . Diagrams showing: (a) arrangement of the oviposition platform in
the cage, and (b) a vertical section of the oviposition platform.
18
Repellent Effects
due to the fact that the mosquitoes had no opportunity to select a site
for oviposition as. they were confined in small vials.
TABLE 6 - Numbers of eggs laid by A aegypti females in the presence
of repellents. Means of four replicates ± standard errors.
Experiment 3
Experiments were also conducted to determine whether anten-
nectomized mosquitoes would discriminate between the control and the
repellent sides of the oviposition site . Twenty female mosquitoes which
had been fed on blood previously were operated upon for each experiment
on a cold petri dish under a binocular microscope after first chilling
them for 1.5 minutes at 14°F. Ten to 12 flagellar segments of the
antennae were excised and the mosquitoes then released in the cage with
the oviposition platform shown in figure 1.
Though sometimes mosquitoes could be seen sitting on the control
side of the egg laying platform, no eggs were laid in any of the experiments
over a week's timeexceptin the experiment with diethyl toluamide where
there were 4 eggs on the control side. A high mortality (70-75%) was
also observed in mosquitoes during this period. The almost complete
absence of oviposition by antennectomized mosquitoes may be due to
lack of orientation of mosquitoes to the water soaked cheese cloth on
account of the great reduction in the number of hygr or eceptor s as a
results of excision and consequently a great increase in the threshold
of moisture perception. The high mortality rate can also be assigned
to the same factor, i. e. , lack of orientation to the water soaked cheese
cloth and hence dehydration. Mosquitoes were seldom seen sitting on
the wet cheese cloth. Most of the time they were found sitting on the
walls of the cage with very little flight activity. The very low activity
in antennectomized mosquitoes confirms the findings of Bar-Zeev (I960)
who found only 4 per cent of mosquitoes could be activated when anten-
nectomized as compared to 60. 1 per cent when intact.
Khan
19
TABLE 7 - Numbers of eggs laid by single A. aegypti females after
painting the tarsi with repellents. Means of four replicates
± standard errors.
Experiment 4
Experiments were also conducted to test oviposition after treating
the terminalia of the females with repellent. The female aegypti
mosquitoes were fed on blood when 7-8 days old, and their terminalia
painted with repellent 72 hours after the blood feed by the same technique
as described in the previous experiments, and then released in the cage.
All the mosquitoes became too crippled to move about or fly
shortlyafter the painting of the tip of the abdomen and died in a few hour s .
Behaviour in Relation to Wind Direction and Speed
Kalmus and Hocking (I960, p. 21) conducted a series of experi-
ments in which target areas were drawn out on the backs of subjects
who wore shirts with the backs cut out. They recorded the distribution
of bites in relation to a small repellent treated area. To demonstrate
the effect of wind direction on the distribution of bites in relation to
repellent, experiments were conducted in the laboratory on Aedes aegypti
using the same technique.
Experiment 1
A circle of 3. 5 cm radius was drawn in hard clear nail varnish
on the bare chest of a subject. Concentric to this another circle of 6. 5
cm radius was drawn. The outer circle was divided into two equal
halves by drawing a diameter. A hair drier was used to produce the
air current and a variable transformer was included in the circuit to
permit adjustment of the speed of the wind. The wind speed was kept
at 43 cm/ sec and its direction at right angles to the drawn diameter.
The source of wind, i. e. , the nozzle of the blower was kept 23 cm away
from the central circle which was coated with repellent. The blower
was kept in such a position as to give a uniform flow of air over the
marked area. The repellent used was diethyl toluamide. One hundred
7-8 day old female Aedes aegypti mosquitoes were taken in a one cubic foot
cage of steel wire with nylon net around it for each experiment. The
mosquitoes were fed on sugar solution only before the experiment. The
cage was placed on the marked area and the portion of skin outside the
marked area was covered with a polyethylene sheet. Mosquitoes soon
started biting through the nylon net on the floor of the cage. An observer
kept a record of the mosquitoes that settled and flew away, or settled
and bit, in the upwind and downwind halves of the circle. The counts
20
Repellent Effects
were made for 5 minutes in each experiment. Controls were run with
the same wind speed without repellent. The number of mosquitoes that
settled or bit in the upwind and downwind halves of the circle are given
in table 8.
The results show that in the control where repellent was not
painted in the central circle, significantly more mosquitoes settled or
bit on the downwind side of the circle than on the upwind side. This is
in conformity with the obs ervations made by Kalmus and Hocking (I960,
p. 4) with field mosquitoes. However, when repellent was painted in
the central circle it was observed that the number of mosquitoes settling
or biting on the downwind half of the outer circle was significantly lower
than the number settling or biting on the upwind half. This was due to
the presence of repellent vapor carried by the wind on the downwind
half of the outer circle.
TABLE 8 - Numbers of A. aegypti lemales settling or biting in relation
to wind direction and D. E. T. on the marked area of skin.
*■ Means of two counts ± standard error
** Means of three counts ± standard error
Experiment 2
In another set of experiments the effect of different wind speeds
was determined on the settling and biting of mosquitoes in relation to
repellent. Experiments were conducted in a similar fashion as described
under the experiments with different wind directions, except that the
portion of the body used was the thigh instead of the chest, which gave
the advantage of the subject himself making notes of the number of
mosquitoes landing or biting. A control was run with each wind speed
and all the controls with different wind speeds were run first in order
to avoid contamination of skin area with repellent vapors. After the
controls were run, different batches of mosquitoes were then used in
experiments with the same wind speeds in relation to repellent painted
in the central circle. The repellent used was diethyl toluamide. The
portion of skin outside the outer circle was covered with polyethylene
sheet and the count of mosquitoes settling or biting in the upwind or
downwind half of the circle was recorded for five minutes in each experi-
ment.
The results are shown in table 9. In previous experiments with
different wind directions the number of mosquitoes settling or biting in
the upwind half of the circle in experiments with repellents was
Kiian
21
significantly higher than the number of mosquitoes in the downwind half
of the circle. The results given in table 9 show that the mosquitoes
continue to showthe strong tendency of settling more on the upwind side
in relation to repellent with different wind speeds.
The maximum wind speed at which mosquitoes were able to settle
on a bluff body was reported to be 95 cm/ sec and that of settling on the
streamlined body to be 55 cm/ sec. Kalmus and Hocking (I960, p. 15).
In this case the maximum speed of wind at which the mosquitoes settled
on the skin was 265 cm/ sec which is very high as compared to the wind
speed with the bluff or streamlined bodies. This is probably due to the
attractant factors of the skin acting on the mosquitoes.
TABLE 9 - The number of A. aegypti females settling or biting in the up-
wind or the downwind half of the circle marked on skin in
relation to different wind speeds and diethyl toluamide.
Orientation to Gravity and Centrifugal Force
Experiment 1
To study the orientation of Aedes aegypti to gravity in relation to
repellents, experiments were conducted in a plastic petri dish of 9 cm
diameter. The lid of the petri dish was perforated with 2 mm diameter
holes, about 9 holes per sq cm to allow the repellent vapors inside the
dish to escape. The floor of the petri dish was lined with a filter paper
which was divided into four quadrants designated top, left, bottom, and
right.
Twenty female mosquitoes, 7-8 days old were taken, chilled for
1.5 minutes at 14°F and then released in the petri dish. On recovery
22
Repellent Effects
of mosquitoes from chill the petri dish was turned with a diameter
vertical and given five complete turns on the horizontal axis through its
center; thereafter the position and the number of mosquitoes was noted
in each quadrant after a minute. The experiment was replicated five
times without repellent as a control. A band of repellent 1 cm wide was
then painted on the outer margin of the top quadrant. Mosquitoes were
chilled and placed in the petri dish and allowed to recover. After the
mosquitoes had completely recovered, the dish was given five complete
rotations as in the control, keeping it vertical and rotating it about its
horizontal axis. The experiment was repeated five times with each
repellent.
In the control the mosquitoes could be seen walking upwards and
most of them collected in the top quadrant. Significantly les s mosquitoes
remained in other quadrants . Almost all the mosquitoes were seen facing
upwards and the root mean square deviation of their body axes from the
vertical axis of the petri dish was found to be zero.
With repellent significantly less mosquitoes entered the top
quadrant. Most of them remained in the left, right, g.nd bottom quadrants .
They were also seen walking at an angle to the repellent or turning away
from it. Their angle of turning (i. e. , the angles which the longitudinal
axes of the bodies formed with the vertical axis of the petri dish) was
noted by marking their position in each quadrant on a separate sheet of
paper and then measuring the angle and direction of inclination to the
vertical.
Table 10 shows the distribution of mosquitoes in the various
quadrants of the petri dish in the presence of repellents, and table 11
shows the root mean square of the angle of inclination of the body axes
of mosquitoes to the vertical in the presence of repellents in the petri
dish.
Results with olive oil werenot found to be significantly different
from those of the plain control.
The effect of the presence of repellent on the head upwards
orientation of the mosquitoes in relation to gravity was highly significant.
Experiment 2
The effect on geotaxis of painting repellent on the mesonotum
and the antennae was also observed. Seven to 9 days old female
mosquitoes were chilled for 1.5 minutes at 14°F and their mesonotaor
antennae were painted with repellent. They were then placed in a 9 cm
petri dish having holes in the lid and lined with filter paper. After
complete recovery of the mosquitoes from chill the dish was held with
its central axis horizontal and rotated slowly about this, one rotation in
2 0 seconds, and the positions of the mosquitoes were noted. Normal
female A. aegypti show a counter rotation to maintain a head upward
under these circumstances (Kalmus and Hocking, I960, p. 8).
The mosquitoes with their mesonota painted oriented facing up-
wards by counter rotation but when the antennae were painted with
repellent, on placing the dish in a vertical position the mosquitoes could
be seen sitting on the vertical surface head upwards cleaning their
antennae with the tarsi of the forelegs. When the dish was rotated slowly
Khan
23
while they were cleaning their antennae, they did not react until they
faced downwards. Then they were found to lose their balance and were
seen to place their forelegs on the vertical surface. Some of them
turned around, faced upwards and started cleaning the antennae again,
but typical counter rotation was absent.
TABLE 10 - Numbers ol A. aegypti females found in different quadrants
in relation to repellents. Means of five replicates ±
standard errors.
Experiment 3
According to Kalmus and Hocking (I960, p. 8), when mosquitoes
were centrifuged in a 9 cm petri dish at 390 rpm and observed under
stroboscopic illumination, they were found facing towards the center of
the dish, and sometimes walking towards it.
In this study of the same behavior in relation to repellents a
plastic petri dish of 9 cm diameter was lined with filter paper on which
one radius was drawn in ink. Its lid was extensively perforated by
small holes. Mosquitoes, bothmalesand females (50 to 60 adults) were
released in this dish and centrifuged at 390 rpm on a turntable and
observed under stroboscopic illumination. Mosquitoes were seen as
reported by Kalmus and Hocking (I960) facing towards the center and
walking towards it. Most of them collected near the center roughly 1
to 1 . 5 cm from it; fewer mosquitoes remained at the periphery. The
centrifugal force at 1 cm from center was 1. 7 g and 1. 5 cm 2. 5 g. As
the dish continued to rotate more mosquitoes could be seen moving
towards the center. For experiments with repellents the mosquitoes
were taken in batches of 15, in a sucking tube, chilled for 1. 5 minutes
at 14°F and then their mesonota painted with repellent on a cold petri
dish under a binocular microscope. All four repellents were tested.
After treatment the mosquitoes were released in a cage and allowed to
24
Repellent Effects
recover. They were then introduced in the petri dish (50-60 of them)
and made to rotate.
Under stroboscopic illumination it was observed that the mos-
quitoes did not collect in greater numbers near the center of the dish
and the movement towards the center was less noticeable. The dish
gave an appearance of a scattered distribution of mosquitoes as compared
to a circular distribution near the center in the control. Quite a few
(10-15%) faced directions other than the center.
TABLE 11 - Root mean square of angles of inclination of the body axes
of. A. aegypti to the vertical in the presence of repellents in
a rotated petri dish in degrees. Means of five replicates
± standard errors.
In another experiment the mosquitoes themselves were not treated
but a disc of 4 cm diameter (centrifugal force 3.5 g) was painted with
repellent in the center of the dish. Mosquitoes (50-60) were introduced
in the aish which was then rotated. It was observed that with an exception
of one or two the mosquitoes remained outside the disc, sometimes
facing towards it and sometimes turning away from it or walking around
it. In yet another experiment when the diameter of the circle painted
with repellent was increased to 6 cm (centrifugal force about 5 g) in 9
cm petri dish the same behavior was observed. Most of the mosquitoes
remained outside the circle, although the non-treated peripheral belt
around the repellent coated circle was only 1. 5 cm wide.
Yisual Responses
The optomotor and visual responses of mosquitoes have been
studied by many workers. Kalmus (1958) reported that A. aegypti shows
responses to the rotation of the plane of polarization of light. In a later
study Kalmus and Hocking (I960, p. 19) observed swarming flight in
A. aegypti close underneath a weak light sour ce placed on top of a darkened
cage, but the same was not observed when a much stronger light was
made to pass through a red filter. Mosquitoes were also observed by
these workers to aggregate near the margins of black objects when these
were placed on top of a weakly illuminated cage.
The visual response of mosquitoes was also studied by Kennedy
(1939) and Rao (1947). Kennedy reported that suspended mosquitoes
orientated accurately towards a vertical black stripe on a white back-
ground. Presented with two stripes the mosquitoes faced one or the
other stripe and not between the two. Rao ( 1 947 ) reported similar findings
with Anopheles maculipennis atroparvus van Thiel, and Culex (Culex) molestus F orskal
Khan
25
rendered flightless by the removal of the wings or by sticking them
together .
To test the effect of repellents on the visual respons e of Aedes aegypti
to black stripes, 20 female mosquitoes were taken in a glass bottle 12
cm tall and with a diameter of 3 cm. The inside of the bottle was lined
with white nylon net to give the mosquitoes a good foothold. This bottle
was placed inside a glass cylinder 14 cm high and with a diameter of 6
cm. The bottle and the cylinder were placed on a thick glass plate which
was resting on a tripod stand. Under the glass was placed a 40 watt
electric lamp which was covered all around with a cylinder of black
paper so that light could go only upwards and light the bottle and the
cylinder outside it uniformly from inside. In order that the inside of
the cylinder be evenly illuminated, a filter paper was placed on the glass
plate on which the outer cylinder and the inner bottle rested. The outer
cylinder was divided into four quadrants and the alternate two quadrants
were covered with black paper strips, each covering 90°. The remain-
ing two quadrants were left uncovered, (figure 2).
As the outer cylinder was placed around the inner bottle contain-
ing mosquitoes and kept there for a short time, the mosquitoes inside
moved and came to rest on the wall of the bottle facing the black stripes.
The outer cylinder was then rotated 90° so that all the mosquitoes now
faced uncovered portions of the cylinder. The mosquitoes moved again
in the direction of the black stripes and again came to rest opposite to
them. This behavior could be observed again and again. However,
when the antennae were painted with any of the four repellents they
showed complete indifference to the black stripes and did not move
towards them as in the control.
The experimental data on the effects of repellents on behaviour
are summarized in table 1 1A.
TABLE 11A - Summary of data on the effect of repellents on responses
to stimuli.
26
Outer cylinder
Black paper strips
pasted on cylinder
Figure 2. Diagrams showing: (a) a vertical section of the
apparatus used for testing the visual response of A.aegypfi
females to black stripes in relation to repellents, and (b)
a cross section of the outer cylinder.
Khan
27
EXPERIMENTAL - LIPOID SOLVENTS
Amongst the advocates of chemical theories referred to in a
previous section, many have suggested lipoid solubility as a basis of
olfaction (Cohn, 1924; Dyson, 1938; Dethier & Chadwick, 1947; Dethier,
1948). Experiments were conducted to examine the effect of fat solvents
applied on the antennal chemor eceptor s of Aedes aegypti females on their
behavior towards a host.
Ten female Aedes aegypti eleven days old were taken for each experi-
ment. The mosquitoes , which were fed on sugar solution only, were
taken in a sucking tube and chilled for 1. 5 minutes at 15°F. They were
then placed on a filter paper on top of a cold petri dish and their antennae
were washed with lipid solvents applied with a fine camel hair brush.
The mosquitoes were then transferred to a clean petri dish lined with
filter paper in a one cubic foot cage and allowed to recover. Thirty
minutes after the operation a hand was introduced into the cage and the
number of landings of mosquitoes on it was recorded for a period of 15
minutes. Mosquitoes were shaken off gently on landing and were not
allowed to feed on blood. The antennae of controls were rubbed with a
clean dry brush.
The observations are recorded in table 12, and show that the
number of landings decreased very significantly on washing the antennal
chemor eceptors with the lipid solvents. But whether the decrease in
landings is due to the loss of lipids from the chemoreceptor s , or due to
the narcotic, anesthetic, or other effect of the solvents is uncertain.
TABLE 12 - Numbers of A edes aegypti females landing on a hand in a 15
minute period after treatment of the antennae with lipoid
solvents. Means of three replicates ± standard errors.
Control Acetone Ether
174 ±8 52 ±12 43 ± 13
DISCUSSION
The action of repellent chemicals on mosquitoes has no specificity
for blood feeding behavior. It has been shown that repellents in the
vapor phase have the following effects on .Aedes aegypti . They inhibit feeding
on both blood and sugars, reduce the mating rate, and cause rejection
of oviposition sites. The repellents also affected orientation to gravity
and centrifugal force and the visual response to black stripes.
Mosquitoes became quiescent and less active when repellents
were applied on them. This slowing down of motor activity suggests
the external stimuli normally acting on the mosquito are perhaps blocked
or interfered with by the repellent. As there is no delay in the effect
of repellents on the behavior of mosquitoes, that is, protection is obtained
immediately these materials are applied, their action on the insect may
be assumed to occur at the surface of the body. Repellents have not
been shown to penetrate rapidly into the body where they could act on
28
Repellent Effects
the nerve synapses or the central nervous system, nor have they been
shown to affect the muscular system directly. It thus seems unlikely
that they act by blocking the nerve impulses or the motor response.
The most probable action seems therefore to be the blocking of reception
of stimuli at the receptor site.
Somewhat different behavior in relation to repellents of another
kind has been described by Kennedy (1947). He studied the effects of
contact with DDT on the activity and distribution of mosquitoes. He
argued from his experiments that a variety of reactions may give rise
to repulsion. Reactions may occur at a distance or only after contact
with a repellent surface. The contact stimuli may be mechanical or
chemical. The reactions may take the form of an increase of merely
random activity or they may be directed away from the surface. They
may be quick or slow to appear and weak or strong in expression. In
contrast to my findings of reduced activity in his work an increase in
activity was found.
The factor s that attract mosquitoes to the host have been reviewed
above. The mode of action of insect repellents can be best understood
when studied in relation to these factors.
The effects of repellents on the evolution of carbon dioxide and
moisture from a human arm, and the correlation of this evolution with
the natural attractiveness of human beings and protection time of
repellents were studied by Gouck and Bowman ( 1959) at Orlando, Florida.
In their experiments, repellents applied to the arms of three subjects
reduced the CO2 emitted by 9 to 14 per cent but they concluded: ''Although
these reductions are considerably greater than the differences between
untreated arms (4%) they are not great enough to indicate that the mode
of action of these repellents is based upon the retardation of carbon
dioxide evolution". The repellents used were, dimethyl phthalate,
diethyl toluamide and ethyl hexanediol. With regard to the moisture
collected from untreated and repellent treated arms they concluded:
"The quantities from the arms of all subjects varied from day today but
in most individual tests the two arms agreed within about 5 per cent
indicating that no real difference in the amount of moisture evolved was
caused by application of repellents. " They believed that the protection
time is governed by the rate of loss of repellent from the skin by absorption
and evaporation. Peters and Kemper (1958) have shown that there are
no considerable temperature differences between repellent treated and
untreated parts of the skin.
In the light of these findings it can be said that repellents affect
the reception of these stimuli rather than the stimuli themselves. This
supports the hypothesis advanced that repellents affect many kinds of
behavior of mosquitoes by interfering in the reception of many different
kinds of stimuli.
Search for chemical factors other than carbon dioxide attracting
mosquitoes to the host has claimed the attention of many workers. The
findings of Shaerffenberg and Kupka (1951) and Bur ges s and Brown (1957)
have indicated that attractive factors other than carbon dioxide are
present in the vapor from mammalian blood and body exudations. A
distillate obtained from mammalian blood by Shaerffenberg and Kupka
Khan
29
(1959) proved highly attractive to Culex pipiens L. Rudolfs (1922) found
benzoic acid, dilute ammonia, phenylalanine, alanine, aspartic acid,
cystine, and hemoglobin to be attractive to Aedes sollicitans Walker and
Aedes cantator Coquillett, but Reuter (1936) found the last six materials
unattractive to Anopheles maculipennis atroparvus . Brown and Carmichael (1961)
reported that E-lysine and L-alanine wer e attractive to Aedes aegypti. The
effect of repellents in association with these chemicals found to be
attractive remains to be studied.
Travis and Smith (1951) evaluated dimethyl phthalate, indalone,
and ethyl hexanediol against Aedes aegypti besides other mosquitoes, and
found average repellent times (i. e. , times in minutes from application
of the repellent to the first bite) as follows: ethyl hexanediol - 331
minutes, dimethyl phthalate - 247 minutes, and indalone - 111 minutes.
Although the results of my experiments are not strictly comparable with
those of Travis and Smith (1951) for I worked with a different culture of
mosquitoes and at a different time and place, the mosquitoes fed on
blood much sooner after treatment when repellents were applied on the
mosquito receptor sites. For example, about33 per cent of mosquitoes
fed on blood within 10 minutes after application of dimethyl phthalate on
both antennae. When diethyl toluamide, indalone, and ethyl hexanediol
were separately applied on both antennae, some of the first bites were
recorded after 10 minutes. The reason for this behavior is perhaps the
more rapid adaptation of the receptors'to the repellents because of the
greater concentration gradient resulting from their application on the
receptors themselves. In this way the threshold for reception of re-
pellents increased greatly but that for other stimuli remained the same.
The s equence of stimuli and responses leading to blood feeding therefor e
remained unaffected. But this is, of course, incompatible with the
hypothesis that repellents block all receptors.
The presence of separate chemor eceptor neurons mediating
acceptance and rejection is assumed from the study of labellar chemo-
receptor cells of Phormia regina . These cells have been the subject of co-
ordinated behavioral, histological, and physiological study. A chemo-
sensory hair of the labellum of this blowfly was described by Dethier
(1955) as a hollow extension of the body cuticle possessing two distinct
lumina. The chemosensory hair has been shown to be associated with
three bipolar neurons, two of which send distal fibers to the terminal
papilla by way of the thick- walled lumen of the hair. Dethier (1955)
concluded that one of these neurons mediates acceptance while the other
mediates rejection. On electrophysiological studies one of the two
neurons was later designated the L fiber (for large spikes which re-
sponded to salts and the other the S fiber (for small spikes) which re-
sponded to sugars (Hodgson et al. 1955; Hodgson and Roeder, 1956).
Wolbarsht and Dethier (1958) were able to detect the spikes of the third
neuron which terminated in a process at the base of the hair. It was
designated M for mechanor eceptor . Evans and Mellon (1962) have now
detected spikes from a fourth neuron which responds to water.
In the course of electrophysiological studies of chemor eceptor
hairs it has been shown that when mixed stimuli are applied there is an
interaction between activity in the L and S fibers (Hodgson, 1956 ,
30
Repellent Effects
1957; Morita, 1959; Sturckow, 1959). Hodgson (1957) found that the
presence of S impulses is accompaniedby a decrease in L impulses and
conversely the S spikes decrease when the L fiber is stimulated. My
experiments show that the repellents block the reception of attract ant
and other stimuli. This assertion needs to be confirmed by electro-
physiological methods.
Mosquitoes with antennae painted with diethyl toluamide landed,
walked around, and even probed on an arm also treated with the same
repellent but did not feed on blood. This may be explained in one of two
ways. It may be that the piercing of the skin by the mosquito is induced
by some chemical factor on the skin which was neutralized by the
application of the repellent or, it may be due to the effect of repellent
on the action of thermoreceptor s or contact chemor ecptor s which in-
duce feeding on blood. The latter explanation would be more in con-
formity with the findings that repellents interfere with the reception of
all kinds of stimuli affecting the total behavior of mosquitoes.
The study on blood feeding when repellents were applied on parts
of the mosquito revealed that of the four repellents dimethyl phthalat e
has the greatest effect on blood feeding behavior when it is painted on
the tar sal receptor s and the smallest effect when it is painted on the
receptors of the antennae. As is known, the olfactory receptors are
located on the antennae and the contact chemor eceptors mostly on the
tarsi of the mosquito. Dimethyl phthalate, which has the highest boiling
point and hence the lowest vapor pressure, may, for this reason, have
more effect than the other repellents through the tar sal chemor eceptor s
in the liquid phase but less than these through the olfactory receptor s of
the antennae where it has to act in the vapour phase which is at a lower
concentration.
That repellents also acted as irritants was evident from the
intense wriggling activity of the mosquito when repellents were applied
on the proboscis and from the vigorous cleaning of repellent from the
antennae with the tarsi of the fore legs. This evident awareness of the
presence of an irritant chemical indicates the existence of receptors
sensitive to it, perhaps those of the common chemical sense. It may
well be that these are the only receptors not blocked by repellents.
In the vapor phase repellents were found to inhibit landing of
mosquitoes. This was observedin experiments on blood feeding, sugar
feeding, oviposition, and air flow. In the liquid phase, however, the
repellents showed more irritant and some toxic effects, and the mos-
quitoes showed considerable decrease in locomotor activity, in part on
account of preoccupation with attempts at cleaning off the repellents.
Repellents have been defined as compounds which elicit an
avoiding reaction (Dethier, 1956b). While the four materials studied
may all do this, this is by no means their only effect and may not, in-
deed, be the most important one.
Khan
31
ACKNOWLEDGEMENTS
I am most thankful to Professor B. Hocking for his keen inter est,
constructive criticism and many valuable suggestions duringthe progress
of this work. I also gratefully acknowledge the support of the Defence
Research Board, Department of National Defence, Government of
Canada, who financed this study.
REFERENCES
Amoore, J. E. 1963. Stereochemical theory of olfaction. Nature, Lond,
198 : 271 - 272.
Amoore, J. E. , Johnston, J. W. Jr. , and Rubin, M. 1964. The stereo-
chemical theory of odor. Sci. Amer. 210 (2) : 42 - 49.
Banks, C. S. 1908. Biology of Philippine Culicidae, Philip. J. Sci.
3 (A) 235 - 258.
Bar Zeev, M. I960. The reaction of mosquitoes to moisture and high
humidity. Ent. exp. appl. 3 : 198 - 211.
Bates, M. 1949. The Natural History of Mosquitoes. Macmillan, New
York. XV. 379 pp.
Bishop, A., and Gilchrist, B. M. 1949, Experiments upon the feeding of
Aedes aegypti through animal membranes, etc. Parasitology 37 :
85 - 100.
Brown, A. W. A. 1951, Studies on the responses of the female Aedes
mosquito. Part IV. Field experiments on Canadian species.
Bull. ent. Res. 42 : 575 - 582.
Brown, A. W. A. 1956. Factors which attract Aedes mosquitoes to humans.
Proc. tenth int. Congr, Ent. 3; 757 - 763.
Brown, A. W. A., and Carmichael, A. G. 1961. Lysine and alanine as
mosquito attractants. J. econ. Ent. 54 (2) : 317 - 324.
Browne, B. L. I960. The role of olfaction in the stimulation of oviposi-
tion in the blowfly, Phormia regina, J, Insect Physiol. 5 : 1 6 - 22.
Burgess, L. , and Brown, A. W. A. 1957. Studies on the responses of
the female Aedes mosquito. Part VIII, The attractiveness of beef
blood to Aedes aegyptiL,. Bull. ent. Res. 48 : 783 - 793,
Christophers, S. R. I960. Aedes aegypti , the yellow fever mosquito - Its
life history, bionomics and structure. Cambridge University
Press. 739 pp.
Cohn, G. 1924. Die Riechstoffe. 2nd Ed. Braunschweig: Verlag
Vieweg. pp. viii + 216.
Davies, J. T. 1962. The mechanism of olfaction, p. 170-179, Symposi;
of the Society for Experimental Biology. No, XVI - Biological
receptor mechanisms, Cambridge University Press.
DeLong, D. M. 1946. Insect behaviour - Mosquito attractancy and
repellency. U, S. Natl. Res. Council, Insect Control Committei
Abstr , Bull. n. s. 11, pp. 515 - 520.
32
Repellent Effects
DeLong, D. M. , Davidson, R.H. , Peffly, R.L. , and Venard, C. E. 1945.
A study of the action of insect repellents in terms of their effect
on insect behaviour and in relation to their properties. U. S. Natl.
Res. Council, Insect Control Committee. Abstr. Bull. n. s.
3, pp. 96 - 102.
Dethier, V. G. 1948. Chemor eceptor s in insects. Physiol. Rev. 28 :
220 - 254.
Dethier, V. G. 1952. The relation between olfactory response and rec-
eptor population in the blowfly. Biol. Bull. Wood's Hole 102 :
111 - 117.
Dethier, V. G. 1955. The physiology and histology of the contact chemo-
receptors of the blowfly. Quart. Rev. Biol. 30: 348 - 371.
Dethier, V. G. 1956a. Attractants and repellents. 43rd Annual Meeting
Proceedings of the Chemical Specialties Manufacturers Assoc-
iation Inc. 1-8.
Dethier, V. G. 1956b. Repellents. Annu. Rev. Ent. 1 : 181 - 202.
Dethier, V. G. 1957. Chemor eception and the behaviour of insects.
Survey Biol. Progr. , Academic Press, N. Y. 3 : 149 - 183.
Dethier, V. G. , and Chadwick, L. E. 1947. Rejection thresholds of
the blowfly for a series of aliphatic alcohols, J. gen. Physiol.
30 (3) : 247 - 253.
Dethier, V. G. , and Yost, M. T. 1952. Olfactory stimulation of blow-
flies by homologous alcohols. J. gen. Physiol. 35 (6) : 823-839.
Dyson, G. M. 1938. The scientific basis of odor. Chem. &; Ind. (Rev.)
57 : 647 - 651.
Evans, D. R. 1961. The effect of odours on ingestion by the blowfly.
J. Insect Physiol. 7 : 299 - 304.
Evans, D. R. and Mellon, De F. 1962. Electr ophysiological studies of
a water receptor associated with the taste sensilla of the blowfly.
J. gen. Physiol. 45 : 487 - 500.
Frings, H. and Hamrum, C. L. 1950. The contact chemor eceptor s of
adult yellow fever mosquitoes, Aedes aegypti. J. N. Y. ent. Soc.
58 : 133 - 142,
Goeldi, E. A. 1905. Os mosquitos no Para Reuniae gutro trabalhos sobre
os mosquitos indigenas, principalmente as aspecies que molestam
homen. Mem. Mus. Goeldi Hist. nat. Ethnogr. 4 : 154 pp.
Gordon, R.M. and Lumsden, W. H. R. 1939. A study of the behaviour
of the mouth parts of mosquitoes when taking up food from living
tissue, etc. Ann. trop. Med. Parasit. 33 : 259 - 278,
Gouck, H. K. , and Bowman, M. C. 1959. Effect of repellents on the
evolution of CO2 and moisture from human arms. J. econ. Ent.
52 (6) : 1157 - 1159.
Haddow, A. J. 1942. The mosquito fauna and climate of native huts at
Kisumu, Kenya. Bull. ent. Res. 33 : 91 - 142.
Headlee, T. J. 1941. New Jersey mosquito problems. Proc. New
Jersey Mosquito Exterm. Assoc. 28 : 7 - 12.
Hocking, B. 1952. Protection from northern biting flies. Mosquito
News. 12 : 91 - 1 02.
Hodgson, E.S. 1956 . Physiology of the labellar sugar receptors of flies.
Anat. Rec. 125, 555.
Khan
33
Hodgson, E. S. 1957. Electrophysiological studies of arthropod chemo-
reception. 11. Responses of labellar chemoreceptor s of the
blowfly to stimulation by carbohydrates. J. Insect Physiol. 1 :
240 - 247.
Hodgson, E.S., Lettvin, J. Y. , and Roeder, K. D. 1955. Physiology
of a primary chemoreceptor unit. Science 122 : 417 - 418.
Hodgson, E. S. , and Roeder, K. D. 1956. Electrophysiological studies
of arthropod chemor eception. 1. General properties of the
labellar chemor eceptors of Diptera. J. cell, comp. Physiol.
48 : 51 - 76.
Howlett, F. M. 1910. The influence of temperature upon the biting of
mosquitoes. Parasitology 2 : 479 - 484.
Huffaker, C. B. , and Back, R. C. 1943. A study of methods of sampling
mosquito populations. J. econ. Ent. 36 : 5 6 1 - 569.
Jones, F. N. , and Jones, M. H. 1953. Modern theories of olfaction: a
critical review. J. Psychol. 36 : 207 - 241.
Kalmus, H. 1958. Responses of insects to polarized light in the pres-
ence of dark reflecting surfaces. Nature, ' Lond. 182 : 1526 -
1527.
Kalmus, H. and Hocking, B. 1960. Behaviour of Aedes mosquitoes in
relation to blood feeding and repellents. Ent. exp. appl. 3 :
1 - 26.
Kellogg, F. E. and Wright, R. H. 1957. The guidance of flying insects.
V. Mosquito attraction. Canad. Ent. 94 (10) : 1016.
Kennedy, J. S. 1939. The visual responses of flying mosquitoes, Proc.
zool. Soc. Lond. (A), 109 : 221 - 242.
Kennedy, J. S. 1947. The excitant and repellent effects on mosquitoes
of sub-lethal contacts with DDT. Bull. ent. Res. 37 : 593 - 607.
Macfie, J. W. S. 1915. Observation on the bionomics of Stegomyia fasciata
Bull. ent. Res. 6 : 205 - 229.
Morita, H. 1959. Initiation of spike potentials in contact chemosensory
hairs of insects. Ill . D. C. stimulation and generator potential of
labellar chemoreceptor of Calliphora. J. cell. comp. Physiol.
54 : 189 - 204.
Parker, A. H. 1948. Stimuli involved in the attraction of Aedes aegypti
L. to man. Bull. ent. Res. 39 : 387 - 397.
Peters, W. 1956. Experimentelle Unter suchungen zur Wirkungsweise
Insectenabwehr ender Mittel. Z. angew. Zool. 1 : 1 - 75.
Peters, W. and Kemper, H. 1958. Uber die Wirkungsweise von
Repellents. Ein Beitrag sum Problem der Wirtsfindung die
blutsaugenden Arthropoden. Proc. tenth int. Congr. Ent.
2 : 389 - 392.
Rahm, U. 1956. Zum Problem der Attraction von Stechmucken durch
den Menschen. Acta trop. 13 (4) : 319 - 344.
Rahm, U. 1957. Zur Bedeutung des Duftes und des Schweisses bei der
Attraktion vonAedes aegypti durch Menschen. Acta. trop. 14(3) :
208 - 217.
Rahm, U, 1958. Die attraktive Wirkung der von Menschen abgegebenen
Duftstoffe aufAedes aegyptiL,, Z. Tropenmed. u. Parasit. 9 : 146-
156.
34
Khan
Rahm, U. 1958. Die Funktion der Antennen, Palpen und Tar sen von
Aedes aegypti L. beim Aufsuchen des Wirtes. Rev. suisse Zool.
65 : 779 - 792.
Rao, T. R. 1947. Visual responses of mosquitoes artificially rendered
flightless. J. exp. Biol. 24 : 64 - 78.
Reuter, J. 1936. Orienterend ondersoek naar de oorzaak van het
gedrag van Anopheles masculipennisM.eigen by de voedselkeuze.
Acad. Proefachr. Rijksuniv. Leiden, 118 pp.
Richards, O. W. 1927, Sexual selection and allied problems in the
insects. Camb. phil.Soc.R ev. Proc. 2 : 298 - 360.
Roth, L, M. 1948. A study of mosquito behaviour. An experimental
study of the sexual behaviour of Aedes aegyptiL,. Amer. midi.
Nat. 40 : 265 - 352.
Roth, L. M. 1951. Loci of sensory end organs used by mosquitoes
( A. aegypti L. and Anopheles quadrimaculatus Say. ) in r eceiving host stimuli.
Ann. ent. Soc. Amer. 44 : 59 - 74.
Roth, L. M. and Willis, E. R. 1952. Possible hygr or eceptor s in
Aedes aegypti L, and Blatella germanica L. J. Morph, 91 : 1-14.
Ross, R. 1898. Report on the cultivation of Proteosoma Labbein grey
Mosquitoes. Indian med. Gaz. 33 : 173-5, 401-8, 448-51.
Rudolfs, W. 1922. Chemotropism of mosquitoes. New Jersey Agric.
Exp. Sta. Bull, 367,
Shaerffenber g, B. and Kupka, H. 1951. Unter suchungen liber die
geruchliche blutsaugende Insekten. 1. Uber die Wirkung
eines Blutduftstoffes auf Stomoxys und Culex . Ost, zool. Z. 3:
410 - 424.
Shaerffenber g, B, and Kupka, H. 1959. Der attraktiv Faktor des
Blutes fur blutsaugende Insekten. Naturwissenschaften 46:
457 - 458.
Shannon, R. C. and Putnam, P. 1934. The biology of Stegomyia under
laboratory conditions. I and II. Proc. ent. Soc. Wash. 36 :
185 - 242.
Slifer, E. H. 1962. Sensory hairs with permeable tips on the tarsi
of the yellow fever mosquito, Aedes aegypti . Ann. ent. Soc.
Amer. . 55 : 531 - 535.
Slifer, E. H. and Sekhon, S. S. 1962. The fine structure of the sense
organs on the antennal flagellum of the yellow fever mosquito,
Aedes aegypti L. J. Morph. Ill (1) : 49 - 68.
Steward, C. C. and Atwood, C. E. 1963. The sensory organs of the
mosquito antenna. Canad, J. Zool, 41 (4) : 577 - 594,
Sturckow, B. 1959. Uber den Geschmacks sinn und den Tastsinn von
Leptinotarsa decemlineata Say. (Chrysomelidae). Z. vergl. Physiol,
42 : 255 - 302.
Thiel, P. H. Van. 1937. Quelles sont les excitations incitant
1 'Anopheles maculipennis atr oparvus a visiter et a piquer l'homme
ou le betail? Bull. soc. Pat. exot, 30 : 193 - 203.
Thiel, P. H. Van. , and Laarman, J. J. 1953. What are the reactions
by which the female Anopheles finds its blood supplier? Docum.
Med. geogr. trop. 6 : 156 - 1 6 1 . Idem Acta leidensia.
1954. 24 : 180 - 187.
Repellent Effects
35
Thompson, R. P. and Brown, A. W. A. 1955. The attractiveness of
human sweat to mosquitoes and the role of carbon dioxide.
Mosquito News. 15 (2) : 80 - 84.
Travis, B. V. and Smith, C. N. 1951 Mosquito repellents selected
for use on man. J. econ. Ent. 44 (3) : 428 - 429.
Wallis, R. C. 1954. A study of oviposition activity of mosquitoes.
Amer. J. Hyg. 60: 135 - 168.
Wiesmann, R., and Lotmar, R. 1949. Beobachtungen und Untersuch-
ungen uber den Wirkungsber eich des neuen Repellent Kik-
Geigy. Acta Trop. 6: 135 - 168.
Willis, E. R. 1947. The olfactory responses of female mosquitoes.
J. econ. Ent. 40 : 769 - 778.
Wolbarsht, M. L. and Dethier, V. G. 1958. Electrical activity in the
chemor eceptors of the blowfly. 1. Responses to chemical and
mechanical stimulation. J. gen. Physiol. 42 : 393 - 412.
Wright, R. H. 1962. Studies in mosquito attraction and repulsion.
V. Physical theory. Canad. Ent. 94 : 1022 - 1029.
36
Book Review
LINDROTH, C. H. 1963. The gr ound-beeltes of Canada and Alaska.
Part 3. Opuscula Entomologica, Supplementum XXIV, pp. 201 - 408,
Figs. 102-207. Zoological Institute, University of Lund, Lund, Sweden.
Price - 35 Swedish crowns.
This portion of this work, the second to be published, includes
the last part of the taxonomic treatment of the genus Trechus , and a rev-
ision of the bembidiine genera Asaphidion Gozis (three species ), Bembidion
Latreille, and the monotypic genus, Phrypeus Casey. The treatment of
Bembidion occupies 200 of the 207 pages. This volume is based on an
examination of the relevant material stored in the major European and
North American museums and private collections, and on the extensive
collections of Lindroth.
As in part 2, Lindroth provides for each species a succinct
synonymy, a synoptic description, and data on type locality , ecology, and
geographical distribution.
The text is straight - forward, simple English. The resulting
clarity of expression illustrates very well the author's thorough knowl-
edge of his subject.
The illustrations are excellent, and those of the entire insects
are among the best ever executed of carabid beetles. For many of the
species, the internal sac of the male genitalia, with its complex folds
and peculiarly shaped sclerites is illustrated, in the infolded position.
Also provided are simple, clear - cut line drawings of various other
structures. All drawings were made by the author himself.
The treatment of the genus Bembidion is the dominant feature
of this volume. The 193 species, 31 of which occur in the United States
only (excluding Alaska), are arrayed in 48 groups. An additional six
extra - limital species are included in the key to species, but are not
treated elsewhere in the text. For each group, a brief diagnosis is
given, as well as the subgeneric name that would apply if the author chose
to use the category subgenus. Twenty-five new taxa are described, of
which four are ranked as subspecies. Of the new species, the type
localities of six are in the United States (excluding Alaska). Although
the work deals primarily with the Canadian and Alaskan fauna, Lind-
roth treated all of the known North American species for a number of
the species groups.
Bembidion has long been regarded as the most difficult and complex
genus of carabids in North America, and the justification for this opinion
is perhaps best illustrated by the large number of synonyms listed -
165-of which 159 were proposed by one author , Colonel Thomas Lincoln
Casey. (By way of contrast, 21 Casey species are recognizedas valid,
and his names are also used fo-r another two species, as a result of the
fir st-used names being junior homonyms). The synonymy is based upon
study of the type specimens by Lindroth, and the facts should settle any
doubt about the value and quality of Casey's work in the Carabidae.
Hayward's revision of 1897 (Trans. Amer. ent. Soc. , vol. 24) was also
grossly inadequate. Lindroth's extensive knowledge of the European
37
species of Bembidion , plus his thorough familiarity with Netolitzky's
fine study are factors which contributed in an important way to the suc-
cess of the study of the North American species . Thanks to this revision,
it is now a relatively simple task to determine any specimen from Can-
ada or Alaska.
The two keys for identification (one to species groups , and one to
the species) are easy to use. This statement is based on personal
experience gained by identifying several thousand specimens, represent-
ing a substantial portion of the species. Each couplet in the keys consists
of a pair of clear - cut alternatives , and there are no complicated "either-
or” statements. One of the features facilitating use of the long key to
species (225 couplets ) is that the numbers of those couplets which set off
a large number of species are in bold face. In spite of these good features
I have three criticisms to make regarding the keys: a. no attempt was
made to relate directly the species-group key to the species key; b.
names of author s of species were not given in the key; c. page references
to the text were not given for the Canadian and Alaskan species . However ,
these are minor points, and the last one is largely taken care of by the
number which is assigned to each species in both key and text.
In a key of this length, it is almost impossible to avoid errors,
and it is with regret that the following omissions of species are noted:
64. nigrum Say; the species of the incrematum group- 103 incrematum LeConte,
104. immaturum Lindroth, and 105. gracilitorme Hayward; and humboldtiense
Blaisdell, p. 305.
The fact that only a few subspecies were described or recognized
may suggest that the author is unaware of current taxonomic theory.
Such, however, is not the case. Lindroth notes carefully geographical
variation where he finds it, but he describes as subspecies only those
populations which are clearly geographically isolated from their closest
relatives, and which differ markedly from them. He avoids naming
populations which are segments of dines, and thus avoids proposing a
lot of trinominals which will subsequently have to be synonymized.
A search through the work for indications of modern techniques
of analysis will prove fruitless. One does not find complex graphs,
charts, or long tables, and only very few simple statistical parameters
are indicated. However, the study does not suffer from this seeming
lack. This seems to me to show that a major attribute of a good taxon-
omist is the ability to interpret correctly carefully chosen, accurate
observations. This is not to say that the study of the genus cannot be
pursued profitably with more sophisticated techniques, but rather that
I doubt that such techniques would have provided, at the present level of
understanding, much more than Lindroth was able to state using the
methods of analysis that were in use in the time of Linnaeus. This
illustrates that the difference is unimportant between 'modern' as opposed
to 'old fashioned' taxonomy; the distinction should rather be made between
'good' and 'poor' taxonomy.
R egarding clas sification of Bembidion , I think the author is mistaken
inusing only a single infra-generic category, namely 'group'. In a genus
of this size, several infra-generic categories are required to point out
the similarities and differences among the species: subgenus, species
38
group and sub-group, at least, However, Lindroth states that such a
classification should be proposed on the basis of a study of the world
fauna, and perhaps he is right.
The work has, so to speak, opened the door to the study of North
American Bembidion . It provides a basic classification, which can be
easily modified, as required. It shows clearly how diver s e the genus is.
The task of completing the revision of the North American species will
be a pleasure. Because of the marked ecological specialization of many
of the species, the genus should provide valuable material for the study
of the origins of adaptations. Also, the numerous species and their
wide distribution in North America, should provide fertile ground for
the development of zoogeographic studies. And, returning to description
of structures, one should remember that the immature forms are vir-
tually unknown. Lindroth has provided an excellent platform from which
to launch further studies, and it is to be hoped that such studies will be
made in the near future.
Carl Lindroth brought to this work a feeling for these fascinating
little creatures which is best described as deep affection. And this,
combined with unrivalled knowledge, superb talent, and hard work on the
part of the author , has provided us with the finest taxonomic treatment of
a group of carabid beetles ever produced.
George E. Ball
Quaestiones
entomologicae
A periodical record of entomological investigations,
published at the Department of Entomology, Uni-
versify of Alberta, Edmonton, Canada.
VOLUME I
NUMBER 2
APRIL 1965
39
QUAESTIONES ENTOMOLOGICAE
A periodical record of entomological investigations, published
at the Department of Entomology, University of Alberta, Edmonton,
Alberta.
Volume 1 Number 2 6 April 1965
CONTENTS
Editorial . 39
Pucat - The functional morphology of the mouthparts
of some mosquito larvae 41
Editorial - Beastly teachers
Teacher s , they say, area necessary evil; beastly people, teacher s ;
pedantic, dogmatic , intolerant. If this is the nature of the beast, should
we not take Wordsworth's advice and 'let nature be our teacher'? There
could be no better field than entomology in which to put this into practice;
at least we should run no risk of a shortage of teachers.
It is difficult to arrive at a reasonable estimate of the world pop-
ulation of entomologists, because they are difficult people to define and
still more difficult to regiment (praises be]). If one supposed that for
every one attending an International Congress, ten stay at home - or
more likely go out collecting - there must be around 20, 000. If Canada
has as many per head of population as any country, as has been claimed,
the figure may be 50, 000. Let us average these two figures; if we have
35, 000 entomologists, this would allow 22 described species of insect
per entomologist, or if we accept C. B. Williams' estimate of the world
population of insects at 10^, about 3 X 10^ insects per entomologist;
a rather unusual staff/ student ratio.
Insects are certainly pedantic, dogmatic, and intolerant, and
should therefore make good teachers. And as teachers of entomology
they must surely be immune to the fashionable accusation directed at
school teacher s -that they are good teachers but have nothing to teach, if not
to the reciprocal retort often aimed at university teachers. Perhaps this
is the proper role of human teachers of entomology - to help the insect
teach the student, or to help the student to learn from the insect. Cer-
tainly if one had to choose between insects , books, and entomologists , from
which to learn, the choice would be in the order given. Perhaps more
than any other science, biology in general and entomology in particular
mustbe taught from the organisms they are concerned with, in the field
and in the laboratory. Many of us get into the bad habit of reaching for a text
when in doubt about some point of insect structure, when we could just as
40
easily reach for an insect - a much less fallible adviser. The habitual
reference of questions back to the insect might even help us in our dif-
ficulties in keeping up with the literature; it would certainly give us a
surer foundation of knowledge from which to judge whether, in any part-
icular paper, we need to read on. In addition to the rather negative
qualities we started out with, insects are ubiquitous, lively, versatile,
unobtrusive, fertile, and unequivocal. There i s little more one could ask
of a teacher.
One of the interesting advantages of an insect teacher of entom-
ology as compared with a human teacher , is that he can fulfil many of his
functions even after death, especially if well preserved. Indeed it is in
large part the readiness with which they may be acquired in the first
place and preserved in the last place, that makes insects so much more
valuable than many other groups of organisms in the teaching of other bran-
ches of biology. Their only limitation lies in their inability to teach the
structural detail of other groups - unfashionable stuff these days anyhow.
There is a tradition of great teachers of entomology extending back
to the early years of the science itself. Surely a place in this roster
has been earned at least by two species of cockroach, by a fruit fly, and
by mealworms and flour beetles.
Brian Hocking
THE FUNCTIONAL MORPHOLOGY OF THE MOUTHPARTS OF SOME
MOSQUITO LARVAE
A.M.PUCAT „
Division of Natural Sciences Quaestiones entomolo gicae
University of Saskatchewan, Regina oo 1 O
Homologies of the parts of the maxilla and the labium of mosquito larvae were studied.
The name cardobasistipes is proposed for the triangular sclerite latero-posterior of the maxilla,
previously known as the cardo or the palpifer. The numbers of serrations on the prementum and
submentum were found to be of taxonomic value. The sequence of mouthpart movements of filter
feeding and browsing species, and the progress of food particles from the feeding current into
the mouth were observed. Differences in stiffness were found among the setae in different posi-
tions on the mouthparts. These differences were confirmed by staining the cuticle with Mallory’s
triple stain and are correlated with the functions of the setae during feeding. Flexible serrations
at the tips of the labral brush hairs are used for raking food particles in most of the browsing
species of Aedes and Culiseta studied. When in pond water neither the browsing nor the
filter feeding larvae select the type of food they ingest. Feeding behaviour of the predatory
larvae of Chaoborus americanus (J ohannsen) and Mochlonyx velutinus (Ruthe)
was observed.
INTRODUCTION
The mouthparts of a mosquito larva occupy a large portion of its
head; their structure is degenerate. In this work emphasis is placed on
the homologies of the parts of the maxilla and the labium, on the structure
and function of the labral brushes and on the type and size of food part-
icles ingested by the larvae.
The problems of homologies of the mouthparts did not occupy the
early biologists who lacked adequate equipment for detailed study of
minute structures. Hooke (1665) drew a mosquito larva, but he did
not interpret all the parts of its anatomy accurately; for example, he
labelled the external opening of the respiratory siphon as the anus. He
further said about the "Water - Ins ect or Gnat": --"It is suppos'd by
some, to deduce its first origin from the putrifaction of Rain Water. . ."
He wrote that the larvae can move gently through the water by moving
their mouthparts, and "eat" their way up through the water.
Reaumur (1738) described and illustrated the external features
of a mosquito larva which seems to be a Culex species ( pipiens according
to Shannon, 1931). He gave an accurate description of the function of
the labral brushes and described browsing and filter feeding activities
of larvae.
The best known studies on mosquito larvae in the 19th century
42
Mouthparts of Mosquito Larvae
are those of Meinert (1886) and Raschke (1887) who discussed larval
morphology, function of mouthparts, and some of the habits of larvae
and adults.
The names used by authors for the mouthparts of mosquito larvae
are summarized in table 1. The following author s also referred to some
mouthparts by specific names: Miall (1895), Johannsen (1903), Mitchell
(1906), Puri (1925), Montchadsky (1945), and Cook ( 1 956) . A more complete
list of literature on this subject is included in my thesis (Pucat 1962).
It is evident that there is disagreement on the homology and nomenclature
of certain mouthparts. There is less disagreement on the function of
these parts, but this has not been studied exhaustively.
Classification of Feeding Habits
The structure of mouthparts, the method of feeding, and the
habitat of the larvae are inter - related. On the basis of these factors
culicine larvae have been classified into filter feeders, browsers, and
predators (Surtees 1959).
It has been found convenient to follow this classification since
it is based on morphological and functional characteristics. The crit-
eria may be summarized as follows;
Filter Feeders - are larvae which strain out food particles from the
water, such particles being sufficiently small to pass directly into the
digestive tract without undergoing any further breakdown. Their salient
morphological characters are: long, fine, unserrated labral brushes,
large maxillae bearing many fine setae, small weakly chitinized man-
dibles, a weakly chitinized submentum possessing a large number of
very small teeth and, associated with these features, large sub-apical
tufts of setae on the antennae (Surtees 1959). These structural features
were recognized by Wesenberg-Lund (1920) in several Danish species of
mosquitoes. Nuttall and Shipley (1901) described in detail the function
of the labral brushes of a filter feeder, an unnamed Anopheles species.
Feeding action similar to that observed by Nuttall and Shipley
was also observed by Bekker (1938a, b) in Anopheles maculipennis Meigen,
and by R enn (1941) in Anopheles quadrimaculatus Say and Anopheles crucians
Wiedemann. Renn referred to the characteristic anopheline feeding
method in which the floating particles are drawn straight towards the
mouth as "interfacial" feeding. However, sometimes anopheline larvae
employ a feeding method common to the larvae of other genera of mos-
quitoes in vhich the particles move in converging curved lines, and this
Renn calls "eddy" feeding.
Browsers - abrade solid material, the particles of which require
further manipulation by the mouthparts before entering the digestive
tract (Surtees 1959). Mouthparts of this type have been describedby
Mitchell ( 1906), Howard, Dyar, andKnab (1912), We senberg- Lund (1920),
Surtees (1959), Snodgrass (1959), Christophers (I960), and Clements
(1963). All authors agree that browsing larvae are usually bottom feeders.
The labral brushes as well as the maxillary andmandibular bris-
tles are shorter and stiffer than in the filter feeders. As Mitchell (1906)
Pucat
43
pointed out, in brushing over debris at the bottom of a pool very long,
slender hair s would be a disadvantage. Mandibles are used to manipulate
any large particles that come into the feeding stream, and the submentum
is used as a secondary grasping organ. The swimming position is
usually at an angle of about 45 ° to the substratum. Morphological
gradations occur between typical filter feeders and browsers
( Wesenber g- Lund 1920, Surtees 1959).
Predators - have the labral brushes strongly chitinized. The role
of the maxillae has been suppressed and the mandibles are the principal
mouthparts. These are very large with strongly chitinized claws and take
upmostof the oral region of the head capsule. As sociated with the strong
claws are large, stiff spines which also aid in grasping the prey. This
is true of the larvae of Chaoborus and Mochlonyx (Schremmer 1950, Peterson
1951, Cook 1956, and others). The submentum in all predatory species
is well developed, the teeth being large and generally pointed. The
increase in the strength of the submentum is associated with a reduction
in the number of teeth and mouth brushes. Predatory larvae have large
prehensile antennae which aid in grasping prey.
Evolution
Montchadsky (1937) has considered the environmental adaptation
of larval and adult structures and behavioral characteristics important
in classification. The type of feeding is a factor correlating the processes
of evolution of larval and adult mosquitoes.
The Anophelinae and Culicinae have mostly plant-feeding larvae
and blood - sucking adults (Montchadsky 1937, Hennig 1950). However,
the Toxorhynchitinae and the culicine subgenus Lutzia have reversed
their type of feeding; the larvae lead a predatory life, but have structures
which indicate a previous adaptation to a vegetarian type of feeding. The
adults of these mosquitoes either feed on plant juices (but carry traces
of previous ability to suck blood), or appear to be optional blood feeders
(Montchadsky 1937). In the Chaoboridae the adults are plant feeding
while the larvae are predatory. Two lines of adaptation to predation
are known: the surface film feeders such as Eucorethra , and the pelagic
feeders such as Chaoborus .
In the initial stages of evolution of the mosquitoes either there
was a change in the type of feeding of the adults (transition to blood
feeding in the subfamily Culicinae), or of the larvae (the transition to
predation in the Chaoboridae). According to Montchadsky (1937) these
changes were provokedby certain changes in the nutritional requirements
for the ripening of the sexual organs. If adequate food containing high
quality protein is eaten by the predatory larvae, it is not then required
to be eaten by the adults which may be vegetarian. On the other hand,
non - predatory mosquito larvae do not obtain adequate high quality
protein, so that the adults of these species must have it from the blood
of vertebrates.
TABLE 1 - Summary of names which have been used for some mouthparts of mosquito larvae.
46
Mouthparts of Mosquito Larvae
MORPHOLOGY OF THE HEAD AND MOUTHPARTS OF MOSQUITO
LARVAE
The mouthparts of mosquito larvae were compared with the mouth-
parts of larvae of other Nematocera, Mecoptera, and other panorpoid
groups, or with published descriptions of them.
Procedures
Two species of mosquito, Aedes aegypti ( L. ) and Culiseta inornata
(Williston) were reared in the laboratory, so that fresh specimens of
these species were almost always available. Rearing methods of
Trembley (1955) and McLintock (1952) were followed. Specimens from
the field were also observed alive and dissected in the laboratory. Since
larvae were available in abundance, dissected heads were mostly studied.
The dissections were done in glycerine. Hoyer's mounting medium and
neutral Canada Balsam were used for mounting the mouthparts. Eosin-
water solution was used for staining dissected muscles , and modified
(Peterson I960 ) Mallory's triple stain for larval head cuticle. The
mouthparts were boiled for 15 minutes in an 8% aqueous solution of
KOH before staining.
Manton ( 1958) commented on the staining reaction of cuticle
with Mallory's. She concluded that sclerotized non- staining exocuticle
is unstr etchable when thick, that orange and red- staining cuticle are
progressively less fully slcerotized, less rigid, and more elastic than
the non- staining cuticle, and that blue- staining cuticle is fully flexible,
more stretchable, but less elastic.
The structure of the heads of the larvae of Aedes fitchii (Felt and
Young) and Culiseta inornata was studied in detail, and other species (table 2)
were compared with them. Larvae of a Chironomus species, and of Mochlonyx
velutinus (Ruthe) and Chaoborus americanus (.Johannsen) were also examined.
The Head Capsule
The largest sclerite in the head capsule of a mosquito larva is the
fr ontoclypeus , which extends over most of the head surface dor sally.
The genae are lateral, the postgenae postero-lateral; they extend vent-
rally to complete the head capsule (figs 1,2). The median ventral part
of the united postgenae, posterior to the mouth, has been given various
names. I consider it as the subgena. It is bounded by two lines of
cuticular thickening ridges which are known variously as the submental-
postgenal sutures (Shalaby 1956 and 1957a, b,c,d) hypostomal sutures
(Menees 1958a, Christophers I960), and thickening ridges (Snodgrass
1959). I agree with Snodgrass' interpretation of the homologies of the
ventral head sclerites. In homologizing these sclerites of the mosquito
larva Snodgras s digresses to discuss the ventral head sclerites of other
insects, especially insects in which a trend toward a ventral elongation
of the postgenae is evident. As examples he cites certain beetles in which
the entire labium with a gular addition to the submentum is enclosed
between the postgenae. He states, however, that this condition is not
47
0. 5 mm
posterior
tormal
apodeme
labral brush
flexors
hypopharyngeal
bar
salivary
duct
maxillary
pr emental
Fig. 1. The head of Aedes fitchii (F. & Y. ) larva, (a) dorsal view showing muscle
origins and extended labral brushes, (b) ventral view with brushes retracted and
mouthparts removed from right hand side. mx. maxillae, md. mandible, sm. sub-
mentum, t.m. tessellated membrane, aul. aulaeum, p. t. posterior tentorial pit.
Muscle attachments stippled.
48
a
0. 5 mm
Fig. 2. (a) Lateral view of the left side of the head of Aedes fitehii (F. & Y. ) larva,
(b) Sagittal section through the mouthparts of Aedes fitehii larva, md. mandible,
mx. maxillae, pm. prementum, sm. submentum, aul. aulaeum, distist. dististipes.
Muscle attachments stippled.
Pucat
49
r epresented in mosquito larvae . Mor e commonly , the postgenae come to-
gether medially and displace the labium. A final stage in the displacement
of the labium is seen in the larvae of Chironomidae where the labium has
become greatly reduc ed and is hidden from below by a median hypostomal
lobe of the united postgenae.
A similar proces s of closure and elongation of the postgenae and
reduction of the labium occurs in nematocerous larvae as discussed by
Anthon (1943), Hennig (1948, 1950, 1952), and Snodgrass (1959). In
the larvae of the primitive rhyphid Olbiogaster the small postgenal lobes
are posterior to the submentum of the labium (Anthon 1943). In tipulid
larvae, described by Vimmer (1906) and other authors, as well as in
other .iematocer ous larvae the genae are completely united ventrally
and the labium is dorsal to the subgenal lobe. In the mosquito larva,
to distinguish the central area between the thickening ridges of the genae
Snodgrass (1959) named it the subgena, and the areas laterad of the
ridges the postgenae. I use this nomenclature.
Cook ( 1944a, b, 1949), following Ferris's (1947) and Henry's
( 1 947) theories of the segmentation of the arthropod head, considered the
postgenae and the subgena as parts of the maxillary segment. Shalaby
(1957) considered the apical part of the subgena as the mentum and the
remainder as the submentum. As evidence for this idea Shalaby
referred to Wheeler's (1893) embr yological work in which the latter
observed that the rudiments of the second pair of maxillae on the sides
of the embryonic body give rise to the labium in the embryos of the locust
Xiphidium ensiferum Scudder, in Gryllus luctuosus Serville, and in Stagmomantis
Carolina ( Johanns en). Shalaby believed that the median suture present
on the ventral sclerite of the head of Culex molestus Forsk. larva is due
to incomplete fusion of the embryonic rudiments of the second maxillae.
That the embryonic second maxillae give rise to the labium has been
shown by Butt (1957) in Oncopeltus fasciatus (Dallas), and by other authors
in other insects. Christophers (I960) also believes that the subgena is
the labial area; he homologizes the subgenal and postgenal areas posterior
to the maxillae with the fused bases of the maxillae (cardo and stipes).
He thus believes that in the larval as in the adult stages of mosquitoes
the bases of the maxillae extend to the occipital foramen, forming the
hypostomal area. However, the sclerite which Christophers considers
as the base of the maxilla serves as the origin of pharyngeal, man-
dibular, and maxillary muscles which in most other insects originate
on the tentorium or on the cranial wall (Snodgrass 1935). In the adult
Aedes vexans (Meigen) the maxillary muscles originate on the tentorium
(Peterson Hoyt 1952). On the other hand, none of the postgenal muscles
of the mosquito larva originates on the tentorium. If the larval post-
genaand subgena are to be considered as the fused maxillary cardo and
stipes, then the origins of the various muscles upon them are difficult
to explain. Menees (1958a), studying the embryonic development of
A. quadrimac ulatus , observed that the median suture on the ventral head
sclerite in this species is the result of incomplete fusion of the postgenae.
Most sutures which are characteristic of the primitive insect
head are absent from the heads of mosquito larvae. Two cleavage lines
extend anteriorly from a short posterior occipital stem (fig 1). These
50
Mouthparts of Mosquito Larvae
cleavage lines may be homologous with the frontal sutures and the epi-
cranial suture of other insects. However, Snodgrass (1947, 1958) and
DuPorte (1953) state that the frontal arms of this suture follow diverse
paths in different insects, and therefore do not define any specific part
of the head. For this reason, in this workhead sclerites and mouthparts
have been named in reference to muscle origins.
Approximately in the center of the frontoclypeus arise the labral
and epipharyngeal muscles (fig. 1) which usually originate on the clypeus,
and posterior to these are the origins of the pharyngeal muscles which
generally occur on the frons. In the head of Aedes fitchii (Felt and Young)
larva and in all the other mosquito species examined, there is no
demarcation between the areas where the different muscles originate.
According to DuPorte (1962) in some insects the boundary between
the clypeus and frons, in the absence of an epistomal suture, is fixed
by the position of the anterior tentorial pits. In the heads of mosquito
larvae, however, the epipharyngeal muscle (usually on the clypeus) orig-
inates much posterior to the anterior tentorial arms.
The tentorium in the mosquito larva is represented by anterior
and posterior arms. The anterior arms originate on the head capsule
medial to the antennae, in the same area where the hypopharyngeal bars
arise (fig. 1). The long, slender anterior tentorial arms connect to the
short posterior arms on the postero - ventral part of the head. There
is no tentorial bridge.
On each side of the head a hypopharyngeal bar connects the
hypopharynx to the side of the cranium (fig. 1).
The Labrum
The labrum of the larva of Aedes fitchii consists of a narrow
transverse sclerite dorsally (fig. 1). Ventrally it is composed of a
membranous area to which three brushes are attached, one median
and two lateral and movable. The median brush is connected to each lateral
labral brush and to the distal part of the dorsal labral sclerite by a
membrane which has ben variously named. In the larvae of Lutzia halifaxi
Theobald, Cook (1944b) referred to it as a "pennicular area., beset
with small oval pits arranged in definite rows. " Because of its appear-
ance Christophers (I960) called it the tessellated membrane, and this
is the name adopted her e (fig. 5). However, this name does not describe
the membrane accurately in all the larvae that I examined. This is
discussed further below.
In both A. aegypti , (Shalaby 1957a) and A edes fitchii , two types of
hair s ar e found on the median brush; long thin br anched hair s posteriorly ,
and short stout hairs with serrated distal ends anteriorly. Both types
are shorter on the sides of the brush than medially.
The lateral labral brushes are composed of three types of hairs
which differ in length, thickness, curvature, and location. The hairs of
the first type are simple, relatively short, thin, soft, without definite
curvature, and are located postero - laterally, dorsally, and ventro-
medially overhanging the pharynx (figs. 1,3). These hairs, which are
attached to the tessellated membrane, do not take part in creating a
feeding current. Hairs of the second type are long, simple, thin ,
Pucat
51
slightly curved at their bases and at their distal ends, and are located
in the lateral posterior two thirds of the brush (fig. 3). Anterior to
them are hairs of type three. Types two and three take an active
part in creating currents. The apices of type three hairs are provided
with serrations (17-20 per hair). The serrations on the lateral type
three hairs are smaller and slightly closer to each other than those
on the more medial hairs.
Three types of hairs were found in all the browsing species of
Aedes and Culiseta except in Aedes cinereus Meigen and A. canadensis ( Theo).
which have only short, simple hairs on their lateral brushes. When the
labral brushes are stained with Mallory's the bases of all the hairs stain
red. Next above the bases a narrow layer of blue appears across the
hairs and above this layer hairs of type one and two stain red to their
tips. Hairs of type three stain partly red above the blue portion but
they stain blue apically, in their serrated regions. A large proportion
of the most median type three hairs stains completely blue above the
red bases. In A. fitchii and the other Aedes larvae, as well as in the
browsing Culiseta larvae that were examined, the apices of hairs of types
one and two are tapered. Also tapered are the apices of all the hairs
ofthelabral brushes of the filter feeders, Culiseta mors itans (Theo.) and
Culex territans Walker. In the brushes of the filter feeding larvae all
the hairs are simple. They all have red- staining bas es , blue- staining
portions above the bases, and red- staining middle and apical portions.
In the filter feeding larvae a large group of hairs, originating medially
on each lateral labral brush, overhangs ventrally, partly covering the
epipharynx, A smaller number of simple hairs extends in this position
in the browsing larvae (fig, l),In all the larvae that were examined these
hairs are red - staining. In the larvae of Chaoborus americanus the labral
brushes consist of a fewhard, short, brown bristle s on the small sclerite.
In the larva, of a Chironomus specie s examined a few labral bristles ar e red-
staining and the remainder are blue- staining. Thus the staining reaction
of the labral brushes of the filter feeding and browsing larvae indicates
that their hair bases are elastic and the portions above the bases are
flexible. Flexibility of these hairs was seen when larvae were observed
feeding and also v/hen the hairs were deflected with a needle.
In the mosquito larvae examined all the hair s of the lateral brushes
except type one are attached to sclerotized rods which extend transversely
across the basal area of the brush (figs. 3 and 4). Salem (1931) seems to
be referring to these rods in Aedes fasciata (Fab.) ( A. aegypti L. ) when
he states that the chitin of the labral brush "exhibits a peculiar striated
appearance." Christopher's term for these rods , "cross bars, " is used
here. On each lateral labral brush of A. fitchii larvae between forty-
five and fifty of these bars are present and each bears approximately
twenty hairs. Thus each lateral brush contains nearly a thousand hairs.
A similar number of hairs is pres ent in each lateral brush of C. inomata
larvae.
The cross bars are cuticular thickenings of the tessellated mem-
brane (fig, 5) with their dorsal ends free in this membrane next to the
dorsal sclerite of the labrum. When the cross bars are torn away from
the tes sellated membrane and the hairs, depressions on them where the
Fig. 3. Ventral view of the labrum of the larva of Aedes fitchii with the
lateral labral brushes extended. Numbers indicate hair types.
Fig. 4. Details of labral hair attachments of the larva of Culex territans .
53
Fig. 5. (a) Forked bases of labral hairs of Aedes larvae; anterior views, (b) The
relationship between hair base, cross bar, and the tessellated membrane, and the
holes and depressions left in this by the removal of hairs and cross bars. Open
stipple stretchable cuticle (stains blue); close stipple, flexible but relatively non-
stretchable cuticle (stains red). (c) Diagram showing how the hairs are brought
together by the increasing angle of movement at greater distances from the brush
sclerite, because of differential stretching between the cross bars and the tessellated
membrane.
54
Mouthparts of Mosquito Larvae
hairs were attached can be seen. The other end of each cross bar is
curved into a hook; it terminates in t-he brush sclerite which is roughly
triangular and is attached to the median part of the torma by an apodeme
(fig. 3). Muscles that move this sclerite are inserted on the posterior
tormal apodeme (fig.l). When the hairs are pulled off the membrane,
their forked bases, the cross bar s, and part of the membrane comes with
them. This leaves holes in the membrane and confirms that the cross bar s
are more strongly attached to the hair bases than to the membrane. The
hole may be rhomboid, square, pentagonal, hexagonal, oval, or roughly
circular and form a mosaic pattern on the membrane which gives it its
names. The cross bar s leave depressions in the tessellated membrane.
When this complex is stained with Mallory's the cross bars and
the hair bases stain red indicating rigidity, while the tessellated
membrane and small parts of the hairs above their bases stain blue ,
indicating str etchability. The edges of the holes may be outlined in red
perhaps because of some change in the character of the material of the
membrane resulting from tearing.
The ends of the epipharyngeal bar are attached to the posterior
parts of both tormae (figs. 1, 3). At the anterior end of each torma a
narrow sclerite projects medially. These sclerites are known as trans-
verse bars (Shalaby 1957a) or palatal bars (Christophers I960).
Their structure in A. fitchii is slightly different from that in A. aegypti
as described by the above authors. The bars of A. aegypti are sleudof
and from each a small curved sclerite projects anteriorly. In A. fitchii
they are stout and curved medially, and are attached by thin sclerites
to the tormae. In Culex territans the bars are straight and have wide
basal parts.
In the species examined only. the posterior apices of the tormae
stain blue; the remainder of these structures with their apodemes
retain their brown color . Thus the, tormae and their apodemes a,re rigid,
highly sclerotized structures. The associated membranes stain light
blue.
The labrum of the predatory Chaoborus americanus larva is greatly
reduced; it lacks brushes but possesses a few short stiff bristles at the
tip of the labral sclerite (Cook 1956). These bristles stain dark red.
The Epipharynx and Preora! Cavity
The epipharyngeal apparatus lies between the posterior ends
of the tormae and combs food particles from brushes to the mandibles.
Schremmer (1949) called it the "Epipharynx - appar at" because it is
musculated and has an active rather than a passive function.
The structure of the epipharynx in the species examined is
very similar to that described by Shalaby (1957) and Christophers (I960)
in A. aegypti , In A. fitchii andthe other browsers the hairs are coarser
than in Culiseta morsitans and Culex territans , The spines and hairs stain
dark red in A. fitchii which indicates medium hardnes s ; they stain lighter
redin C. morsitans and C. territans and are probably softer in thes e species ,
The epipharyngeal bar stains medium blue in all specimens. That this
flexible structure can move anteriorly and posteriorly has been observed
in living larvae of A. aegypti and C. territans
Pucat
55
The post- epipharyngeal area consists of a membrane between the
epipharynx and the pharynx. It is similar to that described by Cook
(1944b) in Theobaldia incidens {-■ Culiseta incidens ). Two pairs of muscle
strands originate on the frontoclypeus , one of these forks before its
insertion in the membrane between the epipharynx and the pharynx.
Since these muscle strands have a common origin on the cranium med-
ially of the antenna (fig 1), I consider them as fascicles of one muscle,
the postepipharyngeal.
The Mandibles
The mandibles of mature /let/es fikAii larvae consist of flattened,
roughly quadrilateral lobes with their mesal ends produced into strongly
sclerotized toothed processes and lower seta-bearing lobes. They are
similar to the mandibles of most culicine larvae which have been des-
cribed by other authors.
On the mesal margin of each mandible is found a fringe of pig-
mented, long,mesally directed setae with stout bases and sharp points.
Shalaby ( 1 957a) called this fringe the mandibular comb when he described
itin A. aegypti , The number of the curved, stout and sharply pointed s etae
varies in fourth instar larvae of the species that I examined. Eleven
were usually found in A. fitchii , nine in C. inornata , and fifteen in
A. aegypti , Another series of setae extends meso-dor sally from the
dorsal side of the mandible, medially of the large lateral bristles; this
series Shalaby names the mandibular brush. In C. inornata it usually
consists of 40 setae; in A. fitchii of 54. The number of lateral bristles
is variable between species, but constant in all the species seen; in
A. fitchii two are present and in C. inornata three. When the mandibular
brush and comb setae of the Aedes and the Culiseta browsing species are
stainedwith Mallory’s their bases stain blue, and thus are soft; the re-
maining parts stain dark red, and are harder. The mandibular setae of
the filter -feeding species , Culiseta morsitans and Culex territans are softer
than those of the browsing species. The lateral bristles remain brown
in all the species examined. All the mandibular bristles and setae in
the mandible of Chaoborus americanus stain dark red or remain brown.
The number of teeth in A. aegypti , as described by Shalaby, is
similar to that in A. fitchii and to the other Aedes species that were
examined. The number of ventral teeth in C. inornata is similar to that
found in the browsing Aedes species, but dorsally only three teeth are
present in C. inornata whereas five are present in all specimens of all the
Aedes species. The extent of heavy scler otization in the tips of the man-
dibles, mainly the teeth, is approximately the same in C. inornata and the
browsing Aedes species. The heavily sclerotized area is smaller in the
filter feeders, and it is largely extended in the predatory Chaoborus
americanus and Mochlonyx velutinus . These characteristics agree with the
characteristics of browsers, filter feeders, and predators that Surtees
( 1 959) discus ses . Medially, on the dorsoventral ridge of the mandible a
group of long spines reachesthe anterior part of the pharynx. Schr emmer
(1949) discusses the function of similar spines on the mandible of
Anopheles maculipennis ■ Anterior and po sterior mandibular articulations are
indicated in fig. 1.
56
Mouthparts of Mosquito Larvae
The Maxillae
Each maxilla of A. fitchii (fig. 7) consists of a rectangular flattened
lobe which bear s a brush of long hair s apically, and a series of three rows
of short hair s medially in an area demarcated by a suture on the oral (dor-
sal) side. Proximal to the palpus is a triangular sclerite about half the
width of the main looe, which is attached to the s e structures and to the post
gena by a membrane. This sclerite bears a spine medially. Baso-ventr-
ally the maxillary palpus bears scler otized processes which articulate with
a postgenal articular proces s inside the head (fig.l). The mandible also
articulates with the postgena and the maxilla at this point. Two muscles
are inserted in the center of the main maxillary lobe;a single strand
originates on the subgena mesally to the posterior tentorial pit, and a
double strand originates on the postgena posterior to the eye (fig. 1),
To decide what parts of the maxilla of A. fitchii larvae are homo-
logous with parts of maxilla of other insects, the r elation between scler-
ites and musculature must be considered. It is generally accepted that
as Imms (1944) states . . the Mecoptera are the nearest living repre-
sentatives of ancestors of Diptera. . , " This view is also expressed by
Applegarth ( 1 939) , Ferris and R ees ( 1 939) , Potter ( 1 948) , Hinton ( 1 958) ,
and other s . We should therefore lookfor homologies of the maxilla of the
mosquito larva in the Mecoptera and in other members of the suborder
Nematocera. The palpus is the only structure on the homology of which
all the authors agree. Since the palpus is connected to the base of the
main maxillary lobe, and since the palpus in all insects is connected to
the stipes , it seems logical to consider this lobe as the stipes. According
toSnodgrass (1936) and Das (1937) the stipes can be distinguished by the
origin of the muscles of the palpus. However, this criterion does not
apply when the palpal muscles are absent as from mosquito larvae and
larvae of Tipula and Bibio as described by Das (1937) and Cook (1944a),
The two muscles that are present in this structure are probably the cranial
flexors of the stipes (rather than of the lacinia). The double strand which
originates on the postgena is one of these, and the adductor of the stipes
which usually or iginates on the tentorium is the other. In the culicid larva
the origin of the latter has shifted to the subgena.
Snodgrass (1935) and Das (1937) hold that the lacinia has a cranial
flexor and the galea has only a stipital flexor in larval and adult stages of
many insects . Das also states that many larvae lack the flexor of the galea,
but when the lacinia is present its cranial flexor is always retained. The
same author adds that the cranial flexor of the lacinia plays an important
role in the interpretation of the lobes. No trace of stipital flexor was
foundin any culicid larva examined. The only cranial flexor present is
inserted so close to the median side of the main lobe that it is almost on
the bri stle- cover ed area which is demarcated by a suture on the oral side
of the lobe (fig. 7), Furthermore, this median bristly area functions as
a lacinia. Therefore I agree with Shalaby (1957a, 1958) that this part of the
maxilla is the lacinia, and that the cranial flexor of the lacinia now
functions as a stipital flexor.
In the larvae of Panorpa both galea and lacinia are present (Das 1 937) ;
in Apterobittacus only the lacinia i s present in the larval stage and the galea
appear s in the pupal stage (Applegarth 1939); in both Tipula (Dasl937)and
Bibio (Cook 1944b) only the lacinia is present in the larval stage. The
57
mandibular brush’
mandibular comb
dorsal teeth
0. 5 mm
brush of dististipes
disti stipe s
Fig. 6. Ventral view of the left mandible of mature larva of Aedes fitchii
Fig. 7. Dorsal view of the left maxilla of mature larva of Aedes fitchii .
58
Mouthparts of Mosquito Larvae
triangular sclerite which is considered as the palpifer by most authors
I believe to be at least a partial vestige of the cardo. In the larva of
Panorpa the car do has a relative size, shape and position similar to that in
the mosquito larva, and it also lacks musculature (Das 1937) . In the lar-
vae of each of Apterobittacus , Bibio . and Tipula species the structure named
as cardo by the respective author s , is proportionately larger than in the
larvae of Aedes , Culex , and Culiseta. In the former three larvae the so-
called cardo extends posterior to the stipes and the palp. If this structure
is homologous with the triangular sclerite in the mosquito larva then this
sclerite must be the cardo and not the palpifer. However Hinton (1958)
points out that the stipes is divided in to a basistipes and dististipes in all
the Panorpoidea except the more specialized Diptera. The same author
further states that failure to recognize the fact that the stipes is sub-
divided in primitive forms of all recent orders of the Panorpoidea has
resulted in the misidentification of the dististipes as the palpifer. Hinton
also states: "in the Panorpoidea in which the cardo has become fused to
the basistipes the combined structure which may be called the cardostipes
has almost without exception been identified as the cardo and the disti-
stipes as the stipes. For instance, the cardo plus basistipes of Bibio is
called the cardo and the dististipes is called the stipes by Imms (1944) and
Cook (1949). . . " In the light of Hinton's statements then I consider the
triangular sclerite of the mosquito larval maxilla as homologous with
the fused cardo and basistipes. The main maxillary lobe is the disti-
stipes plus the lacinia. In addition Hinton mentions that within the
Nematocera a fusion of the cardostipes with the dististipes takes place for
example in the Culicidae, but he does not specify in vdiat group of the
Culicidae. He may be referring to the genus Anopheles , for in that genus
there is no triangular sclerite proximal to the maxillary palp and the
dististipes as in the genera Aedes, Culex, and
Essentially the same structural arrangement of the maxilla
was found in all the Aedes , Culex , and some Culiseta larvae that I
examined. Some difference from the browsers was found in the shape
of the maxillae of Culex territans , Culiseta morsitans , Aedes canadensis , and
A. cinereus Each maxilla in these species is cone- shaped, wide at the
base and narrow at the apex where a brush of simple hairs is attached.
The maxillae of most browsers are similar in shape to those of
Aedes fitchii . Between the browsers and filter feeders differences occur
in the number and length of hairs on the distal end of the dististipes and
on the lacinia. In the maxillae of both filter feeder s and browser s the
apical brush hairs of the dististipes are longer than the lateral hairs of
the lacinia, and in the filter feeders all these hairs are proportionately
longer than in the browser s . The longest maxillary hairs in Culex territans
and Culiseta morsitans are approximately one and a half times as long as
the dististipes; whereas the homologous hairs in A. fitchii and the other
Aedes browsers are only approximately as long as the dististipes, and
in both Culiseta inornata and C. impatiens (Walker) they are half the length of
the dististipes. The maxillary brushes of the browsing Aedes species are
composed of more hairs than those of the filter feeding species. The
maxillae of C. inornata and C. impatiens larvae have brushes consisting
of very few hairs, thus resembling the maxillae of predatory larvae.
Pucat
59
Another similarity of the maxillae of these two Culiseta species to the
predatory larval maxillae is the fusion of the palps with the cardobasi-
stipites.
With Mallory's stain the bases of the maxillary brush hairs of
browsing larvae stain blue and the remaining parts red, but the whole
hairs stainblue in filter feeders. Thus the maxillary bru shes of browsers
are stiff, a feature of obvious value in their activity.
The short medial bristles of the lacinia are arranged in three
rows in all the species that I studied; they are more numerous in
browser s than in filter feeder s. These hairs are longer in A. fitchii and
the other Aedes browsers than in C. inornata and C. impatiens . In all the
browsers these hairs stain red, indicating moderate stiffness. The
hairs of the lacinia of the filter feeders stain blue and thus are soft.
The Labium and Hypopharynx
I consider the labium of the larva of A. fitchii to consist of the
prementumand the submentum. This view is in agreement with Cook's
(1944b) interpretation for other genera. The prementum (fig. 2) is a
rectangular membranous area bearing a series of serrated sclerites
and papillae, and is situated between the hypopharynx and the mouth
opening dor sally, and the triangular serrated submental plate ventrally.
Dorso-ventrally two long sclerites extend through the centre of
the prementum and dorsally terminate ventral to six small serrated
sclerites which project ventrally from the membranous base. On the
sides of the membrane three serrated plates are situated ventrally.
These three plates are connected to each other, and dorsally to the
small central serrated sclerites. Each plate has a different number of
serrations, which vary in different species. In A. fitchii . the dorsal
plate has four serrations, the median plate nine, and the ventral plate
five. Six larvae of each of two closely related species, Aedes hexodontus
and A. pun c tor were also examined, and the aver age number s of serrations
were found to be; dorsal plate 5 serrations in A. hexodontus , 4 in
A. punctor ; median plate 6 in A. hexodontus , 9 in A. punctor ; ventral plate
6 in A. hexodontus , 10 in A. punctor . This may be a useful taxonomic char-
acter for separating closely related species. Considerable car e is
required in preparing the slides if the serrated plates are to be seen
clearly.
Since these plates in all the species of Aedes , Culiseta , and Culex,
that were examined stain light red basally and dark red to orange distally,
they are quite hard. This is understandable because the mandibular
teeth which are of similar hardness strike against them. The hardness
of both structures could be felt with dissecting needles. In the Aedes
species a group of broad, apically serrated hairs originates on the mid-
ventral side of the premental lobe. Broad, but not serrated hairs occur
in the same position in the Culiseta and Culex species. These hairs are
numerous in Aedes and Culiseta but very scarce in Culex. In all the
species examined they stained medium red with Mallory's.
On the premental lobe laterally, between the central and the
lateral serrated plates four small papillae are present on each side in all
the species of Aedes , Culiseta , and Culex that I examined. The most
60
Mouthparts of Mosquito Larvae
posterior papillae are double on each side; the more anterior two arise
singly. Two similar papilla-like processes are present in the membrane
dor sally between the serrated plates and the salivary duct opening.
In all the species considered the papillary structures stained red, and
the basal membranes light blue. In feeding larvae, food often collected
in the spaces between the papillae and the serrated plates.
It is difficult to homologize the structures of the labium because
of its degenerate nature, but since a pair of muscles attaches the
rectangular lobe to the subgena medially to the posterior tentorial
pits (fig. 1), these muscles are considered as the premental muscles
by Cook (1944b, 1949), Snodgrass (1959), and others, Snodgrass
refers to the lobe as the labial plate. I agree with Cook in calling it
the prementum.
The premental membrane is dorsally suspended from the hypo-
pharyngeal bars. A weak suture continues between these bars and
dorsally of the premental membrane, thus demarcating an oval
membranous hypopharyngeal area above the prementum. The opening
of the salivary duct is located between the premental and hypopharyngeal
lobes. This was so in all the species examined including A. aegypti
although Christophers (I960) shows it in the center of the prementum.
The triangular serrated sclerite below the prementum has been
variously named (table 1). I agree with Cook (1944b, 1949) that it
represents the submentum. Salem (1931) considered it homologous with
the submentum, but thought that the customary name, mentum, should
be retained. Snodgrass (1959) believed it to be an extension of the sub-
gena. Jones (I960), following Snodgrass, calls it the hypostomium in the
larvaof Anopheles quadrimaculatus . My main r eas on for disagr eeing is that in
all the species examined this sclerite articulates with the subgena, and
therefore is unlikely to be an extension of it. Generally the submentum
of insects articulates with the ventral part of the cranium (Snodgrass 1933) .
Snodgrass (1959) however, does not mention that this triangular structure
arcticulates with the subgena. He states that it is continuous with the
subgena, as in the head of Chironomus described by Grouin (1959) who calls
it the hypochilum. Miall and Hammond (1891) indicate that this plate in
Chironomus seems to correspond to the submentum of orthopterous insects.
The submentum stains orange basally with Mallory's and remains
dark brown apically in all the Aedes , Culex , and Culiseta larvae I
examined. It is thus a very hard structure. In the species examined
the number of serrations on it in mature larvae is usually constant;
data are given in table 2.
The lightly sclei otized fringe of hair s (figs. 1, 2) attached to the
submentum ventrally stains similarly; I consider it a part of the sub-
mentum since it is very intimately connected with this structure.
Cook (1944b) calls it the aulaeum.
The Pharynx
The structure and musculature of the pharynx of A. fitchii and
C. inomata larvae are similar to those of Theobaldia incidens [- Culiseta
incidens ) described by Cook (1944b). The large dor sal and vent-
ral sclerites stain light orange in all the Aedes , Culex , and Culiseta larvae
Mouthparts of Mosquito Larvae
61
TABLE 2 - The numbers of serrations on the submentum of the larvae
of mosquito species.
* average ± S. D. of the mean (where applicable);
number of specimens examined in parentheses.
62
Mouthparts of Mosquito Larvae
examined. The lateral dor sal hair s stain light red, and the inner filtering
hairs stain light blue in most species. Schremmer (1949) described
the filtering function of the pharyngeal hairs in Anopheles maculipennis larva.
Discussion
It is difficult to decide on the homologies of degenerate structures
like t h e maxilla and labium of mosquito larvae. Shalaby's(1957d)
interpretation of the triangular labial sclerite as the paraglossa, and the
aulaeum as the glossa is unique, and seems unreasonable. The areas
which I consider as the hypopharynx and the prementum Shalaby regards
as the hypopharynx. Medio- laterally on the premental lobe a pair of
muscles is inserted. These muscles originate on the ventral sclerite
of the head which Shalaby considers as the submentum and which I
regard as the subgena. It is difficult to agree with Shalaby 1 s interpretation
of the labium and the hypopharynx for the following reasons: firstly,
as far as is known, the hypopharynx in insects is not connected
with the paraglos sa, but in the mosquito larva, in Shalaby' s interpretation
the " hypopharynx " is firmly attached to the "paraglossa". Secondly,
other authorities on the morphology of insect larvae (Cook 1944, 1949;
Hinton 1958) state that the retractor muscles of the hypopharynx are
absent in Diptera. Thirdly, when the retractors of the hypopharynx are
present they arise on the postoccipital ridge in the Trichoptera, and
on the tentorial bridge in the Lepidoptera (Hinton 1958), but not on the
"submentum" where these muscles originate in the mosquito larva
according to Shalaby' s interpretation.
Very few muscles which could serve as guides to homology
are present, and this is partly why disagreements exist among the
various morphologists who have studied mosquito larval mouthparts.
Ferris ( 1 948) postulates the following principle: " the evolutionary
changes ,are first to be accounted for by modifications of pre-existing
structures , or by loss of pre-existing structures; Only after these
possibilities have been exhausted will we assume that a completely new
structure has been developed. ..." This principle can be applied to
mosquito larvae and to the larvae of other primitive Nematocera when
we compare them with panorpoid larvae. In mosquito larvae noticeable
modification from Panorpa is seen in the labrum and in the mandibular
teeth. Losses and fusions of pre-existing structures are evident in the
mosquito larval maxilla and the labium.
A difference was found in the hardness and flexibility of the
cuticle of the mouthparts of the filter feeding, browsing, and predatory
mosquito larvae. Essentially, the mouthparts of the filter feeders are
rather soft except for the labral brush hairs and the mandibular teeth;
the mouthparts of the browsers are harder, and the mouthparts of the
predatory larvae are the hardest of all, especially the mandibles, which
are highly sclerotized.
The tips of the simple labral brush hairs of the filter feeding
and browsing larvae are softer than the main parts of the hairs. The
labral brush hairs of these groups of larvae are much harder than they
appear to be since they are refractory to stain until after boiling in a
relatively strong (8%) solution of KOH. It was interesting to find that
Pucat
63
the serrated ends of the lateral labral brush hairs of the browsing larvae
stain blue and thus are soft combs rather than hard ones as they might be
expected to be when their function is considered. Since they are soft
it is probable that when they rub over surfaces soft particles are detached
and then dir ected towards the mouth. The physical characteristics of the
cuticle were estimatedby manipulating the mouthparts, and the impres-
sions obtained agreed with the indications from staining.
The serial row attachment of the labral brush hairs to their
respective bar s is similar in the browsing and the filter feeding larvae.
Christophers (I960) also noted that the hair attachment is similar in
the larvae of a Culex species and of A. aegypti .
In table 3 it is indicated that a reduction occurs in the numbers
of hairs or bristles on the various mouthparts from the filter feeders
to the predators. In the same series an increase in the scler otization
of the mandibular teeth is evident.
TABLE 3 - Similarities and differences in the mouthparts of filter
feeding, browsing, and predatory mosquito larvae.
It is interesting to note that the same genus is represented by
filter feeding ( Culiseta morsitans ) and browsing larvae ( C. inornata and
C. impatiens ) whose mouthparts tend towards the predatory type. Most
64
Mouthparts of Mosquito Larvae
of the Aedes species that were studied are browsers, but the larvae of
Aedes cinereus Meig. and A. canadensis lack serrations on their labral
brushes, have more weakly sclerotized mandibular teeth than the other
Aedes species, and their maxillae are similar to those of the filter
feeders. Thus morphologically these species seem to be intermediate
between the filter feeders and the browsers.
From table 3 it is also evident that the plane of action of the
mandibles in the predatory larvae tends towards that of the longitudinal
axis of the body which is a character common both among the larvae
of the higher flies, according to Cook (1949), and among predators
generally.
FUNCTION OF THE MOUTHPARTS OF MOSQUITO LARVAE
Procedures
The movements of the mouthparts of mosquito larvae and actions
resulting from thes e movements were studiedintwo situations: behaviour
of larvae (mostly Aedes ) was observed in the muskeg pools in the Flatbush
area (100 miles north of Edmonton) in the summers of I960 and 1961;
more extensive observations were made on active larvae in artificial
containers in the laboratory.
After being collected the larvae were kept in pint glass jars,
and in order to retard their development when not being observed they
were kept in the refrigerator at 40°F. The larvae were observed in
groups and individually in the glass jars and some details of movements
of their mouthparts were seen with the aid of a 1 OX hand lens. Individual
larvae were placed in small vials and their mouthparts were observed
from the side with a hand lens. A viscous solution of an inert material
such as methyl cellulose was also used to slow down the motions of the
mouthparts so that details of their actions could be studied.
Larvae of A. aegypti and Culiseta inornata reared in the laboratory
were observed. Other species of Aedes and Culiseta were collected in the
areas of Flatbush, Edmonton, Lake Hastings , Banff, and Seebe, Alberta.
The larvae were identified with the keys of Rempel (1953) and Carpenter
and La Casse (1955).
Since the mouthparts are ventral it was desirable to observe
larvae from the ventral side; three methods were used for this. For
all the methods a container was made by cutting a 1 in long piece of
aplastic vial of 1 in diameter, and gluing it to a microscope slide which
formed the bottom. The container was filled with either pond water or
distilled water and food was added. Usually one larva was studied at
a time, but sometimes two were observed in the same dish.
By means of two concave mirrors, light from two microscope
lamps was directed on the larva through the bottom of the container.
An image of the ventral surface of the larva was reflected by two plane
mirrors at 45°, one below the container and one below the objective of a
stereo-binocular microscope. A satisfactory view of the mouthpart
Pucat
65
movements was obtained in this apparatus. The movements were most
clearly seen at magnifications of six or twelve diameters. More detail
was seen under 25X and 50X, but the images were blurred, especially
at 5 OX.
A second method of observing the mouthparts was by turning the
body and eyepiece of the binocular microscope upside down and focussing
on the larva above the microscope. The best image was obtained by this
method which was used most often. Fluorescent light from above and
tocussed light from below were used separately and in combination.
A third and most convenient method of observing the movements
of larval mouthparts was through a metallurgical binocular microscope
with the stage above the objective lens. In this method it was possible
to have the light coming only from above.
Particles of activated charcoal or methyl red were placed in the
containers with the larvae to show the directions of the currents set up
by the mouthparts.
Observation of the Mouthparts in Action
The operation of the lateral labral brushes was studied by direct
observation of living larvae and by manipulation of prepared material.
The mechanism of action in each type of mouthpart is described sep-
arately below.
Browsers
In this group contraction of the labral muscles exerts tension
on the brush sclerite which in turn pulls the tessellated membrane
and the cross bar s by their hooks. This causes the hairs of the brush to
move ventr o-medially. The hairs spring back outwardly through the
elasticity of the tessellated membrane. The inward and outward move-
ment of the hairs is thus caused by the differential elasticity of the
tessellated membrane and the cross bars. The bases of the hairs are
connected with the cross bars, and forjc on either side of them (fig. 5).
The bifurcations are short, and their ends terminate in the tessellated
membrane belowthe cross bars. The stretch of the tessellated membrane
allows the part of the hair which is attached to the rigid cross bar to move
more than the tips of the fork, so that the hair pivots about this attach-
ment to the cross bar, and its tip swings ventr o-medially. Relaxation
of the labral muscles allows the hair s to return to their original positions
through the elasticity of the tessellated membrane.
The angle through which a hair swings should increase with its
distance from the brush sclerite since it is separated from this by a
greater length of the elastic membrane. This would have the effect of
bunching the hairs together in the median position and allowing them to
fan out in the lateral position, which was repeatedly observed to happen.
The main feeding current, produced by the lateral labral brushes,
is directed toward the epipharynx and the mouth by the median labral
brushes. When creating a current the lateral labral brushes vibrate from
66
Mouthparts of Mosquito Larvae
TABLE 4 - Mean frequency and duration of movements of the lateral
labral brushes of larvae over one minute periods at 24 to
27°C.
postero-medially to anter o-later ally. The brushes of A. aegypti may
vibrate for as long as 2. 5 min without stopping. Then they usually stop
for 5-10 sec before resuming. The more usual timing is vibration for
50 sec, stop for 5-10 sec. and then vibration again. In Culiseta inornata
and in the browsing Aedes species the duration of movement is shorter.
Frequency and duration of movements for C. inornata and A. aegypti and
indicated in table 4. Table 5 shows activity of individual 4th instar
C. inornata larvae, each of which was observed for 3 0 minutes. During
each 30 minute period the activity of the whole body and of the mouthparts,
was observed, and the percentage of time spent in each observable activity
was calculated.
Feeding and locomotory activities of approximately 50 C. inornata
larvae were observed individually for various periods of time throughout
the period of the study, and many more were observed in group behaviour.
Much similarity was noticed in the pattern of behaviour of the various
individuals, and almost any larva could be chosen to represent the common
sequence of activities. The following is a summary of the activities of a
4thinstar C. inornata actively browsing larva (no. 6 in table 5), observed
for 20 min at a magnification of 25X. The container was filled with
pond water.
Pucat
67
During the fir st minute the larva was stationary; it was suspended
from the water- surface with labral brushes extended. For the next 10
sec the labral brushes, the maxillae, and the mandibles created a
current, then a 10 sec period of rest followed with the mouthparts
retracted and the whole body still. During the first 5 min period
such a succession of currents in which all the mouthparts participated
was produced four times, and each time the labral brushes moved about
15 times. The mandibles and the maxillary brushes also moved approx-
imately as many times as the labral brushes.
TABLE 5 - Percentage of time spent by 4th instar C. inornata larvae in
different activity states over a 30 min period.
The larva browsed on a filamentous piece of plant for ten
seconds. The piece of plant was enclosed by the labral brushes and
the mandibular teeth struck it. Then the larva moved to a chickweed
leaf and browsed on its edges for 18 sec. The median hairs (type 3)
of the lateral labral brushes held the edges of the tissue while the more
lateral hairs (type 3) of the brushes produced a current which moved the
larva forward along the leaf. The mandibular teeth struck the tissue.
Then the tissue was left and further currents were produced by the mouth-
parts. Pieces of debris passed into the current which was produced con-
tinuously for approximately 20 sec. Mandibular teeth chopped off small
pieces of decayed material, some of which went into the mouth and the
remainder moved out with the current. Again a piece of plant tissue was
browsed upon, and was then propelled posteriorly. When one end of the
plant was at the submentum the aulaeum clung to it for a few seconds, but
with the subsequent current the tis sue was for ced posteriorly and towards
the bottom of the container.
During the next ten minutes continuous movements of the mouth-
parts occurred 15 times, each time the duration of the current was
approximately 15 sec.
The amount of brush movement and body movement varies among
larvae of different ages and different species. Fourth instar larvae are
more sluggish than younger ones, and 4th instar Culis eta inornata and
Aedes fitchii larvae are more sluggish than the corresponding instars
68
Mouthparts of Mosquito Larvae
of A. aegypti (table 4) . Shannon (1931) and Christopher s (I960) also noticed
that A. aegypti larvae moved considerably faster than the larvae of most
other species of mosquitoes. Fourth instar A. aegypti larvae can ingest
char coal particles faster than 4th instar C. inornata larvae. When activated
char coal was placed in a container with 3 A. aegypti larvae and in another
container with 3 C. inornata larvae (all 4th instar), the guts of the former
were filled in from 90 to 105 min, whereas the guts of the latter species
were only filled after 3.5 hr. Larvae of all the species observed moved
faster and more frequently when they were stimulated to activity by other
organisms ( Daphnia , Cyclops etc. ).
When the brushes are not rhythmically beating to create a feeding
or a locomotory cur rent they remain extended and separated into rows of
four or five layer s (fig. 7) , or they are retracted (fig. 7). Particles which
have been brought close to the mouth by the current continue streaming
towards the mouth through the spaces between the rows of hairs, or
if the brushes are retracted the particles come to rest on the maxillary
brushes. If these are extended the particles stream into the mouth and
some settle on the hairs of the pharynx, the mandibles , the maxillae and
the prementum. The separation of the labral brush hairs into several
rows (fig. 7) is possible because of the basal structure of the brush. Each
row of hairs can be moved about the axis of its cross bar. Several rows
can move in one direction together, and thus water can flow through the
spaces between these groups of hair s. It also seems that the water currents
can for ce the labral brushes to close. The muscles that insert on the
tormal apodemes (fig. 3) extend the brushes by contraction. Relaxation
of the labral muscles allows the hair s to return to their original positions
through the elasticity of the tessellated membrane. This can be demon-
strated in preserved specimens. The contraction of these muscles and
of the epipharyngeal muscles was observed in living larvae of a filter
feeder, Culiseta morsitans .
The feeding currents of Aedes and Culiseta browsers are fast and
can carry large as well as small particles. Objects about one third the
size of a larval head can be circulated in the stream (fig. 8), the current
and the particles reach as far posteriorly as the fourth and fifth abdominal
segments and extend about the same distance in front of the larva. Such
circulation of particles can be observed when the larva is suspended in
water and also when it lies on its dorsal side in an observation cell.
When a larva feeds just above a loose sediment (fig. 8) or browses
its way forward through debris in a container, the particles that do not
enter the mouth fall to the bottom of the container or cling to the brushes;
they do not return to the feeding current. The feeding current is effective
only in front of the larva, and it is slowed down behind the larval head.
The water flows ventrally rather than posteriorly below the body of the
larva. When the larva leaves the browsing area many particles remain
on its labral and maxillary brushes , since what does not fall to the bottom
of the container sticks to the brushes . Some filtering is done by the labral
brushes , especially by the median serrated ends of the lateral labral brush
hairs, quite large particles are found clinging to them. Since particles only
slightly smaller than these have been found in the pharynx and in the
intestine, and since most food seems to come into the mouth via the
Pucat
69
labral brush current, it seems reasonable to assume that the particles
which pas sed into the mouth and eventually into the gut were filtered out
by the brushes. The serrated brush hairs are useful in browsing, for as
they move along surfaces they detach particles from them, many of which
are consumed.
With its labral brushes a browsing larva can attach itself to a grass
stem, to the side of a container, or to a body of a pupa or another larva.
While the labral brushes cling to surfaces the maxillary brushes produce
a current. The browser 's maxillary brushes can create currents that are
as strong as those of the labral brushes. This was observed in fourth
instar larvae of the following species:4ec/es cataphyllaT)yB.r ,A. sticticus[NLeigen),
A. communis (De Geer), A. fitchii ,A. punctor ,A. riparius,A. canadensis , Culiseta inornata and
C. impatiens . The larvae can also browse on parts of their own body,
especially on the posterior regions of the abdomen. This was observed
particularly in containers where Aedes and Culiseta larvae were crowded.
Many times larvae, especially C. inornata and Aedes canadensis ., were seen
browsing on the tips of their own abdomens and creating currents
at the same time. They were in loop-like positions and moved in
circulating paths of the water surface. This was particularly notice-
able in the laboratory with the larvae of A. canadensis ; on one occasion in
June I960, 20 to 30 larvae turned in this manner for several minutes,
individual larvae turning for as long as five to six minutes. Christophers
(I960) states that larvae browse on parts of their own bodies, especially
on the posterior parts, when they are starving. My observations
agree, for in situations where this behaviour took place little food was
present.
Interfacial feeding (Renn 1941, and fig. 8) is a common method of
feeding in the Anopheles filter feeding larvae. Third and fourth instar
C. inornata , A. aegypti , A. fitchii , A. punctor , and A. riparius larvae also brow-
sed at the water surface without browsing on their siphons at the
same tim'e. In this second type of filter feeding only the head of the
larva was at the water surface and the rest of the body remained under
water .
In most browsing activities all or most of the mouthparts are
employed. When an object such as a long thin piece of decaying grass
comes into the feeding current, it comes in contact with the mouthparts
as follows :firstly, the serrated lateral labral brush hairs (median type 3)
hold a part of it, and push the remainder posteriorly; second, it slides
over the central labral brush; third, it passes between the epipharyngeal
bristles; fourth, the mandibular denticles strike it as it passes by, and
if a small piece of it is thus torn off it may go posteriorly with the current,
it may be drawn into the mouth, or it may settle on the prementum; fifth,
it passes between the maxillary brushes; sixth and finally, the particle
of grass touches the submentum and the aulaeum. During this process
some of the median labral brush hairs hold the particle while the remaining
hair s of the brush produce currents.
Sometimes parts of the lateral labral brushes move only slightly
(median type 3 hairs) whereas the hairs of their most posterior (types
2 and 3) move more actively. More commonly, all the hairs on the
brushes move simultaneously when producing a current. When a larva
comes to a stop after moving about in a container , it will gradually extend
70
surface view of currents
Fig. 8. Movements of labral brush currents of browsing larvae; (a) interfacial sur-
face feeding current, (b) current produced under the water surface, (c) current used
to stir up debris from the bottom.
Pucat
71
or contract the brushes.
Most of the observations on the coordination of moving mouthparts
were on Aedes aegypti and A. fitchii larvae which had been slowed down in a
2 0 to 3 0% solution of methocel of 400 centipoises. The larvae were watched
in white porcelain spot plates with their ventral sides turned up. The
following combinations of mouthparts were observed in action: 1. The
lateral labral brushes moved in their usual antero-posterior oblique
direction, and the long apical setae of both maxillae moved backwards
and forwards at the same time. 2. The lateral brushes moved in their
usual direction while the setae of one maxilla remained stationary,
directed posteriorly, and the setae of the other maxilla continued their
antero - posterior motion. The epipharyngeal bars also moved. 3.
The lateral labral brushes were motionless . At the same time either one
or both maxillary brushes waved and thus kept the current in motion.
4. The lateral labral brushes came to rest on the epipharynx and at the
same time the other mouthparts moved in one of the following ways: one
or both maxillae moved in the transverse plane; one or both mandibles
moved in the transverse plane, striking against the hypopharynx; or,
one mandible and one maxilla on the same or the opposite side moved.
The same type of combination of mouthpart movements was observed in
the larvae of the following species: Aedes cataphylla, A. excrucians , Aedes fitchii,
A. hexodontusYAy&v , A. punctor , A. riparius , A. sticticus , A. vexans , andCuZtseia inornata.
Aedes aegypti larvae also browsed on poplar and elm leaves in
the laboratory. For two weeks ten larvae were given no other food but
dried leaves of Ulmus sp. and no mortality occurred. At the end of the
two week period all the larvae had pupated. The larvae of this species
are also rear ed on leaves of a species of poplar in South Africa (Hocking,
personal communication).
Browsing larvae of Aedes and Culiseta were observed in deep water
pools (approximately 1. 5 to 2. 5 ft. deep) and in shallow pools (four to 12
in. deep) . In shallow pools with clear water it was possible to see larvae
browsing on submerged rotting leaves and other objects for as long as
three minutes without coming to the surface for air. When the larvae
came to the surface they sometimes remained there for one to five
minutes and they either moved slowly or continued in one position before
submerging again. Sometimes the wind disturbed the surface of the pool
and some of the larvae that were at the surface moved with the wind, while
others swam against this. In situations of this type, however, most
larvae went to the edge of the pool, where a stable resting position was
found.
Several observations of larval activity were made at a pool 1.-5-
2 ft. deep, and the courses of larval movements were recorded.
Aedes excrucians and Culiseta inornata larvae were able to remain in a stationary
position at the surface for from a few seconds to four minutes. During
this time they probably produced currents with their mouthparts as did
the larvae of these and other browser s when observed in a glass container
in the laboratory. The approximate mean distance that any one larva
covered in four minutes was between four and five feet. In a larger pool
some larvae coveredmore space than this before submerging. The larvae
wentunder either of their own accord, or wehn they came in contact with
72
Mouthparts of Mosquito Larvae
another animal in the water such as a snail, a water beetle, a crustacean,
ora dead insect floating on the surface. When larvae submerged without
coming in contact with something fir st, after detaching their siphons from
the surface film, they were pulled downward by the currents of their
mouthparts .
In pools populated with browsing larvae and located in areas which
were partly shaded, the shaded areas were muchmore crowded with rest-
ing larvae, although the sunny areas were used for moving about and brow-
sing by a few larvae. This behaviour can be interpreted as orthokinetic
(Fraenkel and Gunn I960).
Filter Feeders
Three filter feeding species are at present known in Alberta,
representing three genera; Anopheles earlei Vargas, Culiseta morsitans , and
Culex territans . All of these species are uncommon, hence it was not pos sible
to study the morphology and function of the mouthparts of their larvae
in much detail.
Feeding larvae of Anopheles earlei were obs erved in the laboratory ,
but most were r ear ed into adults and none was preserved for morpholog-
ical study. Anopheles earlei larvae are small and are usually found in deep
water, hence it is difficult to observe the action of their mouthparts in
their natural habitat. In the laboratory all were usually at the water
surface It was common to see some larvae resting with parts of the
abdomen or thorax or both against the side of the container, while others
moved in circular paths around the container. Often two or three larvae
moved side by side in one direction, while one or more other larvae
moved in an opposite direction. Sometimes two larvae, moving towards
each other, collided and then both moved together in the direction initially
travelled by one or the other. It is not known what determined the final
direction of movement;perhaps the larva producing the stronger current
overrode the other. Anopheles larvae turned their heads through 180 degrees
so that the ventral side of the head was at the surface for period of
25 to 30 sec at a time as compared with approximately 10 sec periods
in the normal position.
According to Clements (1963), The area of surface that can be
cleared of particles by an Anopheles larva in a given time varies with the
size of the larva, density of particles, and the rate of filtration, which
is affected by temperature. The effect of these factors on larval move-
ment was not considered in this study.
The movements and feeding behaviour of Culex territans and Culiseta
morsitans are similar. The two species are found in the same type of
habitat, and their mouthparts are similar in form. Since the labral
brushes in these species are longer than in the Aedes or Culiseta browsers,
the currents they create cover a larger area than do those of the browsers.
Also these filter feeders extend their labral brushes mainly laterally,
whereas the browsers extend theirs antero-laterally. Several Culiseta
morsitans larvae were observed in a glass jar in the laboratory.
They moved rapidly by means of the labral currents and fed at the same
time; the pharyngeal movements could be seen through the head cuticle.
Sometimes minute crustaceans were brought to the mouth with the current.
Pucat
73
but they were not ingested. The food of the larvae consisted mainly of
mos s particles which floated in the pool water, and settled on the bottom
of the jar. The particles on the bottom of the jar were agitated by browsing
Aedes cinereus or Culiseta inornata larvae often collected with C. morsitans .
Occasionally the C. morsitans moved their labral brushes just above the
sedimented particles on the floor of the jar in the same manner as the
browsing species. Sometimes two, three, or more of these filter feed-
ing larvae rested in one location close to each other, clinging to the water
surface film with their siphons, and moving their labral brushes. Most
frequently the larvae stayed in such a position between two and three
minutes before being disturbed by a moving larva or crustacean. When
disturbed, the larvae either submerged, or moved horizontally on the
water surface to another location. The first course was followed by about
two thirds of the larvae . After submerging, each larva went in a different
direction and stayed under the water surface for 10 to 15 sec. Upon coming
to the surface the larvae either resumed their stationary positions for two
to three minutes or until disturbed, or they moved horizontally, propelling
themselves by the feeding current. In submerging when disturbed and in
returning to the water surface the characteristic wriggling motion of the
abdomen was used.
In the laboratory C. morsitans larvae assembled in the shady rather
than the sunny part of a container. This observation is in agreement with
that of Hocking (1953) on Aedes communis.
Predators
Three species of predatory larvae, Chaoborus americanus , Mochlonyx
velutinus , Euc or ethra underwoo di Underwood were collected near Flatbush,
Alberta during the summers I960 and 1961. C. americanusldLrva.e were ob-
served feeding on the larvae of several species of Aedes in the laboratory.
The feeding behaviour of Chaoborus species has been studied in detail by
Montchadsky (1945) and by Schremmer (1950). Both authors discussed
the modification of the larval mouthparts for their predator y function. The
mandibles in the larvae of this genus are the important movable mouth-
parts. The maxillae are fused with the ventral part of the cranium, and
prementum is reduced to a wedge-like plate. The mandibles do not have
a primarily crushing function, but their sharp strongly chitinized teeth
have a holding and pushing function (Schremmer 1950). These larvae
also use their prehensile antennae for catching prey. They ingest their
prey whole. The main features of the mouthparts of Chaoborus americanus are
indicated in table 3, The posterior occipital parts of the head capsule of
Chaoborus larvae are connected to the subgena by membranes (Cook 1956) ;
this permits the mouth opening to become enlarged whenever necessary.
In Mochlonyx velutinus larva the ventral part of the head is sclero-
tized, but a large mouth opening is present, as the head capsule is wider
than in Chaoborus . Cannibalism was observed among the M. velutinus larvae
in a jar in the laboratory. The raptorial function of the mandibles and
antennae was observed when the larvae caught their prey tail first. Then
the prey seemed to be held by the maxillae while the mandibles continued
striking it and pushing it further into the mouth. In the specimens that
I observed the process of ingestion lasted approximately two hours.
74
Mouthparts of Mosquito Larvae
Digestion may take as long as three hours (Montchadsky 1945). Some-
times a feeding larva lost its prey, even if this was half ingested, if it
was disturbed by other organisms. James (1957) observed that M. velutinus
larvae are occasional predators on other mosquito larvae. I observed
M. velutinus feeding on larvae of various Aedes species. A similar habit
was observed in M. culiciformis De Geer by Montchadsky (1953) and
Montchadsky and Berzina (1959). Cannibalism was also observed in
Cryophila lapponica Mart, by Montchadsky ( 1953) .
Discussion
The larvae that I studied in this investigation can be classified
as filter feeders, browsers , and predators . There are more similarities in
the structure and in the function of the various mouthparts of filter feeders
and browser s than between either one of these types and the predators.
The labral brushes of filter feeders and browsers are used for
bringing food to the larvae by means of currents which they produce by
vibrations . By means of thes e vibrations the larvae also move through the
water. The labral brushes of the predatory larvae are reduced to a few
bristles and do not produce currents.
The epiphar ynx of the browsing and filter feeding larvae is believed
to have the function of covering the mouth opening (Schremmer 1950).
This was not observed in the larvae that were studied in this project.
The epipharyngeal hairs were erected by the muscle which moves the
epipharyngeal bar, and when these hairs came in contact with the labral
brush hairs, food from the brush hairs was transferred to them. The
epipharyngeal hairs were in turn scraped by the mandibular hairs, and
this food was thus passed towards the mouth opening. If the food did not
go into the mouth, as often happened, particles of it remained on the pre-
mentum and on the hairs of the lacinia.
Mandibles of the browsing larvae were observed in actions of biting
while the larvae browsed. Those of predators were seen grasping and
pushing the captured prey into the mouth. The mandibles of the filter
feeders and the browsers move in a dor so- ventral plane, but those of
the predator s move in an oblique plane which is nearly parallel to the long-
itudinal axis of the body.
LARVAL FOOD AND MOUTHPARTS
As a final step in investigating the function of the mouthparts the
nature of the food of the functional groups of larvae and the relationship
between the size of the food particles and the dimensions of the mouth-
parts were studied.
Procedures
The gut contents of several species of Aedes, Culiseta , and Culex lar-
vae were examined and measured. Most of these contents were dissected
out and mounted in glycerine jelly, a suitable preservative for plant mater-
Pucat
75
ials (Sass 1940). Particles of activated charcoal were made available to sev-
eral A. fitchii and C. inornata larvae, and ingested as well as uningested
particles were measured.
The following measurements were taken of the larvae of available
species, including Anopheles, Chaoborus , and Mochlonyx: head width (between
the bases of the antennae), head length (between the median labral brush
and the occiput) , mean length of the right lateral labral brush (at the center
of the brush), width of the right lateral labral brush (width at the
base of the brush) , and the width. of the epipharyngeal constriction (space
between the most posterior, longest teeth on the transverse bars of the
epi pharynx) .
An examination was also made of the material suspended in the
water of a larval habitat. Ten litres of water was taken from a pool near
Edmonton where C. inornata larvae were collected in September, 1961.
This water was passed through a series of sieves. Material that did not
go through the fir st sieve was examined, and a rough estimate of its comp-
osition was made. These fractions of material were then dried at 100°C
to constant weight;they were ashed in a muffle oven at 575°C; the ash was
weighed and the percentage loss was calculated.
Results
Table 6 contains a summary of the sizes of particles that were found
in the guts and in the environment of the larvae of Aedes fitchii, Culiseta inornata
and Culex territans . Particles that were identified from the guts of 4th
instar larvae of these species are listed in table 7. From this
table it is seen that the gut contents in the three species were similar.
The relationship between the structure of some mouthparts and
the feeding habits of larvae is shown in fig. 9. The points on the graph
were derived in the following manner: (1) for the position on the abscissa
the mean length of the right labral brush was multiplied by its mean width
to give the area swept by the brush. This product was divided by the pro-
duct of the head lenght and the head width, to relate this to the size of the
larva. (2) for the position on the ordinate the width of the epipharyngeal
constriction was divided by the head width to represent the maximum
relative size of particles which could be swallowed. Each point represents
the mean value for a species. A separation between filter feeders and
browsers is shown on this graph.
In fig. 9 the intermediates fall closer to the browsers than to the
filter feeder s. Typical filter feeder s may be tentatively defined as larvae
in which both the ratio of the epipharyngeal constriction to the head width
and the relative area sv/ept by the lateral labral brushes exceed 0. 14. In
browsers and intermediates both of these ratios are less than 0.14. In
typical predator s the first ratio is more than 0.14, but the second is less.
On the basis of morphology representatives of all types of feeders fall
within the range of browsers.
From table 7 it is seen that the gut contents were similar in the
three species, Aedes fitchii , Culiseta inornata , and Culex territans . The guts of
a few Chaoborus americanus larvae that were examined were filled with mus -
cle tissue; some of this was from other mosquito larvae.
76
Mouthparts of Mosquito Larvae
TABLE 6 - Size ranges of particles in the guts and in the environments
of 4th instar mosquito larvae. Percentage by number.
measurements
TABLE 7 - Or ganic particles in larval habitat and gut contents of 4th in-
star larvae of Culiseta, Aedes , and Culex scarce, xx common,
xxx abundant, xxxx very abundant.
Pucat
77
F ungi
of water from a pool occupied by C. inornata larvae.
The following items were retained from the water taken from a
pool were C. inornata larvae were collected by a sieve with 45 meshes
per inch: 60% Cyclops sp. and other copepods, alive;20% decaying animal
and plant material including mosquito eggs, egg cases beetle abdomens,
and mosquito wings; 20% algae, mainly Spirogyra sp. The dry and ash
weights and percentage of organic matter in the material held by sieves
of finer mesh are given in table 8.
Discussion
In examining the gut contents of browsing, filter feeding, and pre-
datory larvae it was found that the browsing Aedes and Culiseta larvae
ingestitems of similar types and sizes. The approximate proportions of
Epipharyngeal constriction width , I
head width x 100 | territans
78
<
o
S
•2
<0
a
SPO
o ,
-d cn
u
^ 1
>n „
ft ^
8
W (M
M (1) rS
.2 £ ®
T3 h o,
<U £ ft
S £ v
<+H /N
d o
<D +3 "
<D -rl CO
> 3 a>
s. cr +;
xi
a>
Xl .
(ti i>N
m ®
a «
<u O
_to 0)
Pucat
79
the numbers of food particles of the different sizes in the guts of Aedes
fitchii and Culiseta inornata la.rvae are: less than 15jJL, one - sixth , 15 - 22 p,
one -third, 22 -4 Op, one -third, 40-7 lp, one - sixth of the measured particles.
In the larvae of A. fitchii 58% of the charcoal particles ingested
were found to be less than 20 microns with the largest percentage (31.7) in
the 10- 15p range ;only 6. 3% of the natural food particles fell in the 10-15 p
range with the largest percentage (27.2) occurring in the 15-20p range.
A similar relationship was found in C. inornata (table 6).
Some plant and animal particles were folded before entering the
mouth of the larva. Also when the larvae browsed on plant surfaces they bit
pieces off plants, scraped surfaces, and thus obtained soft particles of
various sizes and shapes. Many plant particles eaten were long, narrow,
and flat, so they were easily car ried into the mouth by the feeding current.
However, when activated charcoal was placed in the water, the larvae
ingested the small particles that were brought to the mouth with the feeding
current, but did not take in the large ones which rapidly settled on the
bottom of the container. Charcoal particles are denser than natural food
and the browser s 1 currents cannot stir up particle s larger than 15 microns .
The particles are filtered by the labral brushes; large hard particles are
rejected, whereas soft food is actively taken in. Occasionally I stirred
the charcoal in the containers. Sometimes the larvae browsed on the
bottom of the container, but long, flat particles were difficult to obtain.
Thus mostly small charcoal particles were scraped into the mouths.
Since the charcoal particles did not remain in water suspension
very long, they were not fed to the filter feeders. Pond food from the
guts of these larvae was measured (table 6). Also measured were the
spaces between the groups of labral brush hair s through which the feeding
current passes. The size of these spaces was found to be similar to that
of the particles in the guts. Thus filter feeding is possible among these
larvae, for if the ingested particles were larger than the spaces between
the hairs, they would not be trapped in the brushes, but would remain on
the surface of the brush. On the other hand, very small particles would
pas s through the brush with the water current without becoming entangled
in it.
Also, most of the food particles found in the guts of filter feeders
were of the same order of size as the charcoal particles ingested by the
browser s , and smaller than the food particles of browsers that fed in the
field. The epipharyngeal constriction width in filter feeders is greater
than in browser s, therefor e it should permit larger particles to pass to-
wards the mouth. However, the mandibular teeth of filter feeders are
weakly sclerotized and cannot crush or "squeeze'1 large particles in the
feeding current. Thus large soft particles by— pass the mouth openings of
filter feeder s , whereas they are pushed into the mouths by the mandibular
teeth of browsers. But the wide epipharyngeal space of filter feeders
allows the passage of more particles in a given time.
According to Bates (1949), Shipitzina ml935 found that 4th instar
larvae of Anopheles messeae Fall, were able to swallow sand particles from
68-165p wide. The mouth openings of this species must be larger than
those of the culicine larvae I studied. The size range of food particles
found in the guts of three English species of Simulium larvae was found
80
Mouthparts of Mosquito Larvae
to be 1 . 7- 15. 1[± by Williams et al (1961), the size of the mouth openings
of these larvae was not given.
McGregor (1963), working with larvae of Opifex fuscus found that
first instar larvae did not develop serrations on their labral brushes if
they were fed on minute particles of dehydrated blood serum. Serrations
did develop when they were given fish food ranging in particle size 0. 1-
0. 6 mm. Similar experiments with larvae of other feeding types should
be revealing.
The browsing larvae whose guts I examined fed on plant particles
and on microscopic animals, whereas the filter feeder Culex territans had
fed only on plant particles (table 7). Also, all the types of particles that
were present in the pool water where the C. inornata larvae were collected
were found in the intestines of these larvae.lt can be said then that these
larvae do not discriminate in the type of food they ingest. Other workers
have come to similar conclusions: Coggeshall in 1926 as reported by
Bates (1949), Howland (1930) , and Jones (I960) who worked with anopheline
larvae, and Becker (1958) who worked with larvae of Culicoides circumscriptus
Kieff. These authors have found algae, diatoms, and other plant part-
icles in the guts of Anopheles and Culicoides larvae. Rempel (1936) found
similar food materials in larvae of Chironomus hyperboreus Staeg. (= C. rempelii
Thienemann, Rempel 1962). Other culicine larvae also ingested
similar food (Horsfall 1955). Bekker (1938b) found living Euglena in the
gut of Anopheles maculipennis
The Aedes and Culiseta browser s show similarities in both function
and morphology. The range of the ratio of epipharyngeal constriction to
head width is from 9 to 12 . 7 , and the ratio of the area swept by the lateral
labral brushes to the head size ranges from 4 to 11.8 (fig. 9). Two
Anopheles filter feeder s , one Aedes intermediate, and two Culex intermed-
iates also fall within these ranges. The second ratio is even higher for
another intermediate feeder; it is 13 for Aedes cinereus
Of the species I examined, two species of Culex and one of Culiseta
are filter feeders in function and morphology. The species of Chaoborus
and Mochlonyx are predators both functionally and morphologically. The
remainder of the species represented in fig. 9 range between these two
types either in function, morphology, or both. Thus the Aedes and Culex
species labelled as intermediates obtain their food by filtering, but the
structure of their mouthparts is intermediate between the typical filter
feeders and typical browsers. The Anopheles species are also filter
feeders. Their mouthparts fit the general description for filter feeders
but the sizes of the mouthparts measured, upon which the div -
ision in fig. 9 is based, are proportionately smaller than the sizes of
corresponding mouthparts of Culex and Culiseta filter feeders.
While this method of separating larvae of Aedes , Culex , and
Culiseta , into filter feeder s and browsers is satisfactory and can be used
to categorize the predatory species of Chaoborus and Mochlonyx’, it is not
reliable for Anopheles . The filter feeding larvae are consider ed to be the
most primitive and the predatory larvae the most advanced (Montchadsky
1937, Surtees 1959). Thus the largest number of the species studied are
in a transitional stage of evolution.
Pucat
81
GENERAL CONCLUSIONS AND DISCUSSION
According to the functions of the mouthparts three types of mos-
quito larvae can be recognized: filter feeders, represented in Alberta by
Anopheles earlei , Culex territans , and Culiseta morsitans ; browsers, including
most of the Aedes and Culiseta species ; and predatory, represented by
species of Chaohorus , Mochlonyx , and Eucorethra . The Culex and Culiseta filter
feeders are characterized by labral brushes consisting of long, thin,
simple hairs, and by lightly sclerotized mandibles. The Anopheles larvae
have thin, simple lateral labral hairs which are shorter than those of
Culex and Culiseta, slightly sclerotized mandibles , and large rectangular
maxillae with short thin hair s . The browsers have shorter labral brushes
with some serrated, thick hairs, rectangular maxillae with shorter,
thicker brushes, and moderately sclerotized mandibles. The predators
bear only a few setae on their reduced labral areas and on their much more
fused maxillae, and they have heavily sclerotized mandibles.
Among the browsers morphological intermediates occur. Aedes
canadensis and A. cinereus , have short labral brushes with simple hairs,
browser-like mandibles, and maxillae similar to those of the filter
feeders, Culiseta impatiens and Culiseta inornata , have typical browsing labral
brushes and mandibles , but have maxillary structures closely related to
those of predators.
Not much variation was observed in the structures of the labral
brushes, mandibles, or maxillae among most of the browsing Aedes larvae
studied. However, specific differences were found in the numbers of
serrations on the sclerotized plates of the prementum, and on the triang-
ular submentum. These characters may be taxonomically useful.
By staining with Mallory 's triple stain it was found that the cuticle
of the mouthparts varies in hardness and flexibility. The median hairs
of the lateral labral brushes of the browsers have hard basal and central
parts, and flexible parts just above the bases, and at the tips.
An examination of larval food revealed that the browsing and filter
feeding larvae are not discriminatory in the type of food they accept, but
there are limits in the size of particles they can ingest.
ACKNOWLEDGEMENTS
Professor B. Hocking supervised this study, and Dr s. J. Sharplin,
W, G. Evans, and G. E. Ball, Department of Entomology, gave valuable
suggestions and encouragement. Dr . L. Kennedy, Department of Botany,
identified intestinal contents of mosquito larvae. The thesis was revised
at the University of Saskatchewan, Regina Campus, and members of the
faculty criticized it.
Dr. J. McLintock, Canada Department of Agriculture, Ottawa,
and Mr. J. A. Shemanchuk, Department of Agriculture, Lethbridge,
provided me with eggs of Culiseta inornata. Professor J. G. Rempel,
Department of Biology, University of Saskatchewan, Saskatoon, supplied
82
Mouthparts of Mosquito Larvae
larvae of Culex species . Canada Defence Research Board provided finan-
cial assistance. I express my thanks to all.
REFERENCES
Anthon, H. 1943. Zum Kopfbau der primitivsten bisher bekannten
Dipter enlarvae: Olbiogaster sp. (Rhyphidae). Ein Beitrag zur
Phylogenie der nematoceren Dipter en. Ent. Medd. 23 : 303-320.
Applegarth, A. G. 1939. The larvae of Apterobittacus apterus MacLachlan
(Mecoptera : Panorpidae). Microent. 4 : 109-120.
Bates, M. 1949. The natural history of mosquitoes. The Macmillan
Co. , New York. 379 p.
Becker, P. 1958. The behaviour of larvae of Culicoides circumscriptus Kieff.
(Dipt. , Ceratopogonidae) towards light stimuli as influenced by
feeding, with observations on the feeding habits. Bull. ent. Res.
49 : 785-802.
Bekker, E. 1938a. The mouth apparatus of the Anopheles larva and its
movements in feeding upon or ganisms of the surface film of water.
(In Russian with English summary). Zool. Zh. 17 : 427-440.
Bekker, E. 1938b. On the mechanism of feeding in larvae of Anopheles.
Zool. Zh. 17 : 741-762.
Butt, F. H. 1957. The role of the pr emandibular or intercalary seg-
ment in head segmentation of insect and other arthropods.
Trans. Amer. ent. Soc. 83 : 1-30.
Carpenter, S. J. and W. J. LaCasse. 1955. Mosquitoes of North
America (north of Mexico). University of California Press,
Berkeley, 360 p.
Chaudonner et, L. 1951. Reflexions sur les pr emandibules des larves
de culicides (Dipteres nematocdr es) . Bull. Soc. zool. Fr.
76 : 399-404.
Christophers, S. R. I960. Aedes aegypti (L. ) The yellow fever mos-
quito. The University Press, Cambridge. 738 p.
Clements, A. N. 1963. The physiology of mosquitoes . Pergamon
Press, New York. 393 p.
Cook, E. F. 1944a. The morphology and musculature of the labrum and
clypeus of insects. Microent. 9 : 1-35.
Cook, E. F. 1944b. The morphology of the larval head of certain
Culicidae (Dipterae). Microent. 9 : 37-84.
Cook, E. F. 1949. The evolution of the head in the larvae of the Diptera.
Microent. 14 : 1-57.
Cook, E. F. 1956. The nearctic Chaoborinae (Diptera : Culicidae).
University of Minnesota Agricultural Experiment Station, Tech-
nical Bulletin 218. 102 p.
Das, G. M. 1937. The musculature of the mouth-parts of insect larvae.
Quart. J. micr. Sci. 80 : 39-80.
DuPorte, E. M. 1953. The protocephalon. A critique of recent inter-
pretations. Can. Ent. 85 : 41-55.
Pucat
83
DuPorte, E. M. 1958. Are there preoral structures in insects. Xth
int. Congr. Ent. Montreal 1956. 1 : 479-485.
DuPorte, E. M. 1962. The anterior tentorial arms in insects and their
significance in interpreting the morphology of the cranium of the
Cicadas. Can. J. Zool. 40 : 137-144.
Edwards, F. W. , H. Oldroyd, and J. Smart. 1939. British bloodsucking
flies. British Museum (Natural History), London. 156 p.
Farnsworth, Mar jorie W. 1947. The morphology and musculature of the
larval head of Anopheles quadrimaculatus Say. Ann. ent. Soc. Amer.
40 : 137-151.
Ferris, G. F. 1947. The contradictions of the insect head. Microent.
12 : 59-64.
Ferris, G. F. 1948. The principles of comparative morphology.
Microent. 13 : 49-56.
Ferris, G. F. and B. E. Rees. 1939. The morphology of Panorpa
nuptialis Ge rstaecker (Mecoptera : Panorpidae) . Microent. 4: 79-
108.
Foote, R. H. 1952. The larval morphology and chaetotaxy of the Culex
subgenus Melanoconion (Diptera, Culicidae). Ann. ent. Soc . Amer .
45 : 445-472.
Fraenkel, G. S. and D. L. Gunn. I960. The orientation of animals.
Dover Publications, New York. 376 p.
Giles, G. M. 1902. A handbook of the gnats or mosquitoes. 2nd edition,
John Bale, Sons and Danielsson, London. 530 p.
Gouin, F. J. 1959. Morphology of the larval head of some Chironomidae
(Diptera, Nematocera). Smithsonian misc. Coll. 137 : 175-201,
Hennig, W. 1948, 1950, 1952. Die Larvenformen der Dipteren. Pt. 1 .
(1948), Pt. 2(1950), Pt. 3(1952). Akademie Verlag, Berlin.
Henry, Laura M. 1947. The nervous system and the segmentation of
the'head in the Annulata. Microent. 12 : 65-82.
Hinton, E. H. 1958. The phylogeny of the Panorpoid orders. A. Rev.
Ent. 3 : 181-206.
Hocking, B. 1953. Notes on the activities of Aedes larvae. Mosquito
News. 13 : 77-81.
Hocking, B. , W. R. Richards, and C« R. Twinn. 1950. Observations
on the bionomics of some northern mosquito species (Culicidae :
Diptera). Can. J. Res. D. 28 : 58-80.
Hooke, R. 1665. Micrographia. Reprint 1961. Hafner Publishing Co.
New York, N. Y. 247 p.
Horsfall, W. R. 1955. Mosquitoes. Their bionomics and relation to
disease. The Ronald Press, New York. 723 p.
Howard, L. O. , H. G. Dyar, and F, Knab. 1912. The mosquitoes of
north and central America and the West Indies. Carnegie Insti-
tution of Washington, Washington, D. C. Publication No. 159.
520 p.
Howland, L. J. 1930. The nutrition of mosquito larvae, with special
reference to their algal food. Bull. ent. Res. 21 : 431-439.
Imms,A, D. 1907. Notes on the structure and behaviour of the larva of
Anopheles maculipennis Meigen (preliminary note) . Proc. Camb. phil.
Soc. 14 : 292-295.
84
Mouthparts of Mosquito Larvae
Imms,A.D. 1908. On the larval and pupal stages of Anopheles maculipennis
Meigen. Part II. The larva (continued) . Parasitology 1 : 103-133.
Imms, A. D. 1944. On the constitution of the maxillae and labium in
Mecoptera and Diptera. Quart. J. micr. Sci. 85 : 73-96.
James, H. G. 1957. Mochlonyx velutinus (Ruthe) (Diptera : Culicidae), an
occasional predator of mosquito larvae. Can. Ent. 89 : 470-
480.
Johannsen, O. A. 1903. Aquatic nematocerous Diptera. New York
State Museum Bull. 68 : 74-448.
Jones, J. C. I960. The anatomy and rhythmical activities of the aliment-
ary canal of Anopheles larvae. Ann. ent. Soc . Amer . 53 : 459-474.
McGregor, D. 1963. Mouth-brush dimorphism in larvae of Opifex fuscus
Hutton (Diptera : Culicidae). Bull. ent. Res. 54 : 325-327.
McLintock, J. 1952. Continuous laboratory rearing of Culiseta inornata
(Will. )(Diptera : Culicidae). Mosquito News. 12 : 195-201.
Marshall, J. F. 1938. The British Mosquitoes. British Museum (Nat-
ural History), London. 341 p.
Matheson, R. 1944. Handbook of the mosquitoes of north America.
Comstock Publishing Company, Ithaca, N. Y. 314 p.
Meinert, F. 1886. De eucephale myggelarver. K. danske vidensk
Selsk. (nat. math. ) 6 : 373-493.
Menees, J. H. 1958a. The maxilla and labium of the larva of Anopheles
quadrimaculatus Say. Bull. Brooklyn ent. Soc. 53 : 25-38.
Menees, J. H. 1958b. The facial areas, labrum, epipharynx, hypo-
pharynx and mandibles of the larva of Anopheles quadrimaculatus Say.
Bull. Brooklyn ent. Soc. 53 : 124-140.
Miall, L. C. 1895. The natural history of aquatic insects. Macmillan
and Co. , New York. 395 p.
Miall, L. C. and A.R. Hammond. 1891. The development of the head
of the imago of Chironomus . Trans. Linn. Soc. Lond. (Zool. ) 2.
5 : 265-279.
Mitchell, E, 1 906. Mouth parts of mosquito larvae as indicative of habits.
Psyche, Camb. , Mass. 13 : 11-21.
Montchadsky, A. S. 1937. The evolution of the larvae and its relation
with the evolution of the adults in the family Culicidae. Bull.
Acad. Sci. U. R. S. S. (math. nat. sci. ) 4 : 1329-1351. (In
Russian with German summary).
Montchadsky, A, S. 1945. About the mechanism of digestion in Chaoborus
larvae (Diptera, Culicidae) . Zool. Zh. 24 : 90-97. (In Russian).
Montchadsky, A. S. 1953. On the seasonal dimporphism of the larvae of
Mochlonyx culiciformis De Geer (Diptera, Culicidae, Chaoborinae) .
Zool. Inst. Acad. Sci. U. R. S. S. 13 : 363-372. (In Russian).
Montchadsky, A. S. 1959. Intraspecific relations of predatory larvae
of the subfamily Chaoborinae (Diptera, Culicidae). Report I.
Rev. Ent. U. R. S. S. 38 : 505-516. (In Russian with English
summary. )
Montchadsky, A. S. and A. N. Berzina. 1959. Intraspecific relations
of predatory larvae of the subfamily Chaoborinae (Diptera,
Culicidae). Report 2. Zool. Zh. 38 : 1554-1558. (In Russian
with English summary. )
Pucat
85
Muirhead-Thomson, R. C. 1951. Mosquito behaviour in relation to
malaria transmis sion and control in the tropics. Edward Arnold,
London. 2 19 p.
Nuttall, G. H. F. and A. E. Shipley. 1901. Studies in relation to malaria
II. J. Hyg. , Camb. 1 : 45-77.
Petersen, Janet. I960 Morphological and behavioral studies on the
resting position in Lepidoptera. Ph. D. Thesis. University of
London. 222 p.
Peterson, A. 1951. Larvae of insects. Part II. Columbus, Ohio.
416 p.
Peterson Hoyt, C. 1952. The evolution of the mouth parts of adult
Diptera. Microent. 17 : 61-125.
Potter, Edith. 1938. The internal anatomy of the larvae of Panorpa and
Boreus (Mecoptera). Proc. R. ent. Soc. Lond. (A) 13 : 117-130.
Pucat, A. M. 1962. The morphology and function of the mouthparts of
mosquito larvae. M.Sc. Thesis. University of Alberta. 146 p.
Puri, I. M. 1925. On the life-history and structure of the early stages
of Simuliidae (Diptera : Nematocera, Part I and II). Parasit-
ology. 17 : 295-334.
Raschke, E. W. 1887. Die Larve von Culex hemorosus . Ein Beitrag
zur Kenntnis der Insekten - Anatomie und Histologie. Arch.
Naturgesch. 1 : 133-163.
Reaumur, M. 1738. Memoires pour servir a 1'histoire des insectes.
De L'Imprinterie Royale. Paris. Tome 4. 636 p.
Rempel, J. G. 19*36. The life - history and morphology of Chironomus
hyperboreus . J. biol. Bd Can. 2 : 209-220.
Rempel, J. G. 1950. A guide to the mosquito larvae of western Canada.
Can. J. Res. D 28 : 207-248.
Rempel, J. G. , J. M. Naylor , K. Rothfels, and B. Ottonen. 1962. The
sex chromosome constitution of Chironomid inter sexes parasitized
by Nematodes. Can. J. Gen. Cytol. 4 : 92-96.
R enn, C. E. 1941. The food economy of Anopheles quadrimaculatus and A.
crucians larvae, p. 329-342. In. A symposium on hydro-biology.
University of Wisconsin Press, Madison.
Salem, H. H. 1931. Some observations on the structure of the mouth
parts and fore-intestine of the fourth stage larva of Aedes
(Stegomyia) fas data (Fab.). Ann. trop. Med. Parasit. ,25: 393-419.
Sass. J. E. 1940. Elements of botanical microtechnique. McGraw-
Hill, London. 222 p.
Schremmer, F. 1949. Morphologische und funktionelle Analyse der
Mundteile und des Pharynx der Larve von Anopheles maculipennis
Meig. Ost. zool. Z. 2 : 173-222.
Schremmer, F. 1950. Zur Morphologie und funktionellen Anatomie des
Larvenkopfes von Chaoborus (Corethra auct. ) obscuripes v. d. Wulp
(Diptera, Chaoboridae) . Ost zool. Z. 2 : 471-516.
Shalaby, A. M. 1956. On the mouthparts of the larval instars of
Anopheles quadrimaculatus (Say). Bull. Soc. ent. Egypte 40 : 137-174.
Shalaby, A. M. 1957a. On the mouthparts of the larval instars of
Aedes aegypti (L, ) (Diptera ; Culicidae). Bull. Soc. ent. Egypte
41 : 145-177.
86
Mouthparts of Mosquito Larvae
Shalaby, A. M. 1957b. On the mouth parts of the larval instars of
Culex quinquefasciatus (Say). Bull. Soc. ent. Egypte 41 : 269-298.
Shalaby, A. M. 1957c. The mouth parts of the larval instars of
Psorophora ciliata (Fabricius). Bull. Soc. ent. Egypte 41 : 429-455.
Shalaby, A. M. 1957d. The homology of the labium-hypopharynx complex
of the larva of Culex moles tus Forsk. (Diptera : Culicidae) . Ann. ent.
Soc. Amer. 50 : 265-270.
Shalaby, A. M. 1 958 . Morphological adaptations in the maxillae of three
species of mosquito larvae. Bull. Soc . ent. Egypte 42:439-448.
Shannon, R.C. 1931. The environment and behaviour of some Brazilian
mosquitoes. Proc. ent. Soc. Wash. 33 : 1 - 27.
Snodgrass, R. E. 1947. The insect cranium and the "epicranial suture. "
Smithsonian misc . Coll. 107, 52 p.
Snodgrass, R. E. 1958. The insect tentorium and its antecedents.
Xth int. Congr. Ent. , Montreal 1956. 1 : 487.
Snodgras s , R . E. 1 959 . The anatomical life of the mosquito . Smithsonian
misc. Coll. 39, 87 p.
Surtees, G. 1959. Functional and morphological adaptations of the larval
mouthparts in the subfamily Culicinae (Diptera) with a review of
some related studies by Montchadsky . Pr oc . R . ent. Soc. Lond. ,
(A) 34 : 7- 16.
Theobald, F. V. 1901- 1910. A monograph of the Culicidae or mosquitoes.
British Museum (Natural History), London. Vols. 1-5.
Thompson, M. T. 1905. Alimentary canal of the mosquito. Proc.
Boston Soc. nat. Hist. 32 : 145-202.
Trembley, H. L. ' 1955. Mosquito culture techniques and experimental
procedures. American Mosquito Control Association, Bulletin
No. 3. 73 p.
Twinn, C. R. , B. Hocking, W. C. McDuffie, and H. F. Cross. 1948.
A preliminary account of the biting flies at Churchill, Manitoba.
Can. J. Res. D. 26 : 334-357.
Vimmer, A. 1906. Comparative study of the mouth apparatus of the
larvae of Pachyrhyna and Tipula . Casopis Ceske Ceskoslovenske
Spolecnosti Entomologicke . 3:37-43 |in Czech), 43-49 (in German) .
Wesche, W. 1910. On the larval and pupal stages of West African
Culicidae. Bull. ent. Res. 1 : 7-53.
Wesenber g- Lund, C. 1920-21. Contributions to the biology of the
Danish Culicidae. K. danske vidensk. Selsk. (nat. math. )
8 : 1-200.
Wheeler, W. M. 1893. A contribution to insect embryology. J.
Morph. 8 : 1-160.
Williams, T. R. , R. Connolly, H. B. N. Hynes, and W. E. Kershaw.
1961. Size of particles ingested by Simulium larvae. Nature,
London. 189 : 78.
Quaestiones
entomologicae
A periodical record of entomological investigations,
published at the Department of Entomology, Uni-
versity of Alberta, Edmonton, Canada.
VOLUME 1
NUMBER 3
JULY 1965
QUAESTIONES ENTOMOLOGICAE
A periodical record of entomological investigations , published at
the Department of Entomology, University of Alberta, Edmonton, Alberta.
Volume 1 Number 3 20 July 1965
A REVISION OF THE NORTH AMERICAN SPECIES OF THE
CIC1NDELA MARIT1MA GROUP WITH A STUDY OF HYBRIDIZATION BETWEEN
CICINDELA DUODECIMGUTTATA AND OREGON A
RICHARD F REIT AG
Department of Entomology Quaestiones entomologicae
University of Alberta 1:87 — 170. 1965
The North American species of the Cicindela maritima group are: C. duodecimguttata
Dejean; C. oregona LeConte; C. depressula Casey; C. repanda Dejean; C. hirticollis Say; C-
limbata Say; C. columbica Hatch; C. bellissima Leng; and C. theatina Rotger. The male genital-
ia of these species are described. The group is diagnosed and two keys are given , one based on
non-genitalic characters, and the other based on features of the male genitalia. For each of the
species duodecimguttata, oregona, and depressula the following are presented: synonymy, analy-
sis of geographic variation in size, coloration, color pattern of elytra, and distribution. Descrip-
tions of the subspecies of oregona and depressula are given. Hybridization between the species
duodecimguttata and oregona is examined quantitatively by means of the hybrid index method and
the data are presented in the form of histograms. The zone of hybridization lies on the eastern
slopes of the Rocky Mountain System from Colorado to the Northwest Territories, and is about
50 miles wide in Alberta but is nearly 1,000 miles wide in northern Canada. Variation of external
characters and shape of the median lobe of the male is greater in the region of inter gradation
than it is within the range of the pure parental forms. Temporal variation occurs in hybrid popu-
lations. Phylogenetic and zoogeographic relationships are postulated to explain the structural
similarities and distribution patterns of the North American species of the maritima group.
INTRODUCTION
The species of the North American tiger beetles of the genus
Cicindela are for the most part fairly well understood taxonomically, arid
it is possible to identify most adult specimens as a result of publications
by Leng (1902), Horn (1915), Cazier (1936, 1948, 1954, I960), and Wallis
(1961). In addition Hamilton (1925) has described many larvae. With the
descriptive phase in this condition attention must now be directed to tax-
onomic studies at the species level. By such studies phylogenetic relation-
ships of species and delimitations of species groups within the genus can
be worked out.
This study began with the discovery of hybridization between
Cicindela duodecimguttata and Cicindela oregona. Variation of phenotypic characters
of hybrids and pure parental forms was analysed. As a result it was
found that the latest definition of oregona (Wallis, 1961) was composite and
included the definition of depressula . This led to a study of dep ressula , The
male genitalia of all the North American species of the maritima group were
then examined. The features of the internal sac proved to be diagnostic
of this group, while shapes of the median lobes were found to be specif-
ically distinct.
88
Cicindela maritima Group
MATERIALS AND METHODS
Materials
Structural features and their variation were studied in approx-
imately 12, 000 specimens of C. duodecimguttata, C. oregona and C. depressula. The
data from these obs ervations are analysed quantitatively by the following
techniques. Descriptions and specimens of larvae of seven species were
also examined but only as additional material for deriving a phylogenetic
scheme for the North American species of the maritima group.
Methods
External Characters
Distribution of hairs on the head, elytral pattern, and color are
very important in the classification of the North American species in
the maritima group.
Hair s may cover the head and frons either very dens ely or sparse-
ly,'or they occur in the form of a patch near the inner edge of each eye.
The post genae may be glabrous or hairy.
Elytral pattern is composed of the following white markings:
humeral lunule, marginal band, middle band, and apical lunule. The
occurrence, shape, and expanse of these markings are used in showing
interspecific and intraspecific variation (figs 11-16).
The six basic color s that occur in duodecimguttata, oregona and depressula
wer e matched with the color standards of Ridgway (1912). They are listed
below with their corresponding Ridgway names in parentheses: black
(Black), brown (Mummy Brown), copper (Liver Brown), purple (Dull
Violet Black), blue (Dusky Green Blue), and green (Danube Green).
These colors may be dull, opalescent, or metallic.
Male Genitalia
North American species of the C. maritima group can be grouped to-
gether and individually identified by characteristics of the male genitalia.
For study of the genitalia the male beetle was relaxed in boiling water.
Then by inserting a pair of fine forceps into the end of the abdomen the
genitalia were grasped and pulled out. These structures were cleared
in a hot 10% solution of potassium hydroxide for about 1 0 minutes anc
then washed in water. The genital structures were finally stored in
glycerine in a corked microvial and pinned beneath the beetle from which
they were extracted. Drawings of the male armature were made with
the aid of a Wild camera lucida and stereoscopic microscope at a mag-
nification of X 62. 5.
Measurements
Intraspecific variation of size and expanse of color pattern was
Fr eitag
89
analysed by means of measurements. A calibrated eyepiece in a Zeiss
stereoscopic microscope at a magnification of X 10 was used. Relative
size is indicated by length of elytra as measured from the apex of the
scutellum to the tip of the elytral spine and width of the elytra as measur ed
from the midline to the margin at the widest point. Expanse of elytral
pattern is represented by the transverse diameter of the apical dot.
Measurements were made of specimens in each large population sample.
Statistical Methods
Linear measurements were treated statistically and tabulated.
The range, mean, standard deviation, coefficient of variation, and stan-
dard error, were determined in each analysis. The Chi- squar e test
was used in evaluating data of annual and seasonal changes in hybrid
indices of population samples. The method was also employed to assess
randomness of mating in the hybrid zone (Simpson et al. I960, p. 306).
Hybrid Index
Variation in a hybrid population can be analysed using a hybrid
index. This method was developed by Anderson (1949) for plant hybrids
and has been successfully used for study of variation in avian hybrid
populations (Sibley 1950, 1954, Sibley & Short 1959a, 1959b, 1964,
Short 1963). The hybrid index method makes it possible to describe
variation in quantitative terms. The hybrid index is constructed as
follows. Characters that separate the parental forms are determined.
Each character of one parent is scored 0. Those of the other pure parent
are each given a high value and intermediate characters are ascribed
values that fall on the scale between the parental scores. The hybrid index
for each specimen is the sum of its individual character values.
The method was used to analyse variation in duodecimguttata - oregona
hybrid population samples. Results are presented in figures 35 to 44.
In addition, geographic variation in elytral pattern of duodecimguttata is
analysed by this technique (table 3). Because of its broader application
the hybrid index is here also referred to as the "compound char acter
index". In figs 35 to 43 average index changes per mile are indicated
between localities of population samples. These roughly illustrate rel-
ative spatial rates of index change, but they do not imply linear trends.
Pictorialized Scatter Diagrams
Pictorialized scatter diagrams, the alignment of symbols in a
two-dimensional field or graph, are used to describe several character
relationships. The positions of symbols are determined by the calibrated
axes sjach of which is a quantitative expression of a single character or
ratio of two characters. More characters can be considered at a time
by adding appropriate tokens to the specimen symbol. This method is
used to illustrate data on intraspecific relationships of oregona populations
(figs 22 - 31). For a more complete description of this method see
Anderson (1949).
90
Cicindela maritima Group
Pie-graph Maps
This method illustrates geographic relationships of populations
with differ ent varying color characters. Pies plotted on a map represent
geographic positions from which population samples were collected.
Numbers of specimens of particular color combinations are indicated
opposite the appropriate pie sections (figs 19, 20).
Field Methods
Because tiger beetles are rather difficult to see in their natural
environment a technique was necessary to facilitate field observations.
At Nordegg, Alberta, in the hybrid zone, specimens were first caught
with an insect net. The sex and hybrid index value of each individual
was translated into a code that was painted on the elytra with a small
brush. The individuals were then released and observed through field
glasses.
Adult tiger beetles, collected for museum material, were killed
in a jar containing potassium cyanide, and pinned the day they were caught.
Larvae were either trapped at the tops of their burrows by rapid-
ly driving a shovel beneath them, or dug out. They were boiled in water
five minutes to preserve their color and then placed in 7 0% alcohol.
Criteria for Species and Subspecies
Two similar forms are regarded as distinct species if their
geographical ranges overlap and if they show no inter gradation in at
least one character (color excluded). If a narrow stabilized hybrid belt
is developed in the region of contact of two forms that are largely
allopatric they are treated as distinct species (Mayr 1963). Two allo-
patric forms that differ only in coloration are judged to be conspecific.
Allopatric forms of a single species are regarded as being subspecifically
distinct if 75 per cent of the individuals of one form are different from
97 per cent of the individuals of the other (Mayr et al 1953). However,
if a clinal series of intermediate populations is intercalated between
two distinct populations that are widely allopatric subspecific names are
not given.
There are two opposing views regarding the subspecies concept.
Wilson and Brown (1953) believe the subspecies concept to be subjective
and arbitrary in the light of discordant variation, variation inmicro-
geographic races , and the artificiality of quantitative methods of defining
the formal lower limits of the subspecies as well as other reasons.
Inger (1961) however, argues that Wilson and Brown tend to magnify
exceptional cases, and that the subspecies concept despite its limitations
has proved useful. It is this latter view that is followed in this text.
Many more opinions regarding the subspecies concept are expressed in
issues of Systematic Zoology (1953-1960).
Fr eitag
94
MORPHOLOGY OF THE MALE GENITALIA OF THE NORTH AMERICAN
SPECIES OF THE CICINDELA MARITIMA GROUP
Introduction
Several papers dealing wholly or in part with the male genitalia
of American tiger beetles have been published (Horn 1930, Papp 1952,
Rivalier 1954, and Rumpp 1957). Horn observes that for some races
of Omus californicus Eschscholtz shape and size of the penis is characteristic.
Papp presents a detailed study of the internal sac from which relationships
of the Nearctic and Palearctic tiger beetles are deduced, while Rivalier
classifies the entir eCicindela fauna of the Americas. Rumpp uses male
genitalia in s eparating more clearly the species Cicindela praetextata LeConte
and Cicindela californica Menetries.
Male genitalia of three or more specimens of eachNorth American
species of the maritima group were examined. The male armature con-
sists of three relatively large sclerites: a median structure called the
median lobe, penis, or aedoeagus (see fig. 1); and a pair of lateral
parameres, one on each side of the median lobe and articulating with
its base. Inverted in the median lobe is the membranous internal sac
that is everted from the dorso-apical portion of the median lobe during
copulation.
Within each species the shape of the median lobe is quite uniform.
There is, however, a considerable amount of interspecific variation,
particularly in form of the apex, that proves useful in distinguishing
species of the maritima group from one another.
The internal sac comprises many folds, dark areas bearing micro-
trichia or aculeae, and sclerites. These fields of aculeae, and sclerites
can be homologized within the species of the maritima group. Number s are
assigned to sclerites and letters are assigned to fields. This system
of nomenclature follows that of Ball (MS) and is not synonymous with that
of Papp.
When retracted in the abdomen the median lobe lies lengthwise,
parallel to the longitudinal plane of the body of the beetle, and the open-
ing of the internal sac is dorsal. When the median lobe is in a copulatory
position the aperture of the internal sac is ventral. For each species
drawings of the retracted median lobe and the inverted internal sac viewed
from the dorsal and left sides are presented. In addition the shape of
the apex of the median lobe is given separately for each species . Includ-
ed also is a table of the various sclerites of the internal sac for each North
American member of the maritima group.
Descriptions
Male Genitalia of Cicindela duo decim guttata Dejean
The median lobe is of average breadth and length (figs la, b, c,
and 10). Two broad, lateral flanges that occupy the apical region of
the median lobe converge apically to form a marked tip which curves
92
Cicindela maritima Group
ventrally.
The inverted internal sac, in which three fields of aculeae are
distinct is clearly visible. These darkened areas are labelled a, b, and
c. Field a , which has a pebbly appearance is apical in the infolded position
but it is basal when the internal sac is everted. Field 6, ventral in pos-
ition, is a finger -like projection of the membrane from which only the
apical end is separate in the form pf a tiny sac. When the internal sac
is inverted b hangs inward whith its free end nearest the apex of the med-
ian lobe. Conversely, when the internal sac is everted, this field pro-
jects outward its blind end remaining oriented toward the apex of the
median lobe. Field c , is three quarters circular, anterior in the infolded
position, and appear s to serve as the apical limits of the everted internal
sac.
Six sclerotized areas are present. Most noticeable is the flagellum
(4) which is a slender strip, pointed apically, and widened and hooked at
the base. A short rectangular sclerite (3) is present to the left of the
base of the flagellum, and sclerite 5, a cuplike structure, lies posterior
to 3. In fig. 1 sclerite 5 lies to the right of the median line beneath
several membranous folds, but it is more clearly shown in fig. 2.
To the left of and lying in part over the basal portion of the flagellum is
sclerite 1, a quadrate plate. Sclerite 6 is oriented to the right of
the sagittal plane. It is twisted basally and resembles an aculeate
field apically. Sclerite 2 is an elongate curved strip with its apical end
near that of the flagellum. A very small triangular sclerite is present
between 2 and 6, but it is not numbered since it may be a disconnected
piece of one of these two sclerites. This sclerite is illustrated with
sclerite 2 in fig. 10.
Male Genitalia of Cicindela oregona LeConte
Shape of the median lobe (figs 2a, b, c, and 10) is quite different
from that of the preceding species. Though the apical, lateral flanges
are about as long as those of duodecimguttata they are rather narrow. The
apex is not markedly curved ventrally. Fields and sclerites, excepting
sclerite 5 which is relatively large, look like their counterparts of
duodecimguttata.
Male Genitalia of Cicindela depressula Casey
Unlike the penes of duodecimguttata and oregona the apical portion of
the median lobe in this species is characterized by wide lateral flanges
that form a blunt tip (figs 3a, b, c, and 10). The flanges are continuous
and not separated from one another by the raised apical section of the
chamber containing the internal sac as they are in the two preceding
species. The median lobe is short and broad. Field a is composed of
several elongate folds that together form a rough area. Field b is com-
paratively light, and c is c-shaped. Sclerites 1, 2, 3, 4, and 6 are
respectively of the shapes and in the positions described for those of
duodecimguttata. At the basal end of sclerite 2 the small triangular sclerite
is elcn gate, Sclerite 5, large and lightly sclerotized, is visible when the
internal sac is everted or inverted.
93
Fig. 1. Median lobe and inverted internal sac of Cicindela duo dec im guttata, la, dorsal aspect; lb, left lateral
aspect; lc, apical portion, dorsal aspect. Numbers = sclerite nos. Lower case letters = fields.
Fig. 2. Median lobe and inverted internal sac of Cicindela ore^ona. 2a, dorsal aspect; 2b, left lateral aspect;
2c, apical portion, dorsal aspect.
94
Cicindela maritima Group
Male Genitalia of Cicindela repanda Dejean
The portion of the median lobe that contains the internal sac is
more apically confined than those of the three discussed species (figs 4a,
b, c, and 10). The lateral flanges are narrow and widely separated
dorsally by the chamber of the internal sac. Field a is small and lightly
aculeate, while b is a distinct area pebbly in appearance. Field c is of
the common shape . Sclerite 1 is large and triangular, while 2, 3, 4, and
6 are like those of duo de cim guttata. There is no small sclerite near the basal
end of 2. Sclerite 5, large and heavily sclerotized, is quite distinct.
Male Genitalia of Cicindela limbata Say
The median lobe is relatively short and narrow (figs 5a, b, c, and
10). The two broad, later al flanges are evenly rounded and together con-
verge to a marked but non-protruding apex. Fields a and b are strongly
aculeate; and c is clearly indicated in the form of one third of a circle.
Variation is evident in the shape of sclerite 1 which is generally smaller
in size than those of the other North American species of themaritima group.
Sclerites 2, 3, 4, and 6 ar e of the common shape, and sclerite 5 is absent.
Genitalia of Cicindela bellissima Leng
The median lobe is of average length but thicker than those of the
preceding species (figs 6a, b, c, and 10). From a dorsal view the lateral
flanges compose a broad apical region that terminates as a sharp pro-
jecting tip. Field a is clearly indicated by its dark compact appearance.
Both a and b have large and scale-like aculeae. Field c is three quarters
of a circle. Sclerites 1, 2, 3, 4, and 6 are each of the common shade
intensity and shape. Sclerite 5 is absent.
Male Genitalia of Cicindela columbica Hatch
The median lobe is relatively long and slender (figs 7a, b, c, and
10), the apical region comprises two fairly wide lateral flanges that are
slightly constricted basally, and an unprojected, rounded apex . Prominent
aculeae occur on field a , which is smaller and less compact than that
of bellissima. Field b is a lightly shaded area, while c is of the common
type. Sclerites 1, 2, 3, 4, and 6 each resemble their counterparts in
other species of the maritima group. The sclerite between 2 and 6 is large
and heavily sclerotized. Sclerite 5 is small and field-like in appearance
which makes it difficult to detect.
Male Genitalia of Cicindela hirticollis Say
The median lobe is elongate and thick (figs 8a, b, c, and 10).
The chamber which contains the internal sac is extended dor so-apically
so that the lateral flanges are widely separated, and the apical portion
of the median lobe is obscured when viewed from the dorsal side.
Field a is composed of several elongate folds; b is sparsely aculeate;
3a
95
Fig. 3. Median lobe and inverted internal sac of Cicindela depressula. 3a, dorsal aspect; 3b, left lateral aspect;
3c, apical portion, dorsal aspect.
Fig. 4. Median lobe and inverted internal sac of Cicindela repanda. 4a, dorsal aspect; 4b, left lateral aspect;
4c, apical portion, dorsal aspect.
96
5c, apical portion, dorsal aspect.
Fig. 6. Median lobe and inverted inter nal sac of C'icintlchi beUissima. 6a, dorsal aspect; 6b, left lateral aspect;
6c, apical portion, dorsal aspect.
F r eitag
97
and c is a semicircle. Sclerite 1 is relatively lightly sclerotized. Scler-
ites 2, 3, and 6 each have the shape characteristic of the maritima group.
The sclerite between 2 and 6 is large and rectangular much like that of
columbica. The apical twirl in s clerite 4 is markedly pronounced. Sclerite
5 is broad and lightly sclerotized.
Male Genitalia of Cicindela theatina Rotger
The median lobe is of average length and breadth (figs 9a, b, c,
and 10) , the apical region somewhat resembles that of duodecimguttata with-
out the protruding tip. A distinct keel is present on the ventral apical
portion of the median lobe. Fields a and b are strongly micr otricheate
and c is normal. Sclerites 1, 2, 3, 4, and 6 are of the general shape
and size. The sclerite between 2 and 6 is large. Sclerite 5 is barely
visible and only occurs as a small roughened area.
Discussion
It is difficult to fix the genitalia of each species in the same
relative position for drawing purposes. Thus sclerites that are of the
same shape but drawn in different positions may appear to be different
from one another. The shapes of sclerites 2, 3, 4 (excepting that of
hirticollis ), and 6 are remarkably constant throughout the North American
maritima group. This uniformity in sclerite shape sets these species apart
as a unit from other Cicindela groups. Some interspecific differences of
the internal sac are evident, however. These are: shape and size of
sclerite 1; presence and size, or absence of the sclerite between 2 and
6; presence and condition, or absence, of sclerite 5. The shape of the
median lobe is diagnostic for each species, particularly the form of the
apical region viewed from the dorsal or ventral sides.
Median lobes and internal sacs of specimens taken in the hybrid
region of duodecimguttata and oregona were examined. It was found that the
form of the apex of the median lobe changed through intermediate shapes
from pure duodecimguttata to pure oregona .
TAXONOMY OF THE NORTH AMERICAN COMPONENTS OF
THE CICINDELA MARITIMA GROUP
Diagnosis of the Group
At the present time there is no generally accepted definition of
the maritima group ( repanda group, in part). Leng (1902) defined the repanda
group on the basis of external character s in which repanda, hirticollis, oregona,
and duodecimguttata , were brought together, but limbata and bellissima were ex-
cluded. Casey (1913) formed the repanda group on the basis of body size
and shape of humeral lunule. The species limbata and bellissima were not
included, and hirticollis was regarded as constituting a closely related
but separate group. Papp (1952) used characters of the internal sac of
98
Fig. 7. Median lobe and inverted internal sac of Cicindela cotumbica. 7a, dorsal aspect; 7b, left lateral aspect;
7c, apical portion, dorsal aspect.
Fig. 8. Median lobe andinverted internal sac of Cicindela hirticolUs. 8a, dor sal aspect; 8b, left lateral
8c, apical portion, dorsal aspect.
aspect;
F r eitag
99
Fig. 9. Median lobe and inverted internal sac of Cicindela theatina. 9a, dorsal aspect; 9b, left lateral aspect;
9c, apical portion, dorsal aspect.
the male genitalia for grouping member s of the repanda complex [sensu Leng)
together with other species, which, I believe, should have been included
in other species groups. The diagnosis of the North American species of
the maritima group made by Rivalier (1954, Group IV) is followed here.
Rivalier united member s of the repanda group with limbata and bellissima , but
columbica and theatina were not mentioned.
The following combination of character s of the internal sac of the
male genitalia is regarded as being diagnostic, and separates the North
American species of the maritima group from other species groups of
Cicindela {see fig. 10): sclerite 1, a quadrate plate lying over the base of
sclerite 4 (flagellum) jsclerite 2, a flat, elongate, curved strip; sclerite
3, short, rectangular , and lying to the left of sclerite 4; sclerite 6, large,
twisted basally, and lightly sclerotized apically; field a apical in the in-
verted position, roughened or dens ely aculeatejfield b , a finger-like pro-
jection, roughened or densely aculeate; field c, semi-circular shape,
terminal in the everted position; chitinous tooth (defined by Papp, 1952)
absent.
100
C.duodecimguttata C.oregona C.depressula C. repanda
5
I
O O
C Jimbefa C.h@iiissima C.coSumbica C .hfrticollis C.theatina
i fa p p Q C?
3
5
1 mm
6
Fig. 10. Sclerites of the internal sac of the North American species of the maritima group,
numbered as in fig. 1.
F r eitag
101
Keys to the North American Species of the maritima Group
Based on Non-genitalic Characters
For species marked *, reference to the median lobe of the male is ad-
visable.
1 Genae glabrous Z
Genae hairy (if hairs of the head or' genae are abraded their
former positions are indicated by tiny setigerous punctures . . 6
Z (1) Posterior tip of humeral lunule (when present) usually with
a slight anteriorly-directed hook; head with frons covered
with numerous hairs C. hirticollis Say
(not treated in detail).
Posterior tip of humeral lunule not hooked; dorsum of head
covered sparsely with hairs;OR glabrous;OR hairs clustered
near the front inner edge of each eye. 3
3 (Z) Marginal band of elytron absent 4
Marginal band of elytron present 5
4 (3) Cluster of one to four hairs near each eye; shoulder of
middle band (usually) smoothly roundedjvertex often with
several very small hairs * C. depressula Casey (p. )
Cluster of eight to 11 hairs near each eye; shoulder of
middle band (usually) raised; vertex usually glabrous .
• • • ,*“* C. oregona LeConte (p. )
5 (3) Frons sparsely hairy; humeral lunule elongate and markedly
oblique; posterior tip of humeral lunule nearly touching
shoulder of middle band C. hellissima Leng
(not treated in detail).
Frons glabrous; humeral lunule short and slightly oblique;
posterior tip of humeral lunule widely separated from shoulder
of middle band C. columbica Hatch
(not treated in detail).
6 (1) Humeral lunule c- shaped or in the form of two dots; elytral
markings narrowly expanded or broken 7
Humeral lunule oblique; elytral markings very broad,
widely connected, or brown pigment of elytra greatly
reduced obliterating basic elytral markings 8
7 (6) Form broader than repanda ; pronotum broad; marginal
band absent or widely separated from humeral lunule
. . . * C. duodecim guttata Dejean (p. )
Pronotum narrow; marginal band connected to humeral
lunule * c. repanda Dejean
(not treated in detail).
8 (6) Elytra predominantly pale, elytral pattern completely
obliterated C. limbata Say
(not treated in detail).
Elytra predominantly dark 9
9 (8) Marginal band broad and widely connected to other elytral
10Z
Cicindela maritima Group
markings; posterior portion of humeral lunule short
...................... C. iheatina R otger
(not treated in detail).
Marginal band short connected only to middle band; posterior
portion of humeral lunule very long C. limbata Say
(not treated in detail).
Based on the Male Genitalia
1 Median lobe with apical lateral flanges narrow 2
Apical lateral flanges of median lobe broad 4
Z (1) Chamber of internal sac extended dor so-apically; sclerite
between 2 and 6 large; sclerite 4 with a pronounced twist
C. hirticollis Say (figs 8, 10)
Chamber of internal sac not extended; sclerite between
2 and 6 small or absent; twist in sclerite 4 normal 3
3 (2) Sclerite 5 large; no sclerite between 2 and 6; part of
median lobe containing internal sac short
C. repanda Dejean (figs 4, 10)
Sclerite 5 normal size; sclerite between 2 and 6 small;
part of median lobe containing internal sac elongate
C. oregona LeConte (figs 2, 10)
4 (1) Apical portion of median lobe with a distinct keel along
median line C. theatina Rotger (figs 9, 10)
Keel absent 5
5 (4) Apex of median lobe produced into a narrow tip 6
Apex of median lobe blunt, not produced 7
6 (5) Sclerite 5 absent; fields a and b densely aculeate
C. bellissima Deng (figs 6, 10)
Sclerite 5 normal size; fields a and b lightly aculeate
C. duodecimguttata Dejean (figs 1, 10)
7 (5) Lateral flanges of median lobe constricted basally; sclerite
between 2 and 6 large C. columbica Hatch (figs 7, 10)
Lateral flanges of median lobe not constricted; sclerite between
2 and 6 normal size 8
8 (7) Sclerite 5 absent; fields a and b densely aculeate
C. limbata Say (figs 5, 10)
Sclerite 5 present; fields a and b lightly aculeate
C. depressula Casey (figs 3, 10)
The Species Cicindela duodecimguttata Dejean
Cicindela duodecimguttata Dejean 1825:73. Type locality - Amerique
septentrionale. Fall 1901:308. Leng 1902:148. Blatchley
1910:34. Casey 1913:28. Horn 1915:374, and 1930:80.
Stainer 1934:247. Papp 1952:515. Rivalier 1954:252.
Lindroth 1955:16. Wallis 1961:20. Graves 1963:498.
Cicindela bucolica Casey 1913:28. Type locality - Aweme,
Manitoba. Wallis 1961:21.
Cicindela hudsonica Casey 1916:29. Type locality - Hudson
Fr eitag
103
Bay Territory. Wallis 1961:21.
Cicindela repanda edmontonensis Carr 1920:218. Type locality -
Edmonton, Alberta. NEW SYNONYMY
Cicindela repanda duodecimguttata , Horn 1930:81 (not Dejean);
Papp 1952:515.
This species is characterized by its dull brown dor sal surface,
and elytral maculations (see figs 11, 13, 15). Specimens of
duodecimguttata are usually distinguishable from specimens of the markedly
similar species repanda. In western areas, where these species are
sympatric, individuals of duodecimguttata have a broad prothorax, dark brown
elytra, and widely interrupted marginal bands, while specimens of repanda
have a narrower prothorax, lighter brown elytra, and the marginal bands
are narrowly, or not interrupted. In eastern Canada the elytral pattern
of duodecimguttata generally is broken but repanda retains full elytral macul-
ations excepting the subspecies novascotiae Vaurie that occurs on the
Canadian Atlantic coast (see Vaurie 1951). Differences in male genitalia
however, are clear and should be used for definitive identification (see
p. 91 & Lindroth 1955: 16-17).
In western Canada, populations of duodecimguttata occur on the edges
of lakes, ponds, rivers, streams, and sloughs wherever the soil is dark
and wet and consists of mixtures of sand and clay, and clay or mud. This
type of habitat is preferred by duodecimguttata on the mainland in eastern
Canada (see Leng 1902, Blatchley 1910, and Graves 1963). Lindroth
(1955) in Newfoundland found duodecimguttata on sand and gravel, as well
as on clay or humus.
Notes on Synonymy
The name bucolica Casey has been given to specimens of duodecimguttata
with full elytral markings. Such specimens are common in the western
prairies. Casey’s hudsonica , the elytral pattern of which is very reduced,
is a variant of duodecimguttata. The name eJmonioneMis Carr was proposedfor
a variant of duodecimguttata that has a narrow elytral pattern. Horn (1930)
treats repanda as a larger race of duodecimguttata and bucolica as alesser form
of repanda . Evidence for this synonymy is pr esented in the following sec-
tion on geographic variation.
Geographic V ariation
This species inhabits a territory that extends from the eastern
front of the R ocky Mountains to the Atlantic s eaboard, and from the North-
west Territories to Alabama (fig. 17). Throughout the range of
duodecimguttata , except for the zone of inter gradation with oregona, two easily
observed characters vary geographically: color of the dorsal surface of
the body and elytral pattern. Variation in both features has been ex-
amined quantitatively. Variation in length of elytra has also been studied.
Average lengths of elytra, from the tip of the scutellum to the end
of the elytral spine, were calculated for males and females from the 20
104
Figs 11 and 12. Anterior views of the heads of C. duodecimguttata and C. oregona. Figs 13 and 14.
Lateral views of the heads of C. duodecimguttata and C. oregona. Figs 15 and 16. Left elytra of
C. duodecimguttata and C. oregona : HL, humeral lunule; MAR, marginal band; MID, middle band;
AL, apical lunule.
Fr eitag
105
localities listed in table 3. Character gradients are irregular and do
not conform with latitudinal, longitudinal, or altitudinal changes. The
mean lengths of elytra range from 6. 57 mm to 7. 60 mm for males, and
7. 11 mm to 8.22 mm for females. Average elytral lengths of males
and females from the island of Newfoundland (Harmon Field) are not
larger than those of corresponding sex'es from coastal localities of the
adjacent mainland such as Bathurst, New Brunswick.
There are no color differences between sexes of duodecimguttata .
Males and females are usually dull brown dor sally, metallic blue-green
ventrally, and the thoracic pleura are coppery. The dorsal surface is
the only area that is subject to color variation. In most regions the dor-
sum is dull brown, but in eastern Canada and United States, color varies.
Specimens from seven localities on or near the Atlantic seaboard
were examined for color of the dor sal surface; the results are listed in
table 1. Brown specimens are most abundant in all of the population
samples, followed in number by brown-green or green and finally blue
individuals .
The most variable maritime population sample is one collected
at Yarmouth, Nova Scotia. The entire color range is represented. Brown
specimens account for 56 per cent of the sample. Green, brown-green,
and blue individuals follow in number in that order. Blue specimens are
absent from the Goose Bay, Labrador, and Harmon Field, Newfoundland
samples both of which are composed mainly of brown members, followed
by brown- green and green. Only brown individuals occur in the Bathurst,
New Brunswick population sample. Except for one green specimen from
Keene Valley, New York, the inland samples are made up entirely of
brown specimens. Green specimens are not uncommon in coastal pop-
ulations of species of the maritima group (see o. oregona and o. novascotiae Vaurie
1951).
The elytral pattern is fully developed in some duodecimguttataindivid-
uals and almost absent in others. The four main components of the
elytral pattern vary independently. There were assigned numerical
values to form a compound character index for analysing variation in
elytral pattern. If all markings are complete a high score is assigned
(maximum value 11), and if the markings are greatly reduced a low value
is assigned (minimum value 0). Markings that range between complete
and reduced are given intermediate value s. The components of the elytral
pattern are illustrated in fig. 15 and their assigned values are given in
table 2. As many eastern duodecimguttata specimens have maculations typ -
ical of oregona, the compound character index used in this section was not
employed in the hybridization section. A compound character index
(hybrid index), based on the elytral pattern, was determined for each
specimen of 20 population samples from different localities. Results
are presented in table 3. The average index value for each sample is
indicated in fig. 17. The samples are arrayed in five transects so that
geographical variation in elytral maculations may be more clearly appre-
ciated. Three transects - A-A1, I3-B', and C-C1 - are run from west to
east, while two - D-D', and E-E1 - are oriented north to south.
106
Cicindela maritima Group
TABLE 1 - Variation in color of dorsal surface of C. duodecimguttata from
seven eastern N. American localities.
F r eitag
107
TABLE 2 - Values assigned to elytral markings of C. duodecimguttata speci-
mens for determination of compound character indices.
An average index reduction from west to east is seen in the A-A'
transect. The mean values 9. 73 and 9. 39 of the samples that represent
Christopher Lake, Saskatchewan and The Pas, Manitoba, respectively,
indicate full elytral markings. The mean index change per mile between
these localities is about 0. 00160. The mean index for the population
sample from Ogoki, Ontario is 6. 22 which is a change of 0. 00488 index
units per mile from that of The Pas , Manitoba. The trend is less marked
between Ogoki, Ontario and Duparquet, Quebec the rate of change being
0. 00320 index units per mile. A change in average index value of 0.00133
occurs between Duparquet, Quebec (4.94) and Bathurst, New Brunswick
(4. 11) , and a change of 0.00127 occurs between Bathurst, New Brunswick
and Harmon Field, Newfoundland (4.54).
Average index values of the six population samples in transect
B-B 1 complement the trend shown in A-A '. Elytral maculations are quite
full in western localities as shown by average index values: Lethbridge,
Alberta, 10.12; Bottineau County, North Dakota, 9.88; and Minnesota,
9.44. The rate of mean index change per mile between Lethbridge,
Alberta and Bottineau County, North Dakota is only 0.00041 and increases
slightly to 0. 00130 between Bottineau County, North Dakota and Minnesota.
However, the average index value for the Cheboygan, Michigan specimens
is markedly less than that of the Minnesota sample, the rate of change
being 0. 00776 index units per mile. This is approximately six times the
rate of change between Bottineau County, North Dakota and Minnesota.
Mean index differences between eastern population samples are slight.
Average values for areas follow: Cheboygan, 5.79; Keene Valley, New
York, 5. 21; and Yarmouth, Nova Scotia, 5. 10, showing a reduction in
the elytral pattern.
The southernmost west- east transect, C-C1, comprises the follow-
ing population samples and average index values: Bennett, Nebraska,
6.03; St. Louis, Missouri, 4,79; and Jeannette, Pennsylvania, 5.77.
The mean index decreases 0. 00355 units per mile between Bennett and
St. Louis while between St. Louis and Jeannette there is a mean increase
of 0. 00158 units per mile.
Transect D-D 1 is oriented north to south near the western limits
of the range of this species. The pattern of the elytra tends to increase
108
Cicindela maritima Group
TABLE 3 - Frequency distribution of compound character index values
of specimens of C. duo decim guttata from 2 0 localities.
north to south in the first part of the transect as follows: Christopher
Lake 9.71, Bottineau County 9.88, and Wolsey, South Dakota 10.00.
The spatial mean change in index units between these population samples
is negligible in contrast to that which occurs between Wolsey and Bennett
(0. 01498). The mean index value at Bennett is 6. 03, and it is 5. 10 for
Texas. The three northern localities therefore have samples with full
elytral markings, while specimens of the two more southerly localities
have reduced maculations , a sharp change occurring between Wolsey and
Bennett.
A clinal north to south fragmentation in maculation of the elytra
is evident in transect E-E 1 . Ogoki, Cheboygan, St. Louis, and Oktibbeha
County, Mississippi have population samples with mean index values of
6. 22, 5. 79, 4. 79, and 3. 40 respectively. The rate of increase in mean
index units per mile is 0. 00101 between Ogoki and Cheboygan, 0. 00170
between Cheboygan and St. Louis, and 0. 00376 between St. Louis and
Oktibbeha County, Mississippi.
The population samples can be separated into two geographic
groups by areas of marked rates of change in mean index values. The
F r eitag
109
greatest differences in average index values are between The Pas,
Manitoba and Ogoki, Ontario; between Minnesota and Cheboygan County,
Michigan;and between Wolsey, South Dakota and Bennett, Nebraska. The
species therefor e may be divided into northwestern populations that have-
complete elytral markings, and southern and eastern groups that exhibit
a more or less interrupted elytral pattern. However , the two aggregates
of populations are not subspecifically distinct. A separation on the basis
of the 75 per cent rule cannot be made becaues of extensive overlap in
range of variation between the two groups of populations.
Breakdown of the elytral pattern has probably occurred independ-
ently in duodecim guttata , repanda, depressula , and oregona and if so this is a good
example of parallel evolution. Perhaps the br eakdown of elytral pattern
in duodecimguttata is the result of a mutation that has spread throughout most
populations except for those in the west.
Canada. ALBERTA: Andrew, 3; Beaver Hill Lake , 1; Bilby, 14; Chin, 8; Cooking Lake, 7; Cypress
Hills, 2; Doussal, 1; Drayton Valley, 4; Edmonton, 95; Falles, 1; Flatbush, 1; Fort Chipewyan, 1; Halfway
House, 3; Jet. Rte. 39 and North Saskatchewan River, 11; Lake Cardinal, 2; Lesser Slave Lake (east end), 1;
Lethbridge (St. Mary's River), 5; Lethbridge (Six-mile Coulee), 310; Louis Bull Reservation, 3; McMurray, 18;
Medicine Hat, 5; mile 7 on Smith- Fitzgerald Road, 1; Redwater, 2; Stirling Lake, 1; Tilley, 1; Tofield, 19;
Vilna, 1; Wabamun, 3. LABRADOR: Goose Bay, 29. MANITOBA: Aweme, 12; Baldur, 1; Beaver Lake,
Riding Mountain, 2; Berens River, 9; Birtle, 1; Brandon (15 miles south), 2; Carberry (5 miles west), 1;
Clear Lake, Riding Mountain, 2; Dauphin, 5; Delta, 3; Douglas, 1; Gladstone, 2; Glenboro, 1; Grunthal, 1;
Hilton (6 miles south), 21; Holland, 2; Husavick, 3; Kelwood, 1; The Pas, 23; Magnuls, 1; Makinock, 5;
Marchand, 1; Max Lake, Turtle Mountain, 10; Melita, 1; mile 360, Rte. 10, 1; Morris, 1; Ninette, 25; Nor gate
(5 miles west), 5; Oak Lake (4 miles west), 1; Red River, 1; Red Rock Lake, 1; Rennie (15 miles east), 1;
Riding Mountain, 7; Rounthwaite, 1; Sandilands, 3; Shilo (5 miles southwest), 4; Shoal Lake, 1; Silver Falls,
2; South Junction, 3; Stonewall, 1; Treesbank, Assiniboine River, 35; Vassar, 1; Victoria Beach, Late
Winnipeg, 8; Wanless, 1; Wasagaming, 1; Watson's Lake, 1; Waugh, 1; Westbourne, 5; Whitemouth, 1.
NEW BRUNSWICK: Apohaqui, 1; Bathurst, 19; Chipman, 6; Penobsquis, 3; St. John, 1; Sackville,3;
Shediac, 10. NEWFOUNDLAND: Bay of Islands, 4; Bay St. George, 13; Codnoy, 11; Deer Lake, Humber
River, 2; Gander, 14; Harmon Field, 59. NORTHWEST TERRITORIES: Fort Smith, 163; Hay River (Great
Slave Lake Shore), 1; Seven Mile Lake (26 miles west Fort Smith) , 6; Resolution, 5. NOVA SCOTIA: Armdale,
3; Baddeck, 2; Barrington Passage, 1; Boisdale, 16; Cape Breton, 35; Cow Bay, 29; Bigby, 1; Great Village,
8; Halifax, 1; Ingramport, 5; Kedgemakooge Lake, 2; Kentville, 11; Lockeport, 2; North Sidney, 9; Port
Maitland, 36; Queens, 2; Truro, 10; Weymouth, 1; Wilmot, 1; Yarmouth, 52. ONTARIO: Agawa Bay, Lake
Superior, 1; Coniston, 3; DeCew Falls, 1; Goderich, 1; Gravenhurst, 1; Hamilton, 2; Hearst (65 miles west),
23; Hudson Bay, 1; Ingolf, 9; James Bay, 1; Kearney, 6; Kenora (14 miles east), 1; Lake of the Woods ,
Harris Hill, 2; Loleo, 7; Minnitaki, 1; Moose Factory, 6; Ogoki, 46; Ojibway, 1; One Sided Lake, 2; Ottawa,
2; Port Arthur, 3; Powasson, 2; Sibley Provincial Park, Middlebrun Bay, Lake Superior, 7; Sudbury, 1;
Toronto, 5; Ventnor, 2; Woodtick Island, 1. QUEBEC: Baie Comeau, 1; Cap Rouge, 1; Cascapedia, 2;
Charlevoix County, 4; Duchesnay, 6; Duparquet, 17; Gaspe, 5; Joliette, 2; Knowlton, 2; Lachute, 2; Ladysmith,
1; Lake Blanch, 12; Lac Opasatika, 1; Mont Joli, 2; Mont Lyall, 2; Montreal, 2; Natashquan, 1; Otter Lake,
12; St. Alexandre, 1; Ste. Anne de Monts, 1; Val Morin, 1. SASKATCHEWAN: Big River, 16; Broadview,
4; Candle Lake, 3; Ceylon, 1; Christopher Lake, 24; Cut Knife, 4; Estevan, 2; Fish Lake, 1; Glaslyn, 3;
Good Spirit, 10; Holbein, 1; Kenosee, 11; Lake Manitou, 1; Neat Frys, 9; Pike Lake, 1; Pike Lake Park, 41;
Regina, 1; Saskatoon, 22; Val Marie, 2; Waskisiu Lake , 10; White Fox, 12; Yorkton, 2.
United States. ALABAMA: Chilton County: Coosa River, 1. Tuscaloosa County, 1. ARKANSAS:
Boone County: Harrison, 2. Lawrence County, 1. Washington County: Fagett, 1. Localities of unknown
counties: Ozark Mountains, 11. COLORADO: Fremont County: Coal Creek, 1. CONNECTICUT: Litch-
field County: Cornwall, 3; Litchfield, 2; Torrington, 3; Twin Lakes, 4; Washington, 2. New Haven County:
Meriden, 7. DELAWARE: New Castle County: Newark, 6. GEORGIA: Fulton County: Atlanta, 8; East
Point, 1. Haber sham County: Cornelia, 1. Localities of unknown counties: Georgia, 1. ILLINOIS: Champaign
County: Champaign, 1; Urbana, 4. Cook County: Chicago, 17; Cook County, 2; Flossmoor, 3; Palos Park, 2;
Summit, 10. Fayette County: Ramsey, 2. Lake County: Antioch, 1; Cedar Lake, 3; Lake County , 1; Waukegan,
3. McHenry County: Algonquin, 3; McHenry, 1; Richmond, 3. McLean County: Bloomington, 1; Normal, 2.
MaconCounty: 1. Marshall County: Taluca, 1. Ogle County: White Pines Forest, 6. Peoria County: Peoria,
4. Perry County: Pyem, 1. Piatt County , Atwood, 1. Putnam County: 2. Randolph County: Chester, 3. Rock
Island County: Moline, 1. Union County: Ware, 3. Will County: New Lenox, 1. Williamson County:
Crab Orchard Lake, 1. Localities of unknown counties: Dune Park, 1. Edgebrook, 16. Funks Grove, 4.
Illinois, 1. Rock, 1. INDIANA: Cass County: 1. Gibson County: 2. Jefferson County: Hanover, 1 .
Knox County: Vincennes, 1. Porter County: Beverley Shores, 1. Posey County: 3. Starke County: North
Judson, 12. Tippecanoe County: Lafayette, 3. Localities of unknown counties: Lake Station, 1. Mineral
Springs, 3. Pine, 1. T.R.P. Indiana, 1. IOWA: Boone County: Boone, 27. Cerro Gordo County: Clear
Lake, 1. Clayton County: Guttenberg, 1. Dickinson County: Lake Okoboji, 3. Henry County: Mount
Pleasant, 1. Howard County: Elma, 1. Johnson County: Iowa City, 4. Lee County: Fort Madison,l.
Story County: Ames, 2. Woodbury County: Sioux City, 1. Localities of unknown counties: Iowa, 1. Silver
Lake, 1. KANSAS: Atchison County: Atchison, 3. Bourbon County: Fort. Scott, 3. Coffey County:
Burlington, 1. Douglas County: 3. Ellis County: 1; Hays, 1. Johnson County: Argentine, 14. Leavenworth
County: Leavenworth, 7. Pottawatomie County: Onaga, 8. Riley County: 1. Saline County: Salina, 1.
Shawnee County: Topeka, 11. Trego County: Wakeenay, 1. Localities of unknown counties: Central Kansas,
1. KENTUCKY: Localities of unknown counties: Kentucky, 3; Kentucky near Cincinnati, Ohio, 1. Maine:
Hancock County: Bar Harbor, 1; Mount Desert, 6; Seal Harbor, 7. Kennebec County: Monmouth, 12.
Piscataguis County: Greenville, 1; Mount Katahdin, 1. York County: Agamenticus, 2. Localities of unknown
110
counties: Bass Harbor, 2; Maine, 1; Pleasant Ridge, 5; Wales, 2. MARYLAND: Allegheny County:
Mount Savage, 4. MASSACHUSETTS: Berkshire County: Benedict Pond, 2; Lenox, 4. Bristol County:
Rehoboth, 2. Middlesex County: Framingham, 19; Sherborn, 9. Norfolk County: Sharon, 1. Plymouth
County: 3. Suffolk County: Cambridge, 1; Medford, 1; Stoneham, 1. Worcester County: Southboro, 2.
MICHIGAN: Alger County: 2; Onota Twp. , 10. Allegan County: Allegan State Forest, 1; Rabbit River, 1.
Alpena County: Alpena, 1; Squaw Bay, 2. Arenac County: White Stone Point, 1. Berrien County: Sawyer
Dunes, 1. Cass County: 1. Charlevoix County: Beaver Island, 1; Thumb Lake, 1. Cheboygan County: 16;
Douglas Lake, 84. Chippewa County: Marquette N. F. , 1; Neebish Island, 4. Clare County: 8-Point Lake, 1.
Delta County: Garden, 1. Genesee County: Davison T.W.P., 1; Flint, 1. Gogebic County: 12; Black
River Park, 4. Huron County: Pte. Aux Barques, 2; Port Austin, 1; Sand Point, 1. Ingham County: 1.
Ionia County, 1. Iosco County: 2; State Game Refuge, 1. Keweenaw County: Copper Harbor, 1; Eagle Harbor,
Lake Superior, 10; Manganese Lake, 11. Lapeer County: Hadley T.W.P.,1. Mackinac County: 1; St. Ignace,
3. Marquette County: Huron Mountains , 12; Marquette, 1. Menominee County: Daggett, 1. Monroe County:
Erie, 1. Montmorency County: 3. Ontonagan County: Gogebic Lake, 15. Otsego County: 2; Vanderbilt, 1.
Schoolcraft County: Germfask, 1. Tuscola County: Bay Park, 2. Wayne County: Detroit, 6. Washtenaw
County: Ann Arbor, 12. Localities of unknown counties: Pen. Ind. , 1; Michigan, 2.' MINNESOTA: Aitkin
County: Aitkin, 1. Anoka County: 2. Becker County: 5. Beltrami County: Waskish, 3. Carlton County:
Moose Lake, 3. Carver County: Lake Waconia, 2. Clearwater County: 14; Gonvick, 1. Hennepin County:
1; Minneapolis, 2. Koochiching County: Rainy Lake, 3. Lac Qui Parle County: 3. Lake County: South
Kawishiwi, 1. Lake of the Woods County: Williams, 1. Le Sueur County: 1. Nicollet County: St. Peter, 5.
Renville County: Bird Island, 1. Roseau County: Roseau, 2. St. Louis County: Duluth, 2. Scott County:
Jordan, 1. Stearns County: Koronis Lake, 6; Rice Lake, 10. Traverse County: 3. Wilkin County: Rothsay,
1. MISSISSIPPI; Oktibbeha County: 12; A & M College, 72. Tippah County: Tiplersville, 2. MISSOURI*
Caldwell County: Hamilton, 13. Carter County: Van Buren, Ozarks Mountains, 2. Greene County: Spring-
field, 2; Willard, 8. Linn County: 1. Pike County: Louisiana, 7. St. Louis County: St. Louis, 19; Valley
Park, 2. MONTANA: Cascade County: Ulm, 1. Chouteau County: Fort Benton, 1. Hill County: Fresno,
2. Roosevelt County: Brocton, 1. Teton County: Chouteau, 3. Toole County: Dunkirk (8 miles south), 5.
NEBRASKA: Dawes County: Wayside, 5. Lancaster County: Bennet, 73; Lincoln, 25; Malcoln, 31. Sarpy
County: Bellevue, 1. NEW HAMPSHIRE: Caroll County: Ellis River Road, Jackson, 1; Wildcat Bank,
Jackson, 1. Cheshire County: Jaffrey, 2. Coos County: Gorham, Peabody River, 4; Jefferson, 25; Israel
River, Jefferson, 10. Grafton County: Twin Mountain, 21. Hillsboro County: Antrim, 1. Sullivan County:
Meriden, 12. Localities of unknown counties: Glen House, White Mountains, 2; Martin Loe'n, White Mountains,
3; New Hampshire, 4; White Mountains , 1. NEW JERSEY: Bergen County: Ramsey, 1. Cape May County:
Ocean City, 3. Essex County: South Orange, 1. Hudson County: Arlington, 6; Snake Hill, 5. Middlesex
County: Jamesburg, 10; Milltown, 2. Passaic County: Paterson, 1. Salem County: Canton, 2. Sussex
County: Lake Hopatcong, 1. Localities of unknown counties: Frieses Mill, 1; Manchester, 1; New Jersey, 2.
NEW YORK: Cortland County: McLean Bogs, 1. Delaware County: Stamford, 3. Erie County: Buffalo, 1.
Ebenezer, 1. Essex County: Ausable Lakes, 1; Elizabeth Town, 1; Heart Lake, 5; Jay Mountains, 1; Keene
Valley, 46; Lake Golden, 1; Mount Whiteface, 3; Wilmington, 4. Franklyn County: Duane, 1. Fulton County:
1. Genesee County: Bergen, 6. Hamilton County: Lake Pleasant, 4; Racquet Lake, 3. Nassau County:
Freeport, 1. New York County: New York City, 5. Niagra County: Lockport, 3. Onondaga County: White
Lake, 5. Orangelo County: Pine Island, 3. Oswego County: Minetto, 2. Queen's County: Far Rockaway,
1. Richmond County: Clover Valley, Staten Island, 1; Hugenot, Staten Island, 1. St. Lawrence County: Cran-
berry Lake, 2. Tompkins County: Ithaca, 16. Warren County: Lake George, 1; Stamford, 3. Localities
of unknown counties: Big Island, 3; Clearwater, 1; Luzerne, 1; Quaker Hill, 2. NORTH CAROLINA: Bun-
combe County: Black Mountains, 1. Guilford County: Jamestown, 1. Mecklenburg County: Charlotte, 2.
Moore County: Manly, 2; Southern Pines, 2. Orange County: Chapel Hill, 2. Localities of unknown counties :
Morrison, 6. NORTH DAKOTA: Benson County: 2. Bottineau County: 73; Lake Metigoshe, Turtle Mountain,
8; Omemee, 3. Burleigh County: 3; Bismark, 1.- Burke County: 4. Cass County: Fargo, 1. Cavalier
County: 4. Divide County: 4. Eddy County: New Rockford, 14; Sheyenne River, 3. Kidder County:
Tappen, 1. Logan County: 4. McHenry County: 14. McLean County: 1. Morton County: 3. Nelson
County: 4; Stump Lake, 3. Pembina County: 1. Ransom County: 3. Rawsey County: 1. Renville County:
8. Richland County: 2. Rolette County: 10; Golden Lake, Turtle Mountain, 2. Sheridan County: 2. Ward
County: 4. Wells County: 1. Williams County: 3. Localities of unknown counties: Jarves Lake, 1;
Mooreton, 3. OHIO: Ashtabula County: Ashtabula, 2; Jefferson, 17. Delaware County: 1. Franklin County:
Columbus, 10. Gallia County: Vinton, 3. Hamilton .County: Cincinnati, 4. Hocking County: 1. Licking
County: Bowling Green Trail, 4; Newark, 1. Localities of unknown counties: Crane Hollow, 1. OKLAHOMA:
Blaine County: Roman Nose State Park, 3. Cleveland County: Norman, 2. Garfield County: Enid, 1.
Johnston County: 1. Kay County: Newkirk, 2. Kingfisher County: Kingfisher, 1. Murray County: Sulphur,
5. Noble County: Perry, 7. Payne County: Lake Carl Blackwell, 1; Stillwater, 1. Localities of unknown
counties: Blue Jacket, 3; Centralia, 1; Wyandotte, 1. PENNSYLVANIA: Allegheny County: 5; Fair Oaks,
2; Westview, 1. Crawford County: Meadville, 1. Cumberland County: Mount Holly, 1. Delaware County:
4; Lansdowne, 4. Forest County: Endeavor, 1. Mercer County: Sharpsville, 1. Montgomery County: 1.
Philadelphia County: Lawndale, 1; Philadelphia, 3. Warren County: Bear Lake, 1. Westmoreland County:
Jeannette, 31. Localities of unknown counties: Castle Rock, 7; Pennsylvania, 1; Springfield, 2; Wali, 1.
RHODE ISLAND: Providence County: Elmwood, 5. Washington County: Misquamicut, 2. SOUTH CAROLINA:
Greenville County: Greenville, 1. Pickens County: Clemson College, 5. Richland County: Columbia, 11.
SOUTH DAKOTA: Beadle County: Wolsey, 17. Fall River County: Hot Springs (5 miles south), 1. Lawrence
County: Deadwood, 1; Savoy, 2. Meade County: Sturgis, 3. Moody County: Colman, 2. Pennington
County: Rapid City, 1. Localities of unknown counties: South Dakota, 1. TENNESSEE: Knox County:
Knoxville, 2. Pickett County: 1. TEXAS: Blanco County: 1. Dallas County: 1; Dallas, 10. Randall
County: Canyon, 1. Washington County: Burton, 2. Localities of unknown counties: Cyp. Mills, 1; Texas,
4. VERMONT: Bennington County: Mount Equinox, 2. Localities of unknown counties: Vermont, 1.
VIRGINIA: Bath County: Warm Springs, 4. Lee County: Pennington Gap, 1. Nansemond County: Suffolk-,-
2. Localities of unknown counties: Black Pond, 1; Virginia, 1. WEST VIRGINIA: Wyoming County: Pine-
ville, 3. WISCONSIN: Bayfield County: Lake Namekagon, 1. Clark County: Wordon Township, 1. Dane
County: 2. Dodge County: Beaver Dam, 10. Douglas County: Superior, 1. Kewaunee County: Kewaunee,
1. Milwaukee County: Milwaukee, 1. Vilas County: Tenderfoot Lake, 1. Walworth County: Allens Grove,
2. Waukesha County: Oconomowoc, 1. Waupaca County: Waupaca, 1. WoodCounty: Cranmoor, 4. Localities
of unknown counties: Walker, 1; Wisconsin, 1. WYOMING: Crook County: Alva (6 miles east), 3; Devil's
Tower, 6. Sheridan County: Sheridan (8 miles north west), 4.
Fr eitag
111
The Species Cicindela oregona LeConfe
Cicindela oregona oregona LeConte 1857:41. Type locality - Oregon Territory
and northern California as far as San Francisco. Fall 1901:308.
Leng 1902:149. Casey 1913:29. Horn 1915:377, and 1930:82.
Varas Arangua 1928:247. Tanner 1929:83. Papp 1952:514, Hatch
1953:41 ( see Hatch 1953 for more references to o. oregona ).
Cazier 1954:242. Rivalier 1954:252. Wallis 1961:22.
Cicindela guttif era , Fall 1901:308.
Cicindela guttifera , Leng 1902:150.
Cicindela depressula scapularis Casey 1909:272. Type locality- California.
Cicindela guttifera sonoma Casey 1913:29. Type locality - California
(maritime regions north of San Francisco). Horn 1915:378.
Cicindela quadripennis Casey 1913:30. Type locality - Hawthorne,
Nevada. Horn 1915:378.
Cicindela ovalipennis Casey 1913:30. Type locality - Hawthorne,
Nevada. Horn 1915:378.
Cicindela oregona scapularis , Horn 1915:378.
Cicindela oregona guttifera LeConte 1857:42. Type locality - New Mexico,
Leng 19 02:41. Horn 1915:378, and 1930:82. Varas Arangua 1928:
250. Tanner 1929:83. Cazier 1954:242. Wallis 1961:22.
Cicindela sterope Casey 1913:28. Type locality - Kansas. Horn 1915:
378.
Cicindela audax Casey 19 13:29. Type locality - Colorado. Horn 1915:
378.
Cicindela guttifera , Casey 1913:29.
Cicindela oregona ore gonella Casey 1924:16. Type locality - Parowan,
Utah.
Cicindela duodecimguttata, Hatch 1953:38 (not Dejean).
Cicindela oregona guttifera x Cicindela oregona maricopa
Cicindela provensis Casey 1924:15. Type locality - Parowan and Provo
Canyon, Utah.
Cicindela provensis mormonella Casey 1924:15. Type locality - Eureka,
Provo Canyon, Parowan and Vineyard, Utah.
Cicindela provensis nephiana Casey 1924:16. Type locality - Parowan,
Utah.
Cicindela oregona maricopa , Tanner 1929:83 (not Leng).
Cicindela oregona maricopa Leng 1902:150. Type locality - Phoenix, Arizona.
Horn 1915:378. Horn 1930:82.
Cicindela guttifera maricopa , Casey 1913:27. Varas Arangua 1928:250.
Cicindela oregona navajoensis Van Dyke 1947:155. Type locality - Kayenta,
Arizona.
On the basis of a patch of hair s confined to the front inner edge of
each eye this species may be distinguished from all other tiger beetles of
the maritima group, except depressula and female scutellaris Say (see figs 12,
14, 16). Female scutellaris and oregona specimens usually can be distinguish-
112
Cicindela maritima Group
edfrom each other simply by noting the geographical location from which
the specimens were taken. The range of scutellaris is east of the Rocky Mount-
ains while oregona occur s in the west limited by the eastern foothills of the
Rockies. Specimens of the subspecies scutellaris scutellaris are present in
Colorado and New Mexico (Shelford 1917) but these forms are quite dif-
ferent from oregona in that they have bright cupr eous to red elytra. Another
subspecies of scutellaris , related to the subspecies criddlei , also occurs in
Colorado (Rumpp 1961) and it is characterized by broad white margins
of the elytra. The species depressula and oregona , on the other hand are
sympatric. Individuals of these two species can be told apart by the
numbers of hairs forming the clusters near each eye. The species
oregona normally has eight to eleven hairs in this area while depressula us-
ually has one to three and rarely four. A more reliable character for
distinguishing between oregona and depressula is the shape of the median lobe
of the male.
Like most other species of the maritima group oregona lives along the
edges of river s , lakes , and sloughs and is found on a variety of substrates.
I have taken oregona on sandy beaches, gravelly banks, and indeed onrock.
This species is more common where there are open patches of beach.
Notes on Synonymy
Casey proposed the names C. quadripennis and C. ovalipennis for male
and female C. o. oregona respectively, that occur in Hawthorne, Nevada.
SimilarlyC. oregorcaspecimens collected north of San Francisco, were
regarded by Casey as a subspecies of guttifera and he applied the name
sonoma to them. Casey also considered a coastal blue form of o. oregona
to be a subspecies of depressula and named it scapu/aris . However scapularis
does not itself occur in uniform geographic populations and consequently
I have not given it taxonomic status (see Wallis 1961).
Casey's audax and sterope arebothforms of o. guttifera . Their orig-
inal descriptions indicate these names refer to typical guttifera in Colorado
and NewMexico. The name o. oregonella Caseyhasbeen given to specimens
from highly variable populations of o. guttifera which occur in north central
Utah. C. provensis Casey refers to blue specimens that were taken in
Parowan and Provo Canyon, Utah. Parowan is located in southwestern
Utah, a hybrid area of o. guttifera and o. maricopa , and Provo Canyon is
situated in north central Utah where o. oregona and o. guttifera intergrade.
The name provensisrepresents hybrid individuals of these regions. Tanner
regarded guttifera x maricopa and guttifera x oregona hybrid specimens in Utah
as being variants of o.. maricopa.
Geographic Variation and Subspecies
The species Cicindela oregona ranges widely in the west, from
Alaska to southwestern California, Arizona , and New Mexico and eastward
to the Rocky Mountains (fig. 18). Five easily observed characters vary
geographically: body size, color of thoracic pleura, color of elytra,
color of pronotum, and expanse of elytral pattern. Length of elytra is
expressedinmillimeters from the tip of the scutellum posteriorly to the
F r eitag
113
tip of the elytral spine along the suture. Width of elytron is similarly
expressed in millimeters from the median line of the elytra through the
transverse portion of the middle band to the elytral margin. These data
are listed in tables 4 and 5 for males and females respectively. The
tables also summarize data on variation in diameters of apical dots.
The measurements illustrate vatiation in expanse of the elytral pattern.
The apical dot was measured transversely across the widest portion.
Size - Before discussing the geographical aspects of size variation,
I would note that females on the average are larger thanmales of the same
population in every locality listed in tables 4, and 5. This is true for
the sexes in the same locality, but is not necessarily true if opposite
sexes of different regions are compared. For example, females from
Trinidad, Colorado have a mean elytral length of 7. 23 mm while the
average elytral length of males from Tanana River, Alaskais7.26 mm.
Data on variation of elytral length in males and females are given in
tables 4 and 5 respectively.
Three geographical routes (A, B, and C in column 1 of tables 4
and 5) have been selected to facilitate description of geographical vari-
ation in the length and width of elytra and expanse of color pattern.
Tanana River, Alaska and Terrace and Oliver, British Columbia serve
as the northern portion for all three routes. The first transect of pop-
ulation samples (A) extends from Alaska, south to New Mexico through
British Columbia, Montana, Wyoming and Colorado. A second line of
samples (B) is from Alaska to Arizona by way of British Columbia, Idaho
and Utah. A third course (C) is from Tanana River, Alaska to southern
Nevada, through British Columbia and Idaho. The data in tables 4 and
5 are arranged to correspond to these routes.
Because the corresponding character gradients of males and fe-
males are generally parallel, only the male samples are discussed in
detail, with occasional reference to female samples. Table 4 indicates
a decrease in the mean length of elytra of males, from north to south
for all three courses. Evach cline is quite irregular and there are sharp
decreases and increases throughout. These abrupt changes in the
character gradients appear to be correlated, at least in part, with changes
in altitude or with geographic barrier s. However , I have noted discrep-
ancies in the dines that cannot be so related.
Through the northern section of the first route there is a south-
ward decrease in average length of elytra of 0. 007 mm per degree of
latitude for males and 0.018 mm per degree of latitude for females.
From Alaska to Lower Medicine Lake, Montana a distance of 14°30f of
latitude, no marked deviations occur in the trend. Between Lower
Medicine Lake and Hardy, Montana however, a distance of only 1°10'
latitude, mean length decreases by 0.20 mm. There is a drop of 1,500
feet in altitude between these two localities. Another irregularity in
the above character gradient occurs between Helena and Gardiner,
Montana - an increase in mean length of elytra of 0. 18 mm with 1°30'
of latitude. Gardiner is 1, 640 feet higher than Helena and contrast in
elevation again seems to be related to the clinal difference. Population
samples from Gardiner, Montana* Yellowstone National Park, Jackson
Hole National Monument, and Moran, Wyoming have elytra of approx-
114
Fig. 18
Geographical distribution of the subspecies of C. oregona .
Fr eitag
115
imately equal length. Immediately southward the slope of the char-
acter gradient decreases markedly between Moran and Labarge,
Wyoming. Both sites are at approximately the same altitude and
there are no obvious geographic barriers between the two localities.
In the Labarge and Green River regions the reduction in body size may
be due to local factor s such as disease, lack of food or marginal habitats
(Mayr 1963). Jelm and Saratoga in western Wyoming are rather isolated
from Fort Bridger and Green River, in eastern Wyoming by the Great
Divide Basin and the Continental Divide which are situated in south central
Wyoming. East to west gene flow betweenpopulationsof Cicindela oregona is
most likely impeded in southern Wyoming by these geographical features
which may account for the shorter elytra in eastern Wyoming. The dif-
fer ence between the average elytral lengths of males from Jelm and Fort
Bridger is statistically significant but that between females is not.
The second arbitrary line of population samples (B) is from Alaska
to Arizona by way of British Columbia, Idaho, and Utah. A clinal de-
crease in length of elytra is evident throughout this route as well. In
Utah, the Alta, Mount Timpanogos, and Provo Canyon populations have
relatively long elytra. The elevation of Alta is 8, 585 feet, Mount
Timpanogos is 11, 750 feet, and Provo Canyon is located in Provo Park
which rises at a height of 11, 068 feet. Samples collected at lower ele-
vations in areas adjacent to the above mentioned, have a shorter mean
elytral length and populations from Stockton and Provo are examples of
these. Population samples taken in Salt Lake City may have been collected
in any of the creeks entering the city from the Wasatch range which serves
as the eastern geographic limits of the metropolis . Although the insects
wer e labelled as being collected in Salt Lake City, they could conceivably
have been taken at a muchhigher altitude nearby. Floy, Utah, and Kayenta,
Arizona have populations with the shortest elytra in the entire span of this
gradient. South of Kayenta the samples taken in Prescott, Phoenix, and
Globe, Arizona are larger and compare in size with those from Idaho.
These large forms in central and southern Arizona are fairly isolated and
common only in these areas (see oregona marie opa p. 127 ).
The third line of population samples (C) extends from Alaska, south
through British Columbia and Idaho to Nevada. Even though elytra are
generally shorter in more southerly latitudes , the Walker Lake, Nevada
population sample has the value for mean length of elytra equal to that of
Tanana River, Alaska. I cannot account for this discrepancy.
Data on the variation in width of elytra are presented for males in
table 4 and for females in table 5. There is a slight decrease in width
of the elytra from Alaska, southward along all three routes . Irregularities
in the dines of elytral widths correspond with changes in the character
gradients of the lengths of elytra. At higher elevations of Montana,
Wyoming and Utah mean values for elytral width are generally slightly
greater than those of Alaska and Terrace, British Columbia. Tiger
beetles of this species living in these lower latitudes at high altitudes
are normally shorter but wider than their counterparts in boreal areas.
This is especially marked in females. For example compare population
samples of Tanana R iver , Alaska, Terrace and Queen Charlotte Islands,
British Columbia with Gardiner , Montana, Alta and Provo Canyon, Utah,
116
TABLE 4 - Variation in male Cicindela oregona.
Length of Elytra
117
118
TABLE 5 - Variation in female Cicindela oregona.
Length of Elytra
119
Width of Elytra Diameter of Apical Dot
120
Cicindela maritime, Group
and Yellowstone National Park, Wyoming in table 5.
Populations in Nevada are generally longer and wider than samples
from Utah, Colorado, and Wyoming at similar elevations; but east-west
clines are very irregular.
Specimens collected in Queen Charlotte Islands , British Columbia
are the only insular members recorded in tables 4 and 5. The oregona fe-
males of these Islands have a higher mean value for length of elytra than
have those from any other locality listed in table 5, and the mean value
for males is slightly le s s than thos e of Alta and Mount Timpanogos, Utah.
However, the male samples from the Queen Charlotte Islands, Mount
Timpanogos, and Alta each are represented by fewer than 10 specimens
and a more accurate comparison can be made with the females . Individuals
from localities of high elevation in Utah, and Wyoming are scarcely
broader than are those of Queen Charlotte Islands (see table 5).
A relationship of body size with latitude or altitude is evident
in many animal species other than Cicindela oregona. North American brown
bears for example, increase in size as the latitude increases (Rausch
1963). This phenomonon has also been shown in many species of birds
(Mayr 1942 , 1963 , Hamilton 196 1) . In insects , honey bees and two species
of European Carcks vary in the same way (Mayr 1963 , p. 326) . New Guinea
dragon flies have been found to increase in size at higher elevations also
(Mayr 1963, p. 326). In ectothermal animals as a whole the largest body
size may just as often be found in the warmest portion of a species
range. Lindroth (1963) noted that in some Carabids of Newfoundland
dwarf forms are frequently confined to high altitudes or marginal
northern areas of the species range. Ball (1959) observed that several
species of the ground beetle genus Dicaelus were larger in southern areas
of their ranges . Likewise the small, flightless grasshopper Melanoplus puer
shows a general southward increase in size (Hubbell 1956). Mayr (1963)
presents a review of evidence for the adaptive nature of geographic
variation in which latitudinal and altitadinal changes are discussed in
relation to geographic variation in body size.
Color pattern of the elytra - While body size of Cicindela oregona deer eas e S
from north to south the breadth of the white markings of the elytra in-
creases. The diameter of the apical dot increases approximately 0. 007
mm for each degree of latitude. Data on the expanse of elytral pattern,
illustrated by the diameter of the apical dot, are presented in table 4
for males and in table 5 for females .
In the first route (A) from Alaska to New Mexico there is a
slight uniform increase in apical dot size through British Columbia and
Montana. In For t Bridger , Wyoming the cline is steeper and continues to
increase through Colorado to New Mexico. Fort Wingate, New Mexico
is represented by individuals having a very wide apical dot with an aver age
measurement of 0. 78 mm. The difference between the mean values for
males from Fort Wingate and those from Pecons, New Mexico is stati-
stically significant, but this is not so for females. The same applies to
material from Hardy, Montana and from Helena and Lower Medicine
Lake, Montana.
Fr eitag
121
A similar latitudinal increase in the diameter of the apical dot
is evident in course (B) from Alaska to Arizona. There is a slight de-
crease in the mean apical dot size of males in southern Idaho (this is not
true for females) but the color pattern expands markedly in Utah. Pop-
ulations with the widest apical dots are present in Floy, Utah, and Prescott
and Globe, Arizona. A similar cline exists in the third path of population
samples that extends from Alaska to Nevada.
Coloration - The elytra, thoracic pleura, pronotum and, indeed, the
entire body are subject to color variation in this species. The elytra may
be brown, green, purple, blue and occasionally very dark brown that is
almost black. The thoracic pleura are coppery, metallic blue, purple
or green. Color of the pronotum and elytra is generally the same on
each individual. Unicolored specimens occur throughout the species
range in scattered localities and they are usually green, less frequently
blue, In Arizona, southern Nevada, and southern New Mexico a green
pronotum is usually associated with purple elytra. North of the limits
of populations that have purple elytra the beetles are usually metallic
blue-green ventrally and brown dorsally. Figure 19 is a pie-graph map
illustrating color variation of the elytra and thoracic pleura in populations
in the northern portion of the range of oregona. Each "pie" represents a
single population sample. Figure 20 is a southwestern continuation of
figure 19 .
Coppery thoracic pleura are prevalent throughout northwestern
British Columbia, Yukon Territory and Alaska. In figure 19 samples
A (Tanana River , Alaska) to K (Helena, Montana) coppery thoracic pleura
are most frequent, followed in numbers by metallic green,blue and
purple, in that order. In Alaska, the Yukon Territory and northern
British Columbia, only the coppery condition exists. In samples D, E,
F and G, in central British Columbia, individuals with metallic blue-
green or purple thoracic pleura are present in low fr equency. Southward,
coppery thoracic pleura are prevalent in central British Columbia, on the
eastern slopes of the Rocky Mountains in Alberta and Montana, and in
eastern Idaho, Utah, Wyoming, Colorado, northeastern Arizona and
northern New Mexico (see data for Fort Bridger, Wyoming and Jemez
Springs, New Mexico in pictorialized scatter diagrams figures 25 and
27). In British Columbia (fig. 19) coppery thoracic pleura are abruptly
replaced by metallic purple thoracic pleura and this condition extends
throughout southern British Columbia from the Pacific coast east to the
Continental Divide. Of the individuals represented in samples L to U,
only eight have coppery thoracic pleura in this region. Specimens with
metallic green thoracic pleura are common in coastal populations of
British Columbia and Washington, and also along the eastern ridge of the
Rocky Mountains . This conditionis less common in Or egon and California
where metallic purple and metallic blue thoracic pleura are dominant
(fig. 20). Thoracic pleura of specimens of Owyhee County, Idaho (L) are
predominantly metallic purple, also. Populations with thoracic pleura
ranging from coppery to metallic purple through metallic blue and metallic
green are found in northern Utah near Alta and southern Utah in the area
of Zion National Park. Across southwestern United States from San
122
Cicindela maritima Group
Diego, California to Mountain Park, New Mexico, specimens with metallic
purple thoracic pleura are most frequent.
Brown elytra are most common throughout most of the range of
oregona. Brown color is entirely replaced by purple color in central Arizona,
southern New Mexico and southern Nevada. This situation is discussed
in detail in the subspecies section. 4 Populations that are highly variable
in color of elytra occur throughout the range of this species. Such pop-
ulations are on the Pacific coast from Alaska to southern California and
almost all individuals with blue or green elytra exist in coastal localities
(seefigs 19, 20). Brown is most common, followed by green , and then
blue. In Garibaldi Park, British Columbia, a very large proportion of
specimens with blue elyTra are present, while Vancouver and Victoria,
British Columbia populations are made up mainly of individuals with
green elytra. In Humbolt County, California and Port Orford, Oregon,
most of the specimens have brown elytra, and some members with
green and blue elytra are also present. Likewise, a few individuals
with blue and green elytra are present in San Francisco and in San Diego,
California. Variation in elytral color is the rule in northern and southern
Utah, where blue, green, and brown specimens are present.
Shelford (1914) studied color and color pattern of tiger beetles
and he found, of the species studied in detail, the more brilliant colors
occur in warm, arid localities , and extended markings in cooler regions
These findings apply only in part to oregona. In Arizona, New Mexico and
southern Nevada, which are warm dry areas, specimens with bright met-
allic dorsal surfaces are prevalent but brilliant blue and green specimens
of oregona also live along the Pacific coast from California to Alaska, and
this is quite a humid zone. The markings of oregona are expandedin warmer
localities and this condition contrasts with the r esults obtained by Shelford
regarding pattern of elytra.
The pattern of variation - Independent character changes have resulted in
discordant variation within C. oregona. Elytra are dark brown, generally,
in the northern, eastern and western areas of the range but they are
purple in the south, and very light brown in eastern Utah. In contrast,
pleura usually blue to purple in the west and south, are coppery in
northern and eastern portions of the range. In addition general body
size decreases clinally from the north to south and also varies from
higher to lower altitudes. Finally extent of white markings on elytra
increases southward. Although recognition of subspecies in species that
show discordant variation is controversial, (p. 90 , and Inger 1961)1 think
it is useful to group into subspecies the population samples of oregona.
Maintenance of variation in this species appears to be largely
dependent on geographical factor s but may also be due to variation in the
season of occurrence of adults. Mature specimens of my own and other
collector s from boreal populations appear to be most plentiful for June,
July, and August (based on specimen label information and personal col-
lecting). This seems to be true for alpine populations in southern regions
as well. Adult specimens from populations in desert areas of Utah,
Arizona, and New Mexico have been collected from March to October
inclusive. This suggests that they are common throughout this time but
123
Fig. 19. Pie-graph map illustrating geographic variation in the color of the
elytra and thoracic pleura of some populations samples of Cicindela oregona . The
numbers of specimens with a given color combination are indicated opposite
the appropriate section. Thus, 20 placed at 1 o'clock signifies that 20 spec-
imens have coppery thoracic pleura and brown elytra.
124
Fr eitag
125
I do not believe so. Because C. oregona is riparian, size of populations, in
arid southwestern regions, is likely to fluctuate with rain. Kendrew (1961)
points out that rainfall is variable in these desert regions with a max-
imum in late summer and winter , and that the mountain ranges which rise
on the southwestern plateau have rather more rain. It follows that activity
peaks of desert populations probably do not occur at the same time year
after year but in differ ent periods in relation to rainfall . Attempts to col-
lect de sert forms during winter months have been unsuccessful. Perhaps
they are most numerous during late summer-later than the peak of alpine
populations. Such asynchronous number fluctuations effect a reduction
in gene flow and thus maintain the variation between desert and alpine
populations of the southwest.
Subspecies - I recognize four subspecies of Cicindela oregona. I have
followed the 75% rule in defining the group taxonomically (see p. 90 , and
Mayr et al. 1953) . Thedifferencesbetween two or more populations in two
or more characters are best illustrated by a pictorialized scatter diagram.
T en such diagrams and a locality map of the population samples compared
in the scatter diagrams are presented as figs 21-31. Subspecies can be
readily distinguished from each other on the basis of one or more exter-
nal characters. Males and females are treated separately. Generally
five localities are represented in each diagram. Each locality is re-
presented by ten specimens or less, and they have been selected randomly.
Fifty specimen symbols are placed on a diagram.
1 Thoracic pleura blue or purple 2
Thoracic pleura coppery 3
2 Elytra purple, elytral pattern broad, pronotum green
o. maricopa
Elytra brown, green, blue, or rarely purple; elytral
pattern narrow; pronotum brown o. oregona
3 Elytra light brown; elytral pattern broad o.. navajoensis
Elytra dark brown; elytral pattern narrow Q. guttifera
The nominate subspecie s Cicindela oregona oregona ranges from south-
ern British Columbia in Canada, to southwestern California, it is
present throughout Washington, Oregon, and California, except for the
southeast portion of that state. The Continental Divide serves as the
eastern limit in the north, from Banff, Alberta south to Yellowstone
National Park, Wyoming. Further south oregona oregona is found as far
east as Owyhee Count, Idaho, western Nevada and finally near the south-
ern portions of the Sierra Nevada Mountains in California (fig. 18). A
combination of green, blue, or brown elytra with metallic purple or
blue thoracic pleura is characteristic of this subspecies. Individuals
with green elytra are numerous in or near the above localities. The
scatter ed occurrence of these blue and green individuals along the Pacific
coast may be evidence of a blue form that was once widespread in these
126
Cicindela maritima Group
coastal regions, but was infiltrated by a more vigorous stock, character-
ized by the possession of brown elytra. Whenever these two aggregates
of populations came into contact introgr ession took place and the presence
of green individuals interpreted as hybrids, marks what once were zones
of contact. In more southern locations these green and blue forms have
been all but completely replaced by brown. On the other hand they may
be recent phenotypes whose gene complex originated in southwestern
British Columbia when the blue phenotypes were relatively common.
Another possibility is that the green and blue forms might be ecopheno-
types. Although Shelford observed that Cicindela tranquebarica Herbst is
green on the coasts and coastal mountains and also that in Cicindela scutellaris
green forms were most common along the Atlantic coast (Shelford 1917),
he did not believe that this was the result of direct influence of the en-
vironment on the phenotypes.
The subspecies o. oregona comes in contact with o. guttifera LeConte
in southern British Columbia and along the slopes of the Rocky Mountains
from Banff, Alberta to Yellowstone National Park, Wyoming. Many
specimens that appear to be hybrids are present in areas of contact of
these two subspecies, and such are distinguished by their metallic
green thoracic pleura (fig. 19).
In San Diego, California a hightly variable group of populations
is present (fig. 20), for in this area specimens typical of both o. oregona
and o. marie opa occur. This situation could be the result of maricopa genes
infiltrating the more numerous oregona population in the region (figs 22,
23). Only five phenotypically maricopa specimens are known from the San
Diego area.
Cicindela o. guttifera ranges the Rockies from Fort Yukon, Alaska to
northern New Mexico (fig. 18). In Alaska and north and central British
Columbia, guttifera ranges from the Pacific coast to the eastern slopes of
the Rockies but continues southward in a very narrow zone to northern
New Mexico. This subspecies also occurs in northern and central Utah.
Coppery thoracic pleura and brown elytra that have a metallic lustre
characterize it. I have already mentioned that hybridization takes place
between oregona guttifera and oregona oregona in much of eastern Idaho and
western Montana, and intermediate specimens with metallic green sides
are not uncommon in northern and central Utah where they are distributed
through oregona guttifera populations . In southwestern Utah a highly variable
series of populations occur s , consisting of individuals ranging from typ-
ical guttifera to typical maricopa . This region is undoubtedly a melting pot
of these two subspecies (figs. 24,25). It may be argued that this var-
iation is a result of hybridization between oregona and maricopa and not as
above. This is not likely since at the present tim e oregona is uncommon in
eastern Nevada and it is not abundant in Utah, but it may have contributed
to this variation in pluvial times.
The ranges of Cicindela o. navajoensis and guttifera come very close in
northwestern New Mexico (fig. 18). Navajoensis is relatively small in size
and has much lighter brown elytra and broader pattern of elytra than
guttifera. Like the latter, navajoensis has coppery thoracic pleura. A
color character gradient occurs from Kayenta, Arizona to Jemez Springs,
New Mexico through an intermediate locality, Fort Wingate, New Mexico.
Fr eitag
127
Color of the elytra in Kayenta is light brown, dark brown in Jemez
Springs and intermediate in Fort Wingate. Fort Wingate specimens are
also intermediate in lengths and widths of elytra and the apical dot (figs
26, 27).
The geographic ranges of ore gona and navajoensis are not m contact;
their morphological relationships are demonstrated in figs 28 and 29.
In southeastern Arizona and southwestern New Mexico are variable
populations that consist of individuals structurally between navajoensis and
maricopa (figs 30, 31). The members of these groups are generally smaller
in body size than maricopa and larger than navajoensis . Their elytra are
mainly purplish brown. The Fort Wingate sample in figures 30 and 31 is
not pure navajoensis but is intermediate between navajoensis and guttifera (figs
26,27). Thus the specimens appear as intermediates between navajoensis
and maricopa in figures 30 and 31.
Specimens of o. maricopa have brilliant purple elytra, brown to met-
allic green pronota, and metallic purple thoracic pleura. Maricopa is dis-
tributed spar s ely through southern California, southeastern Nevada, and
southern New Mexico , but it is common in central and southern Arizona
(fig. 18) .California " maricopa " may be a minor element in predominantly
o. oregona populations and if so they are maricopa in a typological sense only.
The form of this and other subspecies of Cicindela oregona has been compared
WIDTH OF ELYTRA / DIAMETER OF APICAL DOT WIDTH OF ELYTRA / DIAMETER OF APICAL DOT
128
Figs. 22 to 31. Pictorialized scatter diagrams illustrating character differences between population samples of C. oregona oregona
(©)> C.o. maricopa (Q)» C.o. guttifera (£) , and C.o. navajoensis (©) . Intermediate populations represented by divided circles ( (0 ©
© ®); elytral color by vertical bars: long - purple, medium - green, short - blue, no bar - brown; pleural color by horizon-
tal bars: long- purple, medium - green, short - blue, no bar - coppery. Males above, females below.
WIDTH OF ELYTRA / DIAMETER OF APICAL DOT m WIDTH OF ELYTRA / DIAMETER OF APICAL DOT
129
WIDTH OF ELYTRA / DIAMETER OF APICAL DOT WIDTH OF ELYTRA / DIAMETER OF APICAL DOT
130
Fig. 26 0*
LENGTH OF ELYTRA MM
Figs. 22 to 31. Pictorialized scatter diagrams illustrating character differences between population samples of C. oregona oregona
(©), C .o. maricopa (O), C.o. guttifera (Q) , and C.o. ncivujoensis . Intermediate populations represented by divided circles (@©
$*)); elycral color by vertical bars: long - purple, medium - green, short - blue, no bar - brown; pleural color by horizon-
tal bars: long- purple, medium - green, short - blue, no bar - coppery. Ivales above, females below.
WIDTH OF ELYTRA / DIAMETER OF APICAL DOT WIDTH OF ELYTRA / DIAMETER OF APICAL DOT
131
LENGTH OF ELYTRA MM
WIDTH OF ELYTRA / DIAMETER OF APICAL DOT WIDTH OF ELYTRA / DIAMETER OF APICAL DOT
132
Fig. 30 O*
Figs. 22 to 31. Pictorialized scatter diagrams illustrating character differences between population samples of C. ore gona oregona
(©), c.o. maricopa (O). C.o. guttifera (£) , and Co. navajoensis (©). Intermediate populations represented by divided circles ( @ 0
© ©); elytra™ color by vertical bars: long - purple, medium - green, short - blue, no bar - brown; pleural color by horizon-
tal bars: long- purple, medium - green, short - blue, no bar - coppery. Males above, females below.
Fr eitag
133
above.
History of Distribution and Subspeciation
Distribution of C. oregona is restricted to western United States
and Canada bounded by Alaska, southern California, Arizona and New
Mexico and the Rocky Mountains. Within this area four geographically
distinct groups of populations exhibit boundaries that are generally bar-
riers such as deserts and mountain ranges (fig. 18). All population
samples of these subspecies that I have examined, have been collected
in the above described range of C. oregona. I have seen two maricopa speci-
mens, however, that are labelled "Texas", but specific localities are
not given. They could have been collected in western Texas near maricopa
localities in New Mexico. Because the total range of C. oregona is well
marked and no populations occur in remote regions outside of the
described range it appears highly likely that subspeciation took place
somewhere in western North America.
Before discussing further the questions of how and when form-
ation of subspecies occurred in oregona it should be emphasized that deter-
mination of evolution of a subspecies without a fossil record is a highly
speculative matter. Fossils are not available, and even if they were it
would be impossible to determine all of the subspecific development in
oregona. because color is rarely preserved in fossils. As a result indirect
evidence must be used. This is provided by a consideration of the effects
Pleistocene climatic changes may have had on bringing about subspecia-
tion in oregona.
It seems that isolation of groups of populations of oregona occurred
at different times in relation to climatic changes. In southern portions
of the species range populations were separated from each other during
interglacial periods because of the formation of deserts. Northern
incipient subspecies were probably separated during glacial times.
Because oregona is riparian, regions void of river systems and
lakes act as geographic barriers. The great deserts of southwestern
United States prove to be barrier s (fig. 18), and distribution of subspecies
is closely linked to wet and cool areas. Consequently it may be deduced
that isolation and subsequent genetic divergence of southern populations
tookplace when southwestern United States was largely desert; perhaps
during the last interglacial period. During glacial times, on the other
hand, river systems were very extensive and many lakes occurred in
the southwest ( Blackwelder 1948, Hubbs and Miller 1948). In these
regions populations were undoubtedly dispersed most widely in glacial
times and presumably gene flow was uninhibited.
Conversely, partitioning of incipient subspecies that existed in
northern regions of this species range probably occurred during glacial
times, while range expansions occurred in interglacial periods. In
glacial periods great ice masses moved down from the north, scarcely
crossing the Canadian-American border in the west. These undoubtedly
obstructed gene flow between aggregates of populations on the eastern
portions of the Rockies and populations further west by way of northern
United States and southern Canada. For example if an ice mass was at
134
Cicindela maritima Group
present established across the northwest o. oregona and o. guttifera would
be spatially isolated because they normally inter grade in Idaho, Montana,
southern Canada and northern Utah. Similarly glaciers that developed
throughout most of the major mountain ranges in glacial times must have
reduced east-west gene flow further south.
Interpretation of present distribution of the subspecies and knowl-
edge of the events of the Pleistocene epoch suggests the following course
of subspeciation (see fig. 32). Two subspecies of C. oregona, oregona and
guttifera as they are defined here, were formed in part during the Iowan
glacial stage. A uniform "pr otoor egona" species was distributed across
northwesternUnitedStates and southwestern Canada prior to this period.
With the advent of the Iowan ice mass and glacier formation all "proto-
oregona" populations north of the Canadian - American border were
probably annihilated, at least as a result of cooling and, two large pop-
ulations were isolated, one on either side of the Continental Divide;
race A on the west and race B on the east. Both races were more wide-
spread in the south than they now are. Race B occupied all of the Great
Basin, Arizona and regions west of the Sierra Nevada. Geographic
variation was pronounced in this race with brown forms predominant in
the north and blue in the south. Restricted to regions east of the Con-
tinental Divide in the north, race A extended southward into Colorado and
New Mexico then swung northward through lower elevations in northwest-
ern New Mexico and northeastern Arizona and eastern Utah. This was
the situation when the Prairie interglacial stage began.
Much division and spatial isolation between southern populations
took place during the Prairie inter glacial as a result of vast desert form-
ation. During this stage southern, blue populations of race B subspeciated
to maricopa and brown western (race B) populations to oregona. Race A
forms became navajoensis and guttifera. Distributions of aggregates of
populations shrank and assumed geographic areas approximately where
the subspecies of C. oregona now exist. C. oregona blue populations in
the south were pinched off from their counterparts in the northwest by
deserts where the Mohave Desert and Great Basin are now located. In
Utah navajoensis population s remained partially isolated from populations
in Colorado by the intervening Rocky Mountains, and were isolated from
the effects of the oregona blue forms in the southwest. In the north oregona
and guttifera reinvaded regions south of the retreating ice mass in northern
United States and Canada and formed marked hybrid areas wherever
their ranges came into contact.
Southern hybrid zones were primarily formed in the Wisconsin
glacial period as a result of expanding subspecific ranges . Pluvial lakes
that were reestablished in desert areas along with revived river systems
served as routes for expanding ranges and inter gradation was widespread.
In the north oregona and guttifera were isolated from each other as in the
previous glacial period.
Since the Wisconsin ice age ranges of subspecies have shrunk in
the southwest. A few specimens phenotypically maricopa have been found
in southern California but true maricopa is abundant only in central Arizona.
Thi s implies that maricopa once was more extensively distributed. On the
other hand oregona and guttifera ranges appear to be expanding in the north.
Fr eitag
135
Furthermore distribution of hybrids in the southwest has recently been
reduced in area. Evidence of this is 'available in southeastern Arizona
and southwestern New Mexico. Pure maricopa specimens along with
maricopa x navajoensis hybrids exist in these regions , but pure populations
of navajoensis arelocatedin northeastern Arizona and northwestern New
Mexico, many miles away from where the hybrids are found.
Period
Subspeciation
Thousands of
years B.R
Recent
guttifera navajoensis maricopa oregona
hyb.
Wisconsin
hyb.
hyb.
hyb.
15
Iowan
Race A
Race B
.v
Sangamon
70
Fig. 32 -Hypothetical phylogeny
of Cicindela oregona subspecies.
The diagram is based on
Pleistoc ene chronology postu-
lated by Karlstrom (1961).
protooregona
136
Cicindela maritima Group
Cicindela duodecimguttata andC. o. guifj/eramay have formed hybrid popu -
lations along the southeastern foothills of the Rocky Mountains during
Pleistocene times. Such hybrid populations were probably subject to
extreme fluctuations as in Nordegg, Alberta at the present time (p.i56 ).
These unstable intermediate forms had no profound effect on the par ental
forms .
This discussion presents one interpretation of subspeciation
in C. oregona that is bas ed on knowledge of distribution pattern of C. oregona
and Pleistocene events in southwestern United States. Undoubtedly
other explanations of the available data are possible.
Distribution
I examined 6,073 specimens. Several specimens appeared to
be labelled wrongly. Two maricopa specimens were labelled Fort Garland,
Castilla County, Colorado. Fort Garland is well into guttifera territory
beyond the northern limits of maricopa at that longitude. Another maricopa
specimen is labelled Sonoma County, California, which is in northwestern
California approximately sevenhundredmiles north of the maricopa speci-
mens in San Diego. Five specimens of o. guttifera were labelled as being
collected in Santa Rita Mountains. This is unlikely but not impossible
since these specimens could represent a relict population which has
survived in these mountains since the end of the Pleistocene.
ia LeConte. Canada. BRITISH COLUMBIA:
Aiyansh, 1; Atbara, 7; Bear Foot, 1; Cherrierville , 2; Chilliwack,
ay, 4; Cranbrook, 3; Creston, 139; Duncan, 3;
1; Fairmont, 1; Field, 1; Fraser River, 1; Gabriola, 2; Garibaldi
dstream , 1; Harrison, 2; Hatzic, 20; Hope, 1; Howser, 7; Huntingdon, 15; Kamloops,
2; Likely, 1; Lillooet, 9; Lynn Valley, 1; Lytton, 2; Mabel Lake, 2; MacGillivray ,
53; McIntyre Creek, 1; Merritt, 22; Miracle Beach, 60; Misson City, 20; Nanaimo, 1; Nanoose, 5; North
Bend, 4; Okanagan Lake, 1; Oliver, 61; Osoyoos, 2;
Grey, 2; Powell River, 2; Radium, 3; Read Bay, 4; Riondel, 11; Rock Creek, 7; Salmon Ai Sanca, 12;
Falls, 1; Stillwater, 1; Summerland, 4; Tood Inlet, 1; Trinity, 1; Vancouver, 79;
l, 32. Del Norte
. Valley, 1; Echo, 1; Fallen Leaf Lake, 1. Fr
Blairs Ranch, 3; Bridgeville, 28; Bridgeville (15 miles east) , 5;
Borax Lake, 5; Hullville, 3; Lake Pillsbury, 8; Lakeport, 2; Lower Lake, 1. Lassen County:
Facht, 1; Goumaz, 4; Madeline, 1; Pine Creek, 1; Susanville, 1. Los Angeles County: Burbank, 2; Covina, 1;
Crystal Lake, 1; El Monte, 3; Gabriel Mountains, 1; Los Angeles, 57; Palmdale, 1; Point San Pedro, 5; San
Pedro, 1; Santa Monica, 10; Tropico, 1. Madera County: Jackass Meadow, 1; Madera, 1; North Fork, 2;
Placer Station, 2. Marin County: Bon Tempe, 4; Dillon Beach, 6; Inverness, 1; Lagunitas, 1; Mill Valley, 1;
1; Yosemite National Park, 22; Yosemite Valley,
r, 1; Eagles Nest, 1; Fort Bragg, 1; Littleriver, 2; Philo, 1; Yorkville, 3.
5, 1; Santa Rita, 2. Modoc County: Cedarville, 2; Goose Lake,
1; Lake City, 1. Mono County: Coleville, 1; Lake Mary, 5; Mono County, 13; Mono Lake, 6; Sonora Pass,
2; Topaz Lake, 8. Monterey County: Bradley, 1; Carmel, 34; Pacific Grove, 1; Salinas River, 6; Soledad,
137
Churchill County: Fallon, 3'. Douglas County: Lake Tahoe, 1; Minden, 24. Elko County: Elko, 4; Lamoille,
1. Mineral County: Hawthorne, 10; Walker Lake, 45. Ormsby County: Carson City, 3. Pershing County:
Lovelock, 1. Washoe County: Gerlach, 2; Mount Rose, 1; Nixon, 1; Pyramid Lake, 18; Reno, 19; Verdi, 1.
White Pine County: McGill, 1. OREGON: Baker County: Pine Creek, 1; Richland, 1. Benton County:
Corvallis, 8; Umatilla, 10. Clackamas County: Estacada, 4. Clatsop County: Cannon Beach, 1; Clatsop
Beach, 2. Columbia County: Rainier, 1. Coos County: Cape Arago, 2; Charleston, 16; Coos Bay, 4; CoQs
Head, 1. Curry County: Humbug Mountain, 6; Port Orford, 12. Grant County: John Day George, 2. Harney
County: Frenchglen, 8; Malheur Lake, 15; P. Ranch, 1; Steens Mountains, 15. Hood River County: Hood
River, 8; Mount Hood, 2; Parkdale, 1. Jackson County: Medford, 10; Rogue River, 1; Ruch, 2. Josephine
County: Grants Pass, 5; Hells Gate Bridge, 2; Murphy, 1. Klamath County: Crater Lake, 2; Klamath Lake, 6;
Lake O Woods, 3; Pinehurst (21.9 miles east), 1. Lake County: Lake Albert, 12; Paisley, 3. Lane County:
Eugene, 13; Florence (3 miles north) , 13. Lincoln County: Depoe Bay, 2; Newport, 6; Waldport, 5; Yachats
(5 miles south), 5. Linn County: Cascadia, 1. Malheur County: Sucker Creek Canyon, 1. Marion County:
Detroit, 1. Multnomah County: Portland, 9. Tillamook County: Pacific City, 1; Tillamook, 1; Woods, 4.
Umatilla County: Echo, 2; Hermiston, 3; Meadow Lake, 1. Wasco County: The Dalles, 7; Tygh Valley, 1 .
Yamhill County: Dayton, 6; McMinnville, 4. Localities of unknown counties: Alvord Hot Springs, 3; Blitzen
Valley, 1; Boiler Bay, 5; Buell, 1; Devils Lake, 1; Durnep, 2; McNair Lake, 1; Moffat Mead, 1; Ocean Park,
1; Oregon (south east), 3; Santiam, 4; Sparks Lake, 1; Whitman, 2. UTAH: Salt Lake County: Alta, 13;
Brighton, 1. Utah County: American Fork Canyon, 9. WASHINGTON: Adams County: Othello, 7. Asotin
County: Asotin, 1; Clarkston, 1. Benton County: Paterson, 1. Chelan County: Leavenworth, 1; Peshastin,
4; Stehekin, 1; Wenatchee, 5. Clallam County: Port Angeles, 1. Columbia County: Huntsville, 3. Douglas
County: Moses Coulee, 3. Franklin County: Kahlotus, 2; Pasco, 1. Grant County: Beverly, 3; Goose Lake.l;
Moses Lake, 1; Stratford, 9. Grays Harbor County: Moclips, 3. Island County: Coupeville, 1; Whidby
Island, 53. Jefferson County: Port Townsend, 17. King County: Auburn, 2; Bothell, 4; Cedar Mountain,
5; Maple Valley, 3; Renton, 13; Seattle, 80; Selleck, 1; Snoqualmie, 1. Kitsap County: Bremerton, 49;
Chico, 34; Gorst, 120; Keyport, 1; Kingston, 13; Manchester, 1. Kittitas County: Ellensburg, 2; Vantage,
12. Klickitat County: Goldendale, 1; Goldendale (32.3 miles north), 22. Lewis County: Chehalis, 2. Lincoln
County: Sprague, 2. Mason County: Lake Cushman, 1; Spillman, 2. Okanogan County: Brewater, 5.
Pacific County: Bay Center, 4; Ilwaco, 1; Long Beach, 1; Nahcotta, 2; North Cove, 1; Ocean Park, 3. Pend
Oreille County: Newport, 1. Pierce County: Buckley, 1; Chinook Pass, 3; Mount Rainier, 1; Summer, 1;
Tacoma, 4. San Juan County: False Bay, 1; Friday Harbor, 9. Skagit County: Anacortes, 18. Snohomish
County: Cicero, 2; Darrington, 6; Everett, 3; Index, 4; Sulton, 1; Verlot, 8. Spokane County: Medical
Lake, 4; Spokane, 13. Stevens County: Wellpinit, 1. Thurston County: Olympia, 1; Tenino, 5. Walla Walla
County: Dixie, 10; Lowden, 1; Touchet, 27; Wallula Gap, 1. Whatcom County: ' Bellingham, 1; Mount Baker,
2. Whitman County: Almota, 3; Pullman, 30; Wawawai, 11. Yakima County: Toppenish, 4; White Swan, 1;
Yakima, 2. Localities of unknown counties: Barkerville, 1; Blue Mountains, 1; Central Ferry, 1; Clifton, 1;
Ginko State Park, 2; Half Moon Lake, 2; Lyone Ferry, 9; Neppel, 10; Paha, 5; Pot Holes, 2; Saratoga Beach,
5; Silverton, 1; Skating Lake, 1; Stillaguamish, 2; Tolsak, 1.
Cicindela oregona guttifera LeConte. Canada. ALBERTA:
Kootenay Plains, 14. BRITISH COLUMBIA: Aiyansh, 1; Blue River, 1; Bucks Bar, 1; Cariboo Road (mile 185),
2; Glenora, 1; McNab Creek, 1; Juskalta, 6; Massett, Queen Charlotte Islands, 10; Moresby Camp, Queen
Charlotte Islands, 9; Queen Charlotte City, 1; Stickeen River, 1; Tlell, Queen Charlotte Islands, 9. NORTHWEST
TERRITORIES: Fort Good Hope, 2; South Nahanni River, 1. YUKON TERRITORY: Kirkman Creek, 1;
Watson Lake, 6.
United States. ' ALASKA: Eagle, 3; Fairbanks, 5; Fort Yukon, 11; Haines, 12; Tanana River, 27;
Valdez, 4; Yukon River, 1. COLORADO: Alamosa County: Alamosa, 3. Boulder County: Boulder, 5;
Jamestown, 2; Lyons, 1; Pinecliffe, 2. Chaffee County: Buena Vista, 2; Salida, 3. Clear Creek County :
Georgetown, 1. Conejos County: Cumbres Pass, 4; La Manga Pass, 4. Douglas County: Larkspur, 2. El
Paso County: Cascade, 1; Colorado Springs, 6; Colorado Springs (10 miles south), 6; El Paso County: 2;
Manitou Springs, 5. Freemont County: Coalcreek, 1. Garfield County: Glenwood Springs, 1. Grand County:
Big Muddy Creek, 1; Fraser, 1. Jefferson County: Golden, 7. La Plata County: Electra Lake, 1. Larimer
County: Estes Park, 25; Fort Collins, 6. Las Animas County: Trinidad, 26. Mineral County: Creede, 4;
Wolf Fall Creeks, 2. Ouray County: Ouray, 1. Park County: Colorado Springs (50 miles west), 4. Pitkin
County: Aspen, 4. Saquache County: Great Sand Dunes National Monument, 4. Teller County: Victor, 1.
Localities of unknown counties: Berkley, 1; Gothic, 1; Rockwood, 2; South Fork, 1; Thomasville, 1. NEW
MEXICO: Bernalillo County: Albuquerque, 1. San' Doval County: Bernalillo, 1; Jemez Mountains, 17;
Jemez Springs, 47; Jemez Springs (9 miles north), 73; Jemez Springs (10 miles north), 43. San Miguel Codnty ;
Beulah, 2. Santa Fe County: Pecos River, 17; Santa Fe (3 miles east), 2. Localities of unknown counties:
San Antone , 1. UTAH: Beaver County: Beaver Creek, 1. Cache County: Logan, 3. Juab County: Levan
(5 miles south), 1. Millard County: Lynndyl, 1. Piute County: Piute Reservoir, 18. Rich County: Bear
Lake, 1. Salt Lake County: Mount Dell Creek, 1; Parley Canyon, 3; Salt Lake City, 18; Salt Lake County, 24.
Sampete County: Sevier Bridge Reservoir, 21. Summit County: Echo, 2; Park City, 3. Tooele County:
Stockton, 15. Uintah County: Power Plant, 2. Utah County: Mount Timpanogos, 20; Payson, 1; Provo, 17;
Provo Canyon, 26; Utah Lake, 19. Wasatch County: Soldier Summit, 3; Wasatch County, 2. Weber County:
Ogden, 13; Ogden (30 miles east), 8. Localities of unknown counties: Emigration, 1; Hillneck Canyon, 1;
Kawara, 6; Red Butte Canyon, 1; Salt Creek Canyon, 1; Silver Lake, 5; Vineyard, 26. WYOMING: Albany
County: Centennial, 1; Jelm, 88. Carbon County: Baggs, 2; Saratoga (8 miles south), 24. Fremont County:
Lander, 1. Lincoln County: Labarge (11 miles south), 21. Sublette County: Big Sandy Reservoir, 11; Half
Moon Lake, 11; Sweetwater River , 2. Sweetwater County: Green River, 54; Green River (26 miles south), 6;
Old Ford on Green River, 34; Sweetwater County, 7. Uinta County: Fort Bridger, 64; Lyman, 4.
Cicindela oregona marie opa Leng. United States . ARIZONA:
Coconino County: Grand Canyon, 1. Gila County: Globe, 56; Roosevelt Lake, 4; San Carlos, 1; Sierra Ancha
Mountains, 7. Graham County: Rylas, 8. Greenlee County: Clifton, 9. Maricopa County: Phoenix 25 ;
Tempe, 1. Pima County: Tuscon, 5. Pinal County: Pinal Mountains, 7. Yavapai County: Cottonwood, 1;
Haslampa, 3; Prescott, 266. Localities of unknown counties: Bad Creek Canyon, 1; Bradshaw Mountains, 1;
Mogollon Mountain, 1; Oak Creek Canyon, 2. CALIFORNIA: Los Angeles County, 1. San Bernardino County:
Barstow, 1. NEVADA: Lincoln County: Caliente, 13; Meadow Valley, 1. NEW MEXICO: Otero County:
Cloudcroft, 2; Mountain Park, 15. UTAH: Iron County: Cedar City Canyon, 2.
Cicindela oregona navajoensis Van Dyke. United States.
ARIZONA: Navajo County: Betatakin, 9; Kayenta, 25; Navajo Mountain, 6. Moffat County: Echo Park, 2;
Massadona, 4. Montezuma County: Four Corners, 4. UTAH: Grand County: Floy, 40; San Juan County:
Blanding (10 miles west), 4; National Monument, 1; Navajo Mountain Trading Post, 7.
138
Cicindela maritima Group
Cicindela oregona oregona x Cicindela oregona guttifera
Canada. ALBERTA: Banff, 52; Laggan (Lake Louise) , 22; WatertonPark, 5. BRITISH COLUMBIA: Athalmer
24; Canal Flats, 8; Cinema, 7; Fernie, 25; Fort Fraser, 13; Kootenay National Park, 6; Moyie, 2; Terrace
35; Wasa, 7; Yoho National Park, 17.
United States. IDAHO: Bannock County: Pocatello, 7. Bear Lake County: Bear Lake, 119
Bloomington Lake, 10. Franklin County: Franklin Basin, 11. Fremont County: Parker, 2. Lemhi County
Salmon (21 miles north), 6. MONTANA: Cascade County: Great Falls, 2. Flathead County: Hungary Horse
11; Kalispell , 1. Gallatin County: Bozeman, 2; Gallatin County, 2; Gallatin River and Highway 10, 14; Lak.
Hebgen, 12; Missouri River (headwaters), 2; Three Forks (3 miles west), 13. Glacier County: Lower Medicin.
Lake, 66. Lewis and Clark County: Craig, 14; Helena, 50; Hardy (15 miles south west) , 33. Lincoln County
Troy, 1. Missoula County: Frenchtown, 1. Park County: Gardiner (5 miles north), 86. Ravalli County:
Darby, 2; Florence, 2; Hamilton, 3; Ravalli County, 2; River Bottoms, 10; Skalkaho, 1. Sanders County:
Whitepine, 1. Silver Bow County: Butte, 1. Localities of unknown counties: Beaver Creek, 8; Bitter Root
Mountains, 2; Lost Horse Canyon, 1; Marias River, 2; Stickney Creek, 34. WYOMING: Teton County:
Black Rock Creek, 4; Grand Teton National Park, 13; Hobach Canyon, 3; Jackson Hole National Monument, 27;
Moran (38 miles east), 46. Yellowstone National Park, 140.
Cicindela oregona oregona x Cicindela oregona maricopa.
United States. CALIFORNIA: San Diego County: San Diego, 45.
Cicindela oregona guttifera. X Cicindela oregona maricopa.
United States. UTAH: Beaver County: Beaver (4 miles east) , 11; Beaver Canyon, 6; North Creek, 1. Iro:
County: Burkshire, 1; Cedar City, 9; Iron Springs, 2; Parowan (5 miles southeast), 7; Pa-rowan Canyon, 12
Kane County: Glendale, 1; Kanab, 1; Orderville, 2. Washington County: Bellvue, 2; Pine Valley, 2; Pintura
3; Saint George, 8; Santa Clara, 3; Zion National Park, 26. Localities of unknown counties: Mount Meadows
1; Weeping Rock, 1.
Cicindela oregona guttifera X Cicindela oregona navajoensis.
United States. COLORADO: Mesa County: De Beque, 4; Palisade, 7. NEW MEXICO: McKinley County
Fort Wingate, 33. San Juan County: Farmington, 3. UTAH: Uintah County: Dinosaur National Monument
1; Vernal, 12.
Cicindela oregona navajoensis x Cicindela oregona maricopa.
United States. ARIZONA: Apache County: White Mountains, 8. Cochise County: Chiricahua Mountains, 2,
Navajo County: Carrizo, 8; Cibeque Creek, 2. NEW MEXICO: Catron County: Luna, 5. Brant County
Silver City, 6.
The Species Cicindela depressula Casey.
Cicindela depressula depressula Casey 1897 :297 . Type locality - Placer County,
California. Leng 1902:150. Rivalier 1954:253.
Cicindela oregona depressula Hatch (not Casey 1897) 1953:42. Horn
1930:82. Wallis 1961:24.
Cicindela depressula eureka Fall 1901. Type locality - Humbolt County, Calif-
ornia. NEW COMBINATION.
Cicindela eureka Fall 1901:307. Leng 1902:149. Horn 1930:82.
Rivalier 1954:253.
Two constant differences set apart depressula from other species
of the maritima group. First, in depressula two or three hairs usually occupy
the small area near the front inner edge of each eye; four hairs are seldom
present. If these hairs are abraded, setigerous punctures indicate
their former positions. Second, the distal end of the median lobe of the
male genitalia of depressula has two distinct, broad flanges that form a
blunt apex (for details see fig. 3, and p. 91 ). A partially diagnostic
character is the form of the middle band of the elytra. In Oregon and
California the middle band of depressula tapers evenly posteriorly. This
contrasts with the sharp bend in the middle band of oregona. On the other
hand in Washington, Canada, and Alaska the middle band of depressula
often appears identical with that of oregona.
I collected oregona depressula eurekaon the same sand bank along
the Van Duzen River near Bridgeville, California; and Ball (personal
communication) collected d. depressula and o. guttifera in the same area at
Ter race, British Columbia, and at Haines, Alaska. There was no evidence
of hybridization or cross-mating at these locations, and this suggests
that depressula and oregona are specifically distinct in spite of their many
shared characteristics.
Fr eitag
139
Notes on Synonymy
The tiger beetles called depressula and eureka are very similar to one
another, differing mainly in color and markings of the elytra and in
seasonal occurrence of adults. They are also allopatric. The differences,
however, are not absolute; that is, range of variation in the diagnostic
character s of the two forms is slightly overlapping. Furthermore, Rumpp
has from the Olympic National For est, Washington, a series of specimens
interpreted as hybrids between depressula and eureka. These considerations
of variation and distribution suggest that depressula and eureka are con-
specific, but are subspecifically distinct. Rumpp and I have reached
this conclusion independently.
Geographic Variation and Subspecies
Eleven population samples, whose geographic positions collect-
ively span the known range of this species were selected on the basis of
geographical location and number of specimens and were examined for
variation. Elytral size, elytral color, pleural color, and the condition
of the middle band vary. There maybe a humeral dot on the elytra.
Variation in these was analyzed and the results are summarized in
tables 6-8.
Length of elytra was examined in a cursory fashion (table 6).
Mean values for females are higher than those for males from the same
locality. Mean values for samples from lowland, coastal regions in
Humbolt County, California are normally higher than mean values for
populations from Mount Rainier, Washington. This is the reverse of
the pattern of the size variation in oregona.
TABLE 6 - Variation in length of elytra of male and female d. depressula
and d. eureka.
Populations of the Cascade Range differ in phenology from those
of the Pacific coast of Oregon and California. At these latitudes pop-
ulations from these regions do not meet, for apparently there is a
differ ence in seasonal occurrence of adults. Most adults in alpine regions
are active during the middle and late summer but in coastal populations
the adults are out earlier in the year. On a collecting trip in June
1963 depressula was not found at high elevations, but specimens of
140
Cicindela maritima Group
coastal populations were collected in northwestern California.
Elytra of depressula have a metallic lustre and are either brown,
green or blue (table 7). There is no evidence of a uniform character
gradient. Brown elytra are prevalent in coastal regions of northern
California, on Mount Hood, Oregon and on Mount Rainier, Washington,
Specimens with green elytra occur throughout the range of the species ,
but are most frequent at both northern and southern extremities of
the range, e, g. Haines, Alaska and Eldorado County, California. Blue
elytra are most common in populations fromWashington and southern
British Columbia and as such parallel those of oregona in southwestern
British Columbia. In California, over 90 per cent of specimens from
Cascade Range localities are green whereas over 90 per cent from the
coast are brown.
Color of thoracic pleura is rather variable in every population
listed in table 7 except for Lassen National Park and Eldorado County,
California. Specimens with green and coppery thoracic pleura are more
common than those with blue.
Data on variation in the humeral dot are presented in table 8.
The humeral dot is present in all specimens collected in lowland localities
near the Pacific coast. Among specimens. from localities in the Cascade
Mountains, the dot may be present or absent. Over 90 per cent of the
specimens from north of and including Mount Baker, Washington have
humeral dots on the elytra, but only 10 per cent of the specimens from
southern Oregon and California have them. In the Mount Rainier, Wash-
ington and Mount Hood, Or egon samples the two conditions occur in about
equal frequencies. The variation appears to be clinal.
The middle band of the elytra may be broken or complete,
and the data on the frequencies of these conditions in various population
samples are presented in table 8. In general the broken condition of
the middle band is more common southward, among montane populations,
but this is a poorly marked trend. All of the specimens from coastal
regions in California have a complete middle band.
Variation in hair between the eyes is not tabulated. However,
between the eyes, in the middle of the head, one to four very fine hairs
are present in coastal specimens' from southern localities. The hairs
rarely appear on specimens from northern montane regions, or on
specimens from montane localities in California. These hairs should
not be mistaken for the setae near the front inner edge of the eyes that
are characteristic of depressula as a whole.
Most northern populations studied exhibit appreciable variation.
The five specimens from Haines, Alaska are less variable but more
material is necessary from northern localities to obtain a better knowl-
edge of the degree of variation. In the central portions of the range,
variation is discordant. Southern populations, however, are remarkably
uniform.
Of the characters considered above, specimens collected in
southern areas of the Cascade Range vary mainly in the condition of
the middle band. Almost all of these specimens have the elytra and
thoracic pleura green, and no humeral dot. These characters are
generally most common in northern populations.
Fr eitag
141
Mountain populations of California and Oregon are distributed in
a thin band along the Cascade Range where they are confined to high
altitudes, but further north they also occur in the Rockies (Mount
Revelstoke, British Columbia) and altitudinal prefer ence is not as marked
as in southern regions (fig. 33).
TABLE 7 - Color variation of elytra and thoracic pleura of population
samples of Cieindela depressula ,
TABLE 8 - The occurrence of elytral humeral dot and variation in the
condition of the middle band of elytra among population
samples of Cicindela depressula
142
F r eitag
143
Coastal specimens from California differ in external character-
istics from alpine individuals at the same latitude . Specimens from these
lowland regions usually have brown elytra, a very long, usually complete
middle band, a humeral dot, and coppery, green, or blue thoracic pleura.
They occupy river banks near the Pacific coast of northern California,
Oregon and Washington. Few specimens of this type have been collected
in Oregon and Washington. They may not be abundant in these regions;
or perhaps the adults come out only for a short period. On the other
hand, more intense collecting in May and June may yield larger numbers
of these forms from these areas.
These differences provide the basis for distinguishing two
subspecies: a southern coastal one, d. eureka Fall and a mountain northern
one, d. depressula Casey.
Variation in some morphological char acter s , phenology, hybrid-
ization, and distribution of depressula and eureka have been discussed above.
However, the apparently restricted distribution of eureka should be dis-
cussed further.
Distribution limits of eureka north of California are poorly under-
stood because material is very scanty. In June 1963, a collecting
expedition was made to the American southwest in order to obtain speci-
mens of depressula and oregona . While in Humbolt County, California we
obtained several large series of eureka near the mouths of Redwood
Creek, and Mad and Van Duzen Rivers. Travelling from west to east
in the Van Duzen River valley, we collected eureka as far east as
Bridgeville which is approximately 1, 000 feet above sea level and 30
miles east of the Pacific coast. Fifteen miles east of Bridgeville at
an elevation of about 2, 000 feet eureka was not present, nor was eureka
present along Redwood Creek, 17 miles east of the Pacific coast. Thus
the eastern limit of eureka is a short distance from the coast. At this
latitude average temperature differences due to altitude may play a part
in limiting eureka so closely to the coast. However, much ecological in-
formation is essential in order to understand the forces which confine
eureka to such a restricted region in California.
C. depressula evolved in wester n North America probably as a cool-
adapted lowland species, having shared a common ancestry with
duodecimguttata . During a fair ly r ecent glacial stage the range of the species
was bisected by mountain glaciers with survivors to the west and to the
east of the Cascades at low elevations. During this period of isolation
differentiation occurred, with the coastal populations evolving the least
in color pattern, but becoming bound to climatic conditions existing at
lower elevations. This group became the subspecies eureka . As the
glacier s retreated and the inland refugium became warmer and drier the
populations isolated there {d. depressula) movedup the mountains or north-
wards or both ways. Differentiation then occurred in d. depressula with a
los s of white markings in the southern member s. In northern Washington
the ranges of the two isolates met and hybridization took place.
Distribution patterns like that of depressula are evident in verte-
brates such as Sorex vagrans Baird (Findley 1955), R ana aurora , Baird and
Girard, and Contia tenuis Baird and Girard (Stebbins 1954).
144
Cicindela maritima Group
Distribution
Of the 922 specimens of depressula examined, one appears to be
incorrectly labelled Berkeley, Alameda County, California. This
specimen does not resemble the subspecies d. eureka but it is phenotyp-
ically d. depressula the distribution of which is restricted to the Cascade
Range at that latitude.
Cicindela depressula depressula Casey. Canada. BRITISH COLUMBIA:
Diamond Head Trail, Garibaldi Park, 14; Jade Lake Trail, Mount Revelstoke, 1; Terrace, 4.
United States. ALASKA: Haines Highway, near Haines, 5. CALIFORNIA: Alpine County: 7.
Eldorado County: Echo Lake 15; Fallen Leaf Lake, 2; Keith Dome, 11; Mount Tallac, 10; Summit, 3. Mariposa
County: Kerrick Meadows, Yosemite National Park, 2. Nevada County: Rucker Lake, 1. Placer County:
Summit, 2. Shasta County: Kings Creek, Lassen Park, 6; Mount Lassen, 18. Siskiyou County: Walker, 3.
Localities of unknown counties: Angora Park, 3; Carson Pass, 2; Charity Valley, 1; Sovoft, Sierra Nevada
Mountains, 1; Warner Valley, 2. NEVADA: Washoe County: Mount Rose, 2. OREGON: Deschutes County:
Three Creeks, 1; Todd Lake Meadows, 1. Douglas County: Diamond Lake, 1; Three Lakes, 1. Hood River
County: Mount Hood, 54. Jackson County: Mount Ashland, 2. Jefferson County: Mount Jefferson, 1.
Josephine County: Rogue Riffles, 1. Klamath County: Crater Lake, 60; Summit Lake, 2. Lake County:
Linton Meadows, near Three Sisters area, 17. Lane County: Obsidian Trail, 1; Scott Lake, 9; Wikiup Plains,
5. Linn County: Big Lake, 1; Hoodoo, 2; Santiam, 1. Yamhill County: McMannville, 1. WASHINGTON:
Clallam County: Forks, 6. Clark County: Vancouver, 4. Cowlitz County: Silverlake, 2. King County:
Enumclaw, 1; Red Mountain, 1. Kittitas County: Cle Elum, 1. Pierce County: 1; Long Mire, 3; Mount
Rainier, 461. San Juan Islands, 1. Skamania County: Little Huckleberry Mountain, 1. Snohomish County:
Arlington, 3; Soda Springs , 1; Sultan, 1. Whatcom County: Mount Baker , 9. Yakima County: NachesPass,
4. Localities of unknown counties: Chinook Pass, 8; Greenwater, 2; Lake Cushman, 3; Mora, 2; MountAdams,
13; Pilchuck Mountain, 1; Stillguamish, 2; Verlot, 1.
Cicindela depressula eureka Fall. United States. CALIFORNIA: Humbolt County: Alton, 2; Areata ,
Mad River, 39; Blue Lake, 4; Bridgeville near Van Duzen River, 19; Fortuna, 1; Orick, Redwood Creek, 12;
Scotia, 1; Van Duzen River, 27. Del Norte County: Requa, 3; Terwah, 1. OREGON: Benton County:
Corvallis, 1. Linn County: Albany, 1. WASHINGTON: King County: Seattle, 1. Whatcom County: Naches
Pass, 1.
Introduction
HYBRIDIZATION BETWEEN C. OREGONA AND
C. DUODECIMGUTTA TA
Cicindela duo decim guttata ranges across most of eastern and central
North America from Texas to northern Canada, and from the eastern slopes
of the Rocky Mountains to the Atlantic (fig. 17) .Cicindela or egona occupies
regions in and west of the Rocky Mountains to the Pacific coast from
southern United States to Alaska (fig. 18).
During the glacial stages of the Pleistocene the two forms were
probably isolated from one another, oregona to the west of the Rockies
and duo decim guttata to the east of that mountain range. Since Pleistocenetim.es
their ranges have expanded and have come together forming a zone of
inter gradation on the eastern slopes of the Rocky Mountains, that extends
from Colorado to northwestern Canada. This hybrid zone is approximately
50 miles wid.e in the North Saskatchewan River valley in western Alberta,
but it is nearly 1, 000 miles wide in northwestern Canada. As far as
is known hybridization between these two species occurs in all areas of
sympatry.
This study is based on 19 population samples comprising 1,731
adult specimens of which 1, 291 were collected in Alberta, 2 05 in the
Northwest Territories, 75 in British Columbia, 70 in Montana, 61 in
Saskatchewan, 27 in Alaska and 3 in Colorado (see fig. 34). Additional
material from areas east and west of the hybrid zone was obtained from
various North American institutions and has been analyzed in the oregona
and duo decim guttata taxonomic sections.
Adult male and female duodecimguttata are alike in color , color pattern,
and distribution of hair on the head. Hairs cover the frons, top of the
145
Fig. 34. Locality map of population samples analyzed by hybrid index method. Dark areas
indicate elevations above 5, 000 feet.
146
Cicindela maritima Group
head and postgenae.In the western portions of the range of this species
individuals are brown dor sally and metallic blue- green ventrally, the
prothoracic pleura are coppery, and they have complete elytral patterns
(for details see p. 102 ).
The species duodecimguttata hybridizes with the subspecies oregona
guttifera. Both sexes of this subspecies have similar external features.
Hair s are not present on the post genae, fro ns or top of the head, but a few
occupy a small area near the front inner edges of the eyes. The color
is identical to that of duodecimguttata- brown dor sally and blue- green ventrally,
and coppery prothoracic pleura. A reduced elytral pattern is character-
istic of oregona with humeral and apical lunules each represented by two
dots, and the marginal band absent.
Elytral pattern, and distribution of hair on the head were used
to develop a hybrid index for duodecimguttata and oregona, Male genitalia were
not used because it seemed preferable to use characteristics occurring
in both sexes. High values were assigned to the characteristics of
duodecimguttata, low to those of oregona, Intermediate expressions of these char-
acteristics were scored with intermediate values. These characteristics
are illustrated in figs. 11 to 16 and details of assignment of values are
given in table 9. In population samples of non-hybrid duodecimguttata from
western localitiespmdex values range from 4 to 7;and in non-hybrid popula -
tions of oregona , values range from 0 to 1.
TABLE 9 - Values assigned to diagnostic characters of C. duodecimguttata
and C. oregona used in determination of compound character
indices.
Characteristics of oregona and duodecimguttata occur in many recomb-
inations in the hybrids. Many specimens are like one of the parental
types except for one character while others have index values of 2 or 3.
Specimens that have hairs on the frons and head often have hairy post
genae. This could be a pleiotropic effect of a single gene, but since
the association is inconstant the post genae and frons are treated as
separate characters.
Variation in Space
A hybrid index value was determined for each of the 1,731 spec-
imens. A histogram shows the percentage of specimens per index value
for each population sample (figs 35-44). The localities from which
population samples were collected are illustrated in fig. 34 except
those that are represented by only a few specimens.
147
0.11/mi.
0.11/mi.
Av. 0.74
0.03/mi.
Av. 0. 02
I
O.OOOl/mi.
I
Av. 0.00
Fig. 35. Changes in the frequency distribution of hybrid index values in population samples of C. diwdecimguttaia and C. oregona bet-
weenSaskatoon and Vancouver. Average hybrid indices and the change in hybrid index per mile on the right, number of speci-
mens and air miles between localities on the left.
148
Cicindela maritima Group
cussion; a transect from Vancouver, British Columbia to Saskatoon,
Saskatchewan, via the Saskatchewan River System; the Belt Creek,
Montana sample; the Boulder, Colorado sample; and a transect from
Tanana River, Alaska to Fort Smith, Northwest Territories through
Norman Wells, Northwest Territories.
Vancouver -Saskatchewan River Drainage Transect
In figure 35 three portions of the transect serve to illustrate
spatial phenotypic changes between uncontaminated populations of oregona
and uncontaminated duodecimguttata through a zone of inter gradation centered
in the upper regions of the North Saskatchewan River valley. The trans-
ect lies on a west-east plane with the geographically extreme points
Vancouver in the west and Saskatoon in the east. The first part includes
population samples from Vancouver and Creston. The second portion of
the transect is along tne western regions of tne North Saskatchewan River
valley and includes population samples from Kootenay Plains, Nordegg,
Saunders, Rocky Mountain House, and Lodgepole. The Edmonton and
Saskatoon samples constitute the eastern portion of the transect. Air
mile distances and index changes per mile between localities are also
given in the figure. Histograms illustrating variation in population samples
from Lethbridge, Saskatchewan Crossing, Garth, Crimson Lake, and
eight miles south of Lodgepole provide additional data.
Specimens collected in Vancouver, British Columbia on a sand
bank bordering an inlet of the Pacific Ocean all score 0 indicating a
pure oregona population.
The sample collected at Creston, 285 air miles east of Vancouver,
is composed of 48 specimens of which 21 were taken by Stace-Smith in
1945 and 27 by Ball in 1957. Only one specimen scores 1, while the
remaining score 0, Average index change per mile, from Vancouver
to Creston is 0. 0001 which is negligible and can be attributed to natural
variation in the Creston population.
In 1962 specimens were collected near the Kootenay Plains, 20
miles down stream from the junction of the Banff- Jasper highway and
the North Saskatchewan River. River banks are steep near the Plains
and are cover ed with under growth to the edge. These banks are therefore
not suitable for riparian tiger beetles such as oregona. However, many
islands formed by sand and clay deposits occupy the river bed and divide
the river into several large streams. Most of these islands are sparsely
covered with grass and shrubs, and river debris such as drift wood
is plentiful. The islands proved to be suitable habitat for oregona and
another riparian species, C. repanda. One specimen scores 1, the other
13 all score 0. The average index value for the sample is 0. 07.
Sand and mud islands that characterize the North Saskatchewan
River near Kootenay Plains are also present 24 miles down stream at
the Forestry Trunk Road crossing near Nordegg. Two islands divide
the river into three large streams. Three road bridges link the islands
with one another and the river banks. The southern island bears large
shrubs which occupy higher central portions, while grass is abundant
throughout. Much of the west half of the island is a clay flat that is
periodically covered with water when the river rises. The soil there
is basic with a pH of 8. 2. There is little organic matter in the soil but
Fr eitag
149
much free lime. Riparian tiger beetles are abundant on the island esp-
ecially along the clay flat. A series of 174 specimens was collected in
1963;the population sample is variable, composed of oregona, duodecimguttata,
and hybrid individuals. The index values range from 0 to 7, with the
average at 0. 74. Of the population sample 76 per cent score 0,
four per cent score 7, and 26 per cent have intermediate values. The
average index value increases 0. 03 units per air mile from Kootenay
Plains to Nordegg, which is 300 times greater than that from Vancouver
to Creston.
Fifteen air miles east of Nordegg near Saunders the North
Saskatchewan R iver is narrow and there are no islands. The north shore
is a sandy strip several to 10 feet wide and littered with drift wood.
The river banks descend sharply to the sandy shore line, and are covered
with undergrowth for the most part but grassy clearings are present.
These clearings are probably periodically inhabited by riparian tiger
beetles when the river is high and covers the sandy shore margins.
C. oregona , duodecimguttata and hybrid individuals ar e abundant on the beach
at Saunders from which a series consisting of 64 specimens was taken
in 1963.
The range in index values is 0 to 7 with the average at 2. 31. This
is a mean index increase of 0. 11 per mile from Nordegg to Saunders,
that is , more than three times the average index increase per mile from
Kootenay Plains to Nordegg. Specimens with an index of 0 comprise
43. 8 per cent of the population sample, while those which score 7 con-
stitute 15. 6 per cent. Individuals which score 2 to 6 inclusive make up
40. 6 per cent of the sample. No specimen has an index value of 1.
Thirty-four miles east of Saunders, at the Highway 11 bridge
near Rocky Mountain House, the river banks are flatter and broader
than those upstream. A mixture of clay, sand and loose gravel forms
the south banknear the bridge. Open patches on which tiger beetles are
active are common.
The index values range from 0 to 7 and the average is 6. 14.
Specimens scoring 7 comprise 64. 7 per cent of the population sample
while individuals with a value of 0 constitute 1. 5 per cent. Most
hybrids r es emble duodecimguttata more closely than oregona. Although non-
hybrid oregona is scarce in this predominantly duodecimguttata p o pu 1 a t i o n
many hybrid individuals are present.
Ten miles south of Lodgepole large sand banks flank the north
side of the Brazeau River adjacent to the Brazeau power house. Because
a dam has been erected further upstream, only a little water is present
near the power house and much of the river bed is exposed. On August
12, 1963 a population sample was collected on the sand banks near the
Brazeau power house.
The specimens have index values that range from 0 to 7 but none
score 4 or 6. The average index is 2.29. Because the Brazeau power
house is located near the North Saskatchewan R iver between R ocky Mount-
ain House and Edmonton, it would appear that the mean index value of a
population sample from that area should fall between 6. 14 and 6. 77, the
average indices of Rocky Mountain House and Edmonton respectively.
150
Cicindela maritima Group
In fact, the average index is approximately 4 less than expected.
This may be in part because the Brazeau power house locality although
downstream from Rocky Mountain House is 17 miles west of it. In
addition, oregona and duo decim guttata exhibit different habitat preferences
(see locality eight miles south of Lodgepole). The pure sand bank near
the power house is particularly suitable for oregona and harbors a pop-
ulation with a lower average index value than that at Rocky Mountain
House.
A sample was collected in 1961 near the Groat Bridge in Edmonton.
The average index value is 6.77 and the range is 4 to 7 inclusive. This
may be the result of hybridization with oregona but is more likely natural
variation.
The mean index value of the Saskatoon sample is 6.76 which is a
change of 0. 00003 index units per mile from Edmonton to Saskatoon a
distance of 310 air miles. A small percentage of the sample with values
of Z and 3 is interpreted as natural variation in a non-hybrid duo decim guttata
population. It is the result of breakdown of the dlytral markings, which
is probably caused by duo decim guttata genes infiltrating from the east, rather
than oregona genes from the west.
Additional localities in western Alberta
Ten specimens were collected near the junction of the North
Saskatchewan R iver and Banff- Jasper Highway about Z0 miles west of the
Kootenay Plains. All the individuals score 0 (fig. 36).
Garth is approximately three miles upstream from Rocky Mount-
ain House. The river on the north side is a clear stretch of sand and
clay. In May, 1963 a sample was taken on this beach. The index values
range from 0 to 7. Most specimens score high and the average index
is 5. Z5 (fig. 37). If this is compared with the histogram for the
Rocky Mountain House sample (fig. 38), that was collected on the same
day, the mean index difference is 1. Z4.
Crimson Lake is about nine miles north of Rocky Mountain House.
At the east end the water front is sandy. For about 500 feet and at the
south end of this beach the sand is light in color and loose and C. repanda
is abundant. Toward the west end of the beach the ground is a hard and
dark mixture of sand and clay and is flanked by a marshy area. Many
duodecimguttata individuals were active in this area of the beach and repanda
was absent. Samples were collected at Crimson Lake in 1961, 196Z,
and 1963. The range of average index values is 6. 14 to 6. 55 (fig. 39).
Several oregona specimens were collected here but they are rare. The
population is mainly duodecimguttata with evidence of oregona gene infiltration.
Between Lodgepole and the Brazeau power house many ponds
are scattered beside the gravel highway. These are water-filled gravel
pits that were dug out for road construction. The soil around the ponds
is normally hardened clay covered with grass. The same day the
Brazeau power house specimens were taken, I also collected 17 specimens
near one of these roadside ponds eight miles south of Lodgepole. The
hybrid index values of the roadside samples range from 3 to 7, and the
average index is 6. 36 (fig. 40). This index average is 4. 07 more than
the mean index of the Brazeau power house population.
The marked average index difference between populations of the
F r eitag
151
Rocky Mtn. House ,
Alto. May, 1963
N 31
38
Av. 6.49
Figs 36-38. Frequencydistribution ofhybridindexvalu.es in population samples of, 36, C. oregona from Saskatchewan Crossing,
Alberta; 37, C. oregona X C. duodecimguttata from Crimson Lake, Alberta. N. = no. of specimens.
Brazeau power house and roadside ponds is indicative of different habitat
preferences of the two species. Populations of oregona are best adapted
to pure sand and clean gravel, commonly found along mountain streams.
The species duodecimguttata normally inhabits edges of lakes, sloughs, and
rivers where there are usually flat clearings of dark sand, clay or mud.
Wherever these two habitats are available together, at least near the
North Saskatchewan River, duodecimguttata and oregona hybridize. Because
the soils around the roadside ponds between Lodgepole and Brazeau power
house are predominantly clay, they were probably first inhabited by
duodecimguttata which invaded these ponds from sloughs etc. nearby. How-
ever the effect of oregona is evident in these roadside populations.
The Lethbridge sample consists of 310 specimens that were
152
Cicindela maritima Group
Crimson Lake/Alta..
Juiy • 1963
N 38
Av. 6.53
Av. 6.55
Av. 6.43
Fig. 39. Frequency distribution of hybrid index values in five population samples of C. oregon a X C. duodecimguttata from Crimson
Lake, Alberta. N. = no. of specimens.
collected throughout the summer of 1962 (fig. 41). More than 84 per
cent of the specimens are pure duodecimguttata. No individuals score
0 or 1. The index range is 2 to 7 inclusive with the average at 6.7.
The histogram indicates that some introgression from oregona is influ-
encing this predominantly duodecimguttata populati on. In I960 two oregona
specimens were collected near Lethbridge. The occurrence of these
specimens supports the supposition that variation in Lethbridge is the
result of hybridization between oregona and duodecimguttata.
High River is 30 miles south of Calgary. Of the four repre-
sentative specimens there are two hybrids, one oregona , and one
duodecimguttata.
In 1925 F. S. Carr collected eleven specimens from Happy
Fr eitag
153
40
8 miles s. Lodge -
pole , Alta. July,
1963
N 19
2
3
Av. 6.06
8 miles s. Lodge -
pole, Alta. August ,
1963
N 17
Av. 6.36
42
Av. 1.2 3
Figs 40-42. Frequency distribution of hybrid index values in population samples of, 40, C. oregona X C. duodecimguttata from 8 mi.
south of Lodgepole, Alberta (2 samples); 41, C. oregona X C. duodecimguttata from Lethbridge, Alberta; 42, X
C. duodecimguttata from Belt Creek near Armington, Montana. N. = no. of specimens.
Valley which is near the Bow River approximately four miles west
of Calgary. The sample consists of six oregona, two duodecimguttata and
three hybrids.
Three specimens were taken in 1961 by Wu near Ricinus along
the Clearwater River, 20 miles south of Rocky Mountain House. Two
individuals are duodecimguttata and one is a hybrid.
Two specimens of duodecimguttata and one hybrid were collected at
Beaver Creek, Alberta. The locality and collector of this small
sample are both unknown.
Belt Creek, Montana
In 1963 a sample was collected along Belt Creek just east of
154
Cicindela maritima Group
Arlington, Montana. Specimens of both oregona and duo de dm guttata were
taken with hybrids. All of the index values are represented, but oregona
specimens represent 64. 3 per cent of the sample while two indiv-
iduals score 7 and constitute 2. 9 per cent of the series. The index mean
is 1. 23 (fig. 42).
There are mean differences in index values between males and
females. Values for males do not exceed 5, and the mean index value
is 0.70. Thirty-four females show a range in values from 0 to 7 and
their average index is 1.79;thatis, 1. 09 more than the males which total
36. In other samples, the differences between males and females is less.
Boulder, Colorado
The southernmost hybrid sample is represented by two
duodecimguttata specimens and one hybrid which were collected four miles
north of Boulder, Colorado in July I960. A histogram is not provided
for this sample.
The Alaska - Fort Smith Transect
This transect, composed of population samples collected at the
Tanana River, Alaska, Norman Wells, Northwest Territories, and Fort
Smith, Northwest Territories extends over a range of about 1, 000 miles
(fig. 34). Index values were determined for all of the specimens and a
histogram is presented for each of the three samples (fig. 43). Included
in the figure are air mile distances and index changes per mile between
the localities.
In 1958 Ball collected a series of specimens at a junction of the
Tanana River and the Alaska Highway in southeastern Alaska. Specimens
that have an index value of 0 constitute 88.9 per cent of the sample, while
two individuals, representing 11. 1 per cent each score 1, because both
have hairs on their heads. The average index value is 0. 11.
Norman Wells, situated near the Mackenzie River, is approxi-
mately 470 miles east - northeast of the Tanana River locality. Index
values range from 0 to 7 with the average at 2. 68. The average index
change per mile from Tanana River to Norman Wells is 0. 00547. The
sample exhibits a great amount of variation and hybrids outnumber the
parental forms. One specimen scores 0 and two members each have
values of 7. Thus the parental specimens together constitute only 6.9
per cent of the sample. In contrast individuals that score 2 occur in the
greatest frequency and make up 69.9 per cent of the sample.lt has been
pointed out earlier that natural variation in an uncontaminated duodecimguttata
sample may include specimens indexed from 2 to 7. Similarly a pure
oregona population can have individuals that score 1 as well as 0. Thus
a true hybrid is considered to have a value of 2 or 3. Because speci-
mens with scores of 2 and 3 dominate the Norman Wells sample it is
regarded as a predominantly hybrid population sample and is the only
one of its kind in this study.
A sample was collected at Fort Smith near the Slave River, 580
miles southeast of Norman Wells. Index values range from 2 to 7 with
the average at 5.97. From Norman Wells to Fort Smith the change per
mile is 0. 00567 units. No specimens have values of 0 or 1. Individuals
that score 2 occur in lowest frequency while those with a value of 7 are
155
Fig
Nor
air
Tanana R., Mi. 1281
Alas. Hway. Alas.
ft-
43
0 12 3 4 5 6 7
Norman Wells,
N.W.T.
N 43
580 miles
Fort Smith, N.W.T.
N 152
■_
A v. 0.11
0.00547/mi.
5 6 7
Av. 2.68
0.00567/ mi.
Av. 5.97
0 12 3 4 5 6 7
. 43. Frequency distribution of hybridindex values in population samples of C. oregona X C. duodecimguttata from Alaska and the
thwest Territories. Average hybrid indices and the change in hybrid index per mile on the right, number of specimens and
miles between localities on the left.
Nordegg , Alta. 1961 .
N 83
Nordegg , Alta. 1962
N 219
Nordegg , Alta . May,
1963
N 62
Nordegg ,Alta.July,
1963
Nordegg , Alta. Aug.,
1963
N 58
44
Av. 0.97
Av. 0.98
Av. 0.72
Fig. 44. Frequency distribution of hybrid index values in population samples of C. oregona X C. duodecimguttata from Nordegg, Alberta.
N. = no. of specimens.
156
Cicindela maritima Group
most numerous. The number of specimens increases with increasing
index values. The sample is therefore a duodecimguttata one affected by
intr ogres sion of oregona characteristics.
Of the five specimens taken at Canol near the Mackenzie River
opposite Norman Wells there are one oregona, one duodecimguttata and three
hybrids.
One oregona specimen and one* hybrid were taken at Fort Simpson
which is 290 miles southeast of Norman Wells near the mouth of the
Liard River.
Variation in Time
Annual and seasonal variation in index values appear to be typical
of most localities in the western section of the North Saskatchewan River.
Variation is illustrated by histograms for population samples collected
in the Nordegg area, near Rocky Mountain House, at Crimson Lake, and
in an area eight miles south of Lodgepole.
Nordegg
Evidence of extensive hybridization is clearly shown in the histo-
gram for the 1961 sample (fig. 44). Index values range from 0 to 7 and
the average is 2. 60. Specimens of oregona are most common, constituting
39. 8 per cent of the sample. Specimens of duodecimguttata follow in number
and are 25. 3 per cent of the sample. The remaining portion of the series
is formed by intermediate specimens which are mainly at the low end of
the index scale.
The situation is markedly changed in 1962. A shift toward the low
end of the histogram occur s. Specimens that score 0 increase to 74.2 per
cent, while specimens with an index of 7 drop to 6. 8 per cent of the sample.
The average index is 0. 97, a drop of 1. 63 from the previous year. There
is a further depletion in duodecimguttata numbers in 1963 but there is very
little general change in the frequency of indices from that of 1962.
This may be a phase in fluctuating populations of the parental
forms. However many more years of sampling at Nordegg would provide
a clearer picture of annual variation in these populations. Analysis of
Nordegg population samples collected in May, July, and August, 1963,
revealed a slight trend in decrease of average index values throughout
the summer.
Values for the May sample, range from 0 to 7, excepting 4, 5,
and 6, with the average at 0. 98. The sample is therefore predominantly
oregona specimens (71 per cent), some hybrids closely resembling oregona,
and five phenotypic duodecimguttata individuals. In the July sample more
classes are present, and only index value 5 is not represented. The
index value is 0.72, which is 0.26 less than that of May. Specimens
with high index values are all but absent from the August sample. Most
of the specimens are at the low end of the scale with the largest number
at 0.
There also appears to be a seasonal change in the relative numbers
of males and females at Nordegg, The ratio of males to females in the
May population sample is approximately 3 to 4(27 males and 35 females),
but males outnumber females in the July sample, 2 to 1 (36 males and
Fr eitag
157
18 females). In the August series, the ratio of males to females is
approximately 3.5 to 1 (45 males to 13 females).
Rocky Mountain House
The range in index values for the May series is 3 to 7 with the
average at 6.49. Of the sample, 24 specimens score 7. Thus the pop-
ulation sample is basically a duocecim guttata one somewhat contaminated by
oregona genes (fig. 38). The August sample is more variable and all the
index values except 2, are represented. The average index value is 5. 83
which is a decrease of 0. 66 from the May value.
R elative member s of males and females also change seasonally,
and parallel the change which occurs at Nordegg. The ratio of males and
females in May is 1 to 1 (15 males and 16 females), while in August there
are approximately three times as many males as females (26 males and 8
females). This difference however is not statistically significant.
Crimson Lake
The histogram for specimens collected at Crimson Lake in 1961
shows a mainly duodecimguttata population (fig. 39). One specimen has an
index value of 2 and one has a value of 3. The mean index is 6. 50. The
range of index values for 1962 is 3 to 7 with the mean value at 6. 53.
The average index change from 1961 to 1962 - 0. 03 - is quite small.
In 1963 the mean value decreased by 0. 19, and the range in index values
is 0 to 7. The population sample however is largely a duodecimguttata one
much like those of the two previous years. The major difference is that
oregona specimens are present in the 1963 series, but they are rare.
From May to August a general decrease occurred in the mean
index values of the Crimson Lake samples. This seasonal index change
corresponds with that of Nordegg and Rocky Mountain House.
Ten males and 23 females in May, 21 males and 17 females in
July, and 25 males and 20 females in August were collected in 1 9 6 3.
The sex ratio is two females to one male in the May sample, but is one
to one for July and August,
Lodgepole - eight miles south
In 1963, small series of 19 and 17 specimens were collected at
this locality in July and August respectively. Seasonal differences
between the two samples do not coincide with those of Nordegg, Crimson
Lake, and Rocky Mountain House but the samples are probably too
small to indicate real differences. However there seems to be a shift
from a lower average index value earlier in the season to a higher average
value later in the season. Both samples consist mainly of duodecimguttata
specimens but with a few hybrids (fig. 40).
The males and females are present in about equal numbers in
both population samples, with nine females and 10 males taken in July,
and nine females and eight males collected in August.
Notes on mating
During the summer of 1962, in the Nordegg area, 26 pairs of
mating tiger beetles were collected. This is a phenotypically varied
158
Cicindela maritima Group
group, including both parental species and hybrids. Hybrid indices were
determined for the specimens. Then, a chi square test for independence
was applied to find out if specimens of a particular index more often
selected a mate of the same index value. It failed to show any selection
in mating.
On several occasions I have taken repanda in copulation with
oregona and also with duo decim guttata , but no hybrids have been found. It
is doubtful that gene exchange takes place between repanda and oregona or
duodecimguttata, to the extent that it does between oregona and duo decim guttata.
Discussion
The kind of hybridization between oregona and duodecimguttata can be
classified as one of introgr ession (Anderson 1949), and secondary inter-
gradation (Mayr 1942). Introgr ession, as described by Anderson, is
the incorporation of genes of one species into the gene complex of
another species. Mayr states that secondary inter gradation has occurred
when"Two units, now connected by a steeply sloping character gradient
were separated completely at one time and have now come into contact
again after a number of differences have evolved.” Mayr (1963) regards
the species involved in this kind of hybridization as semispecies in that
they show some of the characteristics of a species and some of subspecies.
Many such cases have been described for birds, mammals , fish, amphib-
ians and some invertebrates. These are too number ous to mention
here butmany are citedby Dobzhansky (1951), Mayr (1942, 1963), Mecham
(1961), and Sibley (1961).
The situation in western Alberta and northwestern Canada des-
cribed above seems to be the result of secondary contact between the
formerly isolated vicariant species oregona and duodecimguttata. Their pheno-
typic differences probably arose under different selective forces acting
on allopatric populations. Breakdown of external barriers between them
allowed their ranges to expand and come into contact. This has resulted
inhybridization. Speciation of oregona and duodecimguttata was probably initia-
ted in early Pleistocene times, but the process has not been completed.
,Climatic changes during the Pleistocene undoubtedly had a pro-
found influence on the distribution of these two forms as they did on other
North American animal species (see Blair 1951). Unlike the vertebrates,
however, neither locations of refugia for these tiger beetles nor their
population movements during the Pleistocene are known because of
the lack of a fossil record. The history of this zone of secondary inter-
gradation is therefore speculative, and is based on the present distri-
bution of both species, and current concepts of events during the
Pleistocene. During the early period of oregona subspeciation, populations
of duodecimguttata were not present in western regions they now occupy.
Shortly after the development of o. guttifera, perhaps duodecimguttata reinvaded
western Canada east of the Rocky Mountains. Because, at the present
time, few southern populations of duodecimguttata reach the eastern front
of the Rockies in Colorado and New Mexico, the present western limits
are presumed to be the extent of the western limits of duodecimguttata during
the late Prairie interglacial. If any hybridization did occur in Prairie
times it took place where the two species are presently sympat-
ric. However, any evidence of pre-R ecent introgr es sion would be masked
by the present hybrid belt. Hybridization probably did occur in southern
F r eitag
159
regions during the Wisconsin glaciation since no indication of introgress-
ion is evident in southern populations.
The hybrid belt between oregona and duo decim guttata is widest in
northwestern Canada and narrowest in western Alberta. Individual
specimens of oregona and duo dec im guttata existinall areas of greatest varia-
bility including the Norman Wells population where they are outnumbered
by intermediates. Width of the zone of inter gradation is recognized
as spatial limits of extreme variation.
There does not seem to be any reduced viability or fertility in
the hybrid tiger beetles and they are present in large numbers in the
Norman Wells sample described above. A composite of isolating mech-
anisms, although hardly pronounced in southern populations of oregona and
duo decim guttata, may have become more completely developed than in north-
ern populations of the two species before they made contact. This may
account in part for the varying width of the zone of inter gradation between
oregona and duodecimguttata.
Somewhat analogous is the inter gradation zone of the European
crows Corvus corone cornix and C. c . corone (Mayr 1942, pp. 265-266), and that
of the North American grackles Quiscalus quiscula quiscula and Q. q. ve rsicolor
(Huntington 1952). Dobzhansky (1951) attempts to explain the irregular
width of the inter gradation belt of the crows. He suggests that oldest
regions of the zone are narrowest where isolating mechanisms have
had more time to become established. Mayr (1942) does not believe
this explanation corresponds with the facts presented by Meise, who
observed the width of the hybrid zone of Corvus is determined by local
ecological factor s. Further , narrow stretches of inter gradation occur in
both recent and older parts of the zone. Because in Quiscalus, Huntington
(1952) observed no reduced viability or fertility in the intermediates,
he feels Dobzhansky 's explanation is inadequate in principle, and suggests
that increased mixing due to migration, and selective forces favouring
the intermediate in a rapidly changing environment are the two main
factors affecting the width of area of inter gradation.
Because the width of inter gradation zones is determined largely
by isolating mechanisms, it is appropriate to discuss variation in the
width of the tiger beetle hybrid zone in the light of two sets of theories
on the origin of isolating mechanisms.
For several hypotheses natural selection is believed to be
responsible for the promotion of isolating mechanisms. One representative
hypothesis postulates that intermediates are of lowered fertility or
viability compared to parental forms. From this it is argued that
individuals which enter into mixed pairs will eventually be eliminated
from both populations because the hybrids they produce are being selected
against. In time, as isolating mechanisms are perfected, the zone of
intergradation is contracted. This is essentially Dobzhansky's view.
A second hypothesis treats the origin of isolating mechanisms
as an incidental by-product of genetic divergence in isolated populations
(Muller, 1940) rather than as the direct result of selection for
reproductive isolation. Mayr (1963) points out that many isolating
mechanisms vary geographically.
Because many isolating mechanisms have ecological components,
any changes in incipient species are certain to affect their isolating
mechanisms. The narrowness of the zone in western Alberta can be
160
Cicindela maritima Group
due in part to different habitat preferences (see p. 157 ). Clay, or
mud, or sand with organic material, seem to be preferred by duodecimguttata,
while soils of pure sand or clean gravel are typical oregona habitat.
In the north, where the inter gradation zone is wider, both species may
be more broadly adapted. The broader northern zone may also be an
effect of better adaptation of intermediates to the northern environment
than to that of the south. However, 'in order to understand this zone of
inter gradation more completely, ecological investigations should be
undertaken.
The elytral pattern of duodecimguttata is complete in western parts
of the species range but it is reduced in eastern and southern regions.
The full pattern also appear in the zone of inter gradation. Eastern
duodecimguttata specimens often have oregona - like elytral markings (see
p. 102), This situation may be interpreted as sympatric character
divergence, which may be described as follows. Two closely related
species of animals overlap geographically. Their differences are
emphasized in areas of sympatry so that both species are easily recog-
nized. In ranges where one of the species occurs alone it closely
resembles the other species.
For several reasons it is doubtful that the variation in the elytral
pattern of duodecimguttata is evidence of character displacement. Some
workers observe that character displacement occurs within regions of
overlap (Brown and Wilson 1956, Mayr 1963). The complete elytral
pattern of duodecimguttata is present in the hybrid belt in western Alberta
but it is also characteristic of populations far beyond the zone of over-
lap (fig. 17). In addition variation in elytral pattern of duodecimguttata is
not complemented by similar dines of other characters. For example
the shape of the median lobe of the male is uniform throughout the
range of duodecimguttata except in the hybrid zone where there are many
intermediate shapes ranging from that of oregona to that of duodecimguttata (see
p. 97). Similarly, hairs are present on the frons, top of the head, and
postgenaeof duodecimguttata throughout the species range except in the area
of intergradation. Furthermore, since there is random interspecific
mating in the zone of hybridization, the difference in markings does not
serve as an isolating mechanism.
PHYFOGENY AND ZOOGEOGRAPHY OF THE NOR TH AMERICAN
SPECIES OF THE MARITIMA GROUP
Phylogeny
The ancestral form of the North American species of the maritima
group is necessarily reconstructed from features that are widespread
among extant species because no fossils are available. The rationale
and principles used in re-constructing the characters of a hypothetical
ancestor are explained in Cain and Harrison (I960). The characters of
the ancestral species were probably as follows: dorsum, brown, opal-
escent; venter, metallic blue-green; thoracic pleura, copper colored;
humeral, apical, and middle lunules, and marginal band, complete;
Fr eitag
161
lunules narrowly expanded as shown by hirticollis or repanda; shapes of
the individual markings like those of repanda ; hairs present on the head
between the eyes; features of the male genitalia as they are now;
flanges of the median lobe comparatively narrow like those of
hirticollis or repanda ; fields a , b , and c of the internal sac lightly aculeate;
sclerites 1, 2, 3, 4, and 6 of the internal sac general size and shape of
extent species;sclerite 5 large like those of hirticollis and repanda; sclerite
between 2 and 6 intermediate size between that of hirticollis and oregona;
pronotum of the larva densely pilose. The species was a riparian form,
and it gave rise to three lineages (fig. 45).
The first derivative stock (1) was perhaps characterized by an
alteration of the elytral pattern in which the posterior portion of the
humeral lunule was produced obliquely towards the median line;
within the male’s internal sac, field a and sclerites probably became res-
pectively densely aculeate and considerably reduced; the pronotum of
third instar larvae was probably quite pilose.
This primary stock ultimately gave rise to the species limbata,
bellissima, theatina, and columbica. The species limbata and bellissima appear
to be most closely allied. The subspecies limbata hyperborea, and bellissima each
have the posterior tip of the humeral lunule extending almost to the middle
band; sclerite 5 of the male internal sac has been lost in these species;
the riparian habit was abandoned and both species are sand dune inhab-
itants. Beside the differences in shape of their median lobes, limbata is
very hairy between the eyes and bellissima is less so. In this respect,
bellissima has departed further from the ancestral stock than has limbata'J he
southern races of the latter species, however, have lost almost all of
the dark pigment of the elytra.
The species theatina and columbica appear to be more closely related
to each other than they are to the other two. A humeral lunule whose
posterior tip is briefly extended is a diagnostic feature of theatina and
columbica; a very large triangular sclerite has evolved between 2 and 6 of
the male internal sac, and sclerite 5 has not completely disappeared.
Theydiffer mainly in two characters: theatina is densely hairy between
the eyes and lives on sand dunes , while columbica is sparsely hairy between
the eyes and has retained the riparian ancestral characteristic.
The proposed course for limbata, bellissima, theatina, and columbica is
not presented in a dichotomous scheme in fig. 45 because different arran-
gements canbe devised on the basis of other similarities among the four
species. Distribution of hairs on the head, condition of elytral pattern,
color, or habitat preferences each could be used to erect a different
hypothetical course, but each of these wo Id imply a greater amount of
parallelism or convergence than is required in the scheme Ihave
suggested.
The second lineage (2) is represented by the species hirticollis which
is somewhat remotely allied to the other existing North American species
of the maritima group. This form evolved: a humeral lunule the posterior
tip of which is distinctly hook- shaped; a comparatively pronounced twirl
in sclerite 4, and a very large sclerite between 2 and 6 of the male internal
sac;andithas retained a densely pilose pronotum of the third instar Larva.
The third ancestral stock (3) probably developed or retained a c-
shaped humeral lunule; field a of the male internal sac remained lightly
162
Cicindela maritima Group
aculeate and sclerite 5 increasedin size jSecondary setae probably sparse-
ly covered the pronotum of the third instar larva;and the species was most
likely riparian. This ancestral stock gave rise to the species repanda,
depressula, oregona, and duodecimguttata.
The species repanda appear s to be les s closely related to the other
three species than they are to one another. Within the male internal sac,
sclerite 5 has become very well developed and the sclerite between 2 and
6 has been lost; repanda ranges across North America but no introgr ession
is evident with the other three species in areas of sympatry.
Evolving from a repanda like anc e stor , the stock which gave rise to
depressula developed a median lobe with broad flanges, lost most of the frontal
hair s, and developed a modified pattern of white elytral markings , which
at first were extensive but subsequently became much reduced. Also the
ancestral brown color of the dorsum was replaced by blue and green in
the stock which developed reduced markings, and in the east the lowlands
were abandoned by this form for life high in the mountains. Simultaneously
the larva of this derivative form lost most of the pronotal hair s charact -
eristic of the pronotum of the ancestral stock.
Another derivative stock from a repanda - like ancestor , was the
Fig. 45. Hypothetical phylogeny of the North American species of the martima group,
Fr eitag
163
progenitor of duodecimguttata&ndi oregona. This stock developed at fir st slightly
reduced elytral markings. Subsequently it became divided into two
geographically isolated groups the western of which lost the frontal hairs,
developed strongly reduced elytral markings, and throughout much of its
range the brown color of the dorsum was replaced by green, blue, or
purple, and the pleural sclerites became blue or green. The pronotum
of the larva gradually lost much of the pubescence evident in the ancestral
larva. This western isolate is the species oregona. In the eastern isolate,
the elytral markings were also reduced, and blue and green color of the
dorsum appeared. The mutations producing reduced marking s became
widespread replacing the ancestral condition throughout most of the range
of the species. Hairs on the pronotum of the larva were reduced in
number. This eastern isolate is the species duodecimguttata. Following
a period of separation too short to permit the development of complete
reproductive isolation the eastern and western stocks met one another and
a narrow zone of hybridization developed in the area of contact.
This scheme requires postulation of an appreciable amount of
parallel evolution. Frontal hair s were lost or reduced four times. Green
or blue color of the dorsum was developed six times. Broad flanges on
the median lobe were developed six times. The primitive elytral pattern
was lost five times , but'by two differ ent phyletic branches. In one of these
extensive reduction of lateral pigmentation took place. The other type of
pattern breakdown was developed by increased pigmentation.
Thus these species, together form a structurally uniform group
inwhicha number of similar structures have arisen independently. This
suggests that the group possesses a good degree of evolutionary homo-
dynamy (Bock 1963). This principle is defined as follows: "The number
of times and ease with which an identical or very similar feature may
arise independently within a group depends upon its degree of evolutionary
homodynamy which in turn depends upon its common genetical- develop-
mental potential. " In the light of this principle similar structures that
have arisen independently in the North American maritima group are con-
sidered to be homologous in the broad sense, which is defined by Bock
as follows : "homologous features (or conditions of the features) in two or
more organisms are ones that can be traced back to the same feature
(or condition) in a group possessing a high degree of evolutionary
homodynamy. "
Zoogeography
The following account of the development of the distribution of
the North American species of the maritima group is hypothetical. Move-
ments, and times and place of origin of extant species are necessarily
constructed on the basis of: distribution and morphological features of
the species, geological and climatic events of the T ertiary and Pleistocene
inNorth America, and rates of evolution in some other groups of insects.
Insects generally develop modifications of structural features at a
slow rate. Many fossil species of the early or middle Tertiary closely
resemble existing species (Linsley 1958, Ross 1958, Becker 1963,
Quate 1963, Sabrosky 1963, and Sturtevant 1963). Most of these are
members of recent genera. However, Zeuner (Sylvester - Bradley,
1963) , by means of analysis of 212 species of fossil Apoidea,Lepidoptera,
164
Cicindela maritima Group
and Saltatoria, reckons that excepting the honey-bee all living species
evolved in the Pleistocene. He estimates half to one million years is
a reasonable time required for the evolution of a full species. Zeuner
further notes that no insect species are known with certainty to have
survived from the Miocene (see Zeuner 1943, for more information of the
time factor in evolution of insects). There is no evidence of recent
vigorous evolution within the North Ametican species of the maritima group.
Indeed these are rather primitive in comparison with other species groups
of Cicindela. The ancestral stock of the maritima group may have been
in existence during the early Tertiary. Living species may have evolved
during the later Tertiary or early Pleistocene.
Historical events which may have effected geographical isolation
and subsequent speciation of tiger beetle populations are of importance.
Thus it is necessary to review briefly geological and climatic changes
in western North America during the Tertiary and Pleistocene (see
Blackwelder 1948, King 1958, MacGinitie 1958, Martin 1958, and Mengel
1964).
The Tertiary was marked by several periods of crustal dis-
turbances. Early Tertiary was a time of extensive mountain building
through the west, and it was then the initial Rocky Mountain system
was thrust up. Crustal folding was renewed in the middle Tertiary (late
Miocene). Gentle folding in the Rockies prevailed. Disturbances were
evident in coastal and southeastern California, and southern Nevada,
while other mountains were widely distributed throughout the American
west. A chain of volcanoes was built up along the east flank of the
Sierra Nevada and Cascade Mountains. Large basins were produced,
many of which became lake basins. At the close of the Tertiary (late
Pliocene)once again crustal folding occurred along the Pacific coast, and
in Nevada and Utah. The modern California Coast Range, Wasatch and
Ruby Mountains and many others were elevated during this period. The
southwestern plateau was raised to its present level, and most of the
interior drainage systems were renewed.
Early Miocene and most of the Pliocene were periods of relative
quiet. Stream systems wore down western mountains to scattered hills,
and extensive plains wereformedon which large lakes drained or were
filled.
The climate in the early Tertiary was warmer than now. Tropical
for ests filter ed into the north Temperate Zone while temperate conditions
prevailed in Rocky Mountain regions. In the Miocene the climate became
cooler and temperatures steadily decreased into the Pleistocene.
Simultaneously climatic zones moved southward and southwestern
regions became drier.
The end of the Tertiary and beginning of the Pleistocene was
characterized by the gradual development of mountain glaciers and con-
tinental ice masses. There were five major glacial stages in North
America, the Nebraskan, Kansan, Illinoian, Iowan, and Wisconsin.
Between these occurred long warm periods, the Aftonian, Yarmouth,
Sangamon, and Prairie.
In glacial periods glaciers extended southward along mountain
ranges. These gave rise to rivers which descended onto open basins
where much sand and glacial till was deposited. Large lakes developed
F r eitag
165
in nearly all western basins.
Climate and vegetation similar to those of the present time were
prevalent in interglacial periods in northern latitudes.
All of the North American species of the maritima group live in
subarctic to warm temperate regions . Perhaps the ranges of hirticollis
and oregona extend for a short distance into Mexico but for the most part
they are northern forms. The species . hirticollis and repanda are almost
transcontinental and inhabit regions from the Cascades in the west to the
Atlantic coast. Ranging from the Atlantic seaboard to the eastern slopes
of the Rockies duo de dm guttata is the only true eastern form. Inhabiting
areas from the R ocky Mountains to the Pacific coast oregona is the western
counterpart of du ode cim guttata. The species depressula is restricted to high
elevations of the Cascade Range and Sierra Nevada, and in river valleys
near the Pacific coast from northern California to southern Alaska.
The species limbata inhabits areas just east of the Continental Divide.
Further south, however, populations are found in Kane County, Utah
( 1. albissima Rumpp). The ranges occupied by bellissima, theatina, and
columbica are rather r estricted:San Luis Valley in south-central Colorado,
is the entire range of theatina; bellissima occurs on sea beaches in western
Oregon and southwestern Washingtonjwhile columbica exists in southeastern
Washington on beaches of the Snake River.
Knowledge of the distribution of the North American species of
the maritima group supports the premises that: the ancestral species was
a cool adapted form, and mountain ranges of western North America are
effective geographical barrier s particularly the Rocky Mountain system.
The relationships of the Nearctic species of the maritima group
to those of the Old Worldmember s are not understood (but see Papp 1952),
so speculation on time and direction of intercontinental movements is
not warranted. However it seems certain that such movements did occur,
probably by way of a Bering land bridge (see Gressitt 1963), The hypothe-
sis which follows is based on the as yet unestablished premise that all the
Nearctic species are more closely related to one another than to any
Palearctic species.
The primitive ancestor of the North American species of the
maritima group may have inhabited cool - temperate: regions of North
America in late Miocene. By virtue of its habits it may have filtered
southward along alpine river systems near revived mountains of western
North America. It may have assumed a reticular distribution among these
mountains and in cooler regions further east. By the continuous folding
of strata, and volcanic eruptions, populations probably became disjunct
and geographically isolated. The first three derivative stocks may have
been established during the course of this unsettled period.
Very little can be said about the place of origin and geographical
movements of hirticollis because of its present vast range and widespread
sympatry with repanda, oregona, and duo de cim guttata. It is probably a rela-
tively old form.
The derivative stock that gave rise to limbata, bellissima, columbica, and
theatina , may have ranged throughout cooler regions of western North
America up to the late Pliocene. Western North America had been worn
down to extensive plains. Mountains were no longer effective geographic
barriers, and sandy habitats occurred abundantly near the coast, near
166
Cicindela maritima Group
lakes and rivers, and in dry areas remote from water. Perhaps during
its existance the ancestral species became more generally adapted and
improvements of functions allowed it to inhabit sandy environments in arid
regions, but it also continued its riparian habits. The renewed crustal
unrest of the later Pliocene probably disbanded and isolated populations,
that evolved into limbata, bellissima, and theatina.
The species limbata may have developed as a sand dune inhabitant
on the northeastern side of the revived Rocky Mountains in late Pliocene
or early Pleistocene. The original form probably resembled the boreal
subspecies /. hyperborea. Southern populations were probably established
during cooler glacial periods. See Rumpp (1961), for some ecology and
mechanism of lost of elytral pigmentation in southern populations of
limbata.
The ancestral stock of bellissima, probably became isolated on the
Pacific coast by the renewed folding of the Coast Range in late Pliocene
or early Pleistocene.
Populations that evolved into columbica probably became locked
in by the Sierra Nevada and Rocky Mountains perhaps in the early
Pleistocene. Within this area they retained the riparian habits of the
ancestral stock.
The species theatina may have originally been isolated from other
related populations to the east of the Continental Divide in Colorado.
It perhaps had a greater range than the San Luis Valley to which it is
now restricted.
The ancestral stock from which repanda, depressula, oregona, and
duo decim guttata evolved may have originally been isolated to the east of the
Rockies. It eventually became transcontinental, probably in early
Pliocene.
The place and time of origin, and subsequent geographical dis-
tribution of repanda is obscured because it ranges throughout most of
temperate North America and is sympatric with several related
species, and perhaps speciated before late Pliocene.
The species depressula may have developed in late Pliocene.
Primitive populations of depressula, represented by d. eureka, on the west
side of the Cascade Range and northern Sierra Nevada probably became
geographically segregated from the form which gave rise to duo decim guttata
and oregona.
The common ancestor of the species duo decim guttata and oregona prob-
ably occupied the entire cool temperate North America during the middle
Pleistocene. The extant species may have been formed during the middle
Pleistocene. Dissection of the range of the ancestral stock took place in
glacial periods of the later Pleistocene when ice masses covered Canada
and glaciers spread southward on high mountain ranges. The species
duodecimguttata evolved in the east and oregona in the west with the R ockies act-
ing as the major geographical barrier. In glacial periods it is doubtful
that populations of duodecimguttata merged with those of oregona in southeastern
regions of the R ockies for no evidence of that exists. Hybridization bet-
ween these species is proof of their close relationship and that their re-
productive isolating mechanisms have not yet become fully developed.
Perhaps hybridization between them was more extensive in earlier inter-
glacial periods and their isolating mechanisms have become gradually
more effective with each successive glacial period.
Fr eitag
167
ACKNOWLEDGEMENTS
I would like to thank the following and their respective institutions
for the loan of specimens. Their generous cooperation made this study
possible. Dr. W. F. Barr, University of Idaho; Dr. E. C. Becker and
Mr. W. J. Brown, Canada Department of Agriculture; Dr. G. W. Byers,
University of Kansas ;Dr. O. L. Cartwright, United States National Museum;
Dr. P. J. Darlington, Jr., Museum of Comparative Zoology at Harvard
Univer sity;Dr . W. A. Drew, Oklahoma State University; Dr . H. S. Dybas,
Chicago Natural History Museum;Mrs. L. K. Gloyd, Illinois State Natural
History Survey; Dr. R. C. Graves, Flint Junior College; Dr. M. H.
Hatch, Univer sity of Washington;Mr. R. Huber , Minneapolis, Minnesota
(personal material) ;Dr . J. D. Lattin, Oregon State College; Mr. H. B.
Leech, California Academy of Sciences; Dr. T. Moore, University of
Michigan;Dr . L. L. Pechuman, Cornell University; Mr. J. A. Shetterly,
Cambridge, Massachusetts (personal material); Mr. P. E. Slabaugh,
Bottineau, NorthDakota (per sonal material);Dr. C. A. Triplehorn, Ohio
State University; Mrs. P. Vaurie and Dr. J. G. Rosen, The American
Museum of Natural History.
I wish to express my thanks to the National Research Council of
Canada for support of this project through Grants PRA 135 and A1399,
which were held by Dr. G. E. Ball, to whom also I am indebted for his
guidance and encouragement throughout this study.
Thanks are due also to the following for their contributions:
Mr. D. K. Duncan, Mrs. G. Freitag, Dr. W. A. L. Fuller, Dr. B.
Hocking, Mr. Lan Lin Wu, Mr. R. Lister, Mr. K. Richards, and Dr .
N. Rumpp.
REFERENCES
Anderson, E. 1949. Introgressive hybridization. John Wiley and Sons ,
Inc. , New York. 109 p.
Ball, G.E. 1959. A taxonomic study of the North American Licinini
with notes on the Old World species of the genus Diplocheila Brulle
(Coleoptera) . Mem. Amer. ent. Soc. 16, 258 p. , 15 pi.
Ball, G.E. The male genitalia of the Canadian species of Cicindela . MS.
Becker, E.C. 1963. Three new fossil Elaterids from the amber of
Chiapas, Mexico, including a new genus (Coleoptera). J. Paleont.
37 : 125-238.
Blackwelder, E. 1948. The Great Basin. 1. The geological background .
Bull. Univ. Utah 38(20), 16 p.
Blair, W.F. 1951. Interbreeding of natural populations of vertebrates.
Amer. Nat. 85 : 9-30.
Blatchley, W. S. 1910. An illustrated descriptive catalogue of the
Coleoptera or beetles (exclusive of the Rhynchophora) known to
occur in Indiana. The Nature Publ. Co. , Indianapolis. 1386 p.
Bock, J.W. 1963. Evolution and phylogeny in morphologically uniform
groups. Amer. Nat. 97 : 265-285.
Brown, W.L., Fr. , and E.O. Wilson. 1956. Character displacement.
Systematic Zool. 5 : 49-64.
Cain, A.J. , and G.A. Harrison. I960. Phyletic weighting. Proc.
zool. Soc. Lond. 135 : 1-31.
168
Cicindela maritima Group
Carr, F.S. 1920. Albertan Coleoptera. Canad. Ent. 52 : 218-220 .
Casey, T.L. 1897. Coleopterological notices, 7. Ann. N.Y. Acad.
Casey, T.L. 1909. Studie s inthe Caraboidea and Lamellicornia. Canad.
Ent. 41 : 253-284.
Casey, T. L. 1913. Studies on the Cicindelidae and Carabidae of America.
Memoirs on the Coleoptera. Lancaster Press Inc., Pa. 4,400 p.
Casey, T.L. 1916. Further studies inthe Cicindelidae. Memoirs on
the Coleoptera. Lancaster Press Inc., Lancaster, Pa. 11, 347 p.
Cazier, M.A. 1936. Review of the willistoni, fulgida, parowana, and senilis groups
of the genus Cicindela (Coleoptera - Cicindelidae). Bull. S. Calif.
Acad. Sci. 35, 8 p.
Cazier, M.A. 1948. The origin, distribution, and classification of the
tiger beetles of lower California (Coleoptera : Cicindelidae). Amer,
Mus. Novit. 1382, 28 p.
Cazier, M.A. 1954. A review of the Mexican tiger beetles of the genus
Cicindela (Coleoptera : Cicindelidae). Bull. Amer. Mus. Nat. Hist.
103, 84 p.
Cazier, M.A. I960. Notes on Mexican tiger beetles belonging to the
genus Cicindela (Coleopte ra , Cicindelidae) . Amer. Mus. Novit. 2025,
12 p.
Dejean, P. F.M.A. 1825. Species general des coleopteres de la
collection de M. le compte Dejean. Paris 1,463 p.
Dobzhansky, T. 1951. Genetics and the origin of species. 3rd. ed.
Columbia Univ. Press, New York. 364 p.
Fall, H.C. 1901. A new Cicindela, with notes on allied species. Ent.
News 12 : 307-310.
Findley, J.S. 1955. Speciation of the wandering shrew. Univ. Kans.
Publ. 9, 68 p.
Graves, R.C. 1963. The Cicindelidae of Michigan (Coleoptera) . Amer.
Midi. Nat. 69(2) : 492-507.
Gressitt, J. L. (ed. ) 1963. Pacific Basin Biogeography . A symposium.
Bishop Museum Press, Honolulu. 563 p.
Hamilton, C.C. 1925. Studies on the morphology , taxonomy and ecology
of the larvae of holarctic tiger beetles (Family Cicindelidae).
Proc. U.S. nat. Mus. 65 : 1-87.
Hamilton, T.H. 1961. The adaptive significance s of intra specific trends
of variation in wing length and body size among bird species.
Evolution 15(2) : 180-195.
Hatch, M.H. 1953. The beetlesof the Pacific northwest. Univ. Wash.
Publ. Biol. 16, 340 p.
Horn, W. 1915. Coleoptera, Family Carabidae , Subfamily Cicindelinae .
Genera Insectorum 82c : 209-486.
Horn, W. 1930. Notes on the races on Omus californicus and a list of the
Cicindelidae of America north of Mexico. Trans. Amer. ent. Soc.
56 : 73-86.
Hubbell, T.H. 1956. Some aspects of geographic variation in insects.
Ann. Rev. Ent. 1 : 71-88.
Hubbs, C.L., and R. R. Miller, 1948. The Great Basin. 2. The
zoological evidence. Bull. Univ. Utah 38(20), 127 p.
Huntington, C.E. 1952. Hybridization in the purple grackle , Quiscalus
quiscula. Systematic Zool. 1 : 149-170.
Fr eitag
169
Inger,R.F. 1961. Problems in the application of the subspecies concept
in vertebrate taxonomy, p. 262-285. In W. F. Blair, (ed.),
Vertebrate speciation. Univ. Texas Press, Austin.
Karlstrom, T.N.V. 1961. Glacial history of Alaska: its bearing on
paleoclimatic theory. Ann. N. Y. Acad. Sci. 95(1) : 290-340.
Kendrew, W.G. 1961. The climates of the continents. 5th ed. Clarendon
Press, Oxford. 608 p.
King, P.B. 1958. Evolution of modern surface features of western
North America , p. 3-60. In C.L. Hubbs, (ed.), Zoogeography.
Amer. Assoc. Adv. Sci. Symp. , Washington, D.C.
LeConte, J. L. 1857. Revision of the Cicindelae of the United States.
Trans. Amer. phil. Soc. 11 : 27-62.
Deng, C.W. 1902. Revision of the Gicindelidae of boreal America.
Trans. Amer. ent. Soc. 28 : 93-186.
Lindroth, C. H. 1955. The carabid beetles of Newfoundland, including the
French islands of St. Pier re and Miquelon. Opu sc . ent. ,Suppl.l2 ,189 p.
Lindroth, C. H. 1963. The fauna history of Newfoundland. Opus c. ent. ,
Suppl. 23, 112 p.
Linsley, E.G. 1958. Geological origins and phylogenetic affinities of
the Cerambycid beetle fauna of western North America , p. 299-320.
In C.L. Hubbs, (ed.), Zoogeography. Amer. Assoc. Adv. Sci.
Symp., Washington, D.C.
MacGinitie, H.D. 1958. Climate since the late Cretaceous, p. 61-79.
In C.L. Hubbs, (ed.), Zoogeography. Amer. Assoc. Adv. Sci.
Symp., Washington, D.C.
Mayr, E. 1942. Systematic s and the origin of species. Columbia Univ.
Press, New York. 334 p.
Mayr, E. 1963. Animal species and evolution. The Belknap Press of
Harvard University Press, Cambridge, Mass. 797 p.
Mayr, E. , E.G. Linsley, and R.L. Usinger. 1953. Methods and
principles of systematic zoology. McGraw-Hill Book Co. , Inc. ,
New York. 335 p.
Mecham, J.S. 1961. Isolating mechanisms in anuran amphibians , p.
24-61. W.F, Blair, (ed.), Vertebrate speciation. Univ. Texas
Press, Austin.
Mengel, R.M. 1964. The probable history of species formation in some
northern wood warblers (Parulidae). The Living Bird 3 : 9-43.
Muller, H. J. 1940. Bearings of the Drosophila work on Systematics, p.
185-268. In J.S. Huxley, (ed.), The new systematics . Clarendon
Press, Oxford.
Papp, H. 1952. Morphologisch und phylogenetische unter suchungen an
Cicindela . Ost. Zool Zeit. 3 : 494-533.
Quate, L.W. 1963. Fossil Psychodidae in Mexican amber , part 2. J.
Paleont. 37 : 110-117.
Rausch, R.L. 1963. Geographic variation in size in North American
brown bears, JJrsus arctos L. , as indicated by condylobasal length.
Canad. J. Zool. 41 : 33-45.
Ridgway, R. 1912. Color standards and color nomenclature. R •
Ridgway, Washington, D.C. 43 p.
Rivalier, E. 1954. Demembrement de genre Cicindela Linne. II. Faune
Americaine. Rev. fran<£. Ent. 21 : 249-268.
Ross, H.H. 1958. Affinities and origins of the northern and montane
insects of western North America, p. 231-252. In C.L. Hubbs,
170
Cicindela maritima Group
Rumpp, N.L. 1957. Notes on the Cicindela praetextata - califomica tiger beetle
complex. Description of a new subspecies from Death Valley,
California (Coleoptera - Cicindelidae) . Bull. S. Calif. Acad. Sci.
56(3), lip.
Rumpp, N.L. 1961. Three new tiger beetles of the genus Cicindela from
southwestern United States. Bull S. Calif. Acad. Sci. 60(3), 23 p.
Sabrosky, C.W. 1963. A new Acalypterate fly from the Tertiary
amber of Mexico. J. Paleont. 37 : 119-120.
Shelford, V.E. 1917. Color and color pattern mechanisms of tiger
beetles. 111. Biol. Monogr. 3, 130 p.
Short, L.L., Jr. 1963. Hybridization in the wood warbler s Vermivorapinus
and V. chrysoptera, Proc. 13th Intern. Ornithol. Congr. 1 : 147-160.
Sibley, C. G. 1950. Species formation in the red-eyed towhees of
Mexico. Univ. Cal. Publ. Zool. 50 : 109-194.
Sibley, C.G. 1954. Hybridization in the red-eyed towhees of Mexico.
Evolution 8 : 25 2-290.
Sibley, C.G. 1961. Hybridization and isolating mechanisms, p. 69-88.
In W.F. Blair, (ed.), Vertebrate speciation. Univ. Texas Press,
Austin.
Sibley, C.G. , and L.L. Short, Jr. , 1959a. Hybridization in the buntings
of the Great Plains. Auk 76 : 443-463.
Sibley, C.G. , andL.L. Short, Jr. , 1959b. Hybridization in some Indian
bulbuls Pychonotus cafer x P. leucogenysB Ibis 101 : 177-182.
Sibley, C.G. , and L.L. Short, Jr., 1964. Hybridization in the orioles
of the Great Plains. The Condor 66 : 130-150.
Simpson, G. G. , A. Roe , and R. C. Lewontin. I960. Quantitative Zoology.
Harcourt, Brace and Co. , New York. 440 p.
Stainer , J. 1934. Insects, p. 274. Clarke, Public School Exploration
In Newfoundland, London.
Stebbins, R. C. 1954. Amphibians and reptiles of western North
America. McGraw-Hill Book Co., Inc., New York. 528 p.
Sturtevant, A.H. 1963. A fossil Periscelid (Diptera) from the amber
of Chiapas, Mexico. J. Paleont. 37 : 121-122,
Sylvester-Bradley , P.C. 1963. Post- Tertiary speciation in Europe.
Nature, Lond. 199 : 126-130.
Tanner , V. M. 1929. The Coleoptera of Utah. Pan-Pacific, Ent. 6:78-87.
Van Dyke, E.C. 1947. New species of Coleoptera from western North
America. Pan-Pacif. Ent. 23(1) : 155-161.
Varas Arangua, E. 1928. Contribucion al estudio de los Cicindelidae.
Notas sobre las variedades de algunas especies de Cicindela de grupe
“formosa purpurea ore gona”' Rev. Chilena Hist . Nat. 32 : 231-251.
Vaurie, P. 1951. Five new subspecies of tiger beetles of the genus
Cicindela and two corrections (Coleoptera , Cicindelidae) . Amer. Mus.
Novit. 1479, 12 p.
Wallis, J. B. 1961. The Cicindelidae of Canada. Univ. of Toronto
Press, Toronto, 74 p.
Wilson, E.O., and W.L. Brown, Jr., 1953. The subspecies concept.
Systematic Zool. 2 : 97-111.
Zeuner, F.E. 1943. On the systematics of Troides and its allies. Trans,
zool. Soc. Lond. 25 : 10.7-184.
Quaest
lones
entomologicae
A periodical record of entomological investigations,
published at the Department of Entomology, Uni-
versity of Alberta, Edmonton, Canada.
VOLUME I
NUMBER 4
OCTOBER 1965
171
QUAESTIONES ENTOMOLOGICAE
A periodical record of entomological investigations, pub-
lished at the Department of Entomology, University of Alberta,
Edmonton, Alberta.
Volume 1 Number 4 1 October 1965
CONTENTS
Guest Editorial 171
Wellington - An approach to a problem
in population dynamics 175
Wada - Population studies on Edmonton mosquitoes 187
Wada - Effect of larval density on the development
of Aedes aegypti (L. ) and the size of adults 223
Announcement 250
Corrigenda 250
Guest Editorial - Two Cultures and the Information Explosion
We live today in a dangerously unstable and incongruous
world. As a travelling scientist in recent years I have dined with
friends whose principal problems were calories and obesity and
hurried through gloomy alleyways where starving children slept on
the pavement for want of abetter home or shelter. I have been plied
with cocktails in foam-padded chairs at near the speed of sound over
the Pacific Ocean and photographed foot-weary peasants, miles from
their village, overburdened by their precious loads of firewood. All
in a world whose population will double by the end of the century;
and more than half the present population is undernourished, despite
a level of science and technology which could probably solve the
problem within a generation.
Has the scientist anything more to offer society than the extra
miles per hour, the new antibiotic, the faster computer, or the
hydrogen bomb? I feel that he has and he must, but he is handicapped
by the weight of his own information explosion and by its effect upon
his education and later professional outlook.
I suggest that in teaching and research we are developing
science too much as a technical tool and tend to ignore its value as
a guide to human thought and relationships. Before the scientist can
play a more effective role in society, he must first put his own house
in order. He must learn to contain and handle his own information
explosion. Surely it is this explosive growth in scientific and technical
172
knowledge which is the really unique phenomenon in the history
of human society. There is abundant evidence that the population
explosion is one consequence of the information explosion, although
perhaps indirectly as a result of an unbalanced application of the
resulting technology.
C. P. Snow identified the information explosion with the
barrier of communication between the scientist and humanist; the
gulf between the two cultures. In his erudite monograph Foskett (1)
has examined the increasing lack of communication between the
scientist and humanist from the point of view of the professional
librarian: One faced with the task of trying to maintain information
retrieval in a world whose boundaries, like those of the expanding
universe, are lost forever to the observer's telescope. He comes
to the conclusion that "Scientists tend to assume airs of arrogant
superiority over non- scientists.. .control over material phenomena
is possible to an extent undreamtof even fifty years ago, and rightly
used, the discoveries of science could bring about that revolution in
our material conditions foreseen by Wordsworth, who put the poet
at the side of the man of science There is no hope of such a
creative partnership while scientists fail to carry out their duty of
making these discoveries familiar". If the scientist does give this
impression of himself it is a reflection of his education; a result of
not seeing his fellow men and his environment in the very perspectives
dictated by the world of science. This, in turn, is because we are
educating technicians rather than scientists. We lose sight of the
wood too easily for the technical trees.
Are not university courses terribly cluttered with unneces-
sary or even obsolete technical knowledge? Are we not attempting
the impossible by trying to contain forever an exploding volume of
knowledge? I suggest that the problems of documentation and infor-
mation retrieval must play a much more vital role in scientific
education. This would facilitate elimination from the syllabus of
certain knowledge once documented and rapidly available through
efficient information retrieval. It would give more scope for original
thought; a chance to examine some of the fundamental problems of
our time.
Let us adjust our perspectives. Geologists tell us that the
earth is of the order of five billion years old. In order to grasp this
time scale let us suppose that the earth was formed on the occasion
of the birth of Jesus Christ, 2,000 years ago. Now on this scale
William Caxton printed his first book just under three hours ago.
The Wright brothers made their first powered flight tenminutes ago
and 90% of all the world's scientists have been born since] The world
population also doubled in the last tenminutes. We exploded the first
atomic bomb less than two minutes ago. On this time scale the
growth of scientific information and technology can indeed be seen
as an explosion. It is an especially sobering thought if we try to
look forward, just ten minutes]
Now let's turn our attention to space. Hoyle (2) considers
it probable that there are one hundred thousand million stars with
173
planetary systems in our Milky Way galaxy alone. Hence "The
probability of there being intelligent life 'out there' is overwhelmingly
high". Hoyle has seriously suggested that with radiotelescopes
little more sophisticated than those already in existence we should
be able to establish a range of communication to embrace the nearest
million stars. Somewhere in the million, Hoyle suspects, there are
planets on which has evolved intelligence comparable or superior
to our own. He has speculated that intelligent radio- communication
may have been in progress for millions of years. If indeed we can
tap such a cosmic reservoir of intelligence, get into the galactic
telephone directory as Hoyle puts it, then our own information
explosion becomes a mere bubble.
Is it easy for the scientist to conjure up feelings of super-
iority or arrogance with this picture of his environment? Certainly
not if this sort of cosmological appreciation were part of his educa-
tion. In this way he could better approach the larger problems of
humanity with essential humility. The humility due to the constant
knowledge of our colossal relative ignorance.
What doweask when trying to assess a candidate for appoint-
ment or promotion? "How many paper s has he published? " Perhaps
we scan the titles or read a few summaries, lest we appoint a
geneticist instead of a taxonomist! How often do we read even one of
his papers from beginning to end? Not frequently. We haven't
time. So the young scientist with an eye to attracting the attention
of his peers gets out as many papers as he can.
We are all familiar with the appearance of substantially the
same article in two or more journals. And there is another form of
duplication: howmany times do we read an almost identical descrip-
tion of some well established experimental procedure such as this:-
twenty grams of tissue were representatively sampled and accurately
weighed into a Soxhlet extraction thimble and extracted for 24 hours
with A. R. benzene. The extract was taken to dryness on a water
bath and the non-volatile residue weighed etc. etc.
Before publishing we should first ask if we will contribute
either to the knowledge which the student should embrace or to that
to which the specialist- should have access. If the answer be no then
we should abstain even though it would give us our century. If yes,
then how can we strip the publication of non-essentials? Is it to be
a work of literature or a scientific communication? Surely the
latter. One way of improving scientific communication would be to
devise a kind of international shorthand. Some abstracting services
have started this but we could go a good deal further so that the
Soxhlet extraction paragraph might read something like: - Weight
non-vol 24hr. C^H^ extr. 20 g tissue. Many consecutive operations
such as those involving extraction, fractionation, detection and assay
might well be indicated by a symbolic flow sheet. An International
Conference to formulate such a shorthand based on English would
be a most valuable contribution. Once terms and expressions were
agreedupon, theycould be publishedby the various learned societies
in their Journals and their use insisted uponas indeed many abbrev-
174
iations already are. What I'm suggesting is that it's high time we
regarded routine scientific publication for what itis: communication
and documentation; not a work of literature.
Society is becoming increasingly dependent upon science and
technology in a world of limited resources and dangerously unstable
international relationships. This is clearly appreciated by politicians
and administrators but the present tendency is merely to impose
administrative or political philosophies on the world of science. The
converse would be more to the point. That is, the philosophy of
science, an absolute respect for the truth, might be profitably
applied to the problems of government and administration and even,
perhaps, to commercial advertising.
International instability has become a universal threat.
These problems are a direct result of the impact of science on
society. They require scientific analysis and control in a spirit of
scientific humanism. Meanwhile, the best we can hope for is to
keep open the communications between the hations. The machinery
for this exists through the United Nations and its scientific or
specialist agencies. The scientists of the world speak a common
language and must subscribe to the same respect for universal
truths. They have the best opportunities for international meetings
and social and professional intercourse.
They must learn to contain their information explosion; to
re-examine urgently the whole structure of scientific publication.
Only then will they have the time to regain sight of the wood for the
trees. The education of every scientist should provide for an objective
scientific appreciation of his human and physical environment and
the impact of his own technology. He will then be in a position to
regulate better the production of "dangerous knowledge and disor-
ganization" and to challenge its political abuse.
Unless every scientist emerges from the swamp of his own
information he may indeed find himself continually on tap but never
on top: an increasingly dangerous world will remain the politician's
F.P.W. Winteringham
Visiting Professor, State College, Raleigh.,
University of North Carolina,
U.S.A.
Present address:
Agricultural Research Council Pest Infestation Laboratory
Slough, Bucks., England
References:
(1) Foskett, D. J. Science, humanism and libraries. Crosby Lock-
wood and Son, Ltd. , London, 1964, p. 31.
(2) Hoyle. F. Saturday Review, 1964.
175
AN APPROACH TO A PROBLEM IN POPULATION DYNAMICS’"
W.G. WELLINGTON
Forest Entomology and Pathology Laboratory Quaestiones entomologicae
Victoria , British Columbia 1 :175—186. 1965
This is the text of a lecture to a group of graduate students in zoology and entomology.
It describes the first stage of an investigation of the population dynamics of Malacosoma pluviale
(Dyar); what led to the problem; how the study was planned, and how it actually developed. Some
examples show that previous experience may be used to advantage during the planning stage of
an investigation, and that it also may help to exploit the first break-through that occurs. But
another example shows that previous experience then may be a handicap, as it may keep one from
seeing things as they really are. Thus, the second break-through in a new field is more likely to
be accidental, no matter how deliberate it may seem in retrospect. In other words, research still
progresses more erratically than our final reports suggest.
This is not the kind of paper one expects to find nowadays in a
scientific journal. It is not a straightforward account of methods, results,
and conclusions. Instead, it is a discursive personal account of the
beginning of one investigation, and its attendant difficulties and mistakes .
It was originally a lecture given to graduate students and faculty of the
Departments of Entomology and Zoology of the University of Alberta in
1961. I chose this approach because I thought students should hear at
first-hand how our investigations really develop chronologically, and not
always in the logical way in which we report them. I wanted to show
what prompted the investigation in the first place, and how its first
important turning-points were reached.
The lecture was to be published, but has been withheld until now
because some of its points depended on data presented in an accompanying
lecture, and this supporting material had to be developed differently for
publication. Now that the data are in print (Wellington 1964, 1965) there
is no longer any restriction on the content of the original address. The
factual material is drawn from my investigation of the population dynamics
of the western tent caterpillar, Malacosoma pluviale (Dyar).
Most research papers show investigator s moving in such straight
lines that one feels they often must have known their conclusions before
they obtained their results! It is unfortunate that published reports so
consistently give this impression. They do so, of course, because space
limitations in journals permit authors to describe only the ideal routes
to discovery. The truly erratic paths that lead there, or the first faint
sign-posts that indicate the most likely route are almost never described.
As the limitations imposed here are not so severe, I can tell you not only
about my destination, but also something of my reasons for going and
my ways of travelling there. There must be some sort of outline to which
we can refer, however, so let us see how a straightforward description
of the early work might be summarized. . . .
* Contribution No. 1163, Forest Entomology and Pathology Branch,
Department of Forestry, Ottawa, Canada.
176
Population Dynamics
In 1955, an outbreak of the western tent caterpillar was nearing
its peak in the Saanich Peninsula of southeastern Vancouver Island,
Because it offered an opportunity to study the effects of behavioral and
climatic variations on the insect’s population dynamics, I collected some
eggs from the outbreak for experimental purposes, and also mapped its
boundaries so that I could follow later changes in its extent.
In 1956, when the eggs hatched, I subjected the emerging larvae
to a very simple activity test that exploited their response to light. This
test revealed several types of larvae that differed in their ability to
perform directed movements when they were separated from their
fellows. Some were well-directed and active, others were disoriented
and less active, and some were so sluggish that they scarcely moved.
Controlled rearings showed that these differences were persistent, and
that they also affected individual development and survival, because the
various types of larvae differed in their ability to find and utilize food.
Artificial colonies composed of varying proportions of active and
sluggish larvae were established, and their habits were compared with
those of natural colonies in the field. These comparisons led to the
identification of different types of natural colonies, and this discovery
in turn enabled me to find areas where either active or sluggish colonies
predominated. Once these areas were located, working hypotheses could
be developed to account for their existence and predict the ultimate fate
of the populations within them.
The first results suggested that behavioral differences may have
a greater effect on an animal's population dynamics than theorists hitherto
have supposed. But to establish this point it was necessary to subject
the deductions arising from this thesis to repeated tests. Such testing
has been the primary objective of the study since 1957 and, to date,
accumulated observations tend to support the thesis in amost consistent
way. For example, active individuals predominate in new infestations,
but the sluggish component of the population increases as infestations
age. Ultimately, most members of one generation are so sluggish that
they cannot survive. Consequently, numbers within infestations so
affected are drastically reduced.
Although very condensed and incomplete, this summary is
sufficient to provide us with a framework for future reference (see also
Wellington 1957, I960). But why should anyone want to study the effects
of individual differences in behavior or activity on a whole population?
And if they must, why use the western tent caterpillar instead of some
other animal? Furthermore, what led to the rather unusual method of
separating the different types of larvae at the beginning of the invest-
igation? And finally, though the summary seems tidy enough, was the
progress of the work really so direct? Or was it sometimes saved
accidentally from ineffectual circling? In the remainder of this lecture,
I will try to answer these questions.
To answer the first three I must go back several years before
1955. Those of you who read population literature know only too well the
continuing debates among the theorists . For those who are less familiar
with this literature, I can summarize its central theme in the following
way. Many animals are alternately scarce and plentiful. Their numbers
Wellington
177
increase tremendously for a few generations, then decrease again. A
major problem for economic zoologists is to find out what prevents their
indefinite increase; and bad weather, exhaustion of food supplies, or
overwhelming attacks by enemies are often given as reasons why pop-
ulations decrease. The situation is not so simple, however, because the
numbers of animals may continue to decline while the weather is
favorable, and while food is abundant and enemies are scarce (Chitty
1960).
Although population theorists often disagree, such conflict would
be welcome if it included suggestions for experiments designed to disprove
hypotheses. More often than not, however, it involves only comparisons
of all-embracing theories. At least this is how it seems to field ecol-
ogists, who also find a disturbing gap between what the major theories
say should happen in the field, and what actually happens there. Many
investigators therefor e have been dis satisfied with population theory for
a long time.
Before 1952, I was too preoccupied with studies of the effects of
weather on the behavior of insects to be concerned with the theory and
practice of population ecology. One cannot study the effects of weather
on insects for long, however, without being drawn into some of the
population controversies. But when I finally began to consider the various
arguments, I found I was less concerned with some of their more evident
misinterpretations of weather processes than I was with the way in which
they neglected the behavior of animals.
My own experience made me notice an operational weakness in
most studies of population dynamics. In many of these studies there was
a tendency to concentrate on the developmental and reproductive pro-
cesses of the animals, and on measurements of their mortality or
survival, to the virtual exclusion of their behavior and activity. But this
approach overlooked the stubbornfact thatan animal that does not behave
properly, or that does not maintain a certain level of activity at critical
periods in its life, simply does not survive, let alone develop and
reproduce.
The morel thought along these lines, the morel felt that the right
kind of observation would show that widespread neglect of the influence
of individual behavior on survival was actually obstructing the develop-
ment of population theory. And this feeling was not just a product of the
scientific chauvinism that might be expected from my studies of behavior;
it arose from the observation that some of the major theories could not
really be falsified in their existing form (c. f. Platt, 1964). This was my
main reason for wanting to study the effects of the activity and behavior
of individuals on the fate of a whole population. But I had to findaninsect
that would be suitable for such a study.
I had one hint from previous work that Malacosoma spp. might be
suitable. In 1948, C.R. Sullivan and I had studied the light reactions of
three species of Malacosoma that were prevalent near Sault Ste. Marie,
Ontario. We were inter ested in the changes in response that might take
place at high temperatures. And we had been following the usual pro-
cedure; scattering larvae at random on the platform of a choice chamber
that had illuminated and darkened sides. The insects were expected to
178
Population Dynamics
take up positions dictated by their initial response to light at room tem-
perature, then move to different locations if their response changed
when the temperature was raised.
We had done virtually the same thing with other kinds of insects
many times before. But when we used newly-emerged fir st-instar larvae
of Malacosoma only a few acted in the expected way. The majority never
moved after they were dropped on the platform. Consequently, we could
not continue the experiment, because we could not tell how they reacted
to light.
To solve this problem, we put the larvae back on their egg mass,
so that they would be in a more natural situation. In effect, we made the
egg mas s the dark- light alternative, with its top illuminated and its bottom
shaded. When all the larvae were allowed to remain together on their
eggs in this way, they moved about very easily. And since this solved
the technical problem, we proceeded with the investigation (Sullivan and
Wellington 1953).
I wondered afterwards, however, why most members of these
young colonies could perform directed movements while they were
touching one another, but not while they were isolated. And if most of
them were so dependent, why were a few so independent that they could
perform directed movements while they were alone? I had to file this
puzzle for future reference, however, because we had used all the
available larvae. And eventually, of course, I stopped thinking about it.
But I remembered it again in 1952, when 1 began to think about
the possible effects of individual behavior on a population. Here,
apparently, was a group of insects that varied in activity and behavior
as soon as they hatched. Besides, all the members of the genus also
experienced great and comparatively regular changes in numbers. And
some species made conspicuous tents, so that they could still be found
without much difficulty when they were scarce. Malacosoma spp. thus had
much to offer as experimental animals.
It was no help to realize this in 1952, however, because the tent-
forming species were too scarce to provide enough material for testing.
Butwhen I saw the outbreak of M. pluviale on the Saanich Peninsula in 1955,
I was again reminded of my earlier intentions , and pleased to see a good
supply of one of the species that had provided the germ of the idea. And
that is how M. pluviale became the experimental animal in the study.
It is worth noting that at this stage I had very little foundation on
which to build a work plan. I knew nothing of the apparent difference
in activity that I have just described, except that it existed. I did not
know whether it was simply an intrinsic part of each individual's make-
up, varying from time to time as the animal passed through different
physiological states, or whether it was a real and persistent difference
among individual Malacosoma larvae, stable enough to be exploited in the
type of study I had inmind. Since it would not take long to find out which
kind of variability was involved, however, I decided to plan the forth-
coming investigation on the assumption that the difference would prove
to be persistent.
The decision to plan the investigation in this way did not depend
entirely on an act of faith. I had recently observed peculiarities in the
Wellington
179
behavior of some arctiid larvae which suggested that such individual
differences might in fact be stable. Also, as I came to realize later,
my various lines of thought had been channeled during a brief conver-
sation with Dennis Chitty just before I saw the tent caterpillars on the
Saanich Peninsula. Thus my ideas concerning individual behavior were
resting comfortably within a larger framework. And larger frameworks
are always reassuring, even when one is scarcely aware of them.
During our conversation, Chitty and I discovered we were both
dissatisfied with current population theories, and disturbed by the ten-
dency of ecologists to treat the populations with which they worked as
though they were monolithic structures, instead of collections of indi-
viduals. But Chitty also was circling an idea he has since stated more
explicitly; namely, that the composition of a population might change
with changing density, and that this qualitative change might have impor-
tant effects on subsequent densities (Chitty I960). Looking back, I do not
believe I had carried my ideas about the effects of individual behavioral
differences on populations quite so far (although my ready response to
Chitty' s well-nigh subliminal prompting showed me later that I had
obviously been ready to do so). A few months afterwards, however, all
that was clear to meat the beginning of my own study was that I not only
had to determine how any variations in behavior might affect the survival
of individuals within a population; I also had to consider these individual
differences in terms of the changes in population quality with which they
might be associated. Still later, when I had some results to interpret,
I suddenly realized that my final plan of attack had been decided, virtually
at the last minute, by that conversation with Chitty: a conversation,
incidentally, that I had "forgotten" in the enthusiasm engendered by
finding the Malacosoma outbreak and planning my investigation.
The first step in that investigation was to ensure that the differ-
ences observed in 1948 were truly persistent between individuals, not
just internal changes within any individual at different times of the day
or between successive days. If the former situation obtained, many
things followed directly. Otherwise, I scarcely had a problem of the
sort I had imagined. To establish the facts, repeated tests of identified
individuals were required. And I needed a very simple and rapid screen-
ing method that would allow me to handle large quantities of material;
e. g« » perhaps more than 15, 000 larvae per generation. It seemed best
to exploit the difference in activity noted during 1948, as it appeared to
be present as soon as eclosion took place. This, then, was one reason
for using the laboratory test employed at the beginning of the investigation.
But there was another reason that requires further explanation.
Some aspects of reality areunusual enough to seem unacceptable
or even unbelievable when we first encounter them. In these days of team
research and elaborate equipment, we tend to forget that explication of
these unusual and often complex aspects of reality does not always require
a complicated attack. In fact, some of our more mechanized attacks
only obscure reality, or the approaches to it. And obscuring the path
to an incredible result does not often encourage others to verify or dis-
prove it.
180
Population Dynamics
A good example of what I mean may be found in Karl von
F risch's work on dancing bees (1950). Some of those early results and
conclusions were quite unbelievable, but the experiments had a truly
beautiful simplicity. Without such simplicity, other scientists might
still be questioning von Frisch's conclusions. Because of it, they
have been busily extending his results; though, unfortunately, not
always with such elegant methods. Present-day biologists have much
to learnfrom Professor von Frisch's approach to problems, there-
fore, and can profit from it in whatever field they intend to explore.
I was prompted by this line of thought to devise a very simple
test for my own purposes. As each egg mass hatched at room tem-
perature, I took its newly-emerged larvae and distributed them in a
long line parallel to a fluorescent lamp, separating the individuals
so that they had to move more than their own body length before they
could touch any of their fellows. The reasoning was that any indi-
vidual capable of independent, directed movement should proceed
directly toward the light, whereas the others should stay where they
were, or not move very far in any direction. This should separate
any colony into at least two components. And the stability of each of
these components then could be assessed by further testing.
The test worked very well. It was in fact my first break-
through, because without such an easy, rapid, and definite means of
identification of persistent differences among individuals, there
would have been little time to do anything else. Because of the test
and its results, however, the first part of the study opened auto-
matically into a series of sub-projects that virtually had to develop
along certain lines, often with results that were quite predictable,
because they were the logical outcome of the existence of the be-
havioral differences.
Consider the results of the rearing experiments, for example.
Larvae that differ in their ability to perform directed movements
must behave in certain predictable ways when they are gathered into
groups and placed near food. Very sluggish larvae should be in-
capable of fending for themselves, no matter how many are grouped
together. And this proved true. Very sluggish larvae had to be
placed on their food because they were incapable of locating it when
there was no active individual to guide them, even when the food
was only a few mm. away. Without proper care, therefore, they
starved. And proper care included frequent inspections to ensure
that they had not fallen from the food, because they could not return
to it unaided.
More active, but still disoriented larvae proved relatively
easy to handle, as long as they were kept in sufficiently large groups.
Then they spun sufficient silk to be protected from desiccation, and
they eventually found food by a sort of group "amoeboid" flow. Thus
they fed and developed, though with some delay.
In contrast, the independent larvae were more difficult to
handle under artificial conditions. They were too independent in
the rearing jars; a predictable result of their ability to orient and
Wellington
181
travel while isolated. Although each could find food very quickly,
individuals tended to remain scattered for hours instead of clustering
together occasionally. Therefore they had few opportunities to form
the common mat of silk that would protect them from desiccation,
so that they often died when only small numbers were kept together
in the jars. Increasing the number of larvae per jar, however,
solved this problem.
As development proceeded, it was clear that the most active
larvae fed more and developed most quickly, whereas the most
sluggish, if they lived atall, fed least and grew most slowly. There
was no evidence within the generation that disease or any malfunction
not attributable to the basic differences was at the root of such
variation. There was plenty of evidence, however, that eggs laid
by some females yielded colonies that had a high proportion of
sluggish larvae, whereas eggs from other females yielded colonies
that had a much greater proportion of active larvae.
Many other differences in behavior and activity were re-
vealed during these studies, which opened endless avenues for further
physiological research. But I must confine my remarks here to the
development of the population studies. The foregoing descriptions
were necessary to emphasize that there were some very marked
differences in development and survival associated with the differ-
ences in activity and behavior, even though the latter were first
revealed as an apparently trivial response.
As the rearing experiments with pure groups progressed
satisfactorily, I began to make up artificial colonies differing in the
proportions of the types of individuals they contained. These were
studied in the laboratory and in the field to determine what differences
in growth or habits they might have. Those which contained numerous
well-directed larvae were active. They formed several tents in rapid
succession, spacing them widely over the available foliage, and
vacating each in turn before they exhausted the food nearby.
In contrast, colonies that contained a high proportion of
sluggish individuals were very inactive. Such a colony seldom made
more than one tent, and the larvae spent much time clustered on it,
because there were not enough active individuals present to disturb
and scatter the other larvae resting in the cluster. The larvae en-
larged the. tent and occasionally fed out from it for short distances,
but even when they had exhausted nearby food they seldom moved on
to spin another tent, though ample food was available only a short
distance away. Consequently, the member s of truly sluggish colonies
usually starved. If they were saved from this fate by unusually
abundant food right at hand, they were still prey to disease. (They
were more exposed to infection than members of active colonies,
because they often touched the remains of diseased larvae during
their prolonged clustering periods.) Very sluggish colonies, there-
fore, soon were lost to the population by one or other of these means.
When I finally obtained adults from the different types of
larvae, I found that activity differences were still recognizable, and
that their classification could depend once more on a very simple
182
Population Dynamics
test. Active adults left in the jars in which they emerged literally
batter ed themselves to pieces in one or two days. From this extreme
there ranged a graded scale of decreasing damage to the other
extreme: the perfect appearance of sluggish adults that remained
unmarked until they died. They never moved after their wings
expanded.
All the findings described above came from straightforward
exploitation of the logical consequences of the original differences
observed among emerging larvae. They were necessary steps in the
study, but most of them could not immediately add to its further
development. As an isolated group of facts they offered no direct
entry into the next stage: the study of the natural population. In
fact, while all these sub-projects were in progress, I had been try-
ing to find away to distinguish the different types of natural colonies
in the field without having to classify every larva within them.
Without a simple and rapid method of classifying the natural colonies,
I could not progress with the field studies.
The artificial colonies finally provided the solution to this
survey problem. For not only did the active colonies among them
make more tents than the sluggish colonies; they also made tents
of a different shape. The "active” tent was longer and thinner --
inmost instances very obviously club- shaped -- whereas tents made
by less active colonies were shorter and squatter; in extreme
instances, definitely pyramidal.
Here I had the potentially perfect sorting method to bring
order out of the apparent chaos of the peak population of 1956,
provided that natural colonies behaved as the artificial ones had.
If they did, I could close the gap between laboratory and field studies
by using differences in tent shape as a simple but reliable survey
tool to classify every colony I examined. With it, I should be able
to see whether there were areas where one type of colony predom-
inated. In addition, I should be able to accumulate statistics on
differences in the sizes of feeding areas, larval numbers, etc. ,
among colonies. I also should be able to identify colonies that had
changed their characteristics during development after losing one
or other of their constituent groups, because these changes should
be revealed by differences between their previous and current tests.
With so many potential benefits due, I was almost afraid to
examine natural colonies again in case the difference did not exist
among them. It was there, of course, as it had been all along. I
had not seen it before, however, even though I had been happily
finding and counting colonies by watching for their tents’ I did not
see it because I had been caught in the snare that lies in every
research path: inability to get outside one's previous conceptual
framework. Because every entomologist knew that tent caterpillars
occupied box-like or pyramidal tents, I had paid no attention to tent
shapes in my earlier surveys. Consequently, I saw them properly
only after I had a strong incentive to look.
This second break-through of the investigation was a happy
accident, therefore, and not the product of deliberate planning that
Wellington
183
the first had been. If it had not occurred, however, not much else
would have happened during that first season of study, and I would
have begun the next with a serious handicap. Consequently, I have
emphasized it and the preceding mistake. In fact, this whole
sequence of events is a good example of the greatest difficulty that
confronts us whenever we engage in frontier research. At the border
of the unknown, one must consciously strive to escape from the mesh
of former frames of reference, and to remain outside the generally
accepted range of opinion concerning one's problem, foravery good
reason: the problem is rarely what accepted opinion says it is] But
the difficulty is that one tries so hard to keep one's thinking free on
larger issues that one overlooks the danger of continuing to think
about apparently smaller issues in terms of older concepts. This
lapse is always dangerous, and sometimes disastrous, because there
is no small issue ata frontier. And howcanone observe what does not
yet exist as a conceptual possibility (Hanson 1958)?
A new survey soon showed that club-shaped tents predom-
inated in areas that were unoccupied by the expanding population
before 1956. In fact, if the new infestations of 1956 were sufficiently
far from previous infestations, only club-shaped tents occurred. On
the other hand, a larger proportion of pyramidal tents occurred
wherever the population had been in residence for several gener-
ations. In such areas, some trees contained only pyramidal tents,
although there were always some club-shaped tents in any locality.
This information led directly to a testable hypothesis con-
cerning the fate of any local population after its first introduction
into an area. It seemed reasonable to suppose that active adults
would, in general, produce active colonies, whereas less active
adults would produce colonies that were decreasingly active, down
to a level where some would be very sluggish. Also, it was already
known that these various types of adults differed in their ability to
fly. Further observation of their movements made it clear that only
the most active could fly far enough to enter remote, previously
uninfested areas. Therefore, in a new, remote locality, only active
colonies should be produced by these first invaders.
Provided that survival within these colonies was adequate,
however, adults that displayed different amounts of activity would
be produced from them (since even active colonies contain some
inactive or sluggish individuals). Of these, only the active adults
would be able to fly away before they oviposited; the less active
would have to oviposit closer to their birthplace. The next generation
in that locality, therefore, should contain some colonies less active
than any of the parent generation. And in subsequent generations,
an increasing proportion of sluggish colonies should appear in the
locality if emigration of active adults exceeded their immigration.
This is what the local differences observed in 1956 suggested, and
it remained to be seen what actually happened after 1956.
As working hypotheses go, this first model turned out quite
well; i. e. , its major statements could not be disproved. Certain
aspects of the general population trend and of the local environment
184
Population Dynamics
affected the situation in any locality. But within these limitations,
only minor amendments to the hypothesis were required. When newly
infested areas were sufficiently remote, the first generation in fact
consisted entirely of active colonies. In contrast, new infestations
established closer to older ones contained some less active colonies
in the places near est the older foci -- a fact, incidentally, that helped
to establish maximal flight distances for less active females. In the
next generation in an isolated area, however, some sluggish colonies
appeared, and their proportion rose during subsequent year s until the
population included many colonies too sluggish to survive. Similar
changes, though further advanced, could be recognized in older in-
festations. The end resultwas always the same: a sudden reduction
in numbers, because most of the colonies had died.
In that last paragraph I hurried through the findings of several
years, after using considerably more space to outline the sequence
of events that led up to them. But this is as it should be, if I am to
fulfil the intention outlined in my introductory remarks. All the
foregoing results have been published, along with many others I
have not mentioned here (Wellington 1957, I960; 1964, 1965). But
until now, I have not described how I reached them. And it is
reasonably correct to say of this, as of all scientific work, that most
of the original thinking had been done by the time the first experi-
ments were completed. After 1956, the speculation and reasoning
that had led to the first tentative proposals were buried by the
pedestrian process of testing them.
Finally, I should point out something not emphasized earlier,
though it is implicit in much of the foregoing description. Although
this was, and is, a field study of a population, the laboratory has had
a strong influence on its inception, direction, and findings. My
original dissatisfaction with population theory and practice stemmed
partly from the fact that laboratory studies of insect behavior paade
me sceptical of some of the ideas and conclusions of population
ecologists. Many of the clues on how to approach the problem I
wanted to investigate came from laboratory observations, as did
the evidence for the initial differences. Similarly, the different
tent- shapes were detected only by studying colonies with controlled
compositions; a method that is still more common in laboratory
studies than it is in the field.
And this brings me to the point I wish to make. I believe
that laboratory studies by themselves often degenerate into the
pursuit of trivia. But I also, believe that field studies that lack the
benefit of the special discipline that comes from laboratory training
and planning are unlikely to advance much beyond the speculations
with which they begin. In other words, the theory and practice of
population ecology should not be exempt from the general rule that
hypotheses are better disciplined by experiment than by faith and
reason (Chitty 1957). Consequently, when we cannot combine labor-
atory and field studies during population research, we should at least
take the discipline of the laboratory with us when we go to the field.
A balanced program of labor atory and field investigations in
Wellington
185
fact has some very practical attractions. In the studies described
here, I was able to do much more during the 1956 season (a matter
of some two months) by keeping the laboratory stocks and tests
slightly ahead of the equivalent stages in the field. Thus I was able
to make anynumber of mistakes during the fir st round of experiments
and observations, and still have time to correct them by using fresh
material as the field population entered each required stage. This
enabled me to exploit the two break-throughs of that first season with
minimal delay.
ACKNOWLEDGMENTS
I want to thank the members of the Department of Entomology
at the University of Alberta for giving me my first opportunity to air
these views. I am especially grateful to Dr. George Ball for all his
help and encouragement. Dr. Dennis Chitty is to be thanked for
allowing me to involve him rather deeply in some of the more con-
troversial issues I have raised. And I also want to thank the many
colleagues who have expressed interest in seeing this lecture in
print. Special thanks are due to Drs. John A. Chapman and Derek
A. Maelzer for their helpful criticism.
REFERENCES
Chitty, D. 1957. Population studies and scientific methodology.
Brit. J. Phil. Sci. 3 : 64-66.
Chitty, D. I960. Population processes in the vole and their rele-
vance to general theory. Can. J. Zool. 38 : 99-113.
Frisch, K. von. 1950. Bees, their vision, chemical senses, and
language. Cornell University Press, Ithaca, N. Y.
Hanson, N. R. 1958. Patterns of discovery. An inquiry into the
conceptual foundations of science. Cambridge University
Press, London.
Platt, J. R. 1964. Strong inference. Science. 146 : 347 - 353.
Sullivan, C.R., and W. G. Wellington. 1953. The light reactions
of the tent caterpillars, Malacosoma disstria Hbn. , M. americanum
(Fab.), and M. pluviale (Dyar). (Lepidoptera: Lasiocampidae).
Canad. Ent. 85 : 297-310.
Wellington, W. G. 1957. Individual differences as a factor in pop-
ulation dynamics: the development of a problem. Canad. J.
Zool. 35 : 293-323.
Wellington, W. G. I960. Qualitative changes in natural populations
during changes in abundance. Canad. J. Zool. 38:289-314.
Wellington, W. G. 1964. Qualitative changes in populations in
unstable environments. Canad. Ent. 96 : 436-451.
Wellington, W. G. 1965, The use of cloud patterns to outline areas
with different climates during population studies. Canad.
Ent. 97 : 617-631.
187
POPULATION STUDIES ON EDMONTON MOSQUITOES*
Y OSH I TO WAD A
Department of Medical Zoology Quaestiones entomologicae
Nagasaki University School of Medicine 1:187—222. 1965
The seasonal fluctuations of each instar larvae and pupae of Culiseta inornata (W illiston)
in a particular pool near the University of Alberta were investigated and an attempt to estimate
the mortality of the aquatic stages was made. The data for the collections of adults and larvae
of 26 species of mosquitoes found around Edmonton indicate that the black-legged mosquitoes of
subgenus Ochlerotatus, genus Aedes are earlier-appearing, species than others. The distribution
pattern of mosquito larvae was firstly demonstrated to follow a negative binomial distribution with
a common value of constant k for various density levels. Based on this distribution pattern, a
sequential sampling technique was applied to classify a mosquito population into one of three pre-
defined density levels. This was considered useful in deciding whether or not control is neces-
sary, and in evaluating whether or not control has been successful over a wide area in a relatively
short time.
GENERAL INTRODUCTION
The City of Edmonton has been engaged in the control of
mosquitoes and has reduced the mosquito population greatly in the
city (see Klassen and Hocking, 1963 and 1964). However, there are
still some problems to be solved. They include precisely when and
how the insecticidal applications should be made for the effective
and economical control of mosquitoes, how far the larvicide should
be applied beyond the city limits, and so on. For the settlement of
them, extensive fundamental studies are required. This report
deals with the studies conducted in 1964 to approach the problems
from an ecological point of view.
BICNOMICS OF EDMONTON MOSQUITOES
Mosquito Surveys and Identification
Three types of mosquito surveys were made in 1964. Firstly
larval (and pupal) surveys were made at pools in various environ-
ments around Edmonton, mostly westward, from April to July. The
number of dips at each pool was not recorded, except for a few pools
for determining the distribution pattern of larvae per dip, which will
be mentioned later. However, care was taken in catching mosquitoes
so as to represent the mosquito fauna there; only a few dips were
made at pools with high mosquito density and many dips, sometimes
more than 50. at pools with low density.
* Contribution from the Research Institute of Endemics, Nagasaki
University No. 473 and Contribution No. 142 from the Department
of Medical Zoology, Nagasaki University School of Medicine.
188
Population Studies
Secondly collections were made of adult mosquitoes, which
came to feed on me, around a particular pool near the University of
Alberta at approximately one week intervals.
Thirdly larval surveys were made at the pool mentioned
above. The pool harbored almost exclusively Culiseta inomata
(Williston) and the seasonal changes of immature stages were
studied.
The larvae collected were reared in the laboratory to the
fourth instar or to adults, and identified. Some specimens were
separately reared to obtain the adults with associated larval skins
to facilitate determining the species.
The identification of larvae followed Carpenter and La Casse
(1955) and Rempel (1950). Adults were identified mostly after
Carpenter and LaCasse (1955) and Rempel (1953). However, it was
often difficult to separate them to species, especially rubbed speci-
mens of black-legged female Aedes . In such cases, and even for good
specimens, the post-coxal scale patch (between the anterior coxa
and the sternopleuron), mesepimeral scale patch, scales of
probasisternum, and tarsal claws were useful characters (Beckel,
1954; Vockeroth, 1954).
Notes on Some Species
Aedes communis (DeGeer) and Aedes inlrudens Dyar
A. communis and A. intrudens are black-legged species lacking
the post- coxal scale patch. The adult female of A. communis is
usually separable by the contrasting stripes on the scutum from
A. intrudens with a uniformly colored scutum. However, in some
specimens of A. intrudens the scutum shows indications of paired median
brown stripes, and those specimens, particularly when the scales
on the scutum are not complete, are sometimes hard to distinguish
from A. communis.
After examinations of 38 females of A. intrudens and 32 of
A. communis , some of which were associated with their larval skins,
it was found that, as described by Carpenter and LaCasse (1955),
mesepimeral scales reach near lower margin in A. communis, but in
A. intrudens the lower third or fourth is bare. This seems to be a most
useful character to separate them. Other characters, which might
be used, are the number of lower mesepimeral bristles and the color
of the base of the costa. The lower mesepimeral bristles vary in
number in both species, but, in the present specimens A. intrudens
has a smaller number of bristles, ranging from 0 to 3, than A. communis ,
which has 2 to 7 bristles. White scales at base of the costa are
absent, or if present very few in number, inA. intrudens', they are
present in A. communis .
Aedes hexodontus Dyar and Aedes punctor (Kirby)
The adults of these two species are very similar to each
other, however the larvae are distinct. According to Beckel (1954)
the probasisternum has white scales and an extensive patch of white
Wada
189
scales is seen at the base of the costa in A. hexodontus taken in the field
at Churchill, Manitoba; on the other hand in A. punctor taken there
scales on the pr obasisternum are reduced to a few and there are no
white scales at the base of the costa or rarely one or two. These
characters were found useful to separate specimens of these species
taken near Edmonton also, by examination of females associated
with their larval skins.
Knight (1951) recognized two varieties in each species: "type
hexodontus " and "tundra" variety in A. hexodontus and "type punc tor" and
"tundra" variety in A. punctor. The scutum of females has a broad
median dark stripe which may be narrowly divided in "type hexodontus "
and "type p^c for", on the other hand in "tundra" variety of both species
the median dark band is absent or not well defined.
Of the females of Edmonton hexodontus collected or reared from
larvae, 9 are "tundra" variety and one is "type hexodontus " variety.
The latter was collected as an adult on June 7, 1964. In addition to
these, I have another female specimen of "type hexodontus " variety,
which was reared from a larva taken near Jasper, Alberta, on May
16, 1964. The associated larval skin shows that head hairs 5 and 6
are both double, which agrees with the description given by Knight
(1951) for "typ e hexodontus" variety.
As for A. punctor , the many larvae and 18 females, which were
collected as adults or reared from larvae, are consider ed all "type
punctor " variety.
Aedes niphadopsis Dyar and Knab
A larva of this species was taken from a collection of small
scattered pools in a pasture near a creek, about 20 miles west of
Edmonton, on June 7, 1964, and reared to a female adult. This
record is new to Canada (Pucat, 1964).
Aedes pullatus (Coquillett)
This is a species that lacks the post-coxal scale patch, and
bears a distinct hypostigial scale patch. The distribution in Alberta
seems to be limited mostly to mountainous regions. I collected
many larvae from snow-melting pools in Jasper National Park on
June 21, 1964, but no specimens were encountered around Edmonton.
Seasonal Fluctuation and Mortality of Immature Stages of Culiseta inornate (Williston)
Observations were made on the changes in abundance of each
instar larvae and pupae of Culiseta inomata (Williston) throughout a
season at approximately one week interval in 1964 at a pool, ca. 10
x 3 m, near the University of Alberta. The pool is situated on the
south bank of the North Saskatchewan river , and receives little sun-
light because of tall vegetation such as poplars around it. For this
reason, ice remained at the bottom of the pool as late as May 8,
and the water temperature was relatively low throughout the summer;
the maximum water temperature was only 18.3 C,on August 17.
On each day, larvae and pupae were sampled with a dipper
usually ten times, but when necessary, 20 or 50 times, and the
190
Population Studies
numbers of each instar larvae and pupae were recorded. The popu-
lation of mosquitoes in the pool consisted of only C. inomata , as far
as the fourth instar larvae were examined. However, from some
egg rafts collected at the pool on July 6, there emerged some adults
of Culiseta alashaensis (.Ludlow) in addition to C. inornata; this indicated
that a small number of egg rafts, probably one, of the former species
was mixed in the collection of the egg rafts. Therefore, some
C. alaskaensis may have bred also in the pool, even so, the number
seems to have been negligibly small.
Egg rafts were first encountered on May 25, and oviposition
continued until August 10. The number of egg rafts per dip and the
observation for the rafts on the water surface of the pool show that
the peak of oviposition activity of C. inomata was in the first half of
June.
The seasonal distribution for each instar larvae and pupae
is shown in Fig. 1, The first individuals of larvae in the first, the
second, the third, and the fourth instar , and pupae were encounter ed
on May 25, June 2, June 16, and June 22, respectively. The peak
in numbers of first instar larvae was June 8, and with the progress
of the development the time of each peak became successively later;
the peak for pupae was on July 6. The period between the peaks of
first instar larvae and of pupae is about one month. This seems to
be the time required forC. inornata to develop from the first instar
larva to the pupa; the mean water temperature was 11 C during the
period.
The emergence of adults is thought to have occurred most
actively shortly after the peak of pupae, that is in the middle of July.
This time of peak emergence was ascertained by the fact that many
pupal skins were observed on the water surface on July 14 and 23.
It has been reported that the duration of the larval stage of
mosquitoes such as Anopheles quadrimac ulatus Say and Aedes aegypti{ L. ) is
affected by temperature, nature and amount of food, and density of
a population (e. g. see Horsfall, 1955). Therefore, the above period
of one month at mean water temperature of 11 C will be changed to
some extent according to the conditions in a pool, even when the
temperature is the same. Also, the remarkable difference in water
temperatures within a pool (Haufe, 1957) may influence the data.
However, the difference does not seem to be great, as most larvae
inhabit similar environments.
The area surrounded by the abscissa and the curve for each
instar larvae and pupae in Fig. 1 is dependent on the relative abun-
dance and also on the duration of each instar. In the laboratory at ca.
23 C, an egg raft of C. inornata was reared to adults, andmean periods
for each instar larvae and pupae were obtained. If it is supposed that
these mean periods are kept unchanged also in the present field data,
we can get the relative abundance by dividing the calculated area from
Fig. 1 by the mean period. The results are given in Table 1. It is
recognized from the table that the reduction in the relative abundance
is remarkable between the first and the second instar larvae, and
between the third and the fourth instar larvae. The survival rate
Mean no. per dip
191
1
Fig. 1. Mean number of each instar larvae and pupae of Culiseta momata
per dip.
192
Population Studies
from the first instar larvae to the pupae is estimated at 63/248 x
100 = 25%. Thus we get a mortality of 75% for the aquatic stages of
C. inornata , or slightly higher, as the mortality in the earlier half of
the pupae is not included in the above calculation.
TABLE 1 - Relative abundance of each instar larvae and pupae of
C. inornata in the field.
The reliability of the above calculation depends on how
effectively the material was sampled from the pool and how close
the relative mean duration for each instar larvae and pupae obtained
in the laboratory is to that in the field. As will be mentioned later,
the number of larvae plus pupae of mosquitoes per dip follows a
negative binomial distribution having a larger variance than a random
distribution. This means that a larger number of dips is required
to estimate the population effectively, and the number of dips may
be too small in the present field data. As mentioned earlier, the
mean duration of larval stage is affected by temperature, food, and
population density, but perhaps little affected in pupae by the last
two factors. Therefore, it is rather difficult to compare the values
in the laboratory with those in the field. Another difficulty is that
the temperature in the field changes daily and seasonally. Never-
theless, the above method of estimating the mortality is of value as
a first approach to this important subject.
In any case, it seems that the mortality in the aquatic stages
of C. inornata is fairly high in the field. The factors responsible for
this are not known. However., physiological disorder or a sort of
disease is supposed, as some dead larvae were found and all attempts
to find predators in the pool failed.
Seasonal Occurrence of Edmonton Mosquitoes
Table 2 gives the number of larvae (and pupae) collected
around Edmonton and the number of collections in which each species
was found. Mosquitoes were encountered at 30 pools out of more
than 60 examined. Since the number of dips varies from pool to
Wada
193
pool, the number of larvae shown in the table does not represent exactly
the relative abundance of each species. However, the main features of
seasonal appearance are clearly seen.
TABLE 2 - The total number of larvae and pupae collected around
Edmonton, and the number of collections (within paren-
theses) in which each species was found.
194
Population Studies
The results of the collections of female mosquitoes, which
came to feed on me, around a pool on the south bank of the North
Saskatchewan river near the University of Alberta are given in Table
3. This table also indicates an aspect of seasonal fluctuations of
mosquitoes.
From these tables and some other data, seasonal occurrence
of mosquitoes in 1964 is given below.
Anopheles
Anopheles earlei Vargas hibernates as an adult female. Many
larvae were found from late May to early July (Table 2), and one
female was collected at the campus of the University of Alberta on
May 26. Most of 58 larvae collected in late May shown in Table 2
were in the second instar and a few were in the first and a few in the
third. Thus it seems that hibernated females appear and oviposit
their eggs from May, and the emergence of adults occurs from June.
Oviposition continued at least until the beginning of July, as two first
instar larvae were encountered in early July.
Culex
The species of Culex found were C. tarsalis Coquillett and
C. territans Walker. Both hibernate as adult females.
Although only one larva of C. tarsalis was collected, the
hibernated females are considered to oviposit late in the season, as
it is reported that in irrigated areas of Alberta the larvae are found
abundantly in July, August, and September (Shemanchuk, 1959), and
in Saskatchewan the first larvae do not appear until early July
(Rempel, 1953).
The larvae of C. territans were collected in late June to late
July(Table 2), and this seems to be also a late-appearing species .
Culiseta
Three species were encountered around Edmonton, namely
C. alaskaensis (Ludlow) , C. inornata (Williston), and C. morsitans (Theobald).
They all hibernate as adult females.
The first egg raft of C. alaskaensis was found on July 6, as
mentioned earlier, and one larva was collected in late July (Table
2). According to Jenkins (1948), overwinter ed females were common
from late April to mid- June and all instars of larvae were found
from May 11 to July 10 in Alaska. Therefore, the larvae may appear
earlier than July also around Edmonton.
Table 2 indicates that the larvae of C. inornata were collected
from early June, and this agrees with the data mentioned earlier.
The peak of oviposition was found to be in early June and the
peak emergence occurred in mid-July. The feeding activity seems
to be limited mainly to the period from late May to early July, as
judged from the number of females attracted to man (Table 3), and
this is justified by the time of the peak of oviposition. Those females
are considered overwintered ones. However, a small number of
females oviposited as late as August 10 as mentioned earlier. It is
Wada
195
not known whether such oviposition was derived from overwintered fe-
males or from newly emerged ones.
TABLE 3 - The number of female mosquitoes collected ground a pool on
the south bank of the North Saskatchewan River near the
University of Alberta, Edmonton.
*
One hour collection was made in the afternoon each day, excepting two
hour collection on June 22.
Two laryae of C. morsitans were obtained in early June. Rempel
(1953) reported the adults in July. This species perhaps spends a similar
life cycle to C. alaskaensis and C. inornaja in Alberta.
Aedes
All Aedes species recorded here hibernate as the egg stage.
196
Population Studies
Black-legged species belonging to the subgenus Ochlerotaius are
generally earlier-appearing species than other mosquitoes. The dates
of the collections of the larvae and adults in those black-legged species
(from Tables 2 and 3), together with the records of the larvae and adults
in Saskatchewan by Rempel (1953) and the dates of emergence near
Edmonton by Klassen and Hocking (1964) are shown in Table 4.
TABLE 4 - Summary of the occurrence of black-legged Ochlerotatus .
(1) Tables 2 and 3; (2) Rempel, 1953; (3) Klassen and Hocking, 1964.
It is apparent from the table that the larvae of most species appear
very early in the season. However, A. pionips Dyar is perhaps a slightly
later species, as indicated by Haufe (1952) and Rempel (1953), and it
seems in A. intrudens and A. punctor that the hatching from eggs continues until
later in the season, or the life span of the adults is longer. As for
Wada
197
A. niphadopsis it is not clear, as only one larva was collected.
For banded-legged mosquitoes of subgenus Ochlerotatus, similar data are
also given in Table 5.
TABLE 5 - Summary of the occurrence of banded-legged Ochlerotatus
(1), (2), and (3) see table 4; (4) a second generation may occur; (5) long-lived
species, occasional specimens may be encountered late in the season.
Of the species given in the table, A. excrucians Walker, A. fitchii (Felt and
Young), A. increpitus Dyar, A. riparius Dyar and Knab, and A. stimulans (Walker) are
considered woodland species and have only one generation a year. The larvae
appear as early as most black-leggedmosquitoes, but the emergence is delayed
because of slower development, as indicated by Haufe (1953 and 1956) and as
recognized by the fact that the black-legged mosquitoes emerged earlier than
the banded-legged ones, when the larvae from the same pool were reared in
the laboratory. The females were collected as late as September 11 in A. fitchii
198
Population Studies
as August 10 in A. increpitus and A. riparius, and as September 16 in
A. stimulans . These facts seem to indicate that the life span of adults
of those species is very long, asRempel (1953) stated that occasional
specimens of A. excrucians may be encountered in mid-summer.
A. canadensis (Theobald) is also a wood-loving species. The larvae
appeared as early as other banded-legged species mentioned above.
Occasionally hatching occurs in the fall in Illinois (Horsfall, 1955).
Other tabulated species, A. campestris Dyar and Knab,
A. dorsalis (Meigen), and A. flavescens (Muller), are grassland-lovers,
and a second generation may occur, when the environment is favor-
able. They seem to be slightly later-appearing species than the
woodland species.
Aedes vexans { Meigen), which belongs to the subgenus
Aedimorphus , is found in the three main ecological zones in Saskat-
chewan, the prairies, aspen grove region, and coniferous forest
(Rempel, 1953). This species seems to have multiple generations
when the conditions are favorable. The larvae were collected in
early May to late June and the adults on July 13 and August 10. It is
apparently a late-appearing species.
Black-legged A. (Aedes) cinereus (Meigen) seems to.be rather
late in appearance, though the first larva was collected inlateApril.
The adults were collected from June 22 to August 25.
DISTRIBUTION PATTERN OF MOSQUITO LARVAE
Introduction
Populations of animals may be effectively estimated on the
basis of their distribution pattern, and much has been published on
this subject with various kinds of animals, among which however
mosquitoes are not included. In applying the sequential sampling
technique, which will be described later, and also in comparing the
population densities at different pools, it is required to establish the
nature of the frequency distribution pattern of mosquito larvae (and
pupae) .
A dipper is usually used for collecting mosquito larvae, and
is considered a handy and reliable tool. Here, an attempt has been
made to analyse the distribution pattern of mosquito larvae in their
habitats by using the number per dip.
Collections Used for the Determination of the Frequency Distribution
Table 6 gives the data of collections of mosquito larvae for
determining the frequency distribution pattern of the numbers per dipt
Collections numbers 9 to 24 in the table are the same data as used
for the seasonal fluctuation of C. inornata described earlier. The table
indicates that the collections were made at various habitats of various
sizes during the period covering May 25 to September 30, and the
mosquito species collected were distributed in the genera Anopheles ,
Culex , Culiseta , and Aedes. The habitats included a grassland pool, a
woodland pool, a collection of scattered small pools, and the marginal
part of a creek, and the mosquitoes were found at some times as a
single species, and at others mixed.
Wada
199
TABLE 6 - Collections of mosquitoes in immature stages for
the frequency distribution pattern.
The Relation Between Mean and Variance of the Numbers Per Dip
2
In Table 7 the mean.(x), variance (s ), and range of the
numbers of mosquitoes per dip are given. The means vary from
0.02 to 39. 40, and the variances from 0. 02 to 3975. 34. Theminimum
value of the range for most collections is zero, and the maximum
value is up to 206. These figures indicate a great variability in
number of mosquito larvae between the pools and also within each
pool.
Several mathematical models have been developed to describe
the distribution pattern of animal counts. When the distribution is
considered random, a Poisson distribution is often applied. In the
200
Population Studies
Poisson distribution, the probability for a given positive integer x,
is given by P(x) = e”mmx/x' (1)
where m is the mean. It is a property of the Poisson distribution
that the variance is equal to themean, and the expression 2 (x-x)^/k
gives a good approximation to x^ with (n-1) degrees of freedom,
where n is a sample size (Andr ewartha, 1961).
TABLE 7 - Mean, variance, and range of the number s of mosquitoes
per dip, together with x^ - test for significant departure
from Poisson distribution.
x^ with n-1 degrees of freedom is calculated by (n-l)s ^/x. For
further explanation see text. For collection number see table 6.
^Discrepancy from Poisson distribution is significant at 1% level.
Wada
201
It is apparent from Table 7 that the variation of variance is
much greater than that of mean, and the Poisson distribution does
not seem to fit the data excepting collection numbers 4, 9, and 24.
To make sure, the values of 2 (x-x)^/x = (n-l)$^/x were calculated
as the fifth column of Table 7; these were highly significant except
for the above three collections. This shows that a random distri-
bution - Poisson distribution - could not be rejected in the number
of mosquitoes per dip, when the population density was as low as
0.20, and discrepancy from Poisson became greater with the in-
crease of the mean, an aggregated type of distribution being indicated.
As stated by Waters ( 1 9 5 9) > there will be some field counts
for which x^ test will show no significant departure from either
Poisson or an aggregated- type distribution such as negative binomial.
It seems that non- significant values of x^ in collections 4, 9 and 24
are attributable to the sparsity of the population and consequent low
expectation of occurrence of mosquitoes in, individual dips.
Go&dness-of-Fit to the Poisson and the Negative Binomial
Insect counts in the field are often fitted fairly well by a
negative binomial distribution (Andr ewartha, 1961; Anscombe, 1949;
Bliss, 1953), which is one of the aggregated-type distributions . The
frequency distribution of the negative binomial is given by expanding
the expression (q-p)”-^, where q-p = 1, p = m/k, m is mean, and
k is a positive exponent. As the variance of a negative binomial
approaches the mean, or the over-dispersion decreases, k— *-o° and
p— ►O. Under these conditions it can be shown that the distribution
converges to that for the Poisson (Fisher et al. , 1943).
Goodness -of-fit to the Pois son and the negative binomial was
tested (Tables 8 to 11) for the data with 100 dips, i.e. collection
numbers 1, 2, 3, and 7.
Theoretical frequencies for the Poisson were calculated
successively by the following formulae. The probability of observing
zero count, P(0), is
P(Q) = e"m (2)
and the probability of observing (x+1), P(x+1), is
P(x+1) = mP(x)/(x+l), (3)
substituting sample mean, x, for population mean, m. The
theoretical frequency is obtained by multiplying each probability by
the sample size, 100.
The formulae to be used for the theoretical values of the
negative binomial (Bliss, 1953) are;
P(0) = (l+m/k)-k (4)
and
P(x+1) = (x+k)mP(x)/ (x+1) (k+m) (5)
The constant k can be computed by a property of the negative binomial
that the variance, (J ^ , is equal to (m+rn^/k), where m is mean,
substituting again sample mean and variance, x and s2, for m and
Cf 2-
In all of four examples shown in Tables 8 to 11, highly
significant departure from the Poisson was demonstrated (p< 0. 001),
202
Population Studies
which indicates that the distributions cannot be considered random.
On the other hand, those distributions agree well with the negative
binomial, except for collection 3, in which some discrepancy from
the negative binomial is apparent. In this case, 15 larvae per dip
were recorded once, which is a very high count compared with the
others. This high count contributes larger variance, whichin turn,
yields rather small value of k responsible for the discrepancy.
Generally speaking, the frequency distribution of the numbers of
larvae per dip seems to agree with the negative binomial. The dis-
agreement with the negative binomial in collection 3 may be attri-
butable to sampling error.
TABLE 8 - Goodness - of - fit of Collection No. 1 to Poisson and
negative binomial distributions.
*P< 0. 001; **0. 50< P< 0. 75 DF 4* 3* *
Fitting the Negative Binomial Distribution with a Common k
Comparison between the means of two or more distributions
are more direct and unequivocal if they have the same relative
dispersion in terms of k, and two approaches to a common k were
described by Bliss and Owen (1948). The fir st of them is a regression
moment estimate applicable to the present data. The following
calculation is based on Bliss and Owen (1958).
Wada
203
TABLE 9 - Goodness - of - fit of Collection No. 2 to Poisson and
negative binomial distributions.
*P< 0. 001; **0. 25< P< 0. 50 DF 2* 1**
Two statistics, x' and y' are computed from the mean and
variance of each component distribution:
x' = x2 - s2/ n (6)
y' = s2 - x (7)
where n is sample size. Their expectations are given exactly by
E(x') = m2 (8)
E(y') = m2/k (9)
Thus (y'-x'/k) has zero expectation. For a single sample, we have
the ratio
1/kl = y'/x' (10)
as an estimate of 1/k. The variance of (y'-x'/k) is given to order
1/ n2 by
V = 2m2 (m-k)2[k(k-l) - (2k-l)/n - 3/n2]/ (n- l)k4 . . . (11)
The invariance w = 1/V is of the nature of a weight. If calculated by
replacing m by x, m2 by x', and k by an empirical trial value of k1,
we can obtain an estimate of 1/k, l/kc, by
l/kc = 2 (wx'y ') / 2 (wx'2) (12)
as the slope of a linear regression of y' on x', the regression line
being constrained to pass through the origin (x‘ =0, y' = 0).
R ef erring back to the data of Table 7, x'and y' were calculated
by formulae (6) and (7) for each collection, and the relation between
them is given in Fig. 2, in log scales so as to show the values with
great variabilities in one chart.
Assumed that a proportional relation holds between the two,
that is given by y' = (l/k)x‘, then the relation is represented by a
straight line with an inclination of one in the figure in log scales,
because log y1 = log (1/k) + log x1. The data of Fig. 2 satisfies the
204
Population Studies
above assumption very well. This indicates that the relation between
x ' and y' is represented by a regression line passing through the
origin, and, in turn the underlying frequency distributions are
suggested to be the negative binomial with a common k. It is
interesting that the same trend seems to be shown in the regres sion
of y' on x' between collection numbers 1 to 8 for various species of
mosquitoes and 9 to 24 for C. inornata (see Table 6), because the
inclination of the regression line gives the estimate of k, which is
considered an intrinsic property of the population sampled (Fisher
et al. , 1943). However, it is likely that the value of k is species
specific, and further studies are required.
It is known that in some cases k increases somewhat as m
increases (Anscombe, 1949; Morris, 1954; Bliss and Owen, 1958).
So, the values of 1/k^ calculated by equation (10) were plotted against
mean, x, in Fig. 3, which indicates, however, no appreciable
relationship between the two. In order to know the exact situation,
however, the number of dips for each collection seems to have not
always been sufficient, and further investigations are required.
TABLE 10 - Goodness - of - fit of Collection No. 3 to Poisson and
negative binomial distributions.
*P< 0. 001; **0. 001< P< 0. 005 DF 2* 1**
Now, a common value of k will be estimated. The statistics
x1 and y' for each of the distributions have already been obtained.
The next step is to get an initial trial estimate of a common k, k'.
As x varies excessively among the collections, a suitable equation
for k' is
k' = g/Z (y'/x1)
(13)
Wada
205
where g is the number of collections. Thus we got k1 = 0.2822. By-
using this value, l/kc, an estimate of 1/k, was obtained by equation
(12) and as its reciprocal kc = 0.2947, which does not differ so much
from the first trial estimate k! = 0.2822. Thus we have estimated
a common value of k at 0. 2947. If kc should differ appreciably from
its trial value, k', recalculation is necessary byreplacing the initial
k' by kc.
The required tests for agreement with a single kc may be
arranged as an analysis of variance:
where B§ = 2 2 (wx'y1) / 2 (wx1^)
[wx'2] = 2 (wx'2) -22 (wx1) / 2 w, C = 2 2(wy')/2 w,
[wx'y']=2 (wx'y')-2 (wx')2 (wy’)/2 w, B2 = [wx'y1]2 / [wx'2]
[wy*2] = 2 (wy'2, - C, 2 2( ) = (2 ( ))2.
TABLE 11 - Goodness -of -fit of Collection No. 7 to Poisson and
negative binomial distributions.
DF 3* 2**
* P< 0. 001; ** 0. 10< P< 0. 25
206
Population Studies
Fig. 2. Relation between two statistics, x' and y1, defined by equations
(6) and (7). H : Collection Nos. 1-8; O : Collection Nos.
9 - 24. Collection Nos. 4 and 9 are not shown in the figure ,
because x' = 0, y' = 0, and also will be excluded in the later
calculations, because of indeterminate values of y'/x'.
If a single kc is justified, the F -value in the first row should
be clearly significant and that in the second rownot significant. The
calculated values are shown below:
Wada
207
Effect of DF
Slope, l/kc 1
Computed
intercept 1
against 0
Error 19
SS MS F
33.7809 33.7809 24.4670**
4.0297 4.0297 2.9186
26.2327 1.3807
Fig. 3. Relation between mean (x) and estimate of 1/k (1/k^), £ :
Collection Nos. 1-8; O : Collection Nos. 9 - 24.
208
Population Studies
The results are highly significant for slope and not significant for
computed intercept against 0, that is a common value of k is justified.
Consideration of Reasons for a negative Binomial Distribution
I have demonstrated that the number of mosquitoes per dip
follows a negative binomial distribution with a common k. The
negative binomial is generated by a distribution that is "contagious"
in the sense that the presence of one individual ina divisionincr eases
the chance of other individuals falling into that division. However,
as Andr ewartha (1961) stated, agreement with the negative binomial
does not itself permit any inference about the biology of the mos-
quitoes, though a significant discrepancy from the Poisson series
disproves the hypothesis of random scatter. In fact, according to
Bliss (1953) the negative binomial may be regarded as being com-
pounded from a number of Poisson series in which the means vary
in such a way that they are distributed like x^, and furthermore it
is possible to imagine a number of other models to explain it.
The present data are not considered the sum of a number of
Poisson series with different means, and other reasons should be
sought.
One of them which might arise is a dipping error, however,
its effect seems to be of little importance, or at least, the negative
binomial distribution is not attributable only to it.
No habitat of mosquitoes in nature is considered so uniform
that all parts of it are equally attractive to them. Marginal parts
of a pool are usually preferable to mosquito larvae, and it is a
common phenomenon that the spatial distribution of the mosquitoes
is related to water-plants or overgrown vegetation. Thus the hetero-
geneity of the environment seems to be a great reason for the con-
tagious distribution - the negative binomial. In fact. Hocking (1953)
observed strong aggregation of the larvae oiAedes communis DeGeer ,
due apparently to the effect of sunlight and temperature gradient in
the pool.
Another reason to be considered here may be a gregarious
habit of mosquitoes. Although this has not been studied extensively,
it seems important in the ecology of mosquitoes. It is commonly
observed in the laboratory that mosquito larvae show some agg-
regated distribution in a tray, in which the environment does not
appear to differ appreciably. This habit of aggregation differs in
intensity with species, and, for example, strong aggregation of
larvae is frequently seen iq Aedes aegypti{l_,, ), but it is hardly ever
seen in Anopheles hyrcanus sinensis Wiedeman. The biological meaning of
this is not clear at the present time, but is interesting in that it
may be related to the level of optimum density of larvae. At any
rate, the intrinsic behaviour of mosquitoes may play some role in
the contagious distribution.
In short, the heterogeneity of habitat and possibly a sort of
gregarious behaviour of mosquito larvae are considered to be res-
ponsible for the negative binomial distribution which is characterized
by a larger variance than mean.
Wada
209
SEQUENTIAL SAMPLING TECHNIQUE
Introduction
Sequential sampling can be used for classifying a population
into one of a number of pre-defined density levels, based on the
accumulated results of each unit sampled. In classifying animal
populations, it has been applied to the spruce budworm (Morris,
1954), whitefish, Coregonus clupeaformis (Mitchell) (Oakland, 1950), the
lodgepole needle miner (Stark, 1952), and an aphid, Myzus persicae
(Sulzer) (Sylvester and Cox, 1961). However, it has never been
applied to mosquitoes.
The great value of this procedure lies in the fact that it in-
volves a flexible sample size in contrast to conventional sampling
procedures, and it would frequently be possible to determine whether
or not a mosquito population requires control, or satisfactory control
has been obtained, with the expenditure of much less time than would
have been required if the number of sampling units was inflexibly
fixed (Knight, 1964). Therefore, it would be reasonable to extend
this technique to the immature stages of mosquitoes.
The procedure given by Morris (1954) is mainly followed by
the present application.
Density Classes
As mentioned above, the sequential sampling technique is
used for classifying a population into pre-defined density levels. It
is desirable that density classes are determined so as to enable us
to know from these classes whether or not the mosquito density is
so high that control operations are necessary, or whether a control
operation has been successful.
The density classes may be differently set up according to
the situation in the city or town concerned. Here, I have classified
density tentatively into three levels indicated by the critical mean
number of larvae per dip as follows:
Density Mean number of larvae per dip
Low 0. 1 or less
Moderate Between 0. 5 and 2. 5
High 12. 5 or more
Density class "high" may be regarded as an indication that the
mosquito density is so high that control is required, or that a control
operation has influenced the population but little, and "low" may
indicate that the density is so low that control is not required, or
that control was satisfactorily done. "Moderate" is the intermediate
situation between the two. Although the density is not so highcontrol
may be desirable if it is early in the mosquito season.
Of course, the necessity of controlling mosquitoes depends
not only on the mosquito density in each habitat, but also on the
relative area of the habitat compared with the whole area, as well
210
Population Studies
as the location of those habitats in relation to city or town to be
protected from mosquitoes . However, it is still true that population
density must be determinedat each habitat before a decision to control
is taken.
Acceptance and Rejection Lines
To apply the sequential sampling technique to the mosquitoes,
of which number per dip is consider ed tofollowthe negative binomial
distribution, it is necessary to find a common value of k fitting all
the data with different levels of mean, and it has been determined as
0. 2947.
The next step is to set up alternative hypotheses, Hq and Hi,
from the density classes. To distinguish between low and moderate
densities at a certain probability level, HQandHi are that the number
of larvae per dip is 0. 1 or less and 0.5 or more, respectively; to
distinguish moderate and high they are that the number is 2. 5 or less
and 12. 5 or more. The values of the constants based on the negative
binomial distribution at the critical densities under these hypotheses
are shown in Table 12.
TABLE 12 - Values of the constants at the critical densities under
the hypotheses of Hq and Hi, based on the negative
binomial distribution.
Each pair of hypotheses is accompanied by two possible
errors: <K and /3 are the probabilities of rejecting Hq and H^ at the
respective critical densities,. Here, both c*. and /3 were setat.O. 10.
A rather large value for error probability seems to be suitable for
rapid mosquito survey, because it reduces the number of dips to be
taken at each habitat and enables us to decide whether or not control
is necessary by a quick evaluation of the population density over a
wide area in a relatively short time.
Formulae for the acceptance and rejection lines then are:
d = sn + hQ (13)
and
d = sn + hi (14)
Wada
211
where d is the cumulative number of larvae in the first n dips. The
slope of the lines, s, is
s = k log (qi/qo) / log (p^/p^) (15)
where qgand q^ are the values of qand Pq and p^ are those of p under
the hypotheses of Hq and (for actual figures see Table 12), and the
intercepts of the equations (13) and (14) on the d-axis are
h0 = log B / log (Piqo/PoTl) W
where B=/3/(l-ot) (17)
and
h± = log A / log (piq0/pQq1) (18)
where A=(l-/3)/ac (19)
Thus we get the following formulae as acceptance and rejection lines
for low versus moderate classes,
d = 0. 2267n - 2.4153
and
d = 0.2267n + 2.4153,
and for moderate versus high
d = 5. 0891n - 24. 9138
and
d = 5. 0891n + 24. 9138,
as shown in Fig. 4. This graphmay be used in the field to determine
how many dips should be takenat each habitat in order to define the
density class within the accepted limits of and . Itis helpful to
visualize each pair of lines as enclosing a band from which the plotted
points must escape before the density c las s is satisfactorily defined.
For example, in collection number 1 mentioned earlier (see
Tables 6 and 7), the first three dips show no larvae. When zero is
plotted over each number of dips 1, 2, and 3, it is seen that they are
within both bands of low-moderate and moderate-high. The fourth
dip yields two larvae and the fifth none, therefore dips 4 and 5 are
still within these bands. The sixth dip shows three larvae, so 2 + 3
= 5 is plotted over dip 6. This is shown to have escaped from the
bands and to have fallen into the moderate zone, so dipping is dis-
continued. Thus collection number 1 is classified into moderate
density. If the plotted points had escaped into the area above the
higher band, the density would be classed as high, and if below the
lower band, the density would be classed as low.
The Operating Characteristic Curves
The operating characteristic curves are useful aids in under-
standing how the plan operates. The curve is calculated from
L (p) =(Ah . l)/(Ah - Bh) (20)
p =[i - (qo/<n)h]/[(Pi'<io/Po<ii)h ' d] (21)
where L(p) is the probability of accepting Hq for any possible level
of the population mean of kp, A and B are taken from equations (19)
and (17), and h is a "dummy variable" which may be assigned con-
venient values.
The operating characteristic curve is shown in Fig. 5 by
plotting L(p) against population mean, kp. The left-hand curve is
for low versus moderate density classes. When the mean, kp, is
0.1, the probability of accepting Hq (low density class) Is 0.9;
212
Fig. 4. The acceptance and rejection lines.
Wada
213
accordingly the probability of accepting Hd (moderate density class)
is 0.1. When kp = 0. 5 L(p) = 0. 1 for Hq and consequently 0. 9 for
Hd. At these two levels of kp, the probabilities correspond, of
course, to those previously set for oc and fi . As kp decreases below
0.1, L(p) for Hq becomes very low. Whenkpisca. 0.23, the chances
of accepting Hq and Hd are equal. The curve on the right is used in
the same way for the moderate versus high density classes. The
overlapping between the two curves is only at negligible probability
levels. Thus the probability of considering a low density class
high, or high density class low, is very small.
The Average Sample Number Curves
The average sample number curves can be drawn by plotting
the values for E(n), the mean number of dips which must be taken,
against kp, the mean number of larvae per dip, as shown in Fig. 6.
For different values of kp, E(n) is calculated from
E(n) =[hd - (ho-hd) L(p)]/(kp-s) (22)
where ho, h^, L(p), and s are taken from equations (16), (13), (20),
and (15), respectively. E(n) does not indicate the number of dips
which must be taken actually at each pool, but its expectation.
As would be expected, the peaks of the curves in Fig. 6 occur
where populations are borderline between low and moderate or be-
tween moderate and high, which indicates that relatively more dips
are required there.
Applications of the Sequential Sampling Technique in the Field
In applying the sequential sampling technique in the field it
is convenient to use tabulations (Table 13) prepared from theaccep-
tanceand rejection lines, rather than the lines themselves . Dipping
is continued until the cumulative number falls into one of the density
classes. It is apparent from the table that at least 11 dips are neces-
sary for the density to be classed into low, and at least six into
moderate; if the number of larvae in the first dip is 31 or more,
the density is classified as high without further dips.
Table 14 gives the results of applications of the sequential
sampling technique to the data shown in Tables 6 and 7. It is demon-
strated that the sequential plan can be used to classify the density
correctly into one of low, moderate, and high density classes. The
number of dips required for determining the class in various col-
lections ranged from 1 to 20. When the density is high, the required
number of dips was rather small, as expected from Fig. 6. This
is of advantage in field work, because it takes much more time to
count larvae dipped when the density is higher.
In sampling, the larvae are required to be dipped all over a
larval habitat. In a large pool, dividing it into a few portions and
applying the sequential plan at each will facilitate the work. Suggested
larval survey form is given in Table 15.
This technique can be used effectively for the evaluation of
the application of larvicides in a relatively short time. If the control
operation is successful, then the densities at all pools will fall into
214
Fig. 5. The operating characteristic curves for low versus moderate
density classes (left) and for moderate versus high (right),
kp = mean no. of larvae per dip; L (p) = probability of accept-
ing Hq hypothesis.
Fig. 6. The average sample number curves for low versus moderate
density classes (left) and for moderate and high (right),
kp = mean no. of larvae per dip: E (n) = mean no. of dips to be
taken.
Wada
215
the low density level. Also, this may be used for determining whether
or not a second larvicide application is required specifically for the
later -appearing mosquitoes. Necessity for mosquito control depends
on the productivity of mosquitoes in a particular area, rather than
the population density at each pool. To approach this, the following
procedures may be appropriate. Firstly we determine the density
class at each pool by the sequential sampling technique. Then, we
take 0, 1, and 10 as indices for low, moderate, and high density
levels, respectively, and multiply the index by the area of the pool
(the area of the marginal parts if the larval distribution is confined
there). If these are summed for a district to be examined, then it
will represent the productivity of mosquitoes there. The sequential
plan may be used for comparing regional differences of mosquito
abundance, which provide us with the knowledge as to which region
should be stressed for larval control operations.
TABLE 13 - Sequential table for use by field parties, prepared
from the acceptance and rejection lines (Fig. 4).
No. of Cumulative number of larvae
dips Low Moderate High
density classes of low, moderate and high.
216
Population Studies
TABLE 14 - Application of the sequential sampling technique to the
data shown in Tables 6 and 7 of Section 3.
During 10 dips made, the density class was not determined.
SUGGESTED STUDIES TOWARD
BETTER CONTROL OF EDMONTON MOSQUITOES
Introduction
In this section, only ecological questions are discussed,
although studies are also needed on the identification of mosquitoes
including the larvae in younger instars, the development of insec-
ticidal resistance, the methods and evaluation of applications of
chemicals, the effective and economical dosages of larvicides and
adulticides, the residual effects of insecticides when applied to the
habitat in the field, and so on.
Wada
217
TABLE 15 - Suggested mosquito larval survey form for the appli-
cation of sequential sampling technique in the field.
MOSQUITO LARVAL SURVEY FORM
Collection No. Collector:
Place: Hour: , a.m. p.m.
Date: , 19
Breeding place
- type: woodland - pool, grassland - pool, roadside - ditch,
small pools in pasture, creek, other ( )
- permanent, temporary
- size
- notes (marginal vegetation; water plants; animals;
temperature, pH, cleanness of water; etc.)
Density class determined: Low, Moderate, High
Instar of larvae:
Species identified:
The Time of Hatching and Emergence
The prediction of the emergence time of mosquitoes is re-
quired to determine the appropriate time for chemical control. The
best time for controlling mosquito larvae is before they begin to
pupate, the pupae being much more resistant to insecticides than
larvae, but not before hatching is complete. Strictly speaking, the
above situation is hard to realize in the field, because the time of
hatching differs between species and also within species so that there
remain some eggs oi fades to be hatched later in the season after
some adults have emerged. Thus the most effective time for insec-
ticidal applications against mosquito larvae is our special concern.
For this purpose, many points remain to be studied. These include
218
Population Studies
the studies on the time of oviposition and the durations of egg and
larval stages.
In mosquitoes belonging to the genera Anopheles , Cutex ,and
Culiseta, which overwinter as adults, the time of oviposition depends
on the time of blood feeding and the duration of egg development.
Blood feeding is certainly related to temperature and possibly to
adult diapause. The temperature apparently influences the matur-
ation of eggs.
All Aedes mosquitoes found around Edmonton overwinter as
eggs. According to Clements (1963), the different Aedes species fall
fairly clearly into those whose eggs enter diapause and require re-
activation, and those whose eggs merely become quiescent and hatch
shortly after exposure to an adequate hatching stimulus, although
they may require a few hours conditioning. Obligatory diapause in
the egg stage is found in Aedes hexodontus (B eckel, 1958), in Aedes squamiger
(Telford, 1958), and in Aedes stimulans (Horsfall and Fowler, 1961),
where exposure to low temperature is required before egg diapause
can be broken. These mosquitoes have only one generation a year.
Multivoltine species have facultative diapause, as in Aedes dorsalis
(Khelevin, 1958), Aedes nigromaculis (Telford, 1963), and Aedes triseriatus
(Baker, 1935), or have no diapause.
Most mosquitoes found around Edmonton have one generation
a year. However, there is a possibility that a second generation
occurs in some species, such as A. campestris , A. dorsalis , or
A. flavescens , perhaps in August when the conditions are favorable.
It is very likely that there is a wide variability in hatching
response of eggs, so that the time of hatching has a wide range, even
for eggs from the same batch.
Beckel (1958), Telford (1963) and others have discussed the
mechanism and ecological significance of egg diapause in mosquitoes,
and much has been published on the hatching response in quiescent
eggs of Ae des aegypti and some other Aedes species (see Telford, 1963).
However, the situation is still not clear for most Aedes mosquitoes.
After hatching from eggs, the development of larvae depends
on various factors. The most important are temperature, quality
and quantity of food, and larval density. It is expected that the
relation between larval period and temperature is described by an
equilateral hyperbola, or the relation between developmental speed
and effective temperature (temperature minus developmental zero
point) is linear, at least within a reasonable temperature range,
provided that other factors than temperature are constant. Based
on this relation, Haufe (1953) and Haufe and Burgess (1956) attempted
to predict dates of emergence in mosquitoes at Fort Churchill,
Manitoba, and stated: "The tundra species of mosquito ( A. impiger
and A. nigripes) had lower thresholds of development approximating
34F; the forest species [A. communis, A. punctor, A. excrucians)h&d.3. range
of 38 - 40 F, excepts, hexodontus . The products of time and temper-
ature for the period of development of both tundra and forest species
were lower for the smaller than for the larger species". Studies of
this sort are desirable for all the mosquito species found abundantly
Wada
219
around Edmonton.
It is to be noted that the threshold of development obtained
from the above relation is slightly higher than the actual value in the
development of most insects, and is not necessarily the same as the
critical temperature for hatching. Also, the developmental speed
differs greatly according to factors such as quality and quantity of
food and larval density. Thus, for the prediction of the date of
emergence, careful investigations are required in the laboratory for
each species. Another aspect to be involved is the relation between
temperature in pools in various situations and meteorological
records, for example see Hauf e and Burgess (1956) and Haufe (1957).
Flight Range
Southwood (1962) stated: "It is suggested that animal move-
ments fall basically into two types: trivial and migratory. Trivial
movements are normally confined to the territory or habitat of the
population to which the animal belongs, migratory movements carry
the animal away from this area. Although there is undoubtedly no
sharp line but a gradation between these two types, they can be
distinguished by various ecological, physiological and behavioural
characteristics", and "The ideal evidence of migratory movement
is that while engaged in it the animal does not respond to food, a
mate or habitat, and moves from the actual territory where it has
developed into an inhospitable terrain: such movement is normally
at the start of adult life". Provost (1957) reported the findings of a
mark - and - release experiment with Aedes taeniorhynchus as follows:
"Migration occurs the night of departure only, therefore twilight
departures will result in longer migrations than middle -of- the -night
departures. Appetential (trivial of Southwood, 1962) flights expand
the range of occupation by a brood much beyond what is established
by the migration. " Thus mosquito dispersal consists of two phases
of movement, and its range depends greatly on the migratory flight
and to a lesser extent on the appetential flight.
In the appetential flight of mosquitoes, the distribution of
breeding, resting, feeding, and oviposition sites, and in some
species overwintering sites, will influence the degree of dispersal,
because inmosquitoes these sites are situated quite often at different
places and it is suggested that, within limits, the closer these are
situated, the shorter the flight range. This should be considered in
the field data, particularly when mark-and-release experiments are
conducted.
As mentioned earlier, 14 species of adult female mosquitoes
were collected around a pool near the University of Alberta. Ex-
cepting Cu liseta inornata , all these mosquitoes are considered to have
entered from outside or marginal parts of the City of Edmonton,
since there are no breeding places for these species in the central
part. This means that they dispersed at least a few miles, and
should be consider ed potential pests for the Edmonton area according
to their abundance. Here, it is required to determine the range of
dispersal for each species. It is not known whether migratory flight
220
Population Studies
was involved in the dispersal or not. It seems reasonable to suppose
that the range of dispersal is longer inmosquitoes where migratory-
flight is involved than in others. The investigation of this subject
will help decide how widely insecticidal application or other control
of larvae must be extended to control mosquitoes near Edmonton.
Fluctuation in Numbers
The habitats of most mosquito larvae are characterized by
their unstableness. In years with small precipitation, the habitats
will be greatly reduced in extent, though the amount of standing
water is influenced by the dryness of the land to some degree, as
indicated by Rempel (1953), and the reverse is also the case. The
change of the habitats determines the area of breeding and oviposition
places available for mosquitoes . Also, larval mortality has a close
association with the amount of precipitation in some circumstances,
since it is often observed that pools dry up before mosquitoes emerge.
Thus mosquito abundance is expected to be closely correlated to the
amount of precipitation. Temperature is also an important factor
influencing mosquito populations, as indicated by Rempel (1953).
From the above statement, it is clear that there exists a
close relationship between mosquito abundance and meteorological
factors. However, the situation will differ from species to species,
as evidenced by the fact that some species appear abundantly in one
year and others in another year. The analysis of these correlations
over a long period will help in studies on the population dynamics.
Another approach to studies on population dynamics is the
analysis of mortality factors in the field, and also the influence of
various environmental conditions such as trophic factor s, population
density, and so on, on the fecundity of adults in the field and also in
the laboratory.
ACKNOWLEDGEMENTS
I wish to express my sincere appreciation to Professor B.
Hocking, for his giving me a chance to come to the Department of
Entomology, University of Alberta and for his many valuable sug-
gestions during the course of this study. My thanks are also due to
Dr. G. E. Ball, Dr. W. G. Evans, Dr. J. Sharplin, and all other
members of the Department of Entomology, who helped me in various
ways for this study, to Dr. S. Zalik of the Department of Plant
Science, who gave me helpful suggestions in statistical analysis,
and to Miss A. M. Pucat of Division of Biology, University of
Saskatchewan, Regina College, who sent me information about the
mosquitoes of Alberta.
I wish to thank the Parks and Recreation Department of the
City of Edmonton for financial support and help in the field work.
Wada
221
REFERENCES
Andrewartha, H. G. 1961. Introduction to the study of animal
populations. The University of Chicago Press, Chicago,
281 p.
Anscombe, F. J. 1949. The statistical analysis of insect counts
based on the negative binomial distribution. Biometrics, 5:
165 - 173.
Baker, F.C. 1935. The effect of photoperiodism on resting, treehole
mosquito larvae. Canad. Ent. , 67: 149 - 153.
Beckel, W. E. 1954. The identification of adult female Aedes mos-
quitoes (Diptera, Culicidae) of the black-legged group taken
in the field at Churchill, Manitoba. Canad. J. Zool. , 32:
324 - 330.
Beckel, W.E. 1958. Investigations of permeability, diapause, and
Hatching in the eggs of the mosquito Aedes hexodontus Dyar.
Canad. J. Zool. , 36: 541 - 554.
Bliss, C. I. 1953. Fitting the negative binomial distribution to
biological data. Biometrics, 9: 176 - 196.
Bliss, C.I. and Owen, A.R.G. 1958. Negative binomial distributions
with a common k. Biometrika, 45: 37 - 58.
Carpenter, S. J. and LaCasse, W. J. 1955. Mosquitoes of North
America. University of California Press, Berkeley and Los
Angeles, 360 p.
Clements, A.N. 1963. The physiology of mosquitoes. The Mac-
Millan Co. , New York, 393 p.
Fisher, R.A., Corbet, A. S. and Williams, C. B. 1943. The
relations between the number of individuals and the number
of species in a random sample of an animal population. J.
Anim. Ecol. 12: 42 - 58.
Haufe, W. O. 1952. Observations on the biology of mosquitoes
(Diptera: Culicidae) at Goose Bay, Labrador. Canad. Ent.,
84: 254 - 262.
Haufe, W. O. 1953. Predicting mosquito emergence. Proc. New
Jersey Mosquito Extermination Assoc. , 1953, 52 - 57.
Haufe, W. O. 1957. Physical environment and behaviour of immature
stages of Aedes communis (Deg. ) (Diptera: Culicidae) in sub-
arctic Canada. Canad. Ent. , 89: 120 - 139.
Haufe, W. O. and Burgess, L. 1956. Development of Aedes (Diptera:
Culicidae) at Fort Churchill, Manitoba, and prediction of
dates of emergence. Ecology, 37: 500 - 519.
Hocking, B. 1953. Notes on the activities of Aedes larvae. Mos-
quito News, 13: 77-81.
Horsfall, W.R. 1955. Mosquitoes: Their bionomics and relation
to disease. Ronald Press Co. , New York, 723 p.
Horsfall, W.R. and Fowler, H. W. 1961. Eggs of floodwater mos-
quitoes. VIII. Effect of serial temperatures on conditioning
of eggs of Aedes stimulans Walker (Diptera: Culicidae). Ann.
ent. Soc. Amer. , 54: 664 - 666.
222
Population Studies
Jenkins, D. W. 1948. Ecological observations on the mosquitoes
of central Alaska. Mosquito News, 8: 140 - 147.
Khelevin, N.V. 1958. The effect of environmental factors on the
induction of embryonic diapause and on the number of gener-
ations in a season of Aedes caspius dorsalis Mg. (Diptera,
Culicidae). Effect of temperature on the induction of em-
bryonic diapause in Aedes caspius dorsalis Mg. Ent. Rev. 37 :
19 - 35 (cited by Clements, 1963).
Klassen, W. and Hocking, B. 1963. Control of A edes dispersing
along a deep river valley. Mosquito News, 23: 23 - 26.
Klassen, W. and Hocking, B. 1964. The influence of a deep river
valley system on the dispersal of Aedes mosquitoes. Bull,
ent. Res. , 55: 289 - 304.
Knight, K. L. 1951. The Aedes (Ochlerotatus) punctor subgroup in North
America (Diptera, Culicidae). Ann. ent. Soc. Amer. , 44:
87 - 99.
Knight, K. L. 1964. Quantitative methods for mosquito larval
surveys. J. med. Ent., 1: 109 - 115.
Morris, R.F. 1954. A sequential sampling technique for spruce
budworm egg surveys. Canad. J. Zool. , 32: 302 - 313.
Oakland, G. B. 1950. An application of sequential analysis to
whitefish sampling. Biometrics, 6: 59 - 67.
Provost, M. W. 1957. The dispersal of Aedes taeniorhynchus. II. The
second experiment. Mosquito News, 17: 233 - 247.
Pucat, A.M. 1964. Mosquito News, in press.
Rempel, J. G. 1950. A guide to the mosquito larvae of western
Canada. Canad. J. Res. D. , 28: 207 - 248.
Rempel, J. G. 1953. The mosquitoes of Saskatchewan. Canad. J.
Zool., 31: 433 - 509.
Southwood, T.R.E. 1962. Migration of terrestrial arthropods in
relation to habitat. Biol. Rev., 37: 171 - 214.
Stark, R.W. 1952. Sequential sampling of the lodgepole needle
miner. For. Chron. , 23(2): 57 - 60*
Sylvester, E.S. and Cox, E. L. 1961. Sequential plans for sampling
aphids on sugar beets in Kern County, California. J. econ.
Ent. , 54: 1080 - 1085.
Telford, A. D. 1958. The pasture Aedes of central and northern
California. Seasonal history. Ann. ent. Soc. Amer., 51:
360 - 365.
Telford, A. D. 1963. A consideration of diapause of Aedes nigromaculis
and other a edine mosquitoes (Diptera: Culicidae). Ann. ent.
Soc. Amer., 56: 409 - 418.
Vockeroth, J.R. 1954. Notes on the identities and distribution of
Aedes species of Northern Canada, with key to the females
(Diptera: Culicidae). Canad. Ent., 86: 241 - 255.
Waters, W. E. 1959. A quantitative measure of aggregation in
insects. J. econ. Ent., 52: 1180 - 1184.
223
EFFECT OF LARVAL DENSITY ON THE DEVELOPMENT
OF Aedes aegypti (L.) AND THE SIZE OF ADULTS*
YOSH1TO WAD A
Department of Medical Zoology Quaestiones entomologicae
Nagasaki University School of Medicine 1 :223—249. 1965
The effect of larval density of Aedes aegypti (L.) on larval development and the size of
resulting adults was studied in the laboratory. High larval mortality, long larval period, and
small size of resulting adults were observed, when the larval density was high, as well as when
the amount of food was small. Although the high larval density is often associated with shortage
of food, it was demonstrated that even only the high larval density could produce these pheno-
mena, when the amount of food per larva was kept constant. The effect of the density is consid-
ered to be expressed through increased stimulation of larvae by mutual contacts.
INTRODUCTION
The effect of population density on the physiology and ecology
of insects has received much attention by many investigators, as it
is of basic importance in the study of population dynamics. As for
mosquitoes, it is known that high larval densities are associated
with high larval mortality, prolongation of the larval period, and
small size of resulting adults with Aedes aegypti (L. ) (Bar-Zeev, 1957;
Shannonand Putnam, 1934), Anopheles gambiae Giles (Gillies and Shute,
1954), and Anopheles quadrimaculatus Say (Terzian and Stabler, 1949).
AlsoSpielman (1957) and Krishnamurthy and Laven (1961) reported
that overcrowding larvae of Culex pipiens L. f. molestus reduces the
rate of autogeny among the resulting adults, and Gillies and Shute
(1954) mentioned the change in maxillary index of Anopheles gambiae by
larval overcrowding.
Although high larval density, or overcrowding, is often
accompanied by a shortage of food, it seems to be advisable to
separate the effect of density itself from that of starvation, since
the two could be quite different processes. Shannon and Putnam
(1934) seem to have made their experiments by increasing the larval
density and keeping the food amount per container constant. If so,
it is very likely that the larvae in high density were affected not only
by the density itself, but also by the shortage of food. Bar-Zeev
(1957) used a constant amount of food per larva in his experiments
to demonstrate the effect of larval density, and said, "When the
amount of food was not too high, and therefore, no film was formed,
there was no undue mortality under crowded conditions; however,
the development of the larvae was greatly delayed". This seems to
have indicated the effect of density. However, he added "The growth
rate was normal, provided that the amount of food per larva was
* Contribution from the Research Institute of Endemics, Nagasaki
University No. 474 and Contribution No. 143 from the Department of
Medical Zoology, Nagasaki University School of Medicine.
224
Larval Density
adequate, and that the water was renewed so as to prevent the
development of a film of yeast. It can, therefore, be concluded that
the inhibitory effect of crowded conditions on larval development is
due to lack of food".
Thus it seems that no conclusion has been established for
the effect of larval density itself in mosquitoes, and therefore, it
was considered worthwhile to explore this further.
The effect of density would be investigated in an experiment
with a constant quantity of food per individual at varying density levels
(Klomp, 1964). On the other hand, if the quantity of food per con-
tainer is kept constant, the larvae at high density will suffer shortage
of food particularly in the latter part of development, as well as the
effect of high density. In order to recognize the effect of food quantity
free from the effect of density, food quantity would have to be changed
at the same density level.
METHOD OF EXPERIMENTS
The mosquitoes used were Aedes aegypti kept at the Departmertt
of Entomology, University of Alberta. The eggs, not older than 15
days from oviposition, were allowed to hatch in water with a small
quantity of dried yeast (Fleichmann's). The larvae which hatched
within 12 hours were put into cups with 100 ml water containing
dried yeast or rabbit pellets (North West Mill and Feed Co. , Ltd. )
or both. These cups were kept at constant temperatures, and the
observations were made at a certain time every day. At each obser-
vation time, distilled water was added to keep a constant volume.
When pupation occurred, the pupae were put into water in small glass
vials with cotton plugs after recording their number, and emergence
was awaited.
Four experiments were performed.
Experiment 1- This was preliminary in nature. Density range
was 1 to 64 larvae per cup, food used was yeast with quantity range
of 1 to 64 units (1 unit = 1. 7 mg) per cup, temperature, 25. 7 ±1. 5 C.
Experiment II- In this experiment, the quantity of food per
cup was kept constant at various density levels. Density range was
1 to 128 larvae per cup, food used was 64 units of yeast plus 100
units of rabbit pellets per cup, temperature, 29.8 ±1.2 C. From
this experiment, the combined effect of food quantity and larval
density will be seen.
Experiment III- This experiment was done to see the effect of
different foods, that is 64 units yeast, 100 units rabbit pellets, 64
units yeast plus 100 units rabbit pellets, and 64 units yeast plus 200
units rabbit pellets. Density was kept constant at 16 larvae per cup,
temperature, 29.8 ±1.2 C.
Wada
225
Experiment IV - In this experiment, the quantity of yeast per
larva was kept at 1 and 4 units, density range 1 to 256, temperature,
26.3 ±0.9 C. Thus the effect of larval density will be seen from the
data based on series of density levels at constant food quantity per
larva. Also, by comparing in the same density level, the effect of
food quantity will be demonstrated.
RESULTS OBTAINED
Effect of Larval Density on Larval and Pupal Mortalities
The larval and pupal mortalities in Experiments I, II, III,
and IV are given in Tables 1, 2, 3, and 4, respectively.
In Experiment I, low larval mortality was observed at the
density levels of 1 and 4 larvae per cup, when 4 to 64 units of yeast
were supplied to each cup. With increasing density particularly when
the amount of yeast was small, larval mortality became higher. No
pupation occurred in the density 16 with 4 units of yeast per cup or
in the density 64 with 4 or 16 units. No appreciable tendency was
recognized in pupal mortality.
TABLE 1 - Mortalities of Aedes aegypti larvae and pupae reared at
different densities with different amounts of yeast (Ex-
periment I).
* 1 unit =1.7 mg
226
Larval Density-
Experiment II gave generally high pupation rate throughout
the density levels of 1 to 128, indicating that the food used, 64 units
yeast plus 100 units rabbit pellets, is suitable for larval survival.
However, the larval mortality is lower at density 16 than at other
densities, and this seems to indicate the optimum density for larval
survival, with this combination of quantity and quality of the food.
TABLE 2 - Mortalities of A edes aegypti larvae and pupae reared at
different densities with a constant amount of food per cup
(Experiment II).
Food used: 64 unit yeast plus 100 unit rabbit pellets per cup
(1 unit = 1.7 mg).
Experiment III, where the density of larvae was 16 per cup,
shows that larval and pupal mortalities decrease from 64 units yeast
to 64 units yeast plus 200 units rabbit pellets. This means that the
lower food shown in the table is the better food for larval and pupal
survival.
TABLE 3 - Mortalities of A edes aegypti larvae and pupae reared with
different foods (Experiment III).
Density: 16 larvae per cup.
*Y: yeast; R: rabbit pellets; accompanied figure: quantity per
cup in units (1 unit =1.7 mg).
**Data are from Table 2.
Wada
227
In Experiment IV, two series of the amount of yeast, that is
1 and 4 units per larva, were used. When the density was 16 or
less, fairly high pupation was obtained, though the mortality is
slightly higher with 1 unit yeast per larva than with 4 units. In
density 64 with 1 unit yeast per larva, that is 64 units per cup,
larval mortality was more than 50%, and only males pupated. In
the density of 256 with 1 unit yeast per larva, that is 256 units per
cup, larval mortality further increased up to 87%, 40% of pupae
failed to emerge, and very low proportion of females was obtained.
Very high larval mortality was observed also in the density of 256
with 4 units of yeast per larva, that is 1024 units per cup. This
amount of yeast seemed to be too much for 100 ml water, because
a film was formed on the water surface and high mortality occurred
in earlier instars, unlike other combinations of density and food
amount. Thus such a very low pupation rate as 2. 9% is not due to
the effect of high larval density, but probably to the film formation
or other unfavorable conditions of the culture medium.
TABLE 4 - Mortalities of A edes aegypti larvae and pupae reared at
different densities with 2 series of a constant amount of
yeast per larva (Experiment IV).
* 1 unit =1.7 mg
In short, larval mortality generally increases with increased
density and decreased food quantity through the shortage of food and
the larval density itself. There seems to be an optimum density for
larval survival, which differs from the minimum density. If the
conditions are not suitable, then the favored sex is the male.
228
Larval Density-
Effect of Larva! Density on Pupation Curve
Frequency curves of pupation by sex in four experiments are
shown in Figs. 1, 2, 3, 4, and 5. Males pupated earlier than females
throughout the experiments. Generally a shorter larval period is
seen in the cups where the density is lower and the amount of yeast
is larger. When larval periods are compared on the basis of the
same density with different amounts of food (see Fig. 1; compare
Figs. 4 and 5), a longer larval period is seen with the decreased
amount of food.
When the amount of food per larva was kept constant and the
density of larvae was increased, the delay in development is clear,
as seenin Experiment IV (Figs. 4 and 5). This is attributable to the
effect of high larval density, not to the shortage of food, because the
comparisons were made on the basis of the same amount of food per
larva.
Here, it is apparent that the larval development is affected
not only by the quantity of food, but also directly by the larval density,
and the effect is more remarkable, when the amount of food per
larva is smaller.
.It is interesting that the longer larval period is usually
associated with increased variation in larval period and with a
tendency to be skewed towards the right. If a pupation curve is
normally distributed, thenitis expected that a cumulative percentage
frequency of pupation in probit will be linear. Now, the normality
of the pupation curves in Experiments II and IV, in which a fairly
large number of larvae was used, was examined.
Cumulative percentage pupation in probit is plotted against
larval period (days) in Figs. 6 to 11. When the density is low and
food amount is large a linear relationis seen, that is, those pupation
curves are shown to follow the normal distribution. The deviation
from the normal distribution becomes remarkable with increasing
density and decreasing food quantity. Thus there is some deviation
from the normal distribution in the pupation curve, particularly
when the conditions are unfavorable for larval development. Even
when conditions are good, a few individuals sometimes pupate very
iate. For this reason, it seems that the median is a better represen-
tative of larval period than the mean.
Effect of Larva! Density on Larva! and Pupa! Periods
Figs. 12, 13, and 14 show the relation between median larval
period and larval density per cup for Experiment I, II, and IV,
respectively. In these figures, the points with the same amount of
food per cup were connected by straight lines. Generally, the median
larval period becomes longer with increasing larval density. This
is rather natural, because the amount of food per larva decreases
with increasing density.
By connecting the points with the same amount of food per
larva, the data for Experiments I and IV are represented in Figs.
15 and 16. In density levels of 256 and 64 of Experiment IV, a
longer median period was obtained than in 1, 4 or 16, in spite of the
Frequency
229
1- 4
1-16
1-64
4-64
Fig. 1. Frequency distributions of larval period of Aedes aegypti (Experi-
ment I). 4-64, for example, indicates that the larval density is
4 and the amount of yeast is 64 units per cup. O : males; • ;
females .
Frequency
230
Fig. 2. Frequency distributions of larval period of Aedes aegypti (Experi-
ment II). Figure shown indicates larval density. O : males;
# : females.
Wada
231
fact that the amount of food available for each larva in higher densities
is the same as, or even slightly larger than, in lower densities.
Here, the effect of high larval density is again suggested. Also in
Experiment I, the tendency of the median to increase is seen at the
density levels of 64 or more. It is interesting that there seems to
exist a valley in median larval period at density 16, particularly
Fig. 3. Frequency distributions of larval period of Aedes aegypti (Experi-
ment III). Y and R and accompanied figure indicate yeast and
rabbit pellets and their amount in units. O : males, # :
females .
232
Larval Density
8
256
16 24
Larval period (days)
32
Fig. 4. Frequency distributions of larval period of Aedes aegypti (Amount
of yeast per larva: 4 units; Experiment IV). Figure shown
indicates larval density. O : males; • : females.
when the amount of food is small, and furthermore, in the food
amount of 1 in Experiment IV, the median becomes again smaller
at density 1 than 4. The reasons for such peculiarities of the curves
are not clear, but it seems that the median is determined by a balance
between the effects of larval density and the amount of food available,
and perhaps some other factors.
Wada
233
No distinct difference in pupal period was recognized among
various amounts of food nor among larval density levels, though
pupal density may affect the period. It seems that the pupal period
is affected only by temperature, or at least, if some other factors
affect it, their effect is. very small. In Table 5, mean pupal periods
in days are given by sex at the three different temperatures. The
female has a slightly longer pupal period than the male.
It would be practically right to suppose that the larval period
is determined by temperature, larval density, and the conditions of
culture medium such as the quality and quantity of food, but the pupal
period is determined only by temperature. The ratio of larval period
seems, then, to indicate the suitability of the conditions for larval
development. Thisratiomay beused to compare the larval period,
even when experiments were made at different temperatures.
The calculated values for the ratio areshownin Tables 6 and
7, and compared on the basis of the same combinations of larval
density and food amount in different experiments. The ratios for
the combinations of D1Y4 (density 1 larva per cup, yeast 4 units
per cup), D4Y16, and D16Y64 agree quite well among experiments,
but those for D4Y4, D16Y16, D64Y16, and D64Y64, are rather
Fig. 5. Frequency distributions of larval period of Aedes aegypti (Amount.
of yeast per larva: 1 unit; Experiment IV). Figure shown
indicates larval density. O : males; # : females.
234
Larval Density
different from one another. The number of larvae used in Experi-
ment I was not sufficient, and the latter combinations are considered
somewhat unsuitable so that very slight differences in the conditions
will make rather great changes in larval development. These would
be responsible for rather great difference of the ratios in the latter
group of combinations.
The above procedure will be valid only if the ratio of larval
period to pupal period is constant over a reasonable temperature
range. For this reason, further studies are required to determine
the usefulness of the ratio. However, it is clear from the tables
that larval period varies greatly with the quantity and quality of food
at the same density level, and also that the same amount of food per
cup, or even per larva, does not give the same larval period at
different density levels. Therefore, care should be taken in attem-
pting to determine the larval period at a certain temperature, or the
developmental zero of mosquito larvae by rearing them at different
temperatures.
Fig. 6. The relation between cumulative percentage pupation (probit
scale) and larval period in males of Aedes aegypti (Experiment II).
Figure shown indicates larval density.
Wada
235
Effect of Larval Density on Body Size of Resulting Adults
In Figs. 17 and 18, thefrequency distributions of wing length
of the resulting adults in Experiments II and IV are given.
In Experiment II (Fig. 17), the wing length increases in both
sexes slightly from density 1 to 16 larvae per cup, and decreases
greatly with increasing density from 16. Fig. 18 shows the similar
situationin Experiment IV, except for density 256 withyeast 4 units
per larva, where the wing length is not consider ed to reflect the effect
of this density, owing to high larval mortalityin the earlier instar s,
as mentioned earlier. However, the changes in wing length are less
remarkable than in Experiment II. This is due to the fact that the
quantity of food per cup was kept constant in Experiment II, on the
other hand in Experiment IV the quantity per larva was kept constant.
Nevertheless, the apparent effect of larval density on the wing length
can be seen in Experiment IV (Fig. 18).
Fig. 7. The relation between cumulative percentage pupation (probit
scale) and larval period in females of Aedes aegypti (Experiment
II). Figure shown indicates larval density.
Larval period (days)
236
Wada
237
It seems that the wing length of females is more sensitively-
affected than that of males with deer easing suitability for the larval
stage, so that considerable overlapping in wing length of both sexes
appears, as for example between densities 64 and 128 in Experiment
II (Fig. 17). When the' conditions become still less suitable, only
males will pupate, as indicated from the densities 64 and 256 with
yeast 1 unit per larva.
It is interesting that the frequency curve becomes steeper
at the right hand side with decreasing suitability in the conditions
for larval development, but the reasons for this are not yet clear.
Figs. 19 and 20 show the frequency curves of thorax length
in Experiments II and IV. The thorax length shows a similar ten-
dency to the wing length, excepting that the steepness of the curves
at the right hand side is not seen, when the conditions become un-
favorable.
Fig. 10. The relation between cumulative percentage pupation (probit
scale ) and larval period in males of Aedea aegypti (Amount of yeast
per larva: 1 unit; Experiment IV). Figure shown indicates
larval density.
Larval period (days)
238
IO Ot
Cumulative % pupation
o* oo
o o o
3
<0
09
Wada
239
TABLE 5 - Pupal periods of Aedes aegypti by sex at different temper-
atures.
TABLE 6 - The ratio of larval to pupal period of the males of Aedes aegypti
(Experiments I - IV).
* See Table 3.
** Unable to pupate.
*** Film was formed on water surface, larval mortality was very high.
240
Figs. 12-14. Median larval period of Aedes aegypti at each density level.
Points for the same amount of yeast per cup are connected by
lines. 12. Experiment I. 13. Experiment II. 14. Experi-
ment IV. Figure shown indicates the units of yeast per cup.
O : males; # : females.
Median larval period (days)
241
Figs. 15 & 16. Median larval period of Aedes aegypti at each density level.
Points for the same amount of yeast per larva are connected
bylines. 15. Experiment I. 16. Experiment IV. O: males;
0 : females.
242
Larval Density
TABLE 7 - The ratio of larval to pupal period of the females of
Aedes aegyp£t(Experiments I - IV).
* See Table 3.
**, *** See Table 6.
** ** Only three females pupated.
In Fig. 21, the relation between mean wing length and mean
thorax length is shown for each density level in Experiment II and
for each type of food in Experiment III. In density level of 16 or
more in Experiment II, both wing and thorax decrease in length
along a straight line with increasing density. However, in densities
lower than 16, decreased wing length and rather unchanged thorax
length are shown, that is, the points for density levels of 1 and 4
are situated above the line through the points for density 16 to 128.
The larvae in lower density receive relatively large amounts of
food, because the amount of food per cup was kept constant in this
Wada
243
experiment. Therefore, it may be said that at these low density
levels the adults resulting from favorable conditions have relatively
shorter wing length than those from less favorable conditions.
The same is seen in Experiment III, where different diets
were given to the larvae with the same density of 16. The point for
the adults from the culture containing yeast 64 units plus rabbit
pellets 200 units, which is more suitable than yeast 64 plus rabbit
pellets 100 used in Experiment II, is situated above the line through
the points for density 16 to 128 in Experiment II, and the point for
the less suitable diet, yeast 64, below the line.
■aaSuft. aA>,
.e&ZS ■ a ■ A/V
Fig. 17. Frequency distributions of wing length of Aedea aegypti (Experi-
ment II). Figure indicates larval density. Q : males;
# : females.
Frequency
244
Vxl^A^<IOoq»#^ 1 1
y:>?0o0<W^*. V\/,V*»A„«i 4 1
a^^VMa.
16-1
64-1
256-1
10
0
# : females.
Wada
245
In Experiment IV, the situation becomes more complicated,
because the experiment consisted of two series of constant amount
of food per larva, and it is not easy to say which combination of
larval density and amount of food is more favorable for the larval
stage, especially at lower density levels. However, it is seen that
the relative wing length to thorax length for larger amounts of food
or lower density of larvae tends to be smaller than others.
Fig. 19. Fr equency distributions of thorax length of Aedes aegypti (Experi-
ment II). Figure shown indicates larval density. O : males;
6 : females.
Frequency
246
10-
0-
20-
10-
0-
Fig. 20. Frequency distributions of thorax length of Aedes aegypti (Experi-
ment IV). 16~4, for example, indicates that the larval density
is 16 and the amount of yeast is 4 units per larva. O : males;
• : females.
Mean thorax length (units) Mean thorax length (units)
247
Figs. 21 8s 22. The relation between mean thorax length and mean wing length of
Aedes aegypti. Figure shown indicates larval density in Experiment II, 6/ftd
and Y, R, and accompanied figure indicate yeast, rabbit pellets, and their
amounts in units in Experiment III. In Experiment IV, 16-4 for example,
indicates larval density — yeast units. O : males; • : females.
248
Larval Density
The adults from the culture with density 1 and yeast 4 per
cup are not considered to have a relatively smaller wing length than
other densities, unlike Experiment II. This is perhaps because of
the fact that the food of yeast 4 in Experiment IV is apparently less
favorable than that of yeast 64 plus rabbit pellets 100. The adults
with relatively small wing length from low larval density seem to
appear only when the amount of food is large.
CONCLUSIONS
The results obtained are summarized in Table 8. The density
in this table is used in a relative sense to food quantity. Actual
density differs according to the amount of food.
TABLE 8 - Summary of the effects of larval density in Aedes aegypti ,
High larval density apparently has detrimental effects on the
mosquito. Interesting is the relation between very low and low
densities. The characteristics seem to indicate that the adults from
very low larval density have a slightly reduced flight ability in
comparison with those from less low density, as far as judged from
the relative wing length. However, repeated experiments are
desired, as the number of mosquitoes used in lower densities was
not very large.
Wada
249
CONSIDERATIONS ON THE MANNER IN WHICH
LARVAL DENSITY PRODUCES ITS EFFECTS
From the preceding sections, the effect of larval density is
apparent, but its process was not particularly investigated. Since
no effects of metabolic wastes of larvae have been demonstrated
(Bar-Zeev, 1957; Shannon and Putnam, 1934), high larval density
seems to influence the mosquitoes through the stimulation of in-
creased mutual contacts.
Shannon and Putnam (1934) stated "DeBuck, Schoute, and
Swellengr ebel (1932) claim . . . that when they (anopheline larvae)
live in overcrowded conditions food may remain undigested in the
alimentary tract from 12 to 24 hours . . . Improper nourishment due
to massing habits of the larvae (of Aedes aegyph)may account for this
(phenomenon at high larval density) ...". However, the situation
seems to be more complex than Shannon and Putnam (1934) thought,
and neuro-physiological processes may be involved.
ACKNOWLEDGEMENTS
I wish to express my sincere thanks to Professor B. Hocking
of the Department of Entomology, University of Alberta, for his
valuable suggestions and encouragement, and to the Parks and
R ecreation Department of the City of Edmonton for financial support.
REFERENCES
Bar-Zeev, M. 1957. The effect of density of the larvae of a mos-
quito and its influence onfecundity. Bull. Res. Coun. Israel
(B), 6B: 220 - 228.
Gillies, M. T. and Shute, G. T. 1954. Environmental influences
and the maxillary index in Anopheles gambiae . Nature, Lond. ,
173: 409 - 410.
Klomp, H. 1964. Intraspecific competition and the regulation of
insect numbers. Annu. Rev. Ent. ,9: 17 - 40.
Krishnamurthy, B.S. and Laven, H. 1961. A note on inheritance
of autogeny in Culex mosquitos. Bull. World Hlth. Org. ,
24: 675 - 677.
Shannon, R.C. and Putnam, P. 1934. The biology of Stegomyia under
laboratory conditions. (1) the analysis of factors which in-
fluence larval development. Proc. ent. Soc. Wash. , 36 :
185 - 216.
Spielman, A. 1957. The inheritance of autogeny in the Culex pipiens
complex of mosquitoes. Amer. J. Hyg. , 65: 404 - 425.
Terzian, L.A. andStahler, N. 1949. The effects of larval population
density on some laboratory characteristics of Anopheles
quadrimaculatus Sa.y. J. Parasit. , 35: 487 - 498.
250
ANNOUNCEMENT
Quaestiones available on Microfilm
Arrangements have been made with University Microfilm
Inc. to have Quaestiones Entomologicae made available on microfilm
for subscribers to the paper edition. Microfilm copies will become
available after the completion of each volume.
For further information concerning this inexpensive space
saving plan, please direct enquiries to University Microfilms, 313
N. First Street, Ann Arbor, Michigan, U.S.A.
CORRIGENDA
Page 101, couplet 4, add page numbers: 138 and 111.
couplet 7, add page number 102.
Page 109, lines 15 8* 16, insert:
Distribution
A total of 3,433 specimens was examined.
Page 151, caption for Fig. 37 should read:
C. oregona XC. duodec imguttata fr om Garth, Alberta.
Add caption for Fig. 38:
C. oregona XC. (/uoflfecimgu^atafrom Rocky Mountain House,Alberta
Page 169, after MacGinitie, insert:
Martin, P.S. 1958. Pleistocene ecology and biogeography of North
America, pp. 375-420, in C.L. Hubbs, (ed.), Zoogeography.
Amer. Assoc. Adv. Sci. Symp. , Washington, D.C.
Publication of Questiones Entomologicae was started in 1965 as part
of a memorial project for Professor E. H. Strickland, the founder of the
Department of Entomology at the University of Alberta in Edmonton
in 1922.
It is intended to provide prompt low-cost publication for accounts of
entomological research of greater than average length, with priority given
to work in Professor Strickland’s special fields of interest including ento-
mology in Alberta, systematic work, and other papers based on work
done at the University of Alberta.
Copy should conform to the Style Manual for Biological Journals
published by the American Institute of Biological Sciences, Second
Edition, 1964, except as regards the abbreviations of titles of periodicals
which should be those given in the World List of Scientific Periodicals,
1964 Edition. The appropriate abbreviation for this journal is Quaest ent.
An abstract of not more than 500 words is required. All manuscripts will
be reviewed by referees.
Illustrations and tables must be suitable for reproduction on a page
size of 9%X63/4 inches, text and tables not more than 7%X43/4 inches,
plates and figures not more than 8^X5 inches. Reprints must be ordered
when proofs are returned, and will be supplied at cost. Subscription rates
are the same for institutions, libraries, and individuals, $4.00 per
volume of 4 issues, normally appearing at quarterly intervals; single
issues $1.00. An abstract edition is available, printed on one side of
paper or on one or both sides (according to length) of 3X5 inch index
cards (at $1.00 per volume) or on 5X8 inch standard single row punched
cards ($1.50 per volume).
Communications regarding subscriptions and exchanges should be
addressed to the Subscription Manager and regarding manuscripts to:
The Editor, Quaestiones Entomologicae,
Department of Entomology,
University of Alberta, Edmonton, Canada.