7.-/ 7

C. 3k

»&**>**

t^T****

S^e

^

number 17

N.C. DOCUMENTS

CL! "HOUSE

FEB II I992

N.C. STATE LIBRARY

RALEIGH

december 1991

EDITORIAL STAFF

Frank J. Radovsky, Editor

Eloise F. Potter, Managing Editor

Sheree Worrell, Production Manager

RESEARCH CURATORIAL STAFF

North Carolina State Museum of Natural Sciences

Alvin L. Braswell William M. Palmer

Curator, Lower Vertebrates Curator of Lower Vertebrates

Mary K. Clark Frank J. Radovsky

Curator of Mammals Director, Research and Collections

John A. Gerwin Rowland M. Shelley

Curator, Birds Curator of Invertebrates

David S. Lee Vincent P. Schneider

Curator of Birds Research Associate, Paleontology

Brimleyana, the Journal of the North Carolina State Museum of Natural

Sciences, appears at irregular intervals in consecutively numbered issues. Contents

emphasize zoology of the southeastern United States, especially North Carolina

and adjacent areas. Geographic coverage is limited to Alabama, Delaware,

Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina,

South Carolina, Tennessee, Virginia, and West Virginia.

Subject matter focuses on systematics, evolution, zoogeography, ecology,

behavior, and paleozoology. Papers stress the results of original empirical field

studies, but synthesizing reviews and papers of significant historical interest to

southeastern zoology are included. Brief communications are accepted.

Appropriate specialists review each manuscript, and final acceptability is

determined by the Editor. Address manuscripts and all correspondence (except

that relating to subscriptions and exchanges) to Editor, Brimleyana, North

Carolina State Museum of Natural Sciences, P.O. Box 27647, Raleigh, NC

27611.

Address correspondence pertaining to subscriptions, back issues, and exchanges

to Brimleyana Secretary, N.C. State Museum of Natural Sciences, P.O. Box

27647, Raleigh, NC 27611.

In citations please use the full name — Brimleyana.

North Carolina State Museum of Natural Sciences

North Carolina Department of Agriculture

James A. Graham, Commissioner

CODN BRIMD 7

ISSN 0193-4406

Recent Catastrophic Decline of Mussels

(Bivalvia: Unionidae)

in the Little South Fork Cumberland River, Kentucky

Robert M. Anderson, James B. Layzer, and Mark E. Gordon

US. Fish and Wildlife Service

Tennessee Cooperative Fishery Research Unit1

Tennessee Technological University

Cookeville, Tennessee 38505

ABSTRACT— During 1987, we sampled 16 sites in the Little South

Fork Cumberland River to assess the status of mussel populations. We

found 21 species, of which 12 were alive. Mean densities in the lower

one-third of the stream had declined from 2.9 to 7.5 mussels/ m2 in

1981 to 0 to 1.1 mussels/ m2 in 1987. Moreover, we found few live

mussels at six other stations in this section of stream. Between 1981

and 1987, two species {Villosa trabilis and Pegias fabula), listed as

federally endangered, appeared to have been extirpated from the lower

one-third of the stream. Viable populations of most mussel species are

now restricted to the middle section of the Little South Fork. Surface

mining of coal has increased in the lower watershed of the Little South

Fork in recent years and may be responsible for the mussel decline.

Unlike the unionid mussel fauna, the density of the exotic Corbicula

fluminea has increased nearly ninefold since 1981. This increase appears

to be related to the general population expansion of C fluminea in the

Cumberland River system.

Harker et al. (1979, 1980) reported on environmental conditions of

the Little South Fork Cumberland River (LSF). They considered the

stream to have relatively high water quality in comparison with other

drainages in the upper Cumberland River basin and to support a fairly

speciose flora and fauna. A total of 24 species of unionid mussels were

later identified from LSF (Starnes and Bogan 1982); one-third of these

species were Cumberlandian endemics, including two that are listed as

federally endangered, Pegias fabula (Lea, 1838) and Villosa trabalis

(Conrad, 1834). At the time of its discovery in 1977, this population of

P. fabula was considered to be the healthiest known, and LSF was

"perhaps the most pristine stream remaining within the entire known

range of Pegias in the Cumberland and Tennessee drainages" (Starnes

and Starnes 1980).

Starnes and Bogan (1982), who made an extensive survey of

unionids in LSF in 1981, reported qualitative information for the entire

'Cooperators are the Tennessee Wildlife Resources Agency, Tennessee Technological

University, and the U.S. Fish and Wildlife Service.

Brimleyana 17:1-8, December 1991

2 R. M. Anderson, J. B. Layzer, and M. E. Gordon

length of unimpounded river and quantitative data for three sites in the

lower 22.8 km of river, which is designated a Kentucky Wild River. The

two most common species in their quantitative samples were the

Cumberlandian endemics Ptychobranchus subtentum (Say, 1825) and

Pegias fabula. Alhstedt (1986), who examined several sites in LSF in

1984 and 1985, found numerous fresh-dead and relic unionid shells but

few living specimens. He observed deposits of silt at one riffle in the

lower section of the river, but the source and effects of this siltation

apparently were not investigated. Our study was made to determine the

present distribution of mussels within LSF, with particular reference to

the status of the two endangered species.

MATERIALS AND METHODS

In 1987, we repeated qualitative surveys at the 16 sites (Fig. 1)

studied by Starnes and Bogan (1982). At each site we searched for

mussels with glass-bottom buckets and by snorkeling. Relic shells (shells

showing chalkiness or algal growth on the nacre) and fresh-dead shells

were retained; live mussels were identified and returned to the river.

Three sites (8, 13, and 16 in Fig. 1) were sampled quantitatively. Ten

samples (0.092 m2) were taken along each of three transects at each site.

The samples were collected by removing the substrate from the stream

bottom to a depth of about 1 1 cm and placing the material into the net

of a Surber sampler. On shore, the collected substrate was separated

with a 4-mm sieve. This procedure differed from the procedure used by

Starnes and Bogan (1982) only in that they used a mask and snorkel to

search the substrate within the 0. 1 m2 frame in situ. Thus, our method

may have been more efficient in recovering small unionids and

Corbicula.

RESULTS

A total of 21 unionid species and the Asian clam, Corbicula

fluminea (M tiller, 1774), were collected (Table 1). The overall species

composition determined from shells and live specimens was similar in

our study to that reported by Starnes and Bogan (1982). Although we

did not collect four species reported by Starnes and Bogan (1982),

namely, Cyclonaias tuberculata (Rafinesque, 1820), Anodonta grandis

Say, 1829, Anodonta imbecillis Say, 1829, and Actinonaias ligamentina

(Lamarck, 1819), we collected Villosa lienosa (Conrad, 1834), a species

not previously reported from LSF. Further, we found more species than

did Starnes and Bogan at sites 4, 5, and 7 (Table 1). In contrast, when

only live mussels are considered, the results of our sampling differed

substantially from those of Starnes and Bogan (1982). We collected

more species alive upstream from site 8 than they had, but found

substantially fewer species alive at all downstream sites. In fact, most

Catastrophic Decline of Mussels

(UL

Lick Creek-

Baker Branch

Kidds Branch

Kennedy Creek

scale

0 12 3 4 5 6

kilometer

N

Route 92

N Route 156

TN

Fig. 1. Map of the Little South Fork Cumberland River, showing the 16

sampling sites. Location of map is shown on inset map of Kentucky.

4 R. M. Anderson, J. B. Layzer, and M. E. Gordon

species were represented by single specimens (which was not true

upstream), and at four downstream sites we found no live unionids.

Unlike the unionids, C. fluminea was found alive at all sites where it

had been previously reported, as well as at four additional sites

upstream. Data were not collected on other invertebrates or on fishes;

however, both were commonly observed throughout the stream.

Our quantitative sampling also indicated that the mussel fauna

significantly declined between 1981 and 1987. Apparently there was an

almost complete kill of the unionid fauna at sites 8, 13, and 16 (Table

2). Ptychobranchus subtentum was the only unionid found alive in the

quantitative sampling; three specimens were collected at site 13. Between

1981 and 1987, the density of C. fluminea increased markedly at all

three sites.

Table 1. Bivalves recorded from the Little South Fork Cumberland River in 1987.

Stations

Species 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16

Alasmidonta viridis (Rafinesque, 1820) R

Villosa taeniata (Conrad, 1834)b

Lampsilis fasciola Rafinesque, 1820

Medionidus conradicus (Lea, 1834)

Ptychobranchus fasciolaris (Rafinesque, 1820)

Ptychobranchus subtentum (Say, 1825)

Strophitus undulatus (Say, 1817)

Villosa iris (Lea, 1829)

Villosa trabalis (Conrad, 1834)

Lasmigona costata (Rafinesque, 1820)

Pleurobema oviforme (Conrad, 1834)

Lampsilis cardia Rafinesque, 1820

Toxolasma lividum (Rafinesque, 1831)

Alasmidonta marginata Say, 1818

Elliptio dilatata (Rafinesque, 1820)

Leptodia fragilis (Rafinesque, 1820)

Pegiasfabula(Lea, 1838)

Obovaria subrotunda (Rafinesque, 1820)

Proptera a/a/a (Say, 1817)

Villosa lienosa (Conrad, 1834)

Actinonaias ligamentina (Lamark, 1819)

Actinonaias pectorosa (Conrad, 1834)

Corbicula fluminea (Miller, 1774)

aR = relic shells; L = live specimens observed; * - live specimens observed by Starnes and Bogan

(1982). Species listing is by first occurrence starting at the most upstream site.

b Represents the form listed as Villosa taeniata punctata in Starnes and Bogan (1982).

Catastrophic Decline of Mussels 5

Table 2. Comparison of mussel and Corbicula densities (No./m2) estimated

from quantitative samples taken in 198 la and in 1987.

DISCUSSION

Downstream from site 7, unionid populations have been devastated.

Although remnants of the fauna persisted in this section, we saw few

live mussels. The decline in these populations is further illustrated by

comparing estimated densities with those of Starnes and Bogan (1982)

(Table 2). During our survey, only site 13 had mussel populations that

were plentiful enough to enable us to estimate population density — 1.1

mussels/ m2 compared with 7.2 mussels/ m2 estimated by Starnes and

Bogan (1982). Despite the disappearance of unionids, the estimated

populations of C. jluminea increased about ninefold at all three sites

from 1981 to 1987. Although not quantitatively sampled in our study,

pleurocerid snail populations were apparently healthy and often dense.

Although the exact cause of the mussel kill in the lower third of

LSF is unknown, it is apparent that not all species have been affected to

the same degree. In 1981 P. fabula was found at densities of 2.2/m2 at

site 8 and was the most abundant mussel at site 16 (Starnes and Bogan

1982). Although we found shells of P. fabula at all sites from site 8

downstream to site 15, we failed to find living specimens at these sites.

Similarly, V. trabalis, which was previously reported alive at two sites,

was not found alive in this area. In contrast, Ptychobranchus subtentum

was the most abundant mussel in both 1981 and 1987. Although it was

the mussel most frequently seen alive in the lower section of the stream,

its densities were much lower in 1987 than in 1981.

The increase in density of C. Jluminea in the lower section of LSF

and its upstream range extension are in marked contrast to the unionid

populations. Corbicula jluminea has been invading the Cumberland

River and its tributaries for many years (Counts 1986). Apparently it is

still colonizing the LSF; however, part of the observed increase in

6 R. M. Anderson, J. B. Layzer, and M. E. Gordon

Corbicula densities may be a result of our greater sampling efficiency,

especially in detecting small individuals. Although high densities of

Corbicula may negatively affect mussels (Clarke 1988), the adverse

effects reported usually have occurred at substantially higher Corbicula

densities than those observed in our study (Gardner et al. 1976).

Corbicula is generally more tolerant than unionids of environmental

stress (Horn and Mcintosh 1979). Tolerance to, and even preference for,

silted areas by C. fluminea has been reported (Belanger et al. 1985). The

increase in Corbicula densities could be a result of an increasingly

favorable substrate.

The source of silt in the LSF has not been documented. However,

strip-mining activities have increased greatly since 1981, and heavy

siltation has been observed at a few sites within the affected area (S. A.

Ahlstedt, personal communication; G. A. Schuster, personal communi-

cation). On several occasions we also observed a very fine sediment at

Ritner Ford (station 15). These silt deposits varied in depth, with the

greatest accumulation occurring along the sides of riffles. Silt deposits

appeared to be transient and were readily flushed during freshets.

If the observed silt was a result of strip-mining, it may have

contained potentially toxic concentrations of metals such as aluminum

or zinc (Dick et al. 1986). "Yellow boy," a ferric precipitate often

associated with mining activity, was observed in Lick Creek, Jones

Hollow, Baker Branch, and several other unnamed tributaries of LSF.

Most of these tributaries are intermittent. Although we never saw

yellow boy in the river itself for more than a few meters below a source

during low flow, we observed these precipitates to be readily transported

during high flows.

The upstream extent of the mussel die-off in LSF appeared to be

between sites 6 and 8. Because site 6 is a commonly used ford, the

absence of live unionids in 1981 and in 1987 is probably a result of

physical disturbance. Moreover, a cursory examination of a riffle just

upstream from site 6 indicated the presence of live unionids, including

P. fabula. All of the tributaries that drain known strip mines in this

watershed enter the LSF downstream from site 6. Moreover, the pattern

of mining in the LSF appears to be correlated with the sequential die-

off of mussels. Intensive mining began in the lower portion of the

watershed and has been progressing upstream. In 1981 a rich mussel

fauna existed throughout the LSF, and there was relatively little mining.

By April 1985, P. fabula had been extirpated from sites 13 and 14 but

persisted at site 8 (Ahlstedt 1986). In November 1984, a permit was

issued to strip-mine 100 ha in the Kidd's Branch sub-basin of the LSF.

Kidd's Branch enters LSF between sites 6 and 8. By 1987, P. fabula and

most other unionids had disappeared from site 8. This pattern of mining

Catastrophic Decline of Mussels

activity and the results of our study implicate strip-mining as the cause

of the mussel kill. Because unionids are among the most sensitive

organisms to acid mine drainage and siltation (Simmons and Reed

1973), their protection may require stricter regulations on strip-mining

or increased monitoring of mining activities and better enforcement of

existing regulations.

ACKNOWLEDGMENTS.— We thank Donald P. Clover, Scott L.

Niemela, and Tina M. Theyel for their assistance in sampling. We thank

Steven A. Ahlstedt (Tennessee Valley Authority), Richard G. Biggins

and Lynn B. Starnes (U.S. Fish and Wildlife Service), Sam M. Call

(Kentucky Division of Water), Guenther A. Schuster (Eastern Kentucky

University), and personnel of the Kentucky Department for Surface

Mining Reclamation and Enforcement for providing unpublished data.

This study was funded by the Tennessee Wildlife Resources Agency and

Kentucky Department of Fish and Wildlife Resources under Section 6

of the Endangered Species Act. Additional funds and support were

provided by the Center for the Management, Utilization, and Protection

of Water Resources, Tennessee Technological University.

LITERATURE CITED

Ahlstedt, S. A. 1986. A status survey of the little-winged pearly mussel, Pegias

fabula (Lea, 1838). Endangered Species Field Off., U.S. Fish Wildl. Serv.,

Asheville, N.C.

Belanger, S. E., J. L. Farris, D. S. Cherry, and J. Cairns, Jr. 1985. Sediment

preference of the Asiatic clam, Corbicula fluminea. Nautilus 99:66-73.

Clarke, A. H. 1988. Aspects of corbiculid-unionid sympatry in the United

States. Malacol. Data Net 2:57-99.

Counts, C. L. 1986. The zoogeography and history of the invasion of the

United States by Corbicula fluminea (Bivalvia: Corbiculidae). Am. Malacol.

Bull. Spec. Ed. 2:7-40.

Dick, W. A., J. V. Bonta, and F. Haghiri. 1986. Chemical quality of suspended

sediment from watersheds subjected to surface coal mining. J. Environ.

Qual. 14:289-293.

Gardner, J. A., Jr., W. R. Woodall, Jr., A. A. Staats, Jr., and J. F. Napoli.

1976. The invasion of the Asiatic clam {Corbicula manilensis Philippi) in

the Altamaha River, Georgia. Nautilus 90: 1 17-125.

Harker, D. F., S. M. Call, M. L. Warren, K. E. Camburn, and P. Wigley. 1979.

Aquatic Biota and Water Quality Survey of the Appalachian Province,

Eastern Kentucky, Vol. 1. Tech. Rep., Kentucky Nat. Preserves Comm.,

Frankfort.

8 R. M. Anderson, J. B. Layzer, and M. E. Gordon

Harker, D. F., M. L. Warren, K. E. Camburn, S. M. Call, G. J. Fallo, and P.

Wigley. 1980. Aquatic Biota and Water Quality Survey of the Upper

Cumberland River Basin, Vol. 1-2. Tech. Rep., Kentucky Nat. Preserves

Comm., Frankfort.

Home, R., and S. Mcintosh. 1979. Factors influencing distributions of mussels

in the Blanco River of central Texas. Nautilus 94: 1 19-1 33.

Simmons, M., and J. R. Reed, Jr. 1973. Mussels as indicators of biological

recovery zones. J. Water Pollut. Control. Fed. 45:2480-2492.

Starnes, L. B., and A. E. Bogan. 1982. Unionid Mollusca (Bivalvia) from Little

South Fork Cumberland River, with ecological and nomenclatural notes.

Brimleyana 8:101-1 19.

Starnes, L. B., and W. C. Starnes. 1980. Discovery of a new population of

Pegias fabula (Lea) (Unionidae). Nautilus 94:5-6.

Accepted November 1989

Ants and Cockroaches Trapped Outside Suburban Houses

in the Area of Raleigh, Wake County, North Carolina1

C. G. Wright

Department of Entomology

North Carolina State University

Raleigh, North Carolina 27695-7613

T. P. Nuhn

Apt. No. 204, 2057 N. Glebe Road

Arlington, Virginia 22207

AND

H. E. Dupree, Jr.

Department of Entomology

North Carolina State University

Raleigh, North Carolina 27695-7613

ABSTRACT.— Ants and cockroaches were collected from pitfall

traps placed close to houses. The traps were unbaited or were baited

with either boiled raisins or bread. Baited traps collected more ants

and cockroaches than unbaited ones. More ants were taken with

raisins and more cockroaches were taken with bread. Significant

differences were not recorded in relation to ground cover or direction

of trap position. Of the 26 ant species trapped, Pheidole dentata and

Camponotus americanus were the most frequently caught. Parcoblatta

uhleriana was the cockroach species most commonly trapped.

Insects present near suburban houses have received little research

attention. More attention has been given to those species found inside

houses and other buildings because of their visibility and pest potential

to inhabitants. Ebeling (1978) and Mallis (1982) discussed some of the

ant and cockroach species found near buildings. One of the most

comprehensive literature surveys on the ecology of these and other

insects in the urban outdoor environment was by Frankie and Ehler

(1978). A public-opinion survey of the principal outdoor pest problems

for upper and lower middle-income families was made by Frankie and

Levensen (1978). Also, studies have been made of the occurrence and/ or

the ecology of specific groups or species of crawling insects present at

ground level in the urban or suburban outdoor environment. Examples

'Paper No. 12379 of the Journal Series of the North Carolina Agriculture Research

Service, Raleigh, NC 27695-7643. Partial support furnished by Chevron Chemical

Company, Richmond, CA 94804.

Brimleyana 17:9-16, December 1991

10 C. G. Wright, T. P. Nuhn, and H. E. Dupree, Jr.

are wood cockroaches (Hebard 1917, Beatson and Dripps 1972), the

smokybrown cockroach (Fleet and Frankie 1974, Fleet et al. 1978),

peridomestic cockroaches (Brenner 1988, Brenner and Patterson 1988,

Hagenbuch et al. 1988, Patterson and Koehler 1989), and ants (Wesson

and Wesson 1940, Gaspar and Thirion 1978, Kondoh 1978, Pisarski and

Czechowski 1978, Vepsalainen and Wuorenrinne 1978, Nuhn and Wright

1979, Kondoh and Kitazawa 1984, Majer and Brown 1986, Richter et al.

1986, Wuorenrinne 1989. Knight and Rust 1990). Various pitfall traps,

such as those designed by Greenslade (1964), Nuhn and Wright (1979),

Reeves (1980), and Porter and Savignano (1990), are commonly used to

survey crawling insects and their relatives found at the soil level. We

present information here on ants and cockroaches that were captured in

soil-level pitfall traps placed adjacent to private dwellings in a suburban

environment while conducting a study on the efficacy of selected

insecticides in band applications around the exterior of houses (Wright

and Dupree 1984).

MATERIALS AND METHODS

Owners of 35 single-family dwellings in Wake Co., N.C., agreed to

participate in the survey. There was no pest control being done outside

these dwellings according to the owners, and no earlier pest control

attempts were reported. Each agreed to trap placement and retrieval

from around their homes five times during an 8-week period. Traps

made from 480-ml, round cardboard containers (8.6 cm wide x 9.5 cm

deep) were placed in a row of three, 2.5 cm apart, against each of the

four sides of the house, unless that placement was prevented by

structural barriers. The containers were buried in the ground with the

lip at the soil line. A mixture of petroleum jelly and mineral oil (1:1)

was applied as a narrow band inside the lip of the container to prevent

escape of animals that had entered.

By random selection each of the three traps per house side was

baited with 1/4 slice of white bread or with one tablespoon of boiled

raisins or was left unbaited. Ground cover around the traps was

assigned to one of three categories: (1) bare ground, (2) mulch (including

pine bark, pine needles, and hardwood leaves), or (3) ivy cover with or

without mulch. A piece of 12-mm hardware cloth (30.5 x 61 cm) was

placed over all three traps and fastened to the soil with a 25-cm spike in

each corner, to prevent squirrels and other small animals from removing

the cloth and taking the baits from the traps. Traps were placed, left for

24 hours, and removed; trapped specimens were put in 70% ethyl alcohol

or pinned. An initial trapping was done prior to application of an

insecticide during the week of 15-27 July, and additional trappings were

Ants and Cockroaches Trapped Outside Houses 1 1

Table 1. Mean numbers of insects collected near houses during the initial

trapping period with baited and unbaited traps.

aMean number of specimens per trap. Numbers followed by different letters

within a column are significantly different (P< 0.01).

bCollembola are excluded because of the large number of specimens that were

often present in the traps, especially following periods of rain.

done 1, 2, 4, and 8 weeks after insecticide application. Ants and

cockroaches were identified to species.

Data were analyzed using a general linear models procedure with a

one-way analysis of variance and unequal numbers of replications.

RESULTS AND DISCUSSION

A wide range of molluscs and arthropods were taken from a total of

1,969 trap collections. Five classes of arthropods (Arachnida, Chilopoda,

Crustacea, Diplopoda, and Insecta) and 14 orders of insects were

represented.

Data for the 420 traps used during the initial trapping period

indicated significant (P = 0.01) differences in bait preferences for the

trapped insects (Table 1). Ants preferred boiled raisins over bread, while

cockroaches selected bread over boiled raisins. Unbaited traps took the

fewest specimens in all comparisons. All insects combined (omitting

Collembola) preferred boiled raisins over bread. Data for later trapping

periods were not analyzed for bait preferences. Owing to a high standard

error, no differences were discernible in the number of specimens

relative to ground-cover type or trap placement by direction (N,E,S,W)

from the dwellings.

Ants were the most frequently trapped group of insects; that also

was reported by Cockfield and Potter (1984), who found them to be the

most common predatory arthropods. There were 7,331 ants trapped,

representing 26 species in 15 genera (Table 2). The five species most

widely distributed and collected in the greatest number of traps were

Camponotus americanus Mayr, Formica subsericia Say, Paratrechina

faisonensis (Forel), Pheidole dentata Mayr, and Prenolepsis imparis

(Say); C. americanus and P. dentata were the most common of these

12 C. G. Wright, T. P. Nuhn, and H. E. Dupree, Jr.

species. Prenolepsis imparls was by far the most common species in

September, when it replaced dwindling numbers of the other species at

the onset of cool weather. It was collected infrequently in July and

August. Other ant species might have been captured if other baits or

collecting techniques had been used. Nuhn and Wright (1979) collected

36 species in 21 genera on the North Carolina State University campus

by using several ant-collecting techniques.

The number of ants taken in a trap may be determined by several

factors, such as size of nest, number of nearby nests, or recruitment. A

relatively less common species may recruit large numbers of individuals

to a trap if they happen to nest near the trap, which was probably the

cause of the large samples of Monomorium minimum (Buckley),

Tetramorium caespitum (L.), and Lasius alienus (Forster). Other species

may have been more common and were collected in more traps, but

were underrepresented in total numbers collected because they do not

recruit as often or in such large numbers.

Pisarski and Czechowski (1978) and Kondoh (1978) in Poland and

Japan, respectively, reported that one or two dominant ant species

occurred in urbanized areas and a relatively small number of species

were present. Pisarski and Czechowski (1978) found tremendous

numbers of the dominant species. Porter and Savignano (1990) observed

that invading Solenopsis invicta Buren decimated the indigenous ant

populations in urban and agricultural areas in Texas, with species

richness dropping 70%. The total number of native individuals dropped

by 90%, concurrent with a large increase in S. invicta. Knight and Rust

(1990) identified ant specimens collected in and around structures by

professional pest control personnel in California. They found more than

25 ant taxa, of which Iridomyrmix humilis (Mayr) was the most

common species. In Western Australia, some ant species favored urban

areas; they were present in urban gardens and absent or uncommon in

adjacent native vegetation (Majer and Brown 1986). Other species were

present in the native vegetation and absent in the gardens. In contrast,

we found no single species to be clearly dominant around structures,

except for P. imparis in September, probably because of the relatively

greater habitat diversity of the more suburban collecting sites.

Cockroach nymphs collected in the traps were not identified to

species. Other than native wood cockroaches (Cariblatta, Ischnoptera,

and Parcoblatta spp.), the order Dictyoptera was represented by Blatta

orientalis L. (2 specimens) and Periplaneta fuliginosa (Serville) (9

specimens), two species that often occur both inside and outside

buildings (Mallis 1982). Periplaneta fuliginosa occurs both indoors and

outdoors in many southern states and is spreading to areas where it

previously was not found (Mallis 1982). Wright (unpublished data)

Ants and Cockroaches Trapped Outside Houses 13

Table 2. Total ants collected from all traps placed around the outside perimeter

of houses in Wake Co., N.C.

found that P. fuliginosa has become an important pest in buildings in

several areas of North Carolina since first being identified in 1964, as a

result of its continuing spread into previously uninfested areas. Around

Florida suburban houses, Brenner (1988), Brenner and Patterson (1988),

and Patterson and Koehler (1989) trapped P. fuliginosa most often and

Eurycotis floridana (Walker) second. Blatta orientalis, the other non-

wood cockroach trapped, was considered an indoor-outdoor species by

Mallis (1982). Beatson and Dripps (1972) reported on three long-term

infestations of B. orientalis, stating that it was usually considered an

indoor species in Great Britain. However, its importance as a domestic,

indoor species in Raleigh and North Carolina seems to be diminishing

(personal observations by the senior author and communications with

various pest control company personnel). The reason for its apparent

decrease is unknown.

14 C. G. Wright, T. P. Nuhn, and H. E. Dupree, Jr.

Identified wood cockroach species with the number trapped in

parentheses are Ischnoptera deropeltiformis (Brunner) (1), Parcobh.tta

bolliana (Saussure and Zehnter) (4), P. fulvescens (Saussure and

Zehnter) (25), P. lata (Brunner) (3), P. pennsylvanica (DeGeer) (10), P.

uhleriana (Saussure) (113), P. virginica (Brunner) (15), and Cariblatta

lutea lutea (Saussure and Zehnter) (1). Hagenbuch et al. (1988) trapped

Eurocytis floridana most frequently and C. lutea lutea second, during a

survey of species around Florida suburban homes.

All of the wood cockroaches trapped in our study have been

previously reported in North Carolina (Hebard 1917, Brimley 1938);

however, this is the first survey that documents their occurrence near

North Carolina houses.

ACKNOWLEDGMENTS. — The authors wish to extend their apprecia-

tion for assistance in the identification of Parcoblatta spp. discussed in

this study to D. G. Cochran of Virginia Polytechnic Institute and State

University, Blacksburg, Va., and to D. R. Nickle of the Systematic

Entomology Laboratory, U.S. Department of Agriculture. Thanks are

given to K. B. Corrette McGiffen, R. A. Diehl, M. K. Henessey, and

D. L. Stephan, of North Carolina State University, who assisted in

identification of specimens to order, family, and subfamily, and to Dr.

L. A. Nelson, Emeritus Professor of Statistics, North Carolina State

University, who suggested the method of statistical analysis and verified

the results obtained from the analyses.

LITERATURE CITED

Beatson, S. H., and J. S. Dripps. 1972. Long-term survival of cockroaches out

of doors. Environ. Health 80:340-341.

Brenner, R. J. 1988. Focality and mobility of some peridomestic cockroaches

in Florida (Dictyoptera: Blattaria). Ann. Entomol. Soc. Am. 81:581-592.

Brenner, R. J., and R. S. Patterson. 1988. Efficiency of a new trapping and

marking technique for peridomestic cockroaches (Dictyoptera: Blattaria). J.

Med. Entomol. 25:489-492.

Brimley, C. S. 1938. Insects of North Carolina. N.C. Dept. Agric, Raleigh.

Cockfield, S. D., and D. A. Potter. 1984. Predatory insects and spiders from

suburban lawns in Lexington, Kentucky. Great Lakes Entomol. 17:179-184.

Ebeling, W. 1978. Urban Entomology. Div. Agric. Sci., Univ. California,

Berkeley.

Fleet, R. R., and G. W. Frankie. 1974. Habits of two household cockroaches

in outdoor environments. Texas Agric. Exp. Sta. Misc. Publ. 1 153.

Fleet, R. R., G. L. Piper, and G. W. Frankie. 1978. Movement of the

smokybrown cockroach, Periplaneta fuliginosa, in an urban environment.

Environ. Entomol. 7:807-814.

Ants and Cockroaches Trapped Outside Houses 15

Frankie, G. W., and L. E. Ehler. 1978. Ecology of insects in urban

environments. Annu. Rev. Entomol. 23:367-387.

Frankie, G. W., and H. Levenson. 1978. Insect problems and insecticide use:

Public opinion, information and behavior. Pages 359-399 in Perspectives in

Urban Entomology (G. W. Frankie and C. S. Koehler, editors). Academic

Press, Inc., New York.

Gaspar, C, and C. Thirion. 1978. Modification des populations d'Hy-

menopteres sociaux dans des milieux anthropogenes. Memorabilia Zool.

29:61-77.

Greenslade, P. I. M. 1964. Pitfall trapping as a method for studying

populations of Carabidae (Coleoptera). J. Anim. Ecol. 33:301-310.

Hebard, M. 1917. The Blattidae of North America north of the Mexican

boundary. Mem. Am. Entomol. Soc. 2.

Hagenbuch, B. E., P. G. Koehler, R. C. Patterson, and R. J. Brenner. 1988.

Peridomestic cockroaches (Orthoptera: Blattidae) of Florida: their species

composition and suppression. J. Med. Entomol. 25:377-380.

Knight, R. L., and M. K. Rust. 1990. The urban ants of California USA with

distribution notes of imported species. Southwest. Entomol. 15: 167-178.

Kondoh, M. 1978. A comparison among ant communities in the anthropogenic

environment. Memorabilia Zool. 29:79-92.

Kondoh, M., and Y. Kitazawa. 1984. Ant communities on the campus of

UOEH and in an adjacent natural forest. J. Univ. Occupational Environ.

Hlth. 6:221-234.

Majer, J. D., and K. R. Brown. 1986. The effects of urbanization on the ant

fauna of the Swan Coastal Plain near Perth, Western Australia. J. R. Soc.

West. Aust. 69:13-18.

Mallis, A. 1982. Handbook of Pest Control. Franzak and Foster Co.,

Cleveland, OH.

Nuhn, T. P., and C. G. Wright. 1979. An ecological survey of ants

(Hymenoptera: Formicidae) in a landscaped suburban habitat. Am. Midi.

Nat. 102:352-362.

Patterson, R. S., and P. G. Koehler. 1989. Peridomestic cockroach suppression

with hydramethylnon bait. J. Agric. Entomol. 6:37-42.

Pisarski, B., and W. Czechowski. 1978. Influence de la pression urbaine sur la

Myrmecofaune. Memorabilia Zool. 29:109-128.

Porter, S. D., and D. A. Savignano. 1990. Invasion of polygyne fire ants

decimates native ants and disrupts arthropod community. Ecology 71:2095-

2106.

Reeves, R. M. 1980. Use of barriers with pitfall traps. Entomol. News

91:10-12.

Richter, K., B. Klausnitzer, and A. Zimdars. 1986. Contribution to the ant

fauna of various urban ruderal sites in the district of Leipzig, East Germany.

(Hym., Formicidae). Entomol. Nachr. Ber. 30:115-120.

Vepsalainen, K., and H. Wuorenrinne. 1978. Ecological effects of urbanization

on the mound-building Formica L. species. Memorabilia Zool. 29:191-202.

Wesson, L. G., Jr., and R. G. Wesson. 1940. A collection of ants from south-

central Ohio. Am. Midi. Nat. 21:89-103.

16 C. G. Wright, T. P. Nuhn, and H. E. Dupree, Jr.

Wuorenrinne, H. 1989. Effects of urban pressure on colonies of Formica rufa

group (Hymenoptera: Formicidae) in the town of Espoo (Finland). Ann.

Zool. (Warsaw). 42:335-344.

Wright, C. G., and H. E. Dupree, Jr. 1984. Insect control around houses. Pest

Control Technology. 12:58-61.

Accepted February 1990

Bats (Chiroptera: Vespertilionidae) of the Great

Dismal Swamp of Virginia and North Carolina

Thomas M. Padgett1 and Robert K. Rose

Department of Biological Sciences

Old Dominion University

Norfolk, Virginia 23529-0266

ABSTRACT— From autumn 1983 through spring 1986, bats were

collected in the Dismal Swamp, a forested wetland located in south-

eastern Virginia and northeastern North Carolina. Before this survey,

only five species of bats were known from the Dismal Swamp, all first

collected during the 1890s. A total of 89 specimens representing five

genera and seven species were collected, including 50 red bats, Lasiurus

borealis. Four species are new records for the Dismal Swamp, and the

one specimen of Seminole bat, Lasiurus seminolus, represents a first

record for Virginia. During winter, the population of red bats consisted

entirely of males, which were active at ambient temperatures > 10° C.

Five species of bats are considered permanent residents; two of these

were active throughout the year and the others hibernated during the

winter months.

Five species of bats from the vicinity of Lake Drummond (Table 1)

were collected between 1895 and 1898 by A. K. Fisher and William

Palmer, of the U.S. Department of Agriculture's Bureau of Biological

Surveys, during a survey of the Great Dismal Swamp, an 80,000-ha

forested wetland on the coastal plain of Virginia and North Carolina.

The first records of bats from the region (although not from the Dismal

Swamp proper) had been made between 1891 and 1894 by the Smithwick

brothers in the Albemarle Sound area of Bertie Co., N.C. (Brimley

1897).

Since the 1890s, no systematic attempt had been made to study bats

in the vicinity of the Dismal Swamp. Handley (1979a), after an

exhaustive review of the literature, compiled a short account of the

species of bats believed to occur in the Dismal Swamp forests (Table 1).

He attributed 10 species of bats to the Dismal Swamp, but only five

species had been collected there; the remainder were collected near

Albemarle Sound by the Smithwicks or in other areas adjacent to the

Dismal Swamp. The four records from the twentieth century include

three specimens from the Dismal Swamp: one Keen's myotis, Myotis

keenii Merriam, in 1930, and two eastern pipistrelles, Pipistrellus

subflavus F. Cuvier, in 1905 and 1964 (Handley 1979a); the fourth is a

Present address: N.C. Wildlife Resources Commission, P.O. Box 2632,

Elizabethtown, NC 28337.

Brimleyana 17:17-25, December 1991 17

18 Thomas M. Padgett and Robert K. Rose

yellow bat, Lasiurus intermedins Allen, found in Norfolk in 1954 (de

Rageot 1955).

Our objective was to determine the species of bats that currently

inhabit the Dismal Swamp compared with the ones there 90 years ago.

Because much of the Dismal Swamp is now incorporated into a federal

wildlife refuge, present-day information regarding the presence and

abundance of species of bats is important in the long-term management

of the Great Dismal Swamp National Wildlife Refuge (GDSNWR).

METHODS

In the Dismal Swamp, previous attempts at mist-netting bats

proved unsuccessful (C. O. Handley, Jr., and D. Lahti, personal

communications), primarily as a result of dispersal of flying bats over

vast areas of flooded forests and the tendency of some bats to forage at

or above the forest canopy. Because our objective was to compile a

record of the bats of the Dismal Swamp, we elected to collect voucher

specimens by shooting. In that way, a permanent record of the bats of

the Dismal Swamp at the present time is established, and the value of

the voucher specimens will increase with time.

We collected bats using 12- and 20-gauge shotguns bored to

improved cylinder (O'Conner 1965) and loaded with standard 8- and

9-shot loads. In good light conditions, we quickly learned to identify red

bats, Lasiurus borealis M tiller, in flight, and therefore were often able to

record their distribution and emergence times without having to collect

additional specimens.

Sampling sites were restricted to open areas within the Dismal

Swamp, usually within the GDSNWR, such as roadways and road

intersections, which offered clear avenues for collecting. Lake

Drummond, a 1,000-ha natural lake located in the approximate center

of the Dismal Swamp, was sampled from shore and from a boat. Two

old abandoned buildings along the shore of Lake Drummond were

searched unsuccessfully for bats.

Collection times were limited approximately to sunset ±0.5 hour.

For each species, we recorded ambient temperature, local weather

conditions, and the time of emergence, defined as the time in minutes

before and after sunset that a species of bat was first observed or

collected. We collected only during the months of September through

early June; from late June through August, bats produce young and

rear them to flying age.

The specimens collected during the survey were preserved either as

study skins and skulls or in 70% ethanol. They have been deposited at

the U.S. National Museum of Natural History (USNM 448240-312;

448314-317; 448319-331).

Bats of the Great Dismal Swamp

19

RESULTS

From 5 October 1983 through 15 March 1986, we spent 77 evenings

collecting bats (Table 2). We collected a total of 89 specimens represent-

ing five genera and seven species. Four of these species are new records

for the Dismal Swamp: the big brown bat, Eptesicus fuscus (Palisot de

Beauvois); the silver-haired bat, Lasionycteris noctivagans (LeConte);

the hoary bat, Lasiurus cinereus (Palisot de Beauvois); and the Seminole

bat, Lasiurus seminolus (Rhoads).

The silver-haired bat was active and relatively common during the

winter months. Eight were collected (Table 2), indicating that this bat is

relatively common in winter. Although we collected only eight specimens,

we observed silver-haired bats on several occasions during December

and January throughout the Dismal Swamp. Sometimes we saw these

bats emerging from hollow bald cypress trees, Taxodium distiehum,

growing in Lake Drummond. In good light, silver-haired bats could be

distinguished by their nearly black coloration, and they often flew in

Table 1 . Historical records of the 10 species of bats previously thought to occur in the Great Dismal

Swamp (Handley 1979a) compared with the species of bats found there during the present study.

Seminole bat

Hoary bat

Yellow bat

Evening bat

Rafinesque's

big-eared bat

Total specimens

collected

Lake Drummond, Dismal Swamp,

1983

Dismal Swamp, 1983

Willoughby, Norfolk, Va., 1954

Lake Drummond, Dismal Swamp,

1895, 1896, 1898

Lake Drummond, Dismal Swamp,

1897

yes

38

89

20 Thomas M. Padgett and Robert K. Rose

pairs. On 12 February 1984, a male and a female were collected at

Jericho Lane, and on 15 March 1986, another male and female were

collected as they flew together at the Lynn-Badger Ditches intersection.

A third pair, but both males, were collected 16 March 1984 on Railroad

Ditch.

The eastern pipistrelle was found throughout the Dismal Swamp.

This bat, moth-like in flight and scarcely larger than a cecropia moth,

was frequently observed foraging at or above the forest canopy, which

made collecting difficult. Although only six were collected (Table 2), it

appears to be a common permanent resident of the Dismal Swamp.

The big brown bat was most frequently found in the vicinity of

Lake Drummond, where two specimens were collected as they emerged

from hollow bald cypress trees, and west of the lake on Interior and

West ditches. One specimen also was taken on Lynn Ditch, north of the

lake.

The red bat, with 50 specimens, was most numerous (Table 2). Red

bats were active throughout the sampling period whenever the mean

temperature was >10°C. However, once (25 January) we collected three

males when the temperature was only 7°C. During autumn and winter

the population of red bats consisted entirely of males. Females were

taken only during the months of March and April (N = 8; Table 2). The

males (N = 42) were collected from September through June.

We collected 17 evening bats, Nycticeius humeralis Rafinesque,

making them second in abundance. None were seen or collected during

December and January, when we presume they were dormant. We

believe that the evening bat is a permanent resident throughout the

Dismal Swamp.

The remaining two species, the hoary bat and the Seminole bat, are

considered rare in the vicinity of the Dismal Swamp. Both are believed

to be highly migratory species, especially the hoary bat (Barbour and

Davis 1969). The one hoary bat was taken in the northwest corner of

the GDSNWR (Jericho and Hudnell ditches) at 1711 hours on 22

November 1983, at an ambient temperature of 1 1°C. The one Seminole

bat, probably also a migrant, was taken at the mouth of Jericho Ditch

at Lake Drummond at 1745 hours (45 minutes before sunset) on 6

October 1983, at an ambient temperature of 24° C (Padgett 1987).

Red bats changed their patterns of emergence during the year (Fig.

1). Although there was no significant correlation between time of

emergence and the ambient temperature on an annual basis (r = 0.054, P

- 0.90), temperature did appear to play a role in the activity patterns of

red bats on a seasonal basis. From September through November, red

bats foraged after sunset. As the winter progressed, they emerged and

foraged earlier, and by March and April, emergence times coincided

Bats of the Great Dismal Swamp 21

with sunset. As spring progressed, red bats emerged later in the evening,

when we would observe them with our truck headlights as we drove out

from the interior of the GDSNWR.

We recorded 10 evenings with no bat activity (Table 2). On four

evenings when the sky was clear and no bats were observed, the

temperatures were 6° and 8°C in December, 21°C in June, and 14°C in

October. In June and throughout the summer, when temperatures were

high, bats appeared very late in the evening, often near last light, when

the exact time of emergence could not be accurately determined. On 6

October, a warm night when no bats were flying, the moon was full.

Moonlight severely restricts the activity of many bat species, because

they are either avoiding predators (Fenton et al. 1977) or responding to

lower insect abundance (Anthony et al. 1981). It rained on half of the

six cloudy evenings when no bats were observed; the temperatures

ranged from 7° to 20° C (x - 14° C) on those six evenings, and only the

7°C recording was <10°C. However, on seven evenings with overcast

skies but no rain, bats were actively foraging and subsequently collected.

DISCUSSION

Four species of bats — big brown, silver-haired, hoary, and

Seminole — constitute new records for the Dismal Swamp and south-

eastern Virginia, and the Seminole bat represents a new record for

Virginia.

Table 2. Species of bats collected by month, from September through June, 1983-1986. a

22 Thomas M. Padgett and Robert K. Rose

In the piedmont of Virginia, Lewis (1940) found red bats to be

active at > 13°C, but Davis and Lidicker (1958) observed red bats in

West Virginia only on evenings when the temperature was > 19° C.

Kunz (1982) noted that temperature is a crucial factor in controlling

seasonal as well as daily activity patterns of many species of bats

inhabiting the temperate zones.

During the winter, red bats emerged early but seemed to forage

only for a short period of time. In the Dismal Swamp, especially in

winter, the temperature fluctuates greatly, sometimes dropping as much

as 10° C within an hour or less. Consequently, emergence periods during

the winter were often brief; bats foraged briefly and then retired before

the ambient temperature dropped below 10° C.

Many specimens collected during the winter had stomachs full of

insects. During the winter we observed numerous moths as well as

swarms of midges (Chironomidae) when many bats were active. The

activity patterns of bats may be correlated with the presence, abundance,

and activity patterns of insects, as well as with temperature. Twice we

observed bats as potential prey themselves when a red-shouldered hawk

(Buteo lineatus) and an American crow (Corvus brachyrhynchos) pursued

a foraging bat at treetop level; in neither instance did we see the

conclusion of the chases. Horsley (1991) reported seeing a blue jay

{Cyanocitta cristata) fall to the ground while trying to subdue a red bat

in Dare Co., N.C., indicating that such chases sometimes are successful

for the predator.

Silver-haired bats are considered to be highly migratory (Barbour

and Davis 1969). Handley and Patton (1947) and Baily (1946) believed

that silver-haired bats are migrants in spring and autumn but may breed

in the mountainous regions of Virginia. However, silver-haired bats

were seen emerging from hollow trees in Lake Drummond during April,

suggesting that they may indeed breed within the Dismal Swamp.

The species not collected during the survey may be equally

significant. No species of Myotis was collected within the GDSNWR,

although two records of Keen's myotis do exist for the Dismal Swamp

(Table 1). We expected to collect little brown bats, Myotis lucifugus

(LeConte), in part because D. Schwab (personal communication) had

taken them less than 400 m from the western boundary of the Dismal

Swamp, 5 km S of the Suffolk, Va., business district. Furthermore, the

little brown bat is considered to be the most abundant and widely

distributed bat in North America (Barbour and Davis 1969). However,

this colonial bat is not known to be a tree-dwelling species and is more

typically found inhabiting buildings and caves. During our survey, we

searched the few remaining buildings within the Dismal Swamp but

found no little brown bats. In the past, the numerous cabins and other

Bats of the Great Dismal Swamp

23

O

30

20

10

10 ■■

-20-

-30

• 15

LEGEND

x EMERGENCE

o TEMPERATURE

20

10

O

en

Ctl

S O N D J F

MONTH

MAM

Fig. I. Mean emergence times and temperatures for red bats {Lasiurus borealis)

recorded from September through May, 1983-1 986, in the Great Dismal Swamp.

Numbers on graph represent total days sampled per month. Positive values

represent emergence times (minutes) after sunset; negative values are emergence

times before sunset.

buildings along the shores of Lake Drummond may have provided

excellent roosting areas for this bat. Since the creation of the GDSNWR

in 1974, most of those structures have been demolished for esthetic

reasons and for public safety. Therefore, it is likely that absence of little

brown bats is associated with loss of buildings. By contrast, Keen's

myotis is a less gregarious species that sometimes roosts under the loose

bark of trees (Barbour and Davis 1969), which increases the possibility

that this bat still exists within the Dismal Swamp.

Another species that could occur in the vicinity is the northern

yellow bat, Lasiurus intermedius H. Allen. At present, the only Virginia

24 Thomas M. Padgett and Robert K. Rose

record of this bat is a pregnant female collected northeast of the Dismal

Swamp in May 1954 (de Rageot 1955). It is not known whether that

animal flew there or accidentally reached the area by ship (Handley

1979a). The northern yellow bat is generally associated with Spanish

moss, Tillandsia usneoides, which is rare in the Dismal Swamp but does

occur in Seashore State Park in Virginia Beach, less than 15 km from

where the specimen was collected by de Rageot.

Perhaps the most sought-after species of bat in the region is

Rafinesque's big-eared bat, Plecotus rafinesquii macrotis Lesson. None

were collected during our survey. At present, only one specimen has

been reported from Virginia. It was taken in 1897 near Lake Drummond,

where big-eared bats supposedly resided in hollow cypress trees (Handley

1979a). Recent circumstantial evidence suggests that this rare species

still resides in the vicinity. In September 1984, a specimen was discovered

on the grille of an automobile in the Pungo section of Virginia Beach,

east of the Dismal Swamp. Although the specimen was discarded, 35-

mm color slides of the animal were identified as being Rafinesque's big-

eared bat by C. O. Handley, Jr. (personal communication). On 4 June

1989, another specimen of Rafinesque's big-eared bat was photographed

and released by D. Schwab (personal communication) less than 2 km W

of the GDSNWR. The recent collection of several specimens of

Rafinesque's big-eared bat in Merchants Mill Pond State Park, Gates

Co., N.C., within 30 km of the Dismal Swamp (Clark et al. 1985), as

well as the specimen photographed by D. Schwab, suggests that the

species probably inhabits the Dismal Swamp. Despite a paucity of

information, the status of Rafinesque's big-eared bat has been changed

from status undetermined (Handley 1979b) to endangered by the

Commonwealth of Virginia.

Although we did not collect all of the species believed to occur in

the Dismal Swamp, we did document the occurrence of four species of

bats not previously recorded there, and we verified the continued

presence of three other species not seen there for 24 to 90 years. No

Myotis were taken in the Dismal Swamp in 77 collecting days. Male red

bats and silver-haried bats were active throughout the winter when the

temperature was greater than 10°C and 13°C, respectively.

ACKNOWLEDGMENTS.— We thank the staff of the Great Dismal

Swamp National Wildlife Refuge for their support and assistance; C. O.

Handley, Jr., for verifying the identifications of the bats we collected;

and M. K. Garrett, S. Powell, D. Schwab, L. Swanner, and R. Trimyer,

all of whom expended vast quantities of munitions during the course of

our survey.

Bats of the Great Dismal Swamp 25

LITERATURE CITED

Anthony, E. L. P., M. H. Stack, and T. H. Kunz. 1981. Night roosting and the

nocturnal time budget of the little brown bat, Myotis lucifugus, in southern

New Hampshire. Ecology 58:775-780.

Bailey, J. W. 1946. The Mammals of Virginia. Privately published.

Barbour, R. W., and W. H. Davis. 1969. Bats of America. Univ. Press of

Kentucky, Lexington.

Brimley, C. S. 1897. An incomplete list of the mammals of Bertie Co., North

Carolina. Am. Nat. 31:237-238.

Clark, M. K., D. S. Lee, and J. B. Funderburg, Jr. 1985. The mammal fauna

of Carolina bays, pocosins, and associated communities in North Carolina:

an overview. Brimleyana 1 1:1-38.

Davis, W. H., and W. Z. Lidicker. 1956. Winter range of the red bat, Lasiurus

borealis. J. Mammal. 37:280-281.

de Rageot, R. H. 1955. A new northernmost record of the yellow bat,

Dasypterus floridanus. J. Mammal. 36:458.

Fenton, M. B., N. G. H. Hoyle, T. M. Harrison, and D. J. Oxley. 1977.

Activity patterns, habitat use, and prey selection by some African

insectivorous bats. Biotropica 9:73-85.

Handley, C. O., Jr. 1979a. Mammals of the Dismal Swamp: an historical

account. Pages 297-357 in The Great Dismal Swamp (P. A. Kirk, Jr.,

editor). Univ. Press Virginia, Charlottesville.

Handley, C. O., Jr. 1979b. Mammals. Pages 483-602 in Endangered and

Threatened Plants and Animals of Virginia (D. W. Linzey, editor). Ctr.

Environ. Stud., Va. Polytech. Inst, and State Univ., Blacksburg.

Handley, C. O., Jr., and C. P. Patton. 1947. Wild Mammals of Virginia.

Comm. Game and Inland Fish., Richmond.

Horsley, B. D. 1991. Blue jay captures bat. Chat 55:30-31.

Kunz, T. H. 1982. Ecology of Bats. Plenum Press, New York.

Lewis, J. B. 1940. Mammals of Amelia County, Virginia. J. Mammal.

21:422-428.

O'Connor, J. 1965. The Shotgun Book. Alfred A. Knopf, New York.

Padgett, T. M. 1987. The first record of the Seminole bat, Lasiurus seminolus,

from Virginia. Va. J. Sci. 38:253.

Accepted July 1990

26

POTENTIAL EFFECTS OF OIL SPILLS ON SEABIRDS

AND SELECTED OTHER OCEANIC VERTEBRATES

OFF THE NORTH CAROLINA COAST

by

David S. Lee and Mary C. Socci

Based primarily on data gathered offshore during the past 15 years by the

staff of the North Carolina State Museum of Natural Sciences, this book

presents information regarding distribution and susceptibility to oil pollution

for 25 oceanic species: 14 birds, 6 mammals, and 5 turtles. An overlay can be

placed on range maps to demonstrate the proximity of species occurrence to the

oil lease sites off Cape Hatteras.

1989 64 pages Softbound ISBN 0-917134-18-4

Price: $8, postpaid. North Carolina residents add 5% sales tax. Please make checks

payable to U.S. currency to NCDA Museum Extension Fund.

Send order to: OIL SPILL BOOK, N.C. State Museum of Natural Sciences,

P.O. Box 27647, Raleigh, NC 2761 1.

Occurrence of an Introduced African Cichlid,

the Blue Tilapia, Tilapia aurea, in a Tidal Creek

of the Skidaway River, Georgia

L. Stanton Hales, Jr.

Department of Zoology and Institute of Ecology

University of Georgia, Athens, Georgia 30602

and

Marine Extension Service Aquarium

Georgia Sea Grant College Program

Skidaway Island, P. O. Box 13687

Savannah, Georgia 31416

ABSTRACT.— The blue tilapia, Tilapia aurea, has not been reported

previously from estuarine waters of Georgia. More than 35 juvenile

blue tilapia were collected in a tidal creek on Skidaway Island,

Chatham Co. Although it is not possible to determine the exact time

and circumstances of the introduction, those specimens escaped from a

raceway on Skidaway Island during the summer of 1989. The raceway

system has been used for aquaculture experiments since the early

1970s, but recent changes in the drainage system (to provide for

emergency containment of spills) may facilitate the establishment and

spread of this population by warming water to the contained marsh,

which enhances overwintering survival of this tropical species. The

same traits of blue tilapia that are desirable for aquaculture have

enabled the species to establish populations in a diverse range of

habitats, including estuaries, in the southeastern United States.

Deleterious effects of this species on some native fishes have been

reported in Florida and Texas, and it seems prudent to eradicate the

local population if possible.

An extensive review of the status of introduced fishes in the United

States (Courtenay et al. 1984) listed the unconfirmed reports of Tilapia

aurea (Steindachner) and Tilapia mossambica (Peters) on St. Simons

Island and Sea Island (Glynn Co.) as the only reports of cichlid fishes in

Georgia. This paper documents the first confirmed report of a cichlid in

the state of Georgia.

Although it now occurs more widely across Africa, the blue tilapia,

T. aurea, is native to Senegal; Middle Niger; Benue, Shari and Logone

rivers; Lake Chad; the lower Nile from Cairo to the Delta Lakes; the

Jordan River system; the Na'aman and Yarkon rivers of Israel; lakes

Huleh and Kinnereth; Ein Feshkha; and the Dead Sea (Trewavas 1983).

j

Brimleyana 17:27-35, December 1991 27

28

L. Stanton Hales, Jr.

This species is among the most widely distributed exotic fishes in North

America, with populations now established in at least seven states

(Courtenay et al. 1984). In the southeastern United States, the species

has reproducing populations in Lake Julian, Buncombe Co., N. C, six

Texas lakes, and more than 18 counties in Florida (Hubbs et al. 1978,

Courtenay et al. 1984, Dolmon 1990). It is also established in Oklahoma,

Arizona, California, and possibly Nevada (Courtenay et al. 1984). The

species is widely used for control of aquatic vegetation (despite little if

any demonstrated success), and its spread has been facilitated by the use

of juveniles as bait (Courtenay et al. 1984).

Approximately 15 juvenile blue tilapia were collected by a local

fisherman in minnow traps baited with blue crab {Callinectes sapidus)

pieces in a tidal creek on the north end of Skidaway Island along the

Skidaway River (Fig. 1). Unfortunately, the identity of the specimens

Fig. 1. Location of site where blue tilapia were collected. The figure on lower

left shows the location of Wassaw Sound in Georgia, the figure on upper left

shows the Skidaway Institute of Oceanography (outlined area in upper left

corner), and the figure on right shows the location of the small, unnamed creek

east of the laboratory docks in the Skidaway River.

Blue Tilapia in Georgia 29

was not known by the fisherman, who released most back into the

Skidaway River. Examination of dead specimens, discarded as bait,

established their identity as a species of cichlid. These specimens were

kept for more detailed examination, which led to the study reported

herein.

METHODS AND MATERIALS

To determine the extent of the distribution of blue tilapia in the

area, numerous collections were made with five minnow traps, which

were fished overnight in numerous places along the creek. Collections

were also made with a cast net (2.4-m diameter, 1.2-cm mesh) in deep

(>1 m) tidal pools. Daily collections with the cast net and minnow traps

were made from 21 July to 3 August 1989. Temperature and salinity

were recorded. All blue tilapia were placed immediately in 5% buffered

seawater formalin. The occurrence of other species and their life history

stages were noted; those specimens were then released alive. The tilapia

were taken to the lab for processing.

All specimens were measured (total length, abbreviated as TL) on

a measuring board to the nearest mm and weighed on an electronic

balance to the nearest 0.1 g. An attempt was made to determine sex of

the fish from macroscopic examination of gonads. Otoliths and scales

were removed for possible age determination. Counts of dorsal fin rays

and spines were made to ascertain the identity of the species (Trewavas

1983).

Voucher specimens have been placed in the collections of the

University of Georgia Museum of Natural History (Athens, Ga.) and

the Marine Extension Service Aquarium (Savannah, Ga.).

RESULTS

A total of 36 blue tilapia, ranging from 45 to 86 mm total length

(mean - 65) and weighing from 1.7 to 1 1.9 g, were collected. The largest

specimens were collected with cast net, but most specimens were

collected in minnow traps. The length frequency distribution for all

specimens is given in Fig. 2. All specimens exhibited the juvenile color

pattern, consisting of 8 to 10 dark vertical bars from head to tail with a

dark spot at the base of the soft portion of the dorsal fin. In most

specimens, the maxillary and opercle were flecked with a metallic blue

iridescence and the distal margin of the caudal fin was red. Counts of

dorsal fin spines and rays ranged from 14 to 16 (mean = 15.4) and 12 to

13 (mean = 12.5), respectively, indicative of T. aurea and not a hybrid of

T. aurea and T. nilotica (L.) (Trewavas 1983).

It was not possible to determine the sex of any individual from

macroscopic examination. Gonads were barely visible to the naked eye.

30 L. Stanton Hales, Jr.

Marks resembling annuli were not visible on the otoliths or scales of

any specimens. The coloration, lack of gonadal development, and

absence of annuli on any scales or otoliths suggest that all blue tilapia

collected in the tidal creek were young of the year.

Temperature and salinity at this site were 20-29° C and 2-20 ppt,

respectively. During at least two rainstorms (22-23 July and 9 August),

salinity dropped below 1 ppt. Other species collected in this creek were

American eel, Anguilla rostrata Lesueur; sheepshead killifish, Cyprino-

don variegatus Lacepede; mummichog, Fundulus heteroclitus (L.);

mosquitofish, Gambusia affinis (Baird and Girard); sailfin molly,

Poecilia latipinna (Lesueur); striped mullet, Mugil cephalus L.; and

freshwater goby, Gobionellus shufeldti (Jordan and Eigenmann).

DISCUSSION

A self-sustaining population of blue tilapia does not appear to be

established at the present time; however, physiological characteristics of

this species may facilitate its establishment in the Skidaway River.

Although different measures of lethal temperature range from 6 to 12°C

(Shaflund and Pestrak 1982, Zale 1984), the blue tilapia tolerates

temperatures as low as 5°C in fresh water for brief periods (McBay

1961) and requires only 20°C for spawning (McBay 1961). Temperatures

lower than 5°C seldom occur in the Skidaway River, and temperatures

above the minimum for spawning occur from April through October in

most years (D. Miller, University of Georgia Marine Extension

Aquarium, personal communication). Blue tilapia are not only tolerant

of high salinity, but also grow well and reproduce in brackish waters

(Loya and Fishelson 1969, Trewavas 1983). In addition, blue tilapia are

tolerant of poor water quality, including high ammonia (Redner and

Stickney 1979). Conditions suitable for the growth and reproduction of

blue tilapia are present throughout coastal waters of Georgia, including

estuaries.

Although blue tilapia are not well studied in estuarine environments,

other aspects of the ecology of the species appear conducive to its

establishment in the Skidaway River. Mature individuals were not

collected during this study, but ripe gonads have been observed in

specimens as small as 58 mm standard length (Chervinski 1968). In

Alabama ponds, females began to mature at 50 days and 10 cm TL

(McBay 1961). Those conditions of maturity, the presence of potentially

mature (> 58 mm SL) individuals in the Skidaway creek, and the

favorable growing season in coastal Georgia indicate that establishment

of a reproducing population in the Skidaway River estuary is a reason-

able possibility. The blue tilapia is established in at least one Florida

estuary, Tampa Bay (Courtenay et al. 1984).

Blue Tilapia in Georgia

31

Additional information about the feeding ecology, growth, and

reproduction of blue tilapia in estuarine waters might indicate how such

populations are established. Zale (1987) attributed the success of blue

tilapia in fresh waters of the southeastern United States to the large eggs

and large initial size of larvae; consequently, larval blue tilapia have

foraging capabilities superior to those of larval centrarchids, which are

smaller. The diet of blue tilapia in estuarine waters has not been

reported, but the diet of blue tilapia in freshwater habitats includes

phytoplankton, zooplankton, detritus, benthic invertebrates, and

macrophytes (McBay 1961, Spataru and Zorn 1978, Hendricks and

Noble 1979). Additional information on the biology of this species in

estuarine waters is necessary before the critical factors in their dispersal

and establishment in those habitats are understood.

The effects of blue tilapia on native aquatic communities are not

well documented, but they are being reported with increasing frequency

as this species spreads into new habitats and increases in abundance.

Abundance of blue tilapia, approaching 2600 kg/ ha in some waters of

the United States (Noble and Germany 1976, Germany and Noble

Size (TL) Frequency Distribution

o

c

TL (mm)

Fig. 2. Size distribution of blue tilapia collected with cast nest and minnow

traps. Size categories given on the abscissa indicate the lower limit of a 5-mm

category.

32 L. Stanton Hales, Jr.

1977), has reached a point in many lakes of Florida and Texas at which

it is adversely affecting reproduction of native centrarchids (Radonski et

al. 1984, Taylor et al. 1984). The blue tilapia is the dominant species in

cooling lakes at several power plants in Texas; several native fishes are

scarce or absent in those lakes (Dolman 1990). Tilapia aurea has been

reported to reduce the organic content of sediments, which eliminated

blooms of the blue-green alga Oscillatoria chalybea in some Israeli

reservoirs (Leventer 1981). The blue tilapia has been observed to depress

populations of large phytoplankton and vulnerable zooplankton while

enhancing small algae and evasive zooplankton (Drenner et al. 1984,

1987). Presumably through competition, the blue tilapia has been

reported to reduce the catch of desirable fish in Lake Kinneret, Israel

(Gophen et al. 1983, Vineyard et al. 1988). The effects on native fishes

through removal of vegetation have not been documented (Taylor et al.

1984). The effects of blue tilapia on estuarine communities, including

recreationally and commercially important fishes, have not been investi-

gated. Given the inherent variability of the estuarine physical environ-

ment and an estuarine fish community that includes seasonal and

transient members, demonstration of the adverse effects of blue tilapia

on estuarine fishes is very difficult. This difficulty suggests that a

conservative approach should be taken and that in the absence of

demonstrated benefits, the reported population should be eliminated, if

feasible.

Eradication of the Skidaway population may be possible given the

present occurrence in only one tidal creek. The recorded juveniles

escaped from a raceway on Skidaway Island that empties into the tidal

creek. The raceway is one of two built in the early 1970s at the

Skidaway Institute of Oceanography and has been used since its

construction for aquaculture experiments and to hold surplus fishes.

The drain in these raceways consists of an uncovered standpipe, from

which raceway overflow dumps into a drainage culvert and a connecting

tidal creek. Until the recent construction of the drainage culvert, the

raceway overflow emptied directly into a tidal creek. With construction

of the drainage culvert, runoff from several labs at Skidaway empties

into the culvert, where it then drains into the tidal creek. Because of the

form of this culvert and its tidal elevation, water from the lab does not

flow immediately into the creek. In winter, these features may enhance

overwintering survival in the creek by providing a large (approximately

13,000 liter) storage pool of warm water.

Eradication of the population may be possible through the applica-

tion of rotenone or other ichthyocides to the small creek, but such a

procedure has not been attempted. Modification of the raceway drain or

elimination of the raceway population would prevent future introduc-

Blue Tilapia in Georgia 33

tions; however, initial attempts to remove the parent population have

been unsuccessful. Unfortunately, runoff from recent storms (including

Hurricane Hugo) may have already dispersed individuals to other parts

of the Skidaway River estuary. Because of the mouth-brooding habits

of this species, a single female, which can carry as many as 1,600 young

(Payne and Collinson 1983), could disperse a considerable number of

individuals to other parts of the river.

In summary, collections to date suggest that a substantial number

of blue tilapia have escaped into a tidal creek of the Skidaway River.

Whether or not the species is established in the Skidaway River estuary

remains to be seen, but establishment appears both likely and un-

fortunate.

ACKNOWLEDGMENTS.— I thank D. Jennings, B. Wilson, and an

anonymous reviewer for their critiques of the manuscript, and A.

Boyette for preparing Fig. 1. Financial support for this research was

provided by NOAA, Office of Sea Grant, Department of Commerce,

Georgia Sea Grant College Program Grant #NA84AA-D-00072 to G.

S. Helfman and the author. The United States government is authorized

to produce and distribute reprints for governmental purposes notwith-

standing any copyright notation that appears hereon. This study is

Contribution #3 of the Sciaenid Project on the River (SPOTR).

LITERATURE CITED

Chervinski, J. 1968. The cichlids of Ein Feshkha Springs. I. Tilapia aurea exul

(Steinitz). II. Tilapia zillii (Gervais). Hydrobiologia 32:150-156, 157-160.

Courtenay, W. R., Jr., D. A. Hensley, J. N. Taylor, and J. A. McCann. 1984.

Distribution of exotic fishes in the continental United States. Pages 22-77

in Distribution, Biology and Management of Exotic Fishes (W. R.

Courtenay, Jr., and J. R. Stauffer, Jr., editors). Johns Hopkins Univ.

Press, Baltimore, Md.

Dolman, W. B. 1990. Classification of Texas reservoirs in relation to

limnology and fish community associations. Trans. Am. Fish. Soc.

119:511-520.

Drenner, R. W., K. D. Hambright, G. L. Vineyard, M. Gophen, and U.

Pollingher. 1987. Experimental study of size-selective phytoplankton

grazing by a filter-feeding cichlid and the cichlid's effect on plankton

community structure. Limnol. Oceanog. 32:1 140-1 146.

Drenner, R. W., S. B. Taylor, X. Lazzaro, and D. Kettle. 1984. Particle

grazing and plankton community impact of an omnivorous cichlid. Trans.

Am. Fish. Soc. 113:397-402.

Germany, R. D., and R. L. Noble. 1977. Population dynamics of blue tilapia

in Trinidad Lake, Texas. Proc. 3 1st Annual Conf. Southeast. Assoc. Fish

Wildl. Agencies. 31:412-417.

34 L. Stanton Hales, Jr.

Gophen, M., R. W. Drenner, and G. L. Vineyard. 1983. Cichlid stocking and

the decline of the Galilee Saint Peter's fish (Sarotherodon galilaeus) in

Lake Kinneret, Israel. Can. J. Fish Aq. Sci. 40:983-986.

Hendricks, M. K., and R. L. Noble. 1979. Feeding interactions of three

planktivorous fishes in Trinidad Lake, Texas. Proc. 33rd Annual Conf.

Southeast. Assoc. Fish Wildl. Agencies. 33:324-330.

Hubbs, C, T. Lucier, G. P. Garrett, R. J. Edwards, S. M. Dean, and E. Marsh.

1978. Survival and abundance of introduced fishes near San Antonio,

Texas. Texas J. Sci. 30:369-376.

Leventer, H. 1981. Biological control of reservoirs by fish. Bamidgeh 33:3-23.

(Not seen.)

Loya, Y., and L. Fishelson. 1969. Ecology of fish breeding in brackish water

ponds near the Dead Sea (Israel). J. Fish Biol. 1:261-278.

McBay, L. G. 1961. The biology of Tilapia nilotica. Proc. 15th Annual Conf.

Southeast. Assoc. Game and Fish Comm. 15:208-218.

Noble, R. L., and R. D. Germany. 1986. Changes in fish populations of

Trinidad Lake, Texas, in response to abundance of blue tilapia. Pages 455-

461 in Fish Culture in Fisheries Management (R. H. Stroud, editor).

Am. Fish. Soc, Bethesda, Md.

Payne, A. I., and R. I. Collinson. 1983. A comparison of the biological

characteristics of Sarotherodon aureus (Steindachner) and S. niloticus (L.)

and other tilapia of the delta and lower Nile. Aquaculture 30:335-351.

Radonski, G. C, N. S. Prosser, R. G. Martin, and R. H. Stroud. 1984. Exotic

fishes and sport fishing. Pages 313-321 in Distribution, Biology and

Management of Exotic Fishes (W. R. Courtenay, Jr., and J. R. Stauffer,

Jr., editors). Johns Hopkins Univ. Press, Baltimore, Md.

Redner, B. D., and R. R. Stickney. 1979. Acclimation to ammonia by Tilapia

aurea. Trans. Am. Fish. Soc. 108:383-388.

Spataru, P., and M. Zorn. 1978. Food and feeding habits of Tilapia aurea

(Steindachner) (Cichlidae) in Lake Kinneret, Israel. Aquaculture 13:67-79.

Shaflund, P. L., and J. M. Pestrak. 1982. Lower lethal temperatures for

fourteen non-native fishes in Florida. Env. Biol. Fishes 7:149-156.

Taylor, J. N., W. R. Courtenay, Jr., and J. A. McCann. 1984. Known impacts

of exotic fishes in the continental United States. Pages 322-373 in Distri-

bution, Biology and Management of Exotic Fishes (W. R. Courtenay, Jr.,

and J. R. Stauffer, Jr., editors). Johns Hopkins Univ. Press, Baltimore,

Md.

Trewavas, E. 1983. Tilapiine fishes of the genera Sarotherodon, Oreochromis

and Danakilia. Cornell Univ. Press, New York.

Vineyard, G. L., R. W. Drenner, M. Gophen, U. Pollingher, D. L. Winkelman,

and K. D. Hambright. 1988. An experimental study of the plankton

community impacts of two omnivorous filter-feeding cichlids, Tilapia

galilaea and Tilapia aurea. Can. J. Fish Aq. Sci. 45:685-690.

Zale, A. V. 1984. Applied aspects of the thermal biology, ecology, and life

history of the blue tilapia, Tilapia aurea (Pisces: Cichlidae). Doctoral

dissertation, Univ. of Florida, Gainesville.

Blue Tilapia in Georgia 35

Zale, A. V. 1987. Growth, survival, and foraging abilities of early life history

stages of blue tilapia, Oreochromis aureus, and largemouth bass,

Micropterus salmoides. Env. Biol. Fishes 20:113-128.

Accepted February 1991

36

THE SEASIDE SPARROW,

ITS BIOLOGY AND MANAGEMENT

Edited by

Thomas L. Quay, John B. Funderburg, Jr., David S. Lee,

Eloise F. Potter, and Chandler S. Robbins

The proceedings of a symposium held at Raleigh, North Carolina,

in October 1981, this book presents the keynote address of F. Eugene

Hester, Deputy Director of the U.S. Fish and Wildlife Service, a

bibliography of publications on the Seaside Sparrow, and 16 major

papers on the species. Authors include Arthur W. Cooper, Oliver L.

Austin, Jr., Herbert W. Kale, II, William Post, Harold W. Werner,

Glen E. Woolfenden, Mary Victoria McDonald, Jon S. Greenlaw,

Michael F. Delany, James A. Mosher, Thomas L. Merriam, James A.

Kushlan, Oron L. Bass, Jr., Dale L. Taylor, Thomas A. Webber, and

George F. Gee. A full-color frontispiece by John Henry Dick illustrates

the nine races of the Seaside Sparrow, and a recording prepared by J.

W. Hardy supplements two papers on vocalizations.

"The Seaside Sparrow, with its extensive but exceedingly narrow

breeding range in the coastal salt marshes, is a fascinating species. All

the authors emphasize that the salt marsh habitat is at peril. . . . The

collection is well worth reading." — George A. Hall, Wilson Bulletin.

1983 174 pages Softbound ISBN 0-917134-05-2

Price: $15, postpaid, North Carolina residents add 5% sales tax. Please make

checks payable in U.S. currency to NCDA Museum Extension Fund.

Send to SEASIDE SPARROW, N.C. State Museum of Natural Sciences,

P.O. Box 27647, Raleigh, NC 2761 1.

Home Range and Substrate Use by Two Family Groups

of Red-cockaded Woodpeckers

in the North Carolina Sandhills

Richard R. Repasky1 and Phillip D. Doerr

Department of Zoology, North Carolina State University

Raleigh, North Carolina 27695-7617

ABSTRACT. — Home range and foraging habitat use by two family

groups of red-cockaded woodpeckers {Picoides borealis) were studied

over the course of a year. Average year-round convex polygon home

range size was 159 ha. One family group selected foraging areas of

relatively high pine density within the home range, whereas the second

exhibited no selection. Most foraging occurred on living pines as has

been reported elsewhere. Overlap between the sexes in foraging niche,

defined in terms of foraging substrates, was low during winter, when

the percentage of activity spent feeding was greatest. These data

suggest that food limitation, if it occurs at all, is most severe during

early winter.

The red-cockaded woodpecker, Picoides borealis (Vieillot), a species

endemic to the coastal plain of the southeastern United States (Murphey

1939, see Hooper et al. 1980 for a composite range map), is endangered

because of the declining availability of suitable habitats (Jackson 1971,

Thompson 1976, Lennartz et al. 1983). Its distribution is correlated with

the distribution of longleaf pine forest types that are 60+ years old

(Lennartz et al. 1983). Old living pines are known to be a critical

resource for nest and roost cavities (Jackson et al. 1979). There has also

been concern, because of the link with older forests, that those forests

(age 60+ years) provide foraging conditions necessary for the survival of

the species (Skorupa 1979, Ligon et al. 1986), but that has not been

demonstrated. We undertook this study in the Sandhills of North

Carolina, where the largest population of red-cockaded woodpeckers in

North Carolina is located (Carter et al. 1983) and where very little is

known of the foraging habits of this species. Our objectives were to

describe foraging habits in the region, to test for selection of forest

characteristics within home ranges that could be related to quality of

foraging habitat, and to identify that time of year when resources are

least abundant and hence habitat quality most critical.

1 Present address: The Ecology Group, Department of Zoology, University of British

Columbia, Vancouver, British Columbia, Canada V6T 2A9.

Brimleyana 17:37-52, December 1991 37

38 Richard R. Repasky and Phillip D. Doerr

METHODS

Study Area

The study was conducted on the Sandhills Game Land, Richmond

Co., N.C. Uplands were predominantly longleaf pine (Pinus palustris)-

scrub oak (Quercus laevis, Q. marilandica, Q. incana, Q. margaretta)-

wiregrass (Aristida strict a) communities. Small stream margins and

seepage slopes were characterized by pond pine (Pinus serotina)

overstories and understories ranging from grass-sedge bog through

swamp hardwoods. The gameland had been managed primarily for the

production of timber and game. Longleaf pine had been harvested on a

100-year rotation and regenerated by seed tree cuts. Pond pine had been

managed similarly but on an 80-year rotation. Prescribed burning had

taken place on a 5-year rotation, except in northern bobwhite (Colinus

virginianus) management areas, which had been burned on a 1- to 2-

year rotation. Areas under the long burning rotation had dense

hardwood understories, whereas the more frequently burned areas had

open, park-like understories. A general description of the vegetation of

the Sandhills region has been presented by Wells and Shunk (1931).

Home Ranges

Two family groups of red-cockaded woodpeckers (group A and B)

were randomly chosen from four family groups that were located within

an area that was not to be logged during the study. All individuals in

each group were color-banded. Once each month from August 1979

through July 1980, we followed one group for five continuous days of

dawn-to-dusk tracking. Thus, each group was followed six times. When

a group split while it was being tracked, we continued to follow the

subgroup with the adult male. Locations taken at 5-minute intervals

were recorded on an aerial photograph (1:12000). We also made zero-

one scan samples (Altmann 1974) of behavior at 5-minute intervals,

yielding an average of 747 observations per tracking period. Behaviors

were categorized as foraging, resting, preening, and social conflict. We

defined home ranges as the convex polygons enclosing all locations

(Odum and Kuenzler 1955), and we estimated year-round home range

for each group. Territories were defined by plotting territorial conflicts.

The resource base that we assumed to be available to each group

was encompassed within its year-round home range. That assumption

provided liberal estimates of resource availability, because convex

polygons include areas of limited use and areas outside of territorial

boundaries. Therefore, our assumption avoided the tautology of defining

resource availability based on group locations and then testing for

habitat selection with those same locations (Johnson 1980). Additionally,

extraterritorial foraging may be important even if it occurs within the

Red-cockaded Woodpecker Home Range 39

territories of other family groups, because the amount of time spent

foraging within a territory and the time spent in territorial maintenance

may depend upon territory quality (Ewald and Carpenter 1978).

A set of vegetation sampling points was selected from each overall

home range by cluster sampling. Parallel transects traversing the home

range were located perpendicular to an axis that was nearly parallel to

most streams. Transects were located randomly along the axis but could

not be less than 60 m apart. A sample point was located randomly

within each 90-m segment of a transect. Distances were paced along a

compass bearing. One hundred sample points were selected for group A,

and 49 were selected for group B.

A random sample of foraging areas was selected from each tracking

period. Foraging areas were defined as locations in the tracking itinerary

in which the scan sample of behavior indicated that the family group

was foraging. We randomly chose 25 foraging locations from each

tracking period, except the final one for which 24 were chosen because

of an error in a computer program. The samples averaged 5% of the

foraging locations within tracking bouts. Each location was relocated in

the field, and a random distance of up to 23 m (75 feet) was paced in a

random compass direction to offset potential investigator bias in

relocating points. As a further safeguard, distance and direction of each

deviation were not ascertained until the point was relocated.

Each sampling point was the center of a Bitterlich variable-radius

sampling plot (Husch et al. 1982) defined with a ten-factor prism. This

method effectively samples trees of different sizes with plot sizes most

suitable for them. For example, trees of 10 cm diameter at breast height

(DBH) are sampled with a plot size of 33 m2, whereas trees of 45 cm

DBH are sampled with a plot size of 691 m2. Species; DBH, rounded to

the nearest 0.25 cm (= 0.1 inch); and tree height, rounded to the nearest

0.3 m (= 1 foot) were recorded for each tree. Pine stems less than 2.5 cm

DBH were excluded from the analysis because the birds were not

observed foraging on them. Hardwoods were considered understory

trees if they were shorter than the mean pine height. Bole surface area

was calculated as the surface area of a cone with base on the ground,

apex at the tree height, and diameter (DBH) at breast height. For each

plot, pine bole surface density (m2/ha) and tree density (trees/ ha) were

calculated (Husch et al. 1982). Means of bole surface per tree (m2/tree),

pine DBH (cm), pine height (m), and understory hardwood height (m)

were calculated as weighted means using the density expansion factor

for each tree as its weight.

We used the Wilcoxon two-sample test (Sokal and Rohlf 1981) for

all comparisons. Rank tests are recommended in resource-use studies

because of the imprecision with which resource availability is measured

40 Richard R. Repasky and Phillip D. Doerr

(Johnson 1980). Additionally, data in this study were severely non-

normal. Calculations were performed with the Statistical Analysis System

(SAS Institute 1982a, 1982b). Wilcoxon two-sample tests were performed

with a user-written program, and home ranges were delineated and

calculated with a procedure written by the senior author.

Behavior

Behavioral observations were made during the final 3 days of each

tracking bout. Three 2-hour observation periods were conducted each

day, beginning 1 hour after sunrise, 1 hour before the solar noon, and 3

hours before sunset. In May, when nestlings were being tended and

tracking was difficult, observation periods were 3 hours long, and the

beginning and ending times of the midday and evening periods were

adjusted accordingly. During each observation period, a series of focal

individuals was selected for sampling. When an observation bout was to

begin, an individual was randomly selected from the birds available and

followed for 5 minutes or until it flew from sight or became lost amidst

the group. Instantaneous samples (Altmann 1974) of behavior and

substrate were spoken into a cassette tape recorder at 15-second intervals

that were timed with an electronic metronome.

Behaviors were lumped into functional categories. Categories

included four types of foraging behavior, namely (1) gleaning,

(2) peering and poking, (3) pecking, and (4) other, and non-foraging.

Gleaning was picking food items from exposed bark surfaces as the bird

moved forward or backward. Peering and poking consisted of peering

into and poking the bill into bark crevices in search of prey. Side-to-side

head movement was considerably greater than for gleaning. Movement

along the substrate was slower, and stops were frequent. Pecking was

subsurface foraging, including the pecking (percussion) and scaling

categories of others (Jackson 1970, Ramey 1980, Hooper and Lennartz

1981). Pecking was perpendicular to the plane of bark when excavating

for prey, or it was parallel to the bark plane to dislodge pieces. Scaling,

the latter behavior, usually followed or preceeded pecking at a foraging

spot. "Other" included obtaining seeds or fruits, drinking, and obtaining

bone fragments. If a prey item was being handled at the time of

sampling, the foraging technique used in capture was recorded as the

current behavior. Post-capture handling of food for fledglings was

classed as feeding of young instead of by the method of capture. The

consequence of this was small because the process was a rare event and

was seldom recorded. Non-foraging activities included all other behaviors.

Substrate classifications consisted of tree type, location on the tree,

and the vitality of each. Tree species were lumped into pines and

hardwoods. Locations were trunk below the tree crown, trunk within

Red-cockaded Woodpecker Home Range 41

the crown, limb, and pine cone. Trees and locations were classed as

living or dead.

Although observations were made of all family members, only data

for the adult male and female of each group were analyzed because

other birds were not present throughout the study. The 5-minute

observation bout was used as the unit of analysis because observations

made at 15-second intervals were not independent (Repasky 1984).

Substrates were divided into mutually exclusive pairs for analysis,

including pine and non-pine substrates, living and dead pines, trunk and

non-trunk surfaces of live pines, trunk-within-crown and trunk-below-

crown of live pines, and dead limb and other components of non-trunk

areas of live pines. The proportion of foraging time spent upon one

category of each pair was calculated for each bout. Least squares means

(SAS Institute 1982b) from analyses of variance were used to estimate

substrate use because of differences between time periods and between

sexes in substrate use and sample size. Calculations were by weighted

least squares regression to satisfy the assumption of homogeneity of

variance (Neter and Wasserman 1974).

Foraging behaviors were analyzed in a manner similar to that used

for substrates. The proportion of foraging time spent in each behavior

was calculated for each observation bout. Least squares means were

estimated using weighted least squares ANOVAs.

The independence of foraging behavior and substrate was tested

with a 2-way chi-square contingency table. A pool of presumably

independent observations was created by randomly selecting one

observation from each substrate in each 5-minute observation bout. All

observations from the month and individual with the least number of

observations were used in the test with equal numbers of randomly

chosen observations for other months and individuals. In this analysis,

substrates were pooled to hardwoods, dead pine surfaces, non-trunk

surfaces of living pines, live pine trunk within the crown, and live pine

trunk below the crown.

Overlap between the foraging niches of the adults of each group

was calculated monthly as a from Levins (1968). Calculations were

based on the proportions of total foraging time spent on the substrates.

These were calculated from the least squares means obtained from the

analyses described above. For example, the proportion of foraging time

spent on dead limbs of living pines was calculated as the proportion of

foraging time spent on pines times the proportion of pine foraging time

spent on living pines times the proportion of live-pine foraging time

spent on non-trunk areas times the proportion of non-trunk foraging

time spent on dead limbs. Categories used in the calculations were dead

pines, live limbs on live pines, dead limbs on live pines, pine cones,

42 Richard R. Repasky and Phillip D. Doerr

trunk within crown of live pines, trunk below crown of live pines, and

hardwoods.

Statistical summaries and tests were performed with the Statistical

Analysis System (SAS Institute 1982a, 1982b). Analyses of variance

were performed with the General Linear Models (GLM) procedure.

Contingency table tests were performed with the frequency (FREQ)

procedure.

RESULTS AND DISCUSSION

Home Range Use

The year-round range of family group A was 180 ha and that for

group B was 139 ha. Territories comprised 35% and 62% of the year-

round home ranges, respectively. The two groups spent 26% and 16%,

respectively, of foraging time outside of the territories.

Pine characteristics of foraging areas were compared with those of

the overall home ranges. Significant differences were found for group A

but not for group B (Tables 1 and 2). For group A, median pine bole

surface density and median tree density were greater in areas used than

within the home range at large. Median DBH was less. Results for

family group B were opposite to those for group A, although the

comparisons were not statistically significant. Median pine bole surface

density and tree density in foraging areas were less in areas available

than in foraging areas, whereas median DBH was nearly identical.

When the same comparisons were made for individual tracking periods,

the results were similar to the overall comparisons for each group

(Tables 1 and 2), although few of the differences were significant.

Two forest stand characteristics that are expected to be positively

related to foraging habitat quality are negatively related to one another

in nature. Tree density is expected to be positively related to foraging

quality because of improved insect habitat quality and decreasing flight

distance with increasing tree density (Wood 1983). Habitat quality is

also expected to increase as tree size increases because larger trees

should provide better prey habitat (Travis 1977, Jackson 1979) and

more foragable surface per distance travelled between trees (Hooper and

Lennartz 1981). In natural stands, however, tree size and density are

inversely related to one another (Wahlenberg 1946). DeLotelle et al.

(1987) demonstrated that red-cockaded woodpeckers prefer stands of

larger tree size when density is held constant and that they prefer stands

of greater density when tree size is held constant. The type of stands

selected for foraging by a family group is likely to depend on the

variation in stand density relative to variation in tree size.

Variation in density was greater than variation in tree size in the

home ranges studied. The coefficients of variation of tree density (group

Red-cockaded Woodpecker Home Range 43

Table 1. Median values and sample sizes (in parentheses) of pine characteristics of the overall

home range and foraging areas used by family group Aa" .

Foraging areas

1979 1980

Home

Measure range Sept. Nov. Jan. Mar. May July Pooled

Bole surface 1007.0 1053.0 886.0 1438.0** 1145.0 1199.0 1234.0 1200.0*

density (nr/ha)

Tree density

(trees/ ha)

DBH (cm)

dWilcoxon two-sample tests were used to compare pine characteristics of foraging areas with

those of the overall home range.

b * = /><0.05; ** = /><0.01; *** = P< 0.001.

Table 2. Median values and sample sizes (in parentheses) of pine characteristics of the overal

home range and foraging areas used by family group Ba .

Foraging areas

1979 1980

Home

Measure range Aug. Oct. Dec. Feb. Apr. June Pooled

Bole surface 1879.0 1385.0*1648.0 1759.0 1641.0 1570.0 1773.0 1644.0

density (m /ha)

Tree density

(trees/ha)

DBH (cm)

aWilcoxon two-sample tests were

those of the overall home range.

b * = P< 0.05; ** = P< 0.01; *** = P< 0.001

44 Richard R. Repasky and Phillip D. Doerr

A: 124, B: 118) were much larger than the coefficients of variation of

tree size (group A: 28, B: 38). Group A selected areas of greater tree

density and smaller tree size, as might be expected.

Group B did not select foraging areas on the basis of any of the

variables that we measured. Perhaps this was due to conditions within

its home range. Tree density was nearly twice that in group A's home

range, and the understory was much lower than it was in group A's

home range. Group B, therefore, may have had less need to select

foraging areas on the basis of habitat quality than did group A.

Foraging Data

Foraging substrates. Most foraging took place on pines (Table 3).

Living pines were used much more than dead pines. Males spent more

time than females foraging on limbs, and when foraging on the trunk,

males generally foraged higher than females.

With two exceptions, these results are qualitatively similar to those

for red-cockaded woodpeckers in other regions (Ligon 1968, 1970,

Morse 1972, Skorupa and McFarlane 1976, Nesbitt et al. 1978, Skorupa

1979, Ramey 1980, Hooper and Lennartz 1981, Porter et al. 1985).

First, the dead trees that were used extensively during December were

different from those used at other times of the year and in other regions.

Dead trees used outside of December were recently dead and retained

pine needles as described by Hooper and Lennartz (1981). By contrast,

dead trees used in December had long been dead and were missing large

limbs and some bark. None had died of lightning strikes during the

previous summer. Extensive use of long-dead pines has not been reported

previously, although Hooper and Lennartz (1981) reported a single

observation. Second, the extraction of seeds from open longleaf pine

cones has not been reported previously. Hooper and Lennartz (1981)

reported use during September and October of green, unopened longleaf

pine cones that contained insect larvae. During November and

December, however, we observed seeds being removed from open cones,

although this activity was a small percentage of foraging time and did

not occur during sampling.

Foraging behavior. Most foraging time was spent peering and

poking (group A: 55%, group B: 53%; Table 4). Less time was spent

pecking (group A: 27%, group B: 36%; Table 4), and the least time was

spent gleaning (group A: 15%, group B: 9%; Table 4).

Foraging substrate and behavior were not independent (group A:

X2 = 43.6, df = 8, P < 0.001; group B: X2 = 38.6, df = 8, P < 0.001),

indicating that some seasonal variation in foraging behavior was

attributable to changes in substrate use. Too few data were available to

estimate the proportion of time spent in various foraging behaviors on

each substrate. Niche overlap was therefore based on substrate use.

Red-cockaded Woodpecker Home Range

45

Table 3. Percentage of foraging time spent on various types of substrates2

a Percentages, expressed as the amount of time on the first-mentioned substrate versus some

alternative substrate, and standard errors were calculated as least squares means in analyses

of variance.

Use of pine cones occurred only during the period from September through December.

During September the male and female used cones for 8 and 10%, respectively, of time

spent off of the trunk. The figures for October were 51 and 71%, respectively. Use of cones

during November and December was too infrequent to be captured by the sampling

scheme.

46 Richard R. Repasky and Phillip D. Doerr

Niche overlap. Maximum overlap between sexes occurred at

different times for the two groups (Fig. 1). For group A, overlap was

greatest in May, when nestlings were tended and the male foraged

uncharacteristically low and upon the trunk. Overlap in group B was

greatest in October, when pine cones were used by both sexes and the

female foraged uncharacteristically high upon the trunk. Minimum

overlap between the sexes occurred in the late fall and early winter and

then again during summer (Fig. 1).

Sex-specific foraging is a means by which sexes can reduce

competition for resources (Selander 1966, Ligon 1968). Overlap is

expected to decrease as food becomes less abundant (Wallace 1974,

Hogstad 1977), although Winkler (1979) noted that Strickland's

woodpecker, Picoides stricklandi (Malherbe), exhibited the least amount

of overlap when opportunistic conditions permitted sex-specific foraging.

Lack (1954) suggested that winter and the post-fledging portion of

summer may be times of food limitation for birds. Reduced overlap

during summer does not seem to reflect opportunistic use of resources,

for no such activity was recorded in the component variables. It may

reflect resource partitioning with increased group size after fledging.

This may seem an unlikely necessity because Hooper et al. (1982)

concluded that some home ranges contain more resources than

necessary. However, the problem of resource depletion within the

proximity of a predator (Charnov et al. 1976) is magnified with

increasing group size, and reduced foraging overlap may be a solution

to that problem. It is also an alternative to changing home range size in

times of relative food scarcity (Selander 1966).

Foraging time. Red-cockaded woodpeckers were active during

most of the available daylight hours. The interval between leaving

cavities in the morning and roosting averaged 93% of the time from

sunrise to sunset, ranging monthly from 83% to 100% (Table 5). Only

the means for January, February, and March were less than 90%. The

percentage of active time spent foraging was least in May and June and

greatest during December and January (Table 5). The number of hours

of foraging per day, calculated as the product of the number of active

hours and the proportion of time spent foraging, corresponded closely

with the number of active hours per day (Table 5), being greatest in

summer and least in winter.

Foraging activities should reflect the availability of food relative to

needs. The proportion of daylight time spent in activity and foraging as

a proportion of active time are expected to be greatest during the period

of resource limitation. Hinde (1952) found that tits (Parus spp.) increased

the proportion of daylight hours in which they were active during the

winter. Gibb (1954) found that the proportion of time spent foraging

Red-cockaded Woodpecker Home Range

47

was greatest in December for four species of tits and greatest in

February for a fifth species. Our finding that red-cockaded woodpeckers

were least active during the winter is not evidence against the hypothesis

that winter is the principal time of resource limitation for red-cockaded

woodpeckers. It may reflect avoidance of unfavorable conditions by

remaining in the cavity. In winter, the birds generally did not leave their

cavities until the sun reached the trees, and then they often basked.

Thus they may avoid pre-sunrise activity that is expensive for small

birds (Morse 1970). The absolute number of hours spent foraging need

not be greatest in the period of food limitation. In winter, activity is

limited by daylight, and sufficient energy must be acquired to survive

the night. As days lengthen, the potential for more activity increased

(e.g. cavity construction, territorial defense, reproduction). Thus the

energy budget increases and the absolute foraging time increases to meet

these needs. Presumably this is offset, to some unknown degree, by

differences in temperature and food abundance. Strain on the energy

budget is probably best reflected by the percentage of the activity

budget spent foraging. This was greatest during December and January.

1.0

uj 0.4

X

O

Z 0.2

0.0

GROUP B

GROUP A /

\ /

i » i | I I I I I I I

A/79 SONDJ FMAMJ J/80

MONTH

Fig. I. Overlap in foraging substrate use between the sexes within 5-minute

observation bouts.

48 Richard R. Repasky and Phillip D. Doerr

Table 4. Mean and standard error of the percentage of foraging time spent using

various foraging methods.

"Other" time was calculated as the percentage complement of peering and

poking, pecking, gleaning; so, it also includes estimation error associated with

these behaviors.

Synthesis

Several pieces of evidence intersect to suggest that if food is

limiting to red-cockaded woodpeckers, it may be least available during

early winter. Foraging occupied the greatest portion of the day in

December and January, and overlap in substrate use between the sexes

was low during November and December. Furthermore, supplanting

attacks for foraging sites among family group members peaked during

December and January, and supplanting attacks by red-bellied wood-

peckers, Melanerpes carolinus (L.), peaked during January and February

(Repasky 1984). Skorupa (1979) has also argued that winter is the

period of resource limitation on the basis of seasonal territory dynamics.

If foraging habitat quality were related to habitat structure, habitat

preference would be expected to be strongest during the period of food

limitation. The family group that exhibited preference in this study did

so most strongly during January, and it preferred areas of higher pine

density and surface area and smaller tree diameter within its home

range.

Red-cockaded Woodpecker Home Range

49

Table 5. Mean daily activity budgets.

Month

Our data do not permit firm conclusions about food limitation and

habitat preference, but they can be used to suggest areas of future

research. We believe that it is worth investigating whether winter

survival of red-cockaded woodpeckers is a significant factor influencing

population size and structure and influenced by winter foraging

conditions, and if so, whether foraging conditions are related to

manageable characteristics of forests.

ACKNOWLEDGMENTS.— We thank W. Hafley, R. Lancia, M.

Reed, E. Seneca, and J. Walters for reviewing earlier drafts of the

manuscript, and we are grateful to S. Allen, R. Blue, J. Carter, G.

Hepp, S. Klause, J. Phillips, I. Rusnak, T. Stamps, and L. Stribling for

assistance in the field. This project was supported under the Endangered

Species Act of 1973 by the U.S. Fish and Wildlife Service in cooperation

with the N.C. Wildlife Resources Commission. Computing resources

were provided by VanLaan and Associates of Cary, N.C, and by the

50 Richard R. Repasky and Phillip D. Doerr

Department of Zoology at N.C. State University. This is Paper No.

10230 of the Journal Series of the North Carolina Agricultural Research

Service.

LITERATURE CITED

Altmann, J. 1974. Observational study of behavior: sampling methods.

Behaviour 49:227-267.

Carter, J. H., III., R. T. Stamps, and P. D. Doerr. 1983. Red-cockaded

woodpecker distribution in North Carolina. Pages 20-23 in Red-cockaded

Woodpecker Symposium II Proceedings (D. A. Wood, editor). Florida

Game and Fresh Water Fish Comm., Tallahassee.

Charnov, E. L., G. H. Orians, and K. Hyatt. 1976. Ecological implications of

resource depression. Am. Nat. 110:247-259.

DeLotelle, R. S., R. J. Epting, and J. R. Newman. 1987. Habitat use and

territory characteristics of red-cockaded woodpeckers in central Florida.

Wilson Bull. 99:202-217.

Ewald, P. W., and F. L. Carpenter. 1978. Territorial response to energy

manipulations in the Anna hummingbird. Oecologia 31:277-292.

Gibb, J. 1954. Feeding ecology of tits, with notes on treecreeper and goldcrest.

Ibis 96:513-543.

Hinde, R. A. 1952. The behaviour of the great tit and some related species.

Behav. Suppl. II, pp. 1-201.

Hogstad, O. 1977. Seasonal changes in intersexual niche differentiation of the

three-toed woodpecker {Picoides tridactylus). Ornis Scand. 8:101-1 1 1.

Hooper, R. G., and M. R. Lennartz. 1981. Foraging behavior of the red-

cockaded woodpecker in South Carolina. Auk 98:321-334.

Hooper, R. G., A. F. Robinson, Jr., and J. A. Jackson. 1980. The red-

cockaded woodpecker: notes on life history and management. USDA For.

Serv. Gen. Rep. SA-GR9.

Hooper, R. G., L. J. Niles, R. F. Harlow, and G. W. Wood. 1982. Home

ranges of red-cockaded woodpeckers in coastal South Carolina. Auk

99:675-682.

Husch, B., C. I. Miller, and T. W. Beers. 1982. Forest mensuration. 3rd ed.

John Wiley and Sons, New York.

Jackson, J. A. 1970. A quantitative study of the foraging ecology of downy

woodpeckers. Ecology 51:318-323.

Jackson, J. A. 1971. The evolution, taxonomy, distribution, past populations

and current status of the red-cockaded woodpecker, Pages 4-26 in The

Ecology and Management of the Red-cockaded Woodpecker (R. L.

Thompson, editor). U.S. Bur. Sport Fish., Washington, D.C.

Jackson, J. A. 1979. Tree surfaces as foraging substrates for insectivorous

birds. Pages 69-93 in The Role of Insectivorous Birds in Forest Ecosystems

(J. G. Dickson, R. N. Conner, R. R. Fleet, J. A. Jackson, and J. C. Kroll,

editors). Academic Press, New York.

Jackson, J. A., M. R. Lennartz, and R. G. Hooper. 1979. Tree age and cavity

initiation by red-cockaded woodpeckers. J. For. 77:102-103.

Red-cockaded Woodpecker Home Range 51

Johnson, D. H. 1980. The comparison of usage and availability measurements

for evaluating resource preference. Ecology 61:65-71.

Lack, D. 1954. The Natural Regulation of Animal Numbers. Clarendon Press,

Oxford, England.

Lennartz, M. R., H. A. Knight, J. P. McClure, and V. A. Rudis. 1983. Status

of red-cockaded woodpecker nesting habitat in the South. Pages 13-19 in

Red-cockaded Woodpecker Symposium 11 Proceedings (D. A. Wood,

editor). Florida Game and Fresh Water Fish Comm., Tallahassee.

Levins, R. 1968. Evolution in Changing Environments. Princeton Univ. Press,

Princeton, N.J.

Ligon, J. D. 1968. Sexual differences in foraging behavior in two species of

Dendrocopus woodpeckers. Auk 85:203-215.

Ligon, J. D. 1970. Behavior and breeding biology of the red-cockaded

woodpecker. Auk 87:255-278.

Ligon, J. D., P. B. Stacey, R. N. Conner, C. E. Bock, and C. S. Adkisson.

1986. Report of the American Ornithologists1 Union committee for the

conservation of the red-cockaded woodpecker. Auk 103:848-855.

Morse, D. H. 1970. Ecological aspects of some mixed-species foraging flocks

of birds. Ecol. Monogr. 40: 1 19-168.

Morse, D. H. 1972. Habitat utilization of the red-cockaded woodpecker during

the winter. Auk 89:429-435.

Murphey, E. E. 1939. Dryobates borealis (Vieillot), red-cockaded woodpecker.

Pages 72-82 in Life Histories of North American Woodpeckers (A. C.

Bent). U.S. Natl. Mus. Bull. No. 174.

Nesbitt, S. A., D. T. Gilbert, and D. B. Barbour. 1978. Red-cockaded

woodpecker fall movements in a Florida flatwoods community. Auk

95:145-151.

Neter, J., and W. Wasserman. 1974. Applied linear statistical models. Richard

D. Irwin, Homewood, 111.

Odum, E. P., and E. J. Kuenzler. 1955. Measurement of territory and home

range size in birds. Auk 72:128-137.

Porter, M. L., M. W. Collopy, R. F. Labisky, and R. C. Littell. 1985.

Foraging behavior of red-cockaded woodpeckers: an evaluation of research

methodologies. J. Wildl. Manage. 49:505-507.

Ramey, P. 1980. Seasonal, sexual, and geographic variation in the foraging

ecology of red-cockaded woodpeckers (Picoides borealis). M.Sc thesis,

Mississippi State Univ., Mississippi State.

Repasky, R. R. 1984. Home range and habitat utilization of the red-cockaded

woodpecker. M.Sc. thesis, North Carolina State Univ., Raleigh.

SAS Institute. 1982a. SAS User's Guide: Basics. SAS Institute, Cary, N.C.

SAS Institute. 1982b. SAS User's Guide: Statistics. SAS Institute, Cary, N.C.

Selander, R. K. 1966. Sexual dimorphism and differential niche utilization in

birds. Condor 68:113-151.

Skorupa, J. P. 1979. Foraging ecology of the red-cockaded woodpecker in

South Carolina. M.Sc. thesis, Univ. of California, Davis.

Skorupa, J. P. and R. W. McFarlane. 1976. Seasonal variation in foraging

territory of red-cockaded woodpeckers. Wilson Bull. 88:662-665.

52 Richard R. Repasky and Phillip D. Doerr

Sokal, R. R., and F. J. Rohlf. 1981. Biometry. 2nd ed. W. H. Freeman, San

Francisco.

Thompson, R. L. 1976. Change in status of red-cockaded woodpecker colonies.

Wilson Bull. 88:491-492.

Travis, J. 1977. Seasonal foraging in a downy woodpecker population. Condor

79:371-375.

Wahlenberg, W. G. 1946. Longleaf Pine. Charles Lathrop Pack Forestry

Foundation, Washington, D.C.

Wallace, R. A. 1974. Ecological and social implications of sexual dimorphism

in five melanerpine woodpeckers., Condor 76:238-248.

Wells, B. W., and 1. V. Shunk. 1931. The vegetation and habitat factors of the

coarser sands of the North Carolina coastal plain: an ecological study.

Ecol. Monogr. 1:465-520.

Winkler, H. 1979. Foraging ecology of Strickland's woodpecker in Arizona.

Wilson Bull. 91:244-254.

Wood, D. A. 1983. Foraging and colony habitat characteristics of the red-

cockaded woodpecker in Oklahoma. Pages 51-58 in Red-cockaded

Woodpecker Symposium II Proceedings (D. A. Wood, editor). Florida

Game and Fresh Water Fish Comm., Tallahassee.

Accepted March 1991

Mating and First-season Births in Interstate

Transplanted River Otters, Lutra canadensis

(Carnivora: Mustelidae)

Peter J. Tango and Edwin D. Michael

Division of Forestry, West Virginia University

Morgantown, West Virginia 26506

And

Jack I. Cromer

West Virginia Department of Natural Resources

Elkins, West Virginia 26541

ABSTRACT. — River otters were transplanted from North Carolina

and Maryland into the West Fork River, W. Va., during February and

April 1987. Birth of young and mating by the released otters occurred

within 2 months of release. This represents the only known occurrence

of birth during the same year following transplanting.

Several states have transplanted river otters (Lutra canadensis)

(Schreber) in recent years in an attempt to establish or reestablish

breeding populations. Birth of young during the same year otters were

released was not documented in any of those states. Releases in Missouri

resulted in reproduction during the second post-release reproductive

period (Erickson and McCullough 1987). Timing of confinement in

Louisiana, the source of the otters, and of subsequent release in

Missouri appeared to be the cause of that reproductive delay (Erickson

and McCullough 1987). Female otters transplanted by one Louisiana

supplier were usually held until after parturition (Serfass and Rymon

1985). In Tennessee, no reproduction was documented during the first

(17-month) post-release period of radio-tagged animals into the Obed

River system (Griess and Anderson 1987). One of 1 1 females transplant-

ed to Pine Creek, northcentral Pennsylvania, and thought to be pregnant

when released in August, reportedly remained near a den the next

spring; area residents indicated a family of otters was present in that

stretch of river that year (Serfass and Rymon 1985). Pollard (1984) in

Oklahoma also did not indicate any first-year post-release reproduction.

In 1986, a program for reintroduction of river otters into northcen-

tral West Virginia was initiated. The first release of four radio-tagged

otters (three males and one female) on the West Fork River was on 5

Brimleyana 17:53-55, December 1991 53

54 Peter J. Tango, Edwin D. Michael, and Jack I. Cromer

February 1987; the animals had been trapped in North Carolina coastal

waters in January 1987. We first observed an otter pup with the female

(designated F10) at 0636 hours on 21 May 1987, 105 days after her

release. At 1712 hours on 23 May 1987, we confirmed that two pups

were with the female. Young otters typically do not leave the den until

they are approximately 2 months old (Toweill and Tabor 1982);

therefore, we estimated birth of young as the fourth week after release.

That represents the shortest known period between release of a

translocated otter and birth of young.

We also observed mating activities among the four radio-tagged

otters present on the river system in March 1987. A second female (F97)

was released on 17 February 1987, following the death of one of the

males that had been released on 5 February. Female F10, when located

on the morning of 9 March 1987, was on the bank of the main river

channel with a male (M50). A local resident had heard and seen the

animals that same morning and described what Liers (1951) termed

caterwauling. Approximately 5 hours later, at 1403 hours, the same two

otters were observed copulating 1.5 km upstream from what we believed

had been the site of mating activity that morning. Copulation was

observed for 13 minutes and was in the water. The two animals

remained in close proximity to one another for the next 7 days. At 0717

hours on 1 1 March 1987, they were seen on the main channel river

bank, where at 0720 the male mounted the female. The female dragged

the male into the water and the two animals remained coupled. At 0739

it appeared they may have separated briefly, but the male remounted

almost immediately and they remained coupled in the water until 0800

hours. Male M50 remained within 1.7 km of the female F10 den site

until the evening of 16 March 1987, but no other copulations were

observed. The copulations had occurred 0.4-1.9 km downstream from

the den site where female F10 had earlier given birth to pups.

On 10 March 1987 at 0721 hours, we observed female F97 and

male M60 mating in Freemans Creek, the first tributary of the West

Fork River downstream of the release site. The duration of this event

was not determined. Copulation took place on a log in the creek until

the otters fell into the water together. The female was heard caterwauling

during the copulation. All copulations for these four otters took place

within 6.2 km of the February 1987 release site.

Though we suspected that female F10 had young because of her

localized movements with respect to one den site, F97 was less site-

specific in her movements, and no young were seen with her. The

observed mating dates for F10 suggest a birth date in agreement with

backdating from the first observations of her pups, as otters are

considered postpartum-estrous animals (Toweill and Tabor 1982).

Transplanted River Otters 55

The occurrence of a confirmed birth during the same year of

transplanting is exceptional. Whether or not timing of trapping and

translocation to occur 6 to 7 weeks prior to the peak birthing period

would result in more such births is speculative, but the method should

be tested. It could reduce the 1- to 2-year lag in first-generation

production that has been observed in translocation programs. However,

delays in reproduction for 1 to 2 years after release should continue to

be expected following late-winter/ spring release of translocated otters,

because individual responses of females to the entire trap-and-transfer

procedure will remain variable and will influence reproduction.

LITERATURE CITED

Erickson, D. W., and C. R. McCullough. 1987. Fates of translocated river

otter in Missouri. Wildl. Soc. Bull. 15:511-517.

Griess, J. M., and B. Anderson. 1987. Reintroduction of the river otter into the

Obed Wild and Scenic River in Tennessee. Tenn. Wildl. Res. Agency,

Crossville (unpublished report).

Liers, E. E. 1951. Notes on the river otter (Lutra canadensis). J. Mammal.

32:1-9.

Pollard, L. 1984. Return of the river otter. Outdoor Oklahoma 40:2-7.

Serfass, T. L., and L. M. Rymon. 1985. Success of river otter introduced in

Pine Creek drainage in northcentral Pennsylvania. Trans. Northeast Sect.

Wildl. Soc. Fish Wildl. Conf. 42:138-149.

Toweill, D. E., and J. E. Tabor. 1982. River otter (Lutra canadensis). Pages

688-703 in Wild Mammals of North America: Biology, Management, and

Ecology (J. A. Chapman and G. A. Feldhammer, editors). John Hopkins

Univ. Press, Baltimore, Md.

Accepted March 1991

56

ENDANGERED, THREATENED, AND

RARE FAUNA OF NORTH CAROLINA

PART I.

A RE-EVALUATION OF THE MAMMALS

Edited by Mary Kay Clark

This book is a report prepared by a committee appointed in 1985

by the North Carolina State Museum of Natural Sciences to re-evaluate

the list of mammals presented in Endangered and Threatened Plants

and Animals of North Carolina (John E. Cooper, Sarah S. Robinson,

and John B. Funderburg, editors. N.C. State Mus. Nat. Hist., Raleigh,

1977), which is now out of print. Committee members were Mary Kay

Clark, David A. Adams, William F. Adams, Carl W. Betsill, John B.

Funderburg, Roger A. Powell, Wm. David Webster, and Peter D.

Weigl. The report treats 21 species listed in the following status

categories: Endangered (5), Threatened (1), Vulnerable (6), and

Undetermined (9). Most species accounts discuss the animal's physical

characteristics, range, habitat, life history and ecology, special sig-

nificance, and status (including the rationale for the evaluation and

recommendations for protection) and provide a range map and an

illustration of the animal's external characters. Ruth Brunstetter and

Renaldo Kuhler illustrated the book. An introductory section contributed

by Ms. Clark discusses the changes in status that occurred in the decade

between 1975 and 1985. It also mentions efforts to protect marine

mammals and includes a checklist of the cetaceans known from North

Carolina.

1987 52 pages Softbound ISBN 0-917134-14-1

Price: $5 postpaid. North Carolina residents add 5% sales tax. Please make

checks payable in U.S. currency to NCDA Museum Extension Fund.

Send order to: ETR MAMMALS, N.C. State Museum of Natural Sciences,

P.O. Box 27647, Raleigh, NC 2761 1.

Habitat Associated With Home Ranges of Female

Odocoileus virginianus (Mammalia: Cervidae)

in Eastern Kentucky

Richard C. Pais1

Department of Forestry, University of Kentucky

Lexington, Kentucky 40546-0073

William C. McComb2

Department of Forestry, University of Kentucky

Lexington, Kentucky 40546-0073

AND

John Phillips

Kentucky Department of Fish and Wildlife Resources

#1 Game Farm Road, Frankfort, Kentucky 40601

ABSTRACT. — Modified minimum-area home ranges were estimated

for eight does of the white-tailed deer, Odocoileus virginianus, relocated

from bottomland hardwood habitat of western Kentucky to the

Cumberland Plateau of eastern Kentucky and for six does resident to

eastern Kentucky. Mean size of home ranges was similar for resident

(642 ha) and relocated (668 ha) does. Data obtained on vegetation,

land use, and topography from a computerized Geographic Information

System (GIS) indicated that home ranges of resident does included

more bottomland habitat than was randomly available (P< 0.03) and

that those of relocated does included more young forest than was

randomly available (P < 0.05). GIS may be an economical tool for

identification of future release sites.

Populations of white-tailed deer, Odocileus virginianus Rafinesque,

have been increasing throughout most of Kentucky (Phillips 1983)

except on the Cumberland Plateau in the east. Previous attempts to

reintroduce 40-50 deer per county in this area have failed to produce a

herd near carrying capacity (Phillips 1983). Stocking 400-500 deer per

county is being attempted on the Cumberland Plateau, by the Kentucky

Department of Fish and Wildlife Resources, to establish viable

populations.

1 Present address: DAFT-McCune- Walker, Inc., 200 E. Pennsylvania Avenue, Townson,

MD 21204.

2 Present address: Department of Forest Science, Oregon State University, Corvallis, OR

97331.

Brimleyana 17:57-66, December 1991 57

58 R. C. Pais, W. C. McComb, and J. Phillips

The release of deer into areas of high-quality habitat may reduce

dispersal and increase the probability of successful stocking (Pais 1987).

The existence in eastern Kentucky of a computerized Geographic

Information System (GIS) based on maps and remotely sensed data

allowed us to characterize the habitat within home ranges of white-

tailed deer. We know of no~data on size or habitat characteristics of

home ranges of these deer in eastern Kentucky prior to this study.

In this study we proposed to compare the size of home ranges of

resident and relocated does in eastern Kentucky, to characterize the

features of the home ranges according to GIS categories, to compare the

relative abundance of these features on actual and randomly available

home ranges for resident and relocated does, and to characterize areas

that should be considered as future release sites.

STUDY AREA AND METHODS

The trapping site of the white-tailed deer that were to be relocated

was the Ballard County Wildlife Management Area (WMA) in western

Kentucky. Ballard Co. topography ranges from flat to moderately

rolling, with a maximum relief of 55 m. The natural forest type in the

WMA, which is adjacent to the Ohio River, is bottomland hardwood.

Millet {Echinochloa walteri), soybeans (Glycine max), and corn (Zea

mays) have been planted to attract waterfowl. Estimated density of

white-tailed deer on the WMA in 1986 was 1/1.5 ha according to the

DPOP2 deer population model for microcomputers (Phillips 1985).

The release sites were in Knott Co., Ky., in the central Cumberland

Plateau. Local relief of 200-350 m is common. Cliffs occur on surface

mines and highway road-cuts. Approximately 80% of the county is

forested, 10% is reclaimed or active surface mines, and 10% is bottom-

land with little agricultural land present. Roads and houses occur

primarily in bottomlands. The forest type is mixed mesophytic, with

most stands 40 to 60 years old.

Resident white-tailed deer came from the University of Kentucky's

Robinson Forest in Breathitt Co. and Knott Co. Robinson Forest is

similar in topography and vegetation to the release sites in Knott Co.,

but there has been no surface mining and it has been closed to the

public for 10 years.

Knott Co. was closed to hunting during the period of deer stocking

in 1983-85 but was opened to regular statewide seasons in the fall of

1985. Prior to the deer releases in 1983, Knott Co. supported <1

deer/ 600 ha according to the DPOP2 model (Phillips 1985).

During 1983-85, 485 deer were relocated to Knott Co.; they were

generally moved in lots of 25. For this study, which was conducted in

1984-85, radio-collared relocated does were released during 1985 in

Home Ranges of Female Deer 59

three locations in the county: 6 on 6 February at Vest, 11 on 22

February at Knob Bottom, and 18 on 13 March at Carr Fork. Only 8 of

these 35 radio-collared does were used in the analysis reported here,

because not all does survived and established home ranges. In addition,

resident does were captured between November 1984 and April 1985 in

Robinson Forest; these does were also radio-collared, and 6 were used

as controls for the relocated does in this analysis.

Does slated to be relocated were captured in rocket nets, in

Stephenson box traps, and with rifle-propelled darts containing

succinylcholine chloride. We attached radio collars at the time of

capture. The does captured at Ballard County WMA were kept for up

to 5 days prior to shipment in a modified barn designed to reduce stress

and limit human contact. The interior of the barn was dark, and does

were provided with food and water through panels removable to the

barn's exterior. They were loaded for shipment by slowly rolling one of

the barn walls toward a loading ramp until all individuals entered a

waiting truck. The does were not immobilized during shipment.

The three release sites were chosen by local conservation officers.

The criteria for their selection were that local interest in the stocking be

high and that the probability for harassment of the does by dogs and

people be low. Consequently, most release sites were remote.

Resident does were captured with rocket-nets in clearings baited

with corn or salt. We immobilized the deer with intramuscular injections

of xylazine hydrochloride (0.01 mg/kg of body weight) so that radio

collars could be attached. The deer were then released at the point of

capture.

Radio collars had a life expectancy of at least one year. Constructed

of brown nylon, they were permanently attached to each doe. A three-

element H-type antenna was used for aerial and ground radio-tracking.

Radio location vectors were taken on each resident and relocated doe

within Knott Co. at least twice weekly from 6 February to 13 November

1985.

The approximate location of individuals was determined by

triangulation (Cochran 1980:517-519). Vectors that crossed at angles

>135° or<22° were not used to record locations. Hence, not all vectors

resulted in fixes, and the mean number of observations/ doe/ week was

18.6.

The accuracy of vectors was determined by triangulating from

varying distances on three transmitters of known positions at Robinson

Forest and averaging the bearing precision over the mean distance. The

average of the error polygons was determined using two bearings per

triangulation according to the procedures of Heezen and Tester (1967;

see also Nams and Boutin 1991).

^c

60 R. C. Pais, W. C. McComb, and J. Phillips

We used the modified minimum-area method to estimate home

ranges because it miminized the chance of including areas not used by

an individual (Harvey and Barbour 1965, Mooty et al. 1987). This

method is most useful with irregularly shaped home ranges, and those

of white-tailed deer are usually elongated (Marchinton and Hirth 1984).

The cumulative (total to date) locations recorded for each individual

were plotted against each estimated home range to determine if the

estimate was accurate. Locations were recorded 24 hours apart to

increase the independence among observations (Swihart and Slade

1985). Home ranges were plotted only for individuals for which an

asymptote was approached (Fig. 1).

The habitat of the study area was characterized with the aid of the

Kentucky Department of Natural Resources' GIS. This computerized

system uses information from satellite and aerial photography, U.S.

Geological Survey 7.5-minute topographic maps, and site inspections to

categorize habitat features over large land areas. Four GIS files, each

consisting of a habitat category, were used in this study (Table 1).

Appropriate sets of categories were assigned to 0.4-ha (1-acre) polygons.

Fifteen 10-ha site inspections were conducted to ascertain the reliability

of the GIS data in Knott Co. No discrepancies were detected.

Modified minimum-area home ranges for does were digitized onto

a computer map for each GIS file. Each home range was then

repositioned at random on the same 7.5-minute quadrangle map to

estimate the random availability of habitat in the vicinity of the actual

home range. The percentage of the total area represented by each

habitat category was calculated for each home range according to the

MAP model (Berry and Tomlin 1981). Because radio locations were

imprecise (error polygons averaged 50 ha) relative to some map features,

and because deer are highly mobile, we wished to identify home-range

selection rather than selection of patches within home ranges (second-

order selection of Johnson 1980). The percentages represented by each

habitat category in actual randomly available home ranges of both

resident and relocated does were compared by Student's /-tests. Data

approximated normality because percentages of habitat types had narrow

ranges of variability in this data set. Transformation was not necessary.

RESULTS AND DISCUSSION

We recorded 925 radio locations from approximately 1,900 bearing

sets over a 3,600-km2 area; 742 locations were of 1 1 relocated does and

183 were of six resident does at Robinson Forest. Most locations of

relocated does revealed that the does had dispersed less than 15 km

from the release site.

Home Ranges of Female Deer

61

800

<

LU

<

LU

o

<

CC

LU

O

X

LU

>

5

o

600 -

400 ~

200

— DOEN0.12

— DOEN0.21

i — i — i — i — i — i — i — i — i — i — i — i — i — T

3 5 7 9 11 13 15 17 19 21 23 25 27 29

CUMULATIVE LOCATIONS (NO.)

Fig. 1. Actual plotting of cumulative locations versus cumulative home-range

area; note curves for defined (Doe No. 12) and undefined (Doe No. 21) home

range. A home range is defined only when additional locations do not increase

its area.

Triangulation was made difficult by the rugged topography of the

study area. Deflection of signals from mountainsides frequently resulted

in inaccurate bearings. The accuracy of bearings was ±7.5° at 2.5 km,

and the average of the error polygons was 50 ha at that distance. In a

Minnesota study, accuracy of bearings ranged from 0 to 40° (Mooty et

al. 1987). Most (>70%) of our locations were estimated from bearings

taken <2.5 km from the animal. Poor roads and rugged topography

hindered movement from one spot to the next while bearings were being

taken, and some animals moved before locations could be pinpointed.

62

R. C. Pais, W. C. McComb, and J. Phillips

Table 1. Habitat categories in Kentucky's Geographic Information System.

Habitat category

Definition

1. Land form

Bottomland

Sideslope

Ridgetop

2. Land use

Agricultural

Human-altered

Natural vegetation

Young forest

A flat surface adjacent to a stream and low-lying land

bounded by hills or sideslopes.

The steeply inclined portions of a plateau.

A flat surface bounded below by sideslopes.

Areas distinguished by geometric field patterns; areas

that lack activity or reflect patterns of grazing; forest

openings maintained specifically for wildlife.

Areas intensively used by humans, with most land

covered with structures; areas committed to residential

use; areas of sparse residential use such as farmsteads;

areas used for the sale of products and services; areas

of light to heavy manufacture; mines, quarries, and

gravel pits.

Areas of undisturbed indigenous vegetation.

Regeneration of areas in which all mature trees have

been cut and removed, and areas of forest lands inter-

laced with mines.

3. Slope gradient

0-20%, 20-35%,

35-50%, >50%

4. Vegetation

Deciduous forest

Disturbed sites

The ratio between vertical rise and horizontal distance.

All forests dominated by trees that lose their leaves at

the end of the frost -free season.

Areas where human endeavor has changed the surface

of the earth. Usually high-use areas and areas where

grasses and forbs predominate. Former cropland or

pastureland, now grown up in shrubs, in transition

back to forest land.

Wetland Areas where the water table is at, near, or above the

land surface for a major part of the year.

Although the estimated precision of fixes was low, estimates of habitat

use ought to provide a conservative comparison with randomly assigned

home ranges. Error will add to among-animal variance estimates of

habitat features and result in a decreased probability of detecting a

difference. It is now apparent that we should have used Lenth's maximum

likelihood estimators to access precision of locations; the technique was

unavailable to us at the time of the study (Nams and Boutin 1991).

Home Ranges of Female Deer

63

Table 2. Percentage of habitat in actual and randomly available home ranges of

radio-collared does in Knott and Breathitt counties, 1985.a

a Standard errors indicated parenthetically.

b Variables significantly different between actual and randomly available home

ranges (P< 0.05).

c Variables significantly different between home ranges of resident and relocated

does (P< 0.05).

Home Range. Home ranges were not defined for all radio-collared

does because some did not provide a sufficient number of locations (as a

result of death, radio failure, or large dispersal) and because some does

apparently did not establish a home range during the study. Eight

relocated and six resident does established defined home ranges. Resident

does averaged 23 (SE = 3.1) locations per defined home range and

relocated does averaged 29 (SE = 7.4). Mooty et al. (1987) felt that 30

locations were sufficient to construct modified minimum-area home

ranges for white-tailed deer in Minnesota. Sizes of home ranges for

resident (x = 642 ha, SE - 132) and relocated (x = 668 ha, SE = 79) does

were not significantly different (P > 0.05). These home ranges were

large in comparison with others reported for white-tailed deer in the

Southeast: 267 ha (Bridges 1968) and 80 ha (Byford 1970) in the East

64 R. C. Pais, W. C. McComb, and J. Phillips

Gulf Coastal Plain; 70 ha in the West Gulf Coastal Plain (Hood 1971);

58 ha in the Piedmont Upland (Marshall and Wittington 1969); 514 ha

in the Ozark Highlands (Cartwright 1975); and 84 ha in the North

Carolina mountains (Marchinton 1968).

Habitat Analysis. The habitat categories for which home ranges of

resident does differed significantly (P < 0.05) from those of relocated

does included sideslopes, natural vegetation, slopes between 35% and

50%, deciduous forest, and disturbed sites (Table 2). Random availability

of these features also differed (P < 0.05) between the two samples, so

resident does were not pooled with relocated does for habitat analysis.

Actual home ranges of resident does contained significantly more

(P< 0.03) bottomland (jc = 10%, SE = 1.2) than did randomly available

home ranges (Jc = 6%, SE = 1.5) (Table 2). Wildlife openings planted

with winter wheat (Triticum aestivum) were located in these bottomlands.

Winter wheat is a preferred forage of white-tailed deer (Whitehead

1967). In eastern Kentucky, open water is found almost exclusively in

bottomlands; its presence may have been a factor in the relatively high

percentage of bottomland within home ranges of resident does.

Actual home ranges of relocated does had significantly more (P <

0.05) young forest (Jc = 3%, SE = 2.6) than did randomly available home

ranges (jc = 1%, SE = 0.3) (Table 2). Reclaimed surface mines or forest

edges created by mining may have provided the does with dense shrubs

for browse and cover (Harlow and Hooper 1971, Knotts 1975).

Mangement Implications. Two factors may have caused does in

eastern Kentucky to have large home ranges. First, the presence of large

tracts of contiguous forest may have induced the does to expand their

home ranges in search of food and cover; from 88% to 95% of the

habitat available to deer was deciduous forest (Table 2). Second,

harassment by dogs, which has been a source of mortality of deer on the

Cumberland Plateau (Anderson 1979, Pais 1987), can increase deer

dispersal and may have done so in this instance.

The future success of reintroducing white-tailed deer on the

Cumberland Plateau may be enhanced by choosing release sites in high-

quality habitat. That the percentages of bottomlands and young forests

in home ranges were higher than expected (10% and 3%, respectively)

suggests that release sites need not be selected on the basis of their

remoteness from human contact. Good habitat could quickly and

economically be determined with the GIS. Young deciduous forest,

which avergaged 3% (~ 20 ha) in the various home ranges of relocated

does, was distributed in patches throughout these ranges. We therefore

suggest that future release sites contain >3% young deciduous forest

well distributed throughout the area in small patches.

Home Ranges of Female Deer 65

ACKNOWLEDGMENTS.— We thank the Wildlife Biologists and

Conservation Officers who assisted with the project, the Robinson

Forest staff for assistance with field work, and R. Kryscio for advice on

statistical analyses. M. Powers and T. Nieman provided GIS data and

use of computer hardware. C. N. Huegel and M. Newton provided

valuable comments on an earlier draft of this manuscript.

The information reported in this manuscript (88-8-171) was supported

by Federal Aid Project W-45 through the Kentucky Department of Fish

and Wildlife Resources and by the Kentucky Agricultural Experiment

Station. It is published with the approval of the Experiment Station

Director. This is Paper No. 2426 of the Forest Research Laboratory,

Oregon State University.

LITERATURE CITED

Anderson, D. T. 1979. The effect of dog harassment on translocated white-

tailed deer {Odocoileus virginianus) on the Cumberland Plateau in

Tennessee. Tenn. Wildl. Resour. Agency Tech. Rep. No. 79-8, Nashville,

Tenn.

Berry, J. K., and C. D. Tomlin. 1981. Fundamental procedures of geographical

information analysis with an established GIS. NASA/ERRSAC

Applications Conf., Danvers, Mass.

Bridges, R. J. 1968. Individual white-tailed deer movement and related

behavior during the winter and spring in northwestern Florida. M.S. thesis,

Univ. of Georgia, Athens.

Byford, J. L. 1970. Movement of white-tailed deer to changing food supplies.

Proc. Annu. Conf. Southeast. Assoc. Game and Fish Comm. 23:63-78.

Cartwright, M. E. 1975. An ecological study of white-tailed deer in north-

western Arkansas: home range, activity and habitat utilization. M.S. thesis,

Univ. of Arkansas, Fayetteville.

Cochran, W. W. 1980. Wildlife telemetry. Pages 507-520 in Wildlife

Management Techniques, 3rd ed. (S. D. Schemnitz, editor). The Wildlife

Society, Washington, D. C.

Harlow, R. F., and R. G. Hooper. 1971. Forages eaten by deer in the

Southeast. Proc. Annu. Conf. Southeast. Assoc. Game and Fish Comm.

24:18-46.

Harvey, M. J., and R. W. Barbour. 1965. Home range of Microtus ochrogaster

as determined by a modified minimum area method. J. Mammal. 46:398-402.

Heezen, K. L., and J. R. Tester. 1967. Evaluation of radio-tracking by

triangulation with special reference to deer movement. J. Wildl. Manage.

31:124-141.

Hood, R. E. 1971. Seasonal variation in home range, diel movement and

activity patterns of white-tailed deer on the Rob and Bessie Welder Wildlife

Refuge. M.S. thesis, Texas A & M Univ., College Station.

Johnson, D. H. 1980. The comparison of usage and availability measurements

for evaluating resource preferences. Ecology 61:65-71.

66 R. C. Pais, W. C. McComb, and J. Phillips

Knotts, R. W. 1975. White-tailed deer movement and distribution about

surface mines in Preston County, West Virginia. M.S. thesis, West Virginia

Univ., Morgantown.

Marchinton, R. L. 1968. Telemetric study of white-tailed deer movement —

ecology and ethology in the Southeast. Ph.D. dissertation, Auburn

Univ., Auburn, Ala.

Marchinton, R. L., and D. H. Hirth. 1984. Behavior. Pages 129-168 in White-

tailed Deer Ecology and Management (L. K. Halls, editor). Stackpole

Books, Harrisburg, Pa.

Marshall, A. D., and R. W. Wittington. 1969. A telemetric study of deer home

ranges and behavior of deer during managed hunts. Proc. Annu. Conf.

Southeast. Assoc. Game and Fish Comm. 22:30-46.

Mooty, J. J., P. D. Karns, and T. K. Fuller. 1987. Habitat use and seasonal

range size of white-tailed deer in north central Minnesota. J. Wildl.

Manage. 51:644-648.

Nams, V. G., and S. Boutin. 1991. What is wrong with error polygons? J.

Wildl. Manage. 55:172-176.

Pais, R. C. 1987. Mortality, dispersal, and habitat use of resident and

translocated white-tailed deer does on the Cumberland Plateau of eastern

Kentucky. M.S. thesis, Univ. of Kentucky, Lexington.

Phillips, J. H. 1983. A status report on Kentucky's deer herd. Kentucky Happy

Hunting Grounds 39:2-6.

Phillips, J. H. 1985. A deer population model for microcomputers. Proc.

Annu. Conf. Southeast. Assoc. Fish Wildl. Agencies 39:365-372.

Swihart, R. K., and M. A. Slade. 1985. Influence of sampling interval on

estimates of home range size. J. Wildl. Manage. 49:1019-1025.

Whitehead, C. J. 1967. Catoosa wildlife management research. Pages 17-18 in

Big Game Surveys, Tennessee. Tenn. Wildl. Resour. Agency Rep. No.

W-35-R-7-1, Nashville.

Accepted May 1991

Cation Concentrations and Acidity in Breeding Ponds

of the Spotted Salamander, Ambystoma maculatum

(Shaw) (Amphibia: Ambystomatidae), in Virginia

Charles R. Blem and Leann B. Blem

Department of Biology, Virginia Commonwealth University,

Richmond, Virginia 23284-2012

ABSTRACT. — For 8 years beginning in 1983, we monitored breeding

activity by spotted salamanders, Ambystoma maculatum, in temporary

ponds in eastern Virginia. In the spring of 1988, 1989, and 1990, a

majority of the ponds examined (67.4%; N = 218) never contained egg

masses or spermatophores. During the 8 years the number of egg

masses declined severely in many ponds that were used as breeding

sites. Ponds with breeding salamanders had significantly higher pH

values than those that lacked breeding activity for all years, but were

similar in pH to those with failed reproduction. Analysis of the 20

major cations indicated that successful ponds with large numbers of

spotted salamander egg masses had lower aluminum, copper, and lead

levels than ponds with declining populations of salamanders. Silicon

levels were significantly higher in successful ponds. Stepwise dis-

criminant function analyses indicate that high aluminum, copper,

silicon, and zinc concentrations in breeding ponds are associated

significantly with the decline in reproductive activity of spotted

salamanders. The combination of elements in ponds maintaining stable

or increasing populations of breeding spotted salamanders was

distinctive; 91.8% were classified correctly as successful or failed ponds

by canonical correlation analyses. Proximity of roads was not

correlated with the concentration of any cation.

In general, it appears that low pH may produce mortality in

amphibian eggs, larvae, and perhaps adults both by acid toxicity and by

causing an increase in the concentration of free ions of toxic elements

such as aluminum in the water column (Freda and Dunson 1985a).

Acidity interferes with the ability of larvae to regulate internal concen-

trations of sodium, chloride, and perhaps other ions (Freda and Dunson

1985b). Additionally, low pH levels affect composition of the perivitelline

fluid and the color and texture of the egg mass (Robb and Toews 1987).

At extremely low pH (e.g. < 4.0) the egg capsule (perivitelline space)

may shrink dramatically, killing the embryos (Pough 1976, Freda and

Dunson 1985a, Blem and Blem 1989). In the spotted salamander,

Ambystoma maculatum (Shaw), 50% or more of embryos die at pH 5.0

to 6.0 (Tome and Pough 1982) and all die at pH 4.0 or less (Pough and

Wilson 1977, Cook 1983). However, in some ponds with relatively low

Brimleyana 17:67-76, December 1991 67

68 Charles R. Blem and Leann B. Blem

pH, substantial numbers of spotted salamanders may breed successfully

(e.g. Cook 1983), suggesting that simple acidification is not the entire

cause of mortality. Furthermore, high concentrations of organic mate-

rials in such ponds may ameliorate the impact of toxic ions by binding

with them (Seip et al. 1984).

The present study documents concentrations of 20 of the more

common cations present in the waters of temporary ponds used as

breeding sites by spotted salamanders in Virginia and analyzes the

relationships of water chemistry to reproductive success of the spotted

salamander in these ponds.

MATERIALS AND METHODS

We counted egg masses in 218 temporary ponds in a 16-county area

around Richmond, Va., in the spring (March-May) of 1988, 1989, and

1990. At least two counts per year were made in each pond, at least one

of which was made well after all breeding activity ceased (usually late

March). We have monitored pH and the number of egg masses in a

majority of these ponds for 8 years (see Blem and Blem 1989). In

addition, we measured the pH of each pond two to four times each year.

In March 1989 we obtained a 20-ml water sample from each of 48

temporary ponds that had contained egg masses over most of the 8

years of the study. The surface area, depth, and pH of each pond were

measured at the same time. All 48 ponds were similar in altitude,

exposure, and surrounding vegetation. All were in forests consisting of

loblolly pine (Pinus taeda) or mixed pine-hardwoods. The two most

widely separated ponds were 64 km apart. The ponds seldom froze

during late spring and only ponds of sufficient depth to protect eggs

from freezing (i. e. > 20 cm deep) were studied. Egg masses of spotted

salamanders were counted in each pond, and pH was measured with an

Orion SA250 portable pH meter. The water samples were filtered

through No. 4 filter paper and analyzed for tannic acid using the

HACH (1975) test. Samples of filtered water were stored in cleaned 35-

ml borosilicate EPA water analysis vials, and were analyzed by the

Chemical Analysis Laboratory at the Institute of Ecology, The Univer-

sity of Georgia, by means of inductively coupled argon plasma analysis.

Twenty cations were quantified (Table 1). Replicate samples were run

from several ponds to test for sampling variation; no determination

differed by more than 8%. All data sets were tested to determine if they

were normally distributed (SAS 1985: UNIVARIATE), and those deviat-

ing from normality were transformed to natural logarithms before

further analyses were performed (Zar 1984).

We categorized ponds as "successful" (> 20 egg masses were

present and produced larvae in all 8 years and the number did not

Breeding Ponds of the Spotted Salamander 69

Table 1. Chemistry of water samples taken from spotted salamander breeding

ponds in March, 1989.a

Variable Successful ponds Failed ponds All ponds

(N = 25) (N = 23) (N = 48)

a All ion concentrations represent total values (soluble + bound). All values for

elements are means + SE ppm; cobalt and molybdenum concentrations were

below the level of detection.

Failed ponds significantly different from successful ones, MANOVA tests of

logarithmically transformed data, P < 0.05.

decline by more than 10% during the study) or "failed" (the number of

egg masses deposited has declined by 80% or more and/ or all egg

masses failed to produce viable larvae throughout the study). We

compared the physical characteristics and the levels of each element in

successful ponds with those in failed ponds by means of multiple

analysis of variance (SAS Institute 1985, MANOVA). Canonical correla-

tion analyses (SAS Institute 1985, CANCORR) were used to determine

whether successful ponds were different in overall element composition

from those that failed. Stepwise discriminant function analyses (SAS

Institute 1985, STEPDISC) were used to identify the elements that were

correlated with success or failure of ponds. A 5% significance level (P<

0.05) was used in all tests, unless noted otherwise.

70

Charles R. Blem and Leann B. Blem

-o

c

a

Q.

C 00

_o o

OS 13

P

U

rsi rt r- o sc

o — — <~^ —

— CM O O ""> —

O O — — (N M

r- -^r o — *""> © r»

p — — : p w> ■* —

<n (N cn) <-n| — — ^r-j

" — o »/"> — • sO >© '?*"> oo

't "t <n ^; o n 30 ,^ n

— Tf Tf m O V) 30 fS 0C (N

_ — — r, rf f, o f' O

m rj- O — ""> r- m Wi sO rt vO

_ 0 r^J — r*] — ; f"5 C ; C I — ; —

f; a f) '^o rs in o^ t fN x 'm xo

O ri ci Tt ^t - o O (N <n ^. o

roO<^ — oOO^; — (N O f^ -

»^O^vOOS3Cr",^w->sOC! — (N h m

o o © p © — o © — — : p p © ©

(N-btx^^nttr-1-.^o^'oo'^

o — -<3- © © r- <n v> © — © — °°. ** ""J

S «^ r3 3J tt M M (N

i/-> <^, m m — (+1 c^t c*-> — r^4 © © "~| ^ ©

^t (N o ~ <N — OO — CMpOOPP^P

ovoO(NOr*">r--r r*"> r- *^ oo £? so © — ^ ^ oo

o — o — oo^p — P — <^ © r-> o — P P

fflmuuuULL^22zza.a.^75N a.

Breeding Ponds of the Spotted Salamander

71

RESULTS

Only 71 of 218 potential breeding ponds that we checked in 1988,

1989, and 1990 contained egg masses. The incidence of egg laying by

spotted salamanders in those ponds declined between 1988 and 1990 by

23.9% (649 egg masses in 1988, 494 masses in 1989, 450 masses in 1990).

By 1990, nearly 75% of all egg masses were in only 20 of the original 71

study ponds. The pH of the 71 ponds was 5.64 ± 0.07 SE in 1988, 5.30

± 0.08 in 1989, and 5.54 ± 0.08 in 1990. The pH of ponds with no

evidence of salamander breeding activity during the 3-year period was

significantly lower in all 3 years (5.32 ± 0.07, 5.01 ± 0.06, and 5.12 ±

0.08, respectively). Acidity varied from year to year, depending upon the

amount and possibly the pH of rainfall; acidity increased as ponds dried

up in late spring.

In the 48 ponds in which we studied water chemistry (25 successful,

23 failed), no significant difference was detected in the pH of successful

versus failed ponds (Table 1). Of the 20 elements measured, 18 were

found at detectable levels. Cobalt and molybdenum were absent or were

at concentrations below detectable levels. Aluminum, copper, and lead

were at significantly higher concentrations in failed ponds than in

successful ones (Table 1). Silicon concentrations were significantly

higher in successful ponds. There was no statistical difference in pond

depth or in pond surface area between successful and failed ponds

(Table 1). Because some ponds were near roads, we were concerned

about the effects of vehicular emissions and of chemical treatment of

highways on element levels. Comparison by /-tests of the 16 ponds near

roads (100 m or less) with 32 remote ponds did not reveal any

significant difference in pH or in concentration of an element.

eg

c

03

O

-4

-3

-2

-1 0

Can1

Fig. I. Canonical correlation analyses of the effect of cations in successful

breeding ponds (closed circles) and failed ponds (open circles).

72 Charles R. Blem and Leann B. Blem

Pearson correlation coefficients (Table 2) indicate that only calcium

concentration was correlated significantly with pH (P < 0.001;. A

number of other elements were correlated positively with each other and

many of these were metals. For example, zinc and copper, calcium and

magnesium, strontium and magnesium, and strontium and calcium have

correlation coefficients > 0.7. Some of these elements apparently

influence one another's abundance. Canonical correlation analyses (Fig.

1) indicate that successful ponds are distinctive; 91.8% of the ponds

were identified correctly by the analyses. Major loading was on

aluminum and sodium (first canonical correlation) and copper and

silicon (second canonical correlation). Aluminum, copper, silicon, and

zinc are the only significant cations in stepwise discriminant function

analyses of successful and failed ponds (Table 3).

DISCUSSION

In the 1970s and early 1980s, the spotted salamander was abundant

locally in the lower piedmont and coastal plain of Virginia, a zone of

relatively high acid deposition from precipitation (Schwartz 1989).

Recently we observed severe declines in the number of egg masses of the

spotted salamander at some sites (Blem and Blem 1989). For example,

one site had more than 500 egg masses in 1984, but the number declined

annually, and in 1990 only 15 masses were deposited. At the same time

the pH of this pond decreased from > 5.0 to < 4.0. During the same

period we detected a general decrease in survivorship of spotted

salamander eggs and embryos throughout eastern Virginia (Blem and

Blem 1989). In general, there has been a trend for only a few ponds with

pH values between 6.0 and 7.0 to support salamanders. However, the

increased mortality has not been uniform. Some acidic ponds continue

to support substantial numbers of egg masses, which have not declined

during that period.

A correlation between decreased pH of temporary ponds and

decreased survival of spotted salamander eggs and larvae has been

recognized for more than a decade (e.g. Pough 1976). Mortality seems

to increase at a pH below 6.0, although substantial proportions of eggs

survive to hatch at pH values from 4.0 to 6.0 (Pough 1976, Pough and

Wilson 1977, Saber and Dunson 1978, Ling et al. 1986, Blem and Blem

1989). Complete lethality to embryos appears to be in the range of 3.5

to 4.5 (Pierce 1985, Freda 1986, Ling et al. 1986, Robb and Toews

1987), although this depends upon levels of certain elements in the

breeding pond (Pough 1981, Dale et al. 1985, Freda and Dunson 1985a,

1986). Recently it was found that mortality may be associated with

concentrations of specific elements such as aluminum (Albers and

Prouty 1987, Clark and Hall 1985, Clark and LaZerte 1985, Freda and

Breeding Ponds of the Spotted Salamander 73

Table 3. Stepwise discriminant function analyses of the importance of elements

in the use of breeding ponds by spotted salamanders.

Element Partial R2 Wilks' Lambda3

Aluminum 0.209 0.808

Copper 0.232 0.650

Silicon 0.240 0.589

Zinc 0.161 0.494

aElements and associated statistics are listed in order of entry in the analyses.

All Wilks' Lambda values are significant at P< 0.05.

Dunson 1985a), calcium (Freda and Dunson 1985a), magnesium (Freda

and Dunson 1985a), and sodium (Freda and Dunson 1984, 1985a).

However, there seem to be few data regarding the ambient levels of

cations in temporary ponds, and little information regarding the

combined effect of a more complex suite of ions (see Gascon and Planas

1986).

The chemistry of temporary waters on the lower piedmont and

coastal plain of Virginia is complex. The chemical composition of the

water in these ponds, along with its physical characteristics and local

biotic influences, may all vary between ponds even when they are

separated by only short distances. Year-to-year variations in water

depth, temperature, and duration are important determinants of

reproductive success and continued use of ponds by spotted salamanders

(see Shoop 1974, Albers and Prouty 1987). The surrounding forest and

soil are important influences on levels of biologically significant cations

(James and Riha 1986). Coastal plain and lower piedmont soils of

Virginia vary widely in composition (Blem and Blem 1989), and it is not

surprising that elements detected in temporary ponds are also extremely

variable. For example, several of our test ponds had aluminum

concentrations of 2-3 ppm; 0.3 ppm is considered extremely high in

Virginia waters (C. Lunsford, personal communication). Furthermore,

the relationship between pH and aluminum toxicity appears to be

complex (Freda et al. 1989). For example, organically bound aluminum

may be harmless (Freda 1986) or less harmful (Seip et al. 1984) than

free aluminum ions; therefore ponds with large amounts of humic acids

could have high concentrations of aluminum without harm to aquatic

life. Also, aluminum may ameliorate acid toxicity at intermediate

concentrations and low pH values (Freda et al. 1989).

Only traces of copper were found in many ponds, but others had

levels as high as 0.179 ppm. Copper is known to be toxic to amphibians

74 Charles R. Blem and Leann B. Blem

(National Academy of Sciences 1974), although the degree of toxicity

has been determined for only a few species. Likewise, lead levels of

some ponds were in the range of concentrations known to be detrimental

(0.75-20 /xg/ liter; USEPA 1980a). On the other hand, zinc levels of only

79 ppm were obtained and levels of 180-540 jug/ liter are considered to

be detrimental to aquatic organisms (USEPA 1980b). The relationship

between silicon levels and successful use of breeding ponds is puzzling.

Silicon concentrations in successful ponds were significantly higher than

in those that failed (Table 1). We do not know of any relationship

between silicon and viability of aquatic life.

Some elements have complex effects on success of hatching of

amphibian eggs, e.g. a few elements appear to be harmful at both high

and low concentrations. For example, at a pH of 4.25, hatching and

survival of embryos of the Jefferson salamander, Ambystoma jeffer-

sonianum (Green), are greatest at magnesium concentrations of about

20 mg/ liter, but the hatching rate declines at both lower and higher

concentrations (Freda and Dunson 1985a). The effects of most other

elements increase with decreased pH, but the degree of synergism with

other elements is unknown. It appears that failed ponds in Virginia have

distinctive cation compostions (Fig. 1). Stepwise discriminant function

analyses suggest that aluminum, copper, silicon, and zinc all may play a

role in mortality of spotted salamander larvae. Failed and successful

ponds did not differ in pH. Rather, it appears that at the present levels

of acidity, ponds become fatal because of their elemental composition.

The few surveys of the chemistry of temporary ponds used for

breeding by amphibians have considered either restricted suites of

chemical elements or have included relatively few ponds (e.g. Dale et al.

1985, Freda and Dunson 1986, Albers and Prouty 1987). Conclusions

about the importance of specific chemicals in temporary ponds have

been mixed. For example, Freda and Dunson (1986) concluded that

mortality may be influenced by significant interactions of pond pH and

other chemical variables, whereas Albers and Prouty (1987) found that

pond longevity, water temperature, and oxygen content were more

significant in spotted salamander reproduction than chemical alterations

brought about by acid precipitation. Those findings are not incom-

patible, given the range of study areas, times, and substances considered.

ACKNOWLEDGMENTS.- We are indebted to the Nongame

Wildlife and Endangered Species Program, Virginia Department of

Game and Inland Fisheries, for financial support of the research that

led to this publication. Karen Blem helped with some technical aspects

of the research, Charles Lunsford provided important information

about water parameters, and Kathy Beal made many helpful suggestions

about statistical methods.

Breeding Ponds of the Spotted Salamander 75

LITERATURE CITED

Albers, P. H., and R. M. Prouty. 1987. Survival of spotted salamander eggs in

temporary woodland ponds of coastal Maryland. Environ. Pollution

46:45-61.

Blem, C. R., and L. B. Blem. 1989. Tolerance of acidity in a Virginia

population of the spotted salamander, Ambystoma maculatum (Amphibia:

Ambystomatidae). Brimleyana 15:37-45.

Clark, K. L., and R. J. Hall. 1985. Effects of elevated hydrogen ion and

aluminum concentrations on the survival of amphibian embryos and larvae.

Can. J. Zool. 63:116-123.

Clark, K. L., and B. D. LaZerte. 1985. A laboratory study of the effects of

aluminum and pH on amphibian eggs and tadpoles. Can. J. Fish. Aquat.

Sci. 42:1544-1551.

Cook, R. P. 1983. Effects of acid precipitation on embryonic mortality of

Ambystoma salamanders in the Connecticut Valley of Massachusetts. Biol.

Conserv. 27:77-88.

Dale, J. M., B. Freedman, and J. Kerekes. 1985. Acidity and associated water

chemistry of amphibian habitats in Nova Scotia. Can. J. Zool. 63:97-105.

Freda, J. 1986. The influence of acidic pond water on amphibians: a review.

Water, Air, Soil Pollut. 30:439-450.

Freda, J., V. Cavdek, and D. G. McDonald. 1989. Role of organic

complexation in the toxicity of aluminum to Rana pipiens embryos and

Bufo americanus tadpoles. Can. J. Fish. Aquat. Sci. 47:217-224.

Freda, J., and W. A. Dunson. 1984. Sodium balance of amphibian larvae

exposed to low environmental pH. Physiol. Zool. 57:435-443.

Freda, J., and W. A. Dunson. 1985a. The influence of external cation

concentration on the hatching of amphibian embryos in water of low pH.

Can. J. Zool. 63:2649-2656.

Freda, J., and W. A. Dunson. 1985b. Field and laboratory studies of ion

balance and growth rates of ranid tadpoles chronically exposed to low pH.

Copeia 1985:415-423.

Freda, J., and W. A. Dunson. 1986. Effect of low pH and other chemical

variables on the local distribution of amphibians. Copeia 1986:454-466.

Gascon, C, and D. Planas. 1986. Spring pond water chemistry and the

reproduction of the wood frog, Rana sylvatica. Can. J. Zool. 64:543-550.

HACH. 1975. HACH Water and Wastewater Analysis Procedures Manual.

3rd ed. HACH Chemical Co., Ames, la.

James, B. R., and S. J. Riha. 1986. pH buffering in forest soil organic

horizons: relevance to acid precipitation. J. Environ. Qual. 15:229-234.

Ling, R. W., J. P. VanAmberg, and J. K. Werner. 1986. Pond acidity and its

relationship to larval development of Ambystoma maculatum and Rana

sylvatica in upper Michigan. J. Herpetol. 20:230-236.

National Academy of Sciences. 1974. Amphibians: guidelines for the breeding,

care, and management of laboratory animals. Washington, D.C.

Pierce, B. A. 1985. Acid tolerance in amphibians. BioScience 35:239-243.

Pough, F. H. 1976. Acid precipitation and embryonic mortality of spotted

salamanders, Ambystoma maculatum. Science 192:68-70.

c

76 Charles R. Blem and Leann B. Blem

Pough, F. H. 1981. Mechanisms by which acid precipitation produces

embryonic death in aquatic vertebrates. U.S. Dept. Interior Office Res.

Tech., Res. Proj. Tech. Compl. Rep., 1-10.

Pough, F. H., and R. E. Wilson. 1977. Acid precipitation and reproductive

success of A mby stoma salamanders. Water, Air, Soil Pollut. 7:307-316.

Robb, L., and D. Toews. 1987. Effects of low ambient pH on perivitelline fluid

of Ambystoma maculatum (Shaw) eggs. Environ. Pollut. 44:101-107.

Saber, P. A., and W. A. Dunson. 1978. Toxicity of bog water to embryonic

and larval anuran amphibians. J. Exp. Biol. 204:33-42.

SAS Institute, Inc. 1985. SAS User's Guide: Statistics. Version 5 edition. Cary,

N.C.

Schwartz, S. E. 1989. Acid deposition: unraveling a regional phenomenon.

Science 243:753-762.

Seip, H. M., L. Muller, and A. Naas. 1984. Aluminum speciation: comparison

of two spectrophotometric analytical methods and observed concentrations

in some acidic aquatic systems in southern Norway. Water Air Soil Pollut.

23:81-95.

Shoop, R. C. 1974. Yearly variation in larval survival of Ambystoma

maculatum. Ecology 55:440-444.

Tome, M. A., and F. H. Pough. 1982. Responses of amphibians to acid

precipitation. Acid rain/ fisheries. Pages 245-254 in Proceedings of an

International Symposium on Acid Precipitation Fishery Impacts in

Northeastern North America. Am. Fish. Soc, Bethesda, Md.

U.S. Environmental Protection Agency. 1980a. Ambient Water Quality

Criteria for Lead. EPA 440/5-80-057. Natl. Tech. Inf: Serv., Springfield,

Va.

U.S. Environmental Protection Agency. 1980b. Ambient Water Quality

Criteria for Zinc. EPA 440/5-80-079. Natl. Tech. Inf. Serv., Springfield,

Va.

Zar, J. H. 1984. Biostatistical Analysis. Prentice-Hall, Inc. Englewood Cliffs,

N.J.

Accepted June 1991

Spawning Activities of Notropis chlorocephalus,

Notropis chiliticus, and Hybopsis hypsinotus,

Nest Associates of Nocomis leptocephalus

in the Southeastern United States,

With Comments on Nest Association

(Cypriniformes: Cyprinidae)

Carol E. Johnston

Center for Biodiversity

Illinois Natural History Survey

607 E. Pea body Drive

Champaign, Illinois 61820

ABSTRACT. — Spawning activity of Notropis chlorocephalus was

observed in 1987 and 1988 and of Notropis chiliticus and Hybopsis

hypsinotus in 1988. Descriptions of spawning have not been published

previously for these three species. Aggregations of males of all three

species were observed over the nests of Nocomis leptocephalus. A

female N. chlorocephalus or N. chiliticus typically initiated spawning

by moving over the nest, where she was pursued by one to several

males. Spawning then occurred in N. chlorocephalus and N. chiliticus

when a female moved or was driven by a male (or males) to the

substrate; the pair vibrated and presumably released gametes. Spawning

often occurred in small pits (one to several) on the tops of nests; these

pits were constructed by male Nocomis for their own spawning. The

spawning act was not observed for H. hypsinotus; however, numerous

tuberculate males were observed holding territories over nests, and ripe

females were collected from the area immediately downstream of the

nests. Some authors have suggested that associates are attracted to the

nests of other species because of a lack of suitable spawning substrate

elsewhere in the stream. Ten artificial nests were constructed in streams

with active Nocomis nests and monitored for the presence of nest

associates. No fish were seen over the artificial nests (41 observations),

which indicates that the associate species are benefited by other factors

specific to the Nocomis nests, such as the parental care of the nest-

building male or the selfish-herd effect of numerous fishes spawning in

one place.

Little published information on life histories exists for Notropis

chlorocephalus (Cope), Notropis chiliticus (Cope), or Hybopsis

hypsinotus (Cope). Notropis chlorocephalus and N. chiliticus, both

species of the subgenus Hydrophlox, occur in small, clear streams where

they are often abundant. The range of N. chiliticus includes the Peedee

drainage and the Dan River (Roanoke drainage) of North Carolina and

Brimleyana 17:77-88, December 1991 77

78 Carol E. Johnston

Virginia (Gilbert and Burgess 1980a). Notropis chlorocephalus is endemic

to the Catawba River (Santee drainage) of North and South Carolina

(Gilbert and Burgess 1980b). Hybopsis hypsinotus inhabits small to

medium-sized rivers in the Peedee and Santee river drainages of North

and South Carolina and Virginia and is not abundant (Jenkins and

Lachner 1980).

During a study of nest association among North American minnows,

1 observed spawning in N. chlorocephalus and N. chiliticus. Spawning

activities, but not the spawning act, were observed in H. hypsinotus.

Spawning behavior has not been described previously for these three

species, all of which spawn in association with Nocomis leptocephalus

(Girard).

Nest association, the habit of spawning over the nest of another

species, is widespread among minnows of the eastern United States

(Table 1). The strategy is less common among other North American

fishes, and has been described for only three species, including longnose

gar, Lepisosteus osseus (L.) (Goff 1984); lake chubsucker, Erimyzon

sucetta (Lacepede) (Carr 1942); and creek chubsucker, Erimyzon

oblongus (Mitchill) (Page and Johnston 1989). Ten species of minnows

are known to spawn in the nests of sunfishes or basses (Latta 1958,

Kramer and Smith 1960, Hunter and Wisby 1961, Snelson 1972, Chew

1974, Pflieger 1975); the other minnows that spawn as nest associates

spawn over nests constructed by other minnows, particularly species of

Nocomis. The nests and parental care of sunfishes and basses differ

from the nests and parental care of nest-building minnows. The nests of

sunfishes and basses are saucer-shaped depressions built in a variety of

substrates. Eggs laid in the nests of sunfishes or basses receive direct

parental care from hosts, which remain with the young until hatching,

in the form of fanning of the eggs and predator defense. The nests of

minnows range from simple pits where the eggs are not covered (e.g.

Campostoma and Luxilus) to large gravel mounds where eggs are

covered by the male after spawning (e.g. Nocomis). The parental care

offered by minnow hosts includes predator defense, but minnow hosts

may remain over nests only as long as spawning opportunities exist.

Although nest association has been known for many years, few in-

depth studies have been conducted, and most of the reports of nest

association have been little more than accounts of the species involved.

The occurrence of nest association raises important ecological,

evolutionary, and behavioral questions. Are associates attracted to nests

alone, which may provide the best available spawning substrate in the

habitat, or are they attracted to some other factor associated with nests,

such as the parental care offered by the host? What are the costs and

benefits of nest association to the associates and to the hosts? Is the

Nest Associates of Nocomis leptocephalus 79

interaction mutualistic, parasitic, or commensal? How did this behavior

evolve?

Several authors have suggested that nest association occurs because

of the absence of suitable spawning substrate elsewhere in the stream or

pond (Latta 1958, McAuliffe and Bennett 1981, Starnes and Starnes

1981.) If that is the case, associates such as the species described in this

paper would be expected to be attracted to and to use suitable nests in

the absence of the host.

The experiment reported here was conducted with spawning

associations of N. leptocephalus in North Carolina to determine whether

nest association is a response to the clean mounds of gravel built by

Nocomis or a response to other factors associated with nests. Gravel

nests, similar to those of Nocomis, were constructed in areas of suitable

depth and flow and checked for use by spawning fishes. The results of

this experiment are discussed, and spawning activities of the three nest

associates (N. ehloroeephalus, N. chiliticus, and H. hypsinotus) are

described.

MATERIALS AND METHODS

Spawning associations of Notropis ehloroeephalus, Luxilus

coccogenis (Cope), Clinostomus funduloides Girard, Campostoma

anomalum (Rafinesque), and Nocomis leptocephalus were observed in

two streams in the Catawba River, McDowell Co., N.C., on 26 and 27

May 1987, from 10 to 13 May 1988, and from 24 to 27 May 1988.

Spawning aggregations of Notropis chiliticus and Hybopsis hypsinotus

were observed over N. leptocephalus nests in the Fisher River, Surry

Co., N.C., on 26 May 1988.

Observations of spawning fishes were made by snorkeling or from

above the water by using binoculars or polarized sunglasses. Height

(distance from substrate to nest top), width (perpendicular to stream

flow), and length (parallel to stream flow) of active and artificial nests

were measured, along with water depth and surface flow (cm/ sec).

Surface flow was measured by timing the progress of an object over a

known distance.

To simulate Nocomis nests, clean gravel of the sizes used in the

nests was collected from the stream bed, rinsed, and piled in mounds

similar in shape and size to nests in areas similar in depth and flow to

those where active Nocomis nests occurred. In some cases artificial nests

were built in the previous locations of Nocomis nests.

RESULTS

Spawning behaviors in Notropis ehloroeephalus and N. chiliticus

were very similar. Spawning occurred at water temperatures of 11-

80

Carol E. Johnston

Table 1. Species of minnows reported to act as nest associates of nest-building

minnows, sunfish, or bass.

Associate

Host(s)

Reference(s)

Campostoma anomalunf Nocomis, Semotilus

Clinostomus funduloides

Clinostomus elongatus

Cyprinella chloristia

Cyprinella pyrrhomelas

Cyprinella lutrensis"

Hy bops is hyps i not us

Hybopsis labrosus

Hybopsis zanemus

Luxilus cerasinus

Luxilus chrysocephalusa

Luxilus coccogenis

Luxilus cornutus"

Luxilus pilsbryf

Luxilus zonatuf

Luxilus zonistius

Lythrurus ardens

Lythrurus bellus

Lythrurus umbratilus

Nocomis

Nocomus, Luxilus,

Semotilus

Nocomis

Lepomis

Nocomis

Nocomis, Campostoma,

Semotilus, Exoglossum

Nocomis

Nocomis, Campostoma,

Semotilus, Exoglossum,

Micropterus

Nocomis, Campostoma

Nocomis, Lepomis

Lepomis

Notemigonus crysoleucas" Lepomis, Micropterus

Notropis amoenus

Notropis bailey i

Nocomis

Nocomis, Campostoma

Notropis chlorocephalus Nocomis

Notropis chiliticus

Notropis cummingsae

Notropis leuciodus

Notropis lutipinnis

Notropis maculalus

Nocomis

Lepomis

Nocomis

Micropterus

(Cope 1868, Reighard 1943,

Raney 1947)

(Cope 1868, Lachner 1952)

(Greeley 1938, Koster 1939)

(Shute, pers. comm.)

(Pflieger 1975)

(Johnston, this publication)

(Shute, pers. comm.)

(Cope 1868, Raney 1947)

(Raney 1940a,b,

Hankinson 1932)

(Outten 1957)

(Adams and Hankinson

1928, Hankinson 1932,

Raney 1940a, b,c, Miller

1963, 1964)

(Pflieger 1975)

(Johnston and Birkhead

1988)

(Cope 1868, Raney 1947,

Yokley 1974)

(Snelson 1972)

(Hunter and Wisby 1961,

Hunter and Hastier 1965)

(Latta 1958, Kramer and

Smith 1960, DeMont

1982)

(Looset al. 1979)

(Bart, pers. comm.;

Folkerts, pers. comm.)

(Johnston, this publication)

(Fletcher 1990)

(Lachner, pers. comm.)

(McAuliffe and Bennett

1981)

(Chew 1974)

Nest Associates of Nocomis leptocephalus

Table 1. Continued.

81

Associate

Host(s)

Reference(s)

Notropis nubilus

No tr op is procne"

Notropis rubellus

Luxilus

Lepomis

Nocomis, Campostoma,

Semotilus, Exoglossum

Notropis rubricroceus Nocomis

Notropis tristis Lepomis

Phenacobius crassilabrum Nocomis

Phoxinus Campostoma

cumberlandensis

Phoxinus erythrogaster Nocomis, Campostoma

Phoxinus ore as

Rhinichthvs atratulus"

Nocomis, Campostoma

Nocomis

(Pflieger 1975, Fowler et al.

1984)

(Looset al. 1979)

(Adams and Hankinson

1928, Hankinson 1932,

Raney 1940a,b, Reed

1958, Miller 1963, 1964,

Pfeiffer 1955)

(Outten 1958)

(Pflieger 1975)

(Johnston, pers. obs.)

(Starnes and Starnes 1981)

(Smith 1908, Raney 1969,

Settles and Hoyt 1978)

(Cope 1868, Raney 1947)

(Cope 1868, Raney 1969)

^Reported to use spawning strategies in addition to nest association.

17°C. In both species large numbers of males (30-100) held positions in

a school directly over nests or towards the back (downstream side),

although a school of N. chiliticus was observed over the front edge of a

nest. It is possible that the positions depended on the other species

present. Individual males were aggressive and defended moving territories

(the space surrounding an individual male no matter what his position),

as opposed to stationary territories (a specific location), because it was

impossible for any one male to hold a position for more than a few

seconds. Positions within small pits on the tops of the nests seemed to

be preferred areas, and the most vigorous competition among males

(hosts and associates) was for temporary positions within pits, because

most spawning occurred in them. These small pits were constructed by

male Nocomis for their own spawning. Usually more than one pit was

present on a given nest because more than one male N. leptocephalus

worked on a nest at the same time. The schools of males were always

close to the substrate. Females remained in schools behind the nests

until ready to spawn. An individual female that moved onto a nest was

immediately pursued by one or more males until she was driven to the

substrate, often into the small pits on the nests, where one or more

82 Carol E. Johnston

males aligned by her side and vibrated; spawning then occurred. The

same female would often spawn in the same place several times in

succession. Spawning acts always attracted a swarm of other males to

the area.

During peak spawning the entire body of a male N. chlorocephalus

was scarlet red, and all fins, especially the pectorals, were milky white.

When spawning activity was not as intense, males lost the scarlet color;

the body was then orange-red, and a lateral stripe was evident. Females

were not as brightly colored as males, and the lateral stripe was always

evident.

Male N. chiliticus in peak spawning coloration had scarlet red

bodies, eyes, and lips. The head and fins were gold, and there was a gold

lateral stripe. Females had white bellies, silver bodies, red lips, and fins

that were orange at the base and yellow at the tips. Both sexes had dark

crescents of pigment on the sides of the body.

The spawning act was not observed in H. hypsinotus, but tuber-

culate males were observed holding territories above the nests of N.

leptocephalus at the same time the spawning observations were made

for N. chiliticus. From 6 to 12 males were observed holding territories

10 to 15 cm above nests. Ripe females were collected downstream of

nests. The position of H. hypsinotus over the nests is similar to that of

C. funduloides, which was observed in spawning associations with N.

chlorocephalus.

Male H. hypsinotus had metallic blue bodies with dark lateral

stripes and orange fins that were yellow at the base. Females were not

brightly colored.

Eggs of different sizes were found in various stages of development

throughout Nocomis nests. More eggs appeared to be present in the

front (upstream side) and middle of nest tops, but samples were not

quantitative.

Notropis chlorocephalus was observed over N. leptocephalus nests

from 10 to 26 May 1988. Luxilus coccogenis, Clinostomus funduloides,

Campostoma anomalum, and N. leptocephalus were also present over

these nests. Spawning fishes were not observed elsewhere in streams.

Thirteen nests were present in 0.4-km sections of streams at two

localities in the Catawba River drainage. Eight of these nests were active

(i.e. had associations of spawning fishes over them), and at least two

were active for the entire observation period (10 to 26 May). Notropis

chiliticus, H. hypsinotus, and N. leptocephalus associations were

observed only on 26 May 1988 in the Fisher River. Five nests were

present in a 0.4-km stream section, and two of these were active.

Male N. leptocephalus were observed spawning and adding gravel

to nests in 18 of 27 observations. (Observations were made daily from

Nest Associates of Nocomis leptocephalus 83

10 to 13 and from 24 to 27 May 1988 on each active nest at two sites in

the Catawba River drainage and on 26 May in the Fisher River.) In

61% of these observations (lasting from 15 minutes to 2 hours), more

than one male N. leptocephalus was present over the nest. In seven

observations, one male was over the nest; in six, two; in three, three;

and in two, four. Males were never observed actively driving nest

associates from the nest, but they did drive away such potential egg

predators as suckers.

Physical measurements were taken for 1 1 active nests. The mean

length of the nests was 69.0 cm (SD = 21.3); mean width was 71.5 cm

(SD = 16.4 cm); and mean height was 15.0 cm (SD = 4.5 cm). Mean

surface flow was 22.9 cm/ sec. (SD =10.1 cm). Mean water depth was 42.5

cm (SD = 16.2 cm), and the mean distance from the nest top to the

water surface was 28.1 cm (SD = 16.3).

Ten artificial nests were constructed near active N. leptocephalus

nests in similar physical conditions. Each of the artificial nests was

checked over a period of 4 to 7 days, a cumulative total of 41 checks for

the 10 nests, and fish were never observed over them, except incidentally.

No eggs were found in these nests. On all days when artificial nests were

checked, spawning activity was occurring elsewhere in the stream.

DISCUSSION

Spawning behavior of N. chlorocephalus and N. chiliticus is similar

to that of Notropis (Hydrophlox) rubricroceus (Cope) (Outten 1958).

Spawning aggregations of Notropis (Hydrophlox) lutipinnis (Jordan

and Brayton) have been described by McAuliffe and Bennett (1981) and

are similar to those formed by other nest associates. Because virtually

nothing is known about the spawning behavior of Hybopsis (s. str.), the

suggestion that H. hypsinotus is a nest associate is particularly

significant. Reighard (1943) found more than one male Nocomis

micropogon (Cope) working on, or stationed over, nests in 47% of his

observations (n = 15). These extra males were never observed spawning,

and Reighard suspected that their contribution to nest building was

negligible. More than one male N. leptocephalus was observed over

nests in 61% of my observations, indicating that, at least in N.

leptocephalus, the typical situation may be that more than one male

may participate in nest construction and spawning. These observations

suggest species differences in social structure during the spawning season.

Further study may reveal differences in spawning behavior among

species of Nocomis.

Spawning aggregations of nest associates did not form over the

artificial nests built in this study, as would be expected if nest structure

or substrate was the most important factor involved in nest association.

84 Carol E. Johnston

Hunter and Hastier (1965) found the nest to be the least important

stimulus in attracting spawning redfin shiners, Lythrurus umbratilis

(Girard), and Kramer and Smith (1960) found that golden shiners,

Notemigonus crysoleucas (Mitchill), would spawn over nests of large-

mouth bass, Micropterus salmoides (Lacepede), built in a variety of

substrates. These observations suggest that something other than the

substrate of the nest causes nest associates to spawn over the nests of

other species, possibly the parental care of the host or benefits from

other factors.

A potential advantage to nest associates of spawning over the nest

of another species is increased egg survivorship with no costs of parental

care (nest building and defense). Increased egg survival can occur in

gravel-mound nests such as those built by Nocomis because of better

aeration and reduced siltation. Survival is also increased because male

Nocomis cover the eggs and defend the nest against predators. A

consequence of the nest-associate strategy, for both associate species

and nest builders, is hybridization. Hybrids between nest builders and

associate species have been described (Greely 1929, Raney 1940b,

Reighard 1943, Miller 1963). Hybridization can be viewed as a cost,

especially if hybrid progeny are inviable or sterile, or as a benefit if new

and adaptive gene combinations are introduced into the populations by

viable hybrid backcrosses (Futuyma 1986).

The advantage to nest builders, such as Nocomis, of having other

species use their nests is difficult to understand, because associates do

not contribute to nest building and most associate species are too small

to drive egg predators away. Evidence suggests that the larger species of

associates aid in driving egg predators from the nest (Hankinson 1932,

Vives 1988). It is possible that nest builders benefit from a selfish-herd

effect (Hamilton 1971), in which the eggs of the nest associate are as

likely to be eaten as those of the nest builders, or more likely to be

eaten. The result would be increased egg survivorship for the host.

Because so many nest associates usually spawn over a given nest, they

may also benefit from the selfish-herd effect; spawning adults may

benefit directly from this effect as well as through the survival of eggs.

In that case the relationship would be mutually beneficial, and the nest

builder would be expected to tolerate, or even encourage, nest associates.

If the reproductive success of the nest builder is lowered as a result of

the intrusion of nest associates, the relationship would be parasitic, and

one would expect the nest builder to actively discourage nest associates

from spawning over his nest. It is possible, however, that nest builders

are energetically incapable of driving nest associates away. If that is the

case, one might expect to find adaptations to prevent associates from

finding nests, temporal shifts in spawning, or other mechanisms to

Nest Associates of Nocomis leptocephalus 85

prevent association in at least some species of nest builders. The

relationship could also be commensal, with the reproductive success of

the nest builder remaining unaltered while that of the associates increases.

Carr (1942) observed eggs of lake chubsucker, Erimyzon sucetta, in

the nests of largemouth bass, Micropterus salmoides, and eggs of golden

shiner, Notemigonus crysoleucas, in the nests of spotted sunfish, Lepomis

punctatus (Valenciennes) (Carr 1946). She compared these relationships

to the parasitism that occurs between the cowbird and its hosts (Carr

1942). However, her data for the chubsucker/ largemouth bass relation-

ship suggest that instead of a parasitic relationship, in which the host's

reproductive success is reduced, a mutually beneficial relationship occurs

(McKaye 1981). Carr (1942) found 22 largemouth bass nests of which 5

had chubsucker eggs. Four of these 5 nests were successful (produced

fry), while only 1 of 14 nests without chubsucker eggs was successful

(Carr 1942, McKaye 1981). The reproductive success of the largemouth

bass was apparently higher in nests with chubsuckers eggs, a finding that

suggests a mutually beneficial relationship (assuming chubsucker eggs

laid outside nests would not survive). Whether the chubsuckers chose to

lay their eggs in largemouth bass nests that would have been the most

successful regardless of the presence of chubsucker eggs is unknown.

Goff (1984) discovered longnose gar, Lepisosteus osseus, eggs in 10

of 69 smallmouth bass, Micropterus dolomieu Lacepede, nests. Of the

nests with gar eggs, 60% were successful, whereas only 32% of the n^sts

without gar eggs were successful. Goff (1984) also showed that clutches

of gar eggs were the most successful when they were intermingled with

smallmouth bass eggs in the nest, and that eggs did not survive outside

nests. These observations suggest that both species benefit from this

relationship.

Results indicating that nests builders and associates both benefit

from the association support the hypothesis of McKaye and McKaye

(1977) in which natural selection is the driving force behind the evolution

of interspecific brood care in fishes, and fishes that care for unrelated

young have greater reproductive success. This hypothesis suggests that

relationships such as nest association are mutually beneficial.

My future work is aimed at understanding the nature of the

relationship between nest associates and their hosts. Is this a mutualistic,

parasitic, or commensal relationship? Answering that question should

lead to a better understanding of the evolution of the nest-associate

strategy.

ACKNOWLEDGMENTS.— This research was supported in part by

a grant-in-aid of research from Sigma Xi, The Scientific Research

Society. L. M. Page and P. A. Ceas assisted in the field. I thank C. A.

Mayer and P. E. Jarrell for critical review of the manuscript.

86 Carol E. Johnston

LITERATURE CITED

Adams, C. C, and T. L. Hankinson. 1928. The ecology and economics of

Oneida Lake Fish. Bull. N.Y. State Coll. For. l(4a). Roosevelt Wildl. Ann.

1(3-4): 235-548.

Carr, M. H. 1942. The breeding habits, embryology, and larval development of

the large-mouth black bass in Florida. Proc. New England Zool. Club.

20:43-77.

Carr, M. H. 1946. Notes on the breeding habits of the eastern stumpknocker,

Lepomis punctatus punctatus (Cuvier). J. Fla. Acad. Sci. 101-106.

Chew, R. L. 1974. Early life history of the Florida largemouth bass. Bull.

Florida Game and Freshwater Fish. Comm. No. 7.

Cope, E. D. 1868. On the distribution of freshwater fishes in the Allegheny

region of southwestern Virginia. J. Acad. Nat. Sci. Phila. 2:207-247.

DeMont, D. J. 1982. Use of Lepomis macrochirus Rafinesque nests by

spawning Notemigonus crysoleucas (Mitchill) (Pisces: Centrachidae and

Cyprinidae). Brimleyana 8:61-63.

Fletcher, Dean E. 1990. Spawning behavior of the dusky shiner, Notropis

cummingsae. Abstract, 70th annual meeting of the American Society of

Ichthyologists and Herpetologists, Charleston, S.C.

Fowler, J. F., P. W. James, and C. A. Taber. 1984. Spawning activity and eggs

of the Ozark minnow, Notropis nubilus. Copeia 1984:994-996.

Futuyma, D. J. 1986. Evolutionary Biology. 2nd ed. Sinauer Associates, Inc.,

Sunderland, Mass.

Gilbert, C. R., and G. H. Burgess. 1980a. Notropis chiliticus (Cope), redlip

shiner. Page 252 in Atlas of North American Freshwater Fishes (D. S. Lee

et al., editors). N.C. State Mus. Nat. Hist., Raleigh.

Gilbert, C. R., and G. H. Burgess. 1980b. Notropis chlorocephalus (Cope),

greenhead shiner. Page 254. in Atlas of North American Freshwater Fishes

(D. S. Lee et al., editors). N.C. State Mus. Nat. Hist., Raleigh.

Goff, G. P. 1984. Brood care of longnose gar {Lepisosteus osseus) by

smallmouth bass {Micropterus dolomieui). Copeia 1984:153-161.

Greeley, J. R. 1929. Fishes of the Erie-Niagara watershed. Pages 150-179 in A

Biological Survey of the Erie-Niagara System. Suppl. 18th Annu. Rep. N.Y.

Conserv. Dept., 1928.

Greeley, J. R. 1938. Fishes of the area with annotated list. Pages 48-73 in A

Biological Survey of the Allegheny and Chemung Watersheds. Suppl. 27th

Annu. Rep. N.Y. Conserv. Dept., 1937.

Hamilton, W. D. 1971. Geometry for the selfish herd. J. Theor. Biol.

31:295-311.

Hankinson, T. L. 1932. Observations on the breeding behavior and habitats of

fishes in southern Michigan. Pap. Mich. Acad. Sci., Arts and Letters

15:411-425.

Hunter, J. R., and A. D. Hastier. 1965. Spawning association of the redfin

shiner, Notropis umbratilis, and the green sunfish, Lepomis cyanellus.

Copeia 1965:265-281.

Nest Associates of Nocomis leptocephalus 87

Hunter, J. R., and W. J. Wisby. 1961. Utilization of the nests of green sunfish

{Lepomis cyanellus) by the redfin shiner (Notropis umbratilis cyano-

cephalus). Copeia 1961:113-115.

Jenkins, R. E., and E. A. Lachner. 1980. Hybopsis hypsinotus (Cope),

highback chrb. Page 187 in Atlas of North American Freshwater Fishes

(D. S. Lee et al., editors). N.C. State Mus. Nat. Hist., Raleigh.

Johnston, C. E., and W. S. Birkhead. 1988. Spawning in the bandfin shiner,

Notropis zonistius (Pisces: Cyprinidae). J. Ala. Acad. Sci. 59:30-33.

Koster, W. J. 1939. Some phases of the life history and relationships of the

cyprinid, Clinostomus elongatus. Copeia 1939:201-208.

Kramer, R. H., and L. L. Smith. 1960. Utilization of nests of largemouth bass

(Micropterus salmoides) by golden shiners (Notemigonus crysoleucas).

Copeia 1960:73-74.

Lachner, E. A. 1952. Studies of the biology of the cyprinid fishes of the genus

Nocomis of the northeastern United States. Am. Midi. Nat. 48:433-457.

Latta, W. C. 1958. The ecology of the smallmouth bass, Micropterus

dolomieui Lacepede, at Waugoshance Point, Lake Michigan. Dissertation

Abstract 188:1905.

Loos, J. J., L. A. Fuiman, N. R. Foster, and E. K. Jankowski. 1979. Notes on

early life histories of cyprinoid fishes of the Upper Potomac River. Pages

94-139 in Proceedings of a Workshop on Freshwater Larval Fishes (R.

Wallus and C. W. Voistlander, editors). Tennessee Valley Authority, Norris,

Tenn.

McAuliffe, J. R., and D. H. Bennett. 1981. Observations on the spawning

habits of the yellowfin shiner, Notropis lutipinnis. J. Elisha Mitchell Sci.

Soc. 97:200-203.

McKaye, K. R. 1981. Natural selection and the evolution of interspecific brood

care in fishes. Pages 173-183 in Natural Selection and Social Behavior

(R. D. Alexander and D. W. Tinkle, editors). Chiron Press, New York.

McKaye, K. R., and N. M. McKaye. 1977. Communal care and kidnapping of

young by parental cichlids. Evolution 31:647-681.

Miller, R. J. 1963. Comparative morphology of three cyprinid fishes: Notropis

cornutus, Notropis rubellus, and the hybrid, Notropis cornutus x N.

rubellus. Copeia 1962:450-452.

Miller, R. J. 1964. Behavior and ecology of some North American cyprinid

fishes. Am. Midi. Nat. 72:313-357.

Outten, L. M. 1957. A study of the life history of the cyprinid fish Notropis

coccogenis. J. Elisha Mitchell Sci. Soc. 74:68-84.

Outten, L. M. 1958. Studies of the life history of the cyprinid fishes Notropis

galacturus and rubricroceus. J. Elisha Mitchell Sci. Soc. 74:122-134.

Page, L. M.. and C. E. Johnston. 1989. Spawning in the creek chubsucker,

Erimyzon oblongus, with a review of spawning behaviors in suckers

(Catostomidae). Env. Biol. Fish. 27:265-272.

Pfeiffer, R. A. 1955. Studies on the life history of the rosyface shiner, Notropis

rubellus (Agassiz). Copeia 1955:95-104.

88 Carol E. Johnston

Pflieger, W. L. 1975. Fishes of Missouri. Mo. Dept. Conserv.

Raney, E. C. 1940a. Nests under the water. Bull. N.Y. Zool. Soc. 63: 127-135.

Raney, E. C. 1940b. Reproductive activities of a hybrid minnow, Notropis

cornutus * Notropis rubellus. Zoologica 25:361-367.

Raney, E. C. 1940c. The breeding behavior of the common shiner, Notropis

cornutus (Mitchill). Zoologica 25:1-14.

Raney, E. C. 1947. Nocomis nests used by other breeding cyprinid fishes in

Virginia. Zoologica 32:125-133.

Raney, E. C. 1969. Minnows of New York. Parts 1 and II. The Conservationist

23(5):22-29and 23(6):21-29.

Reed, R. J. 1958. The early life history of two cyprinids, Notropis rubellus and

Campostoma anomalum pullum. Copeia 1958:325-327.

Reighard, J. 1943. The breeding habits of the river chub, Nocomis micropogon

(Cope). Pap. Mich. Acad. Sci., Arts and Letters 29( 1942):397-423.

Settles, W. H., and R. D. Hoyt. 1978. The reproductive biology of the

southern redbelly dace, Chrosomus erythrogaster Rafinesque, in a spring-

fed stream in Kentucky. Am. Midi. Nat. 99:290-298.

Smith, B. G. 1908. The spawning habits of Chrosomus erythrogaster Rafinesque.

Biol. Bull. 14(6):9-18.

Snelson, F. F. 1972. Systematics of the subgenus Lythrurus, genus Notropis

(Pisces: Cyprinidae). Bull. Fla. State Mus., Biol. Sci. 17:1-92.

Starnes, L. B., and W. C. Starnes. 1981. Biology of the blackside dace

Phoxinus cumherlanciensis. Am. Midi. Nat. 106:361-371.

Vives, S. P. 1988. The reproductive behavior of minnows (Pisces: Cyprinidae)

in two reproductive guilds. Ph.D. dissertation. Univ. Wisconsin-Madison.

Yokley, P., Jr. 1974. Habitat and reproductive behavior of the rosefin shiner,

Notropis ardens (Cope), in Lauderdale County, Alabama (Osteichthyes,

Cypriniformes, Cyprinidae). ASB Bull. 21(2):93.

Accepted June 1991

Consumption Rates, Evacuation Rates, and Diets

of Pygmy Killifish, Leptolucania ommata, and

Mosquitofish, Gambusia affinis (Osteichthyes: Atheriniformes)

in the Okefenokee Swamp

J. Douglas Oliver1

Institute of Ecology, University of Georgia

Athens, Georgia 30602

ABSTRACT. — I studied feeding dynamics of Leptolucania ommata

and Gambusia affinis in the Okefenokee Swamp. Both fishes mainly

ate insect larvae (such as Chironomidae) and Cladocera. Evacuation

rates ranged from 0.143 (L. ommata in winter) to 0.279/ hour (L.

ommata in summer). Daily food consumption (dry weight) ranged

from 24.2 (L. ommata in winter) to 148.3 mg/g/day (G. affinis in

summer). Maximum consumption by both species was estimated at

26.31 mg/m2/day, in summer. These values are consistent with other

observations supporting a hypothesis that invertebrate prey production

is substantial in these blackwater wetlands.

Relatively little is known about the diets and feeding dynamics of

fishes of freshwater wetlands (Weller 1981). Studies of the diets of

wetland fishes have concerned those in salt and estuarine marshes (e.g.

Harrington and Harrington 1961, Wetzel 1971, Kjelson et al. 1975). I do

not know of a previous study on consumption or evacuation rates of

any fish in any saltwater or freshwater wetland.

This study was based on field data from the Okefenokee Swamp,

Georgia-Florida, the largest entirely freshwater wetland in the United

States (1800 km ). My objectives were to determine the diets and rates

of consumption and evacuation under field conditions of two of the

most abundant fishes in a marsh on the west side of the Okefenokee

Swamp: that is marsh C, the "Control" marsh of Stinner (1983) and

Oliver and Schoenberg (1989). To quantify dynamics at the ecosystem

level and to estimate minimum invertebrate prey production, I then

used measures of daily food consumption to estimate area-based

consumption (consumption per m2) by these fish. These invertebrate

production estimates may help to resolve whether blackwater habitats

have low secondary productivity, as suggested for tropical blackwaters

(Janzen 1974, Fittkau et al. 1975, Araujo-Lima et al. 1986), or substantial

productivity (Freeman and Freeman 1985). Methods are field-based to

obviate elaborate laboratory feeding studies and to obtain data under

natural conditions.

1 Present address: Bureau of Aquatic Plant Management, Florida Department of Natural

Resources, 3917 Commonwealth Boulevard, Tallahassee, FL 32399.

Brimleyana 17:89-103, December 1991 89

90 J. Douglas Oliver

The two fishes examined were Leptolucania ommata Jordan, the

pygmy killifish, and Gambusia affinis Baird and Girard, the mosquitofish,

small members of the order Atheriniformes. Leptolucania ommata lives

in quiet, densely vegetated fresh waters from southern Georgia and

Alabama to Florida (McLane 1955, Laerm et al. 1980). Gambusia

affinis is native from southern Illinois to Texas and Georgia, and it has

been introduced to warm waters around much of the world, primarily to

consume mosquito larvae (Hess and Tarzwell 1942, Krumholz 1948,

Hurlbert and Mulla 1981). In marshes on the west side of the Okefenokee

Swamp, L. ommata, G. affinis, Enneacanthus gloriosus Holbrook, and

Elassoma okefenokee Bohlke are by far the most numerous fishes (in

that order at C and R sites, Oliver and Schoenberg 1989). The invididuals

used in this study were adults that came within the common distributions

of length (L. ommata of 13-20 mm, G. affinis of 15-25 mm standard

length). Both species were studied in the field for a 24-hour period in the

summer and in the winter.

The study site is a subtropical marsh that has large daily temperature

fluctuations. It lies approximately 200 m E of the entrance to the

Suwannee River sill (an earthen dam that borders the west side of the

Okefenokee Swamp). This blackwater area (mean depth 43-1 13 cm) had

floating and submersed macrophytic vegetation (mainly Nuphar luteum

and Utricularia spp.). Daily water temperature ranges were 4-20° C

during winter observations, and 26-37° C during summer observations.

The site differed from the Okefenokee LCP site of Freeman and

Freeman (1985) in that the latter was shallower (10-55 cm) and was

dominated by floating, submersed, and emergent macrophytes

(Nymphaea odorata, Eriocaulon compressum, and Rhynchospora

inundata).

METHODS

I obtained evacuation rates by two related methods done

simultaneously, which allowed comparison of results (compare with

single-method analysis, e.g. Sainsbury 1986). In the "tank" method, the

decline in gut contents of fish held without food was converted to an

evacuation rate (Staples 1975, Garcia and Adelman 1985). Clear

immersed tanks at the edge of the marsh were used to track natural light

levels and water temperatures (measured by a standard mercury

thermometer). Okefenokee water was filtered into tanks through a

screen (<63 mm mesh) to remove potential food items. At approximately

4-hour intervals, large fish captured by seine were placed into two tanks

and smaller fish into two other tanks, so that the larger fish would

neither consume nor frighten the latter. I preserved about half the

seined fish immediately in buffered formalin and preserved the rest

Pygmy Killifish and Mosquitofish 91

approximately 4 hours after they were put in the tanks, for comparison

of gut contents. In the "field" method, the decline in gut contents

between field samples taken at intervals during non-feeding times of day

yielded a second measure of evacuation rates. (See the section on

Feeding Dynamics for an example of evacuation analyses.)

In the laboratory, lengths of foods in foreguts were converted to

weights. Foreguts were analyzed because their contents declined con-

sistently with time in tanks, whereas hindguts continued to receive food

from foreguts in some cases. I examined foregut contents under a

dissecting microscope and measured lengths of food items. Length-

weight regressions in Dumont et al. (1975) were used to calculate dry

weights of Ostracoda, Harpacticoida, nauplii, and most Cladocera. Hall

et al. (1970) gave macrothricid weights. Insect head-capsule widths were

converted to weights (Smock 1980). Ruttner-Kolisko (1977) and J.

Gerritsen and H. S. Greening (personal communication) gave rotifer

length-weight conversions. A regression by Gerritsen (personal com-

munication) for the Okefenokee Swamp was used to derive cyclopoid

weights. Maximum carapace widths of araneids were converted to

weights (Edgar 1971, Barber 1983). Weights of Acari were estimated

from the regression of Oribatei by Persson and Lohm (1977).

I calculated a length-weight regression to yield weights of Gambusia

affinis within Gambusia foreguts. Foregut content (S) was expressed in

relative units, i.e. mg dry food/g dry fish, assuming 20% dry to wet

weight conversion for fish (e.g. Lagler et al. 1977).

Evacuation rates, daily food consumption, and area-based con-

sumption were calculated for both fishes. Area-based consumption

equals the dry weight equivalent of fish density (Oliver and Schoenberg

1989, method of density measurement adapted from Freeman et al.

1984) times calculated daily food consumption (Staples 1975, adjusted

in Elliott and Persson 1978; Persson 1982; Garcia and Adelman 1985).

Results are based on collections of fish at about 4-hour intervals for 24

hours, in summer (L. ommata and G. affinis, 19-20 August 1984) and

winter (L. ommata on 7-8 March 1984, and G. affinis on 16-17

February 1985, when it was common).

ANALYSES AND RESULTS

Diets

Chironomids and Cladocera dominated the diet of Leptolucania

ommata. Major prey types were Chironomidae, unidentified Insecta,

and Cladocera (Table 1). Oribatid mites, not usually found in fish guts

(B. J. Freeman, personal communication), were eaten by both L.

ommata and G. affinis.

oo

ON

I<

•2

S

on o — m o

«s <n «n o o

o ©

o o

— <N NO © © © © ©

oooooooo

o o o o o

o o

oooooooo^tooo«noooo

rslCMOOOOOO^ooOO

oo — ooooot^ooo

^ o o

OO o o

O O CnI o

o o

vO - (N fn N

cm © © no r^

<Nomoo— ' oo

m o

o o o o

OJ

03

a O

03

O ™

•o a x 5 -a -s

o o l, o a 51-

js JH * « *

u S O O ffi

o

I

03

sr

c

03

T3

O

is a

O 3 +- M C

>> aj o o a

U Z C* £ H

2 a

1) 03 \£

S o ."2

<U T3 C

X O D

-J

1 1 ts s <

1 « 5 a B

< O O h h

Pygmy Killifish and Mosquitofish 93

In Gambusia affinis, insects were dominant in the diet, but other

arthropods and some mosquitofish were also eaten. In summer,

Tanypodinae and odonate nymphs were quantitatively significant foods

(Table 1). Cladocera, Cyclopoida, and Araneae composed more of the

diet in summer than in winter. In both winter and summer, non-

tanypode Chironomidae and unidentified Insecta were dominant food

items. Gambusia affinis showed some cannibalism (in summer, about

3%).

Feeding Dynamics

Gambusia affinis. The balance between consumption and evacua-

tion may be inferred from diel changes in foregut content (solid lines on

Fig. 1). When the slope of the line is positive, consumption rate is

greater than evacuation rate during the specified time period; when

negative, consumption is less than evacuation.

Analysis of evacuation rates is based on comparisons of gut-

content trends obtained by field and tank methods, i.e. the solid versus

the dashed lines of Fig. 1 and 2. During certain periods (e.g. 1055 to

1455 hours), fish in the field actively fed, which resulted in an increase

in the foregut content (Fig. 1). For such periods, it is necessary to use

the fish held in food-free tanks (dashed lines) to calculate evacuation

rates. At other times (e.g. 1850 to 2250 hours), the foregut content

declined in both field fish and tank-held fish. For these periods, I

considered field fish, which were unconfined and egesting in their

natural environment, to provide the better estimate of natural evacuation

rate. The evacuation rate is computed by

r= ln(S0+l)-ln(St+l)

in which SQ is relative foregut content before and St is relative foregut

content after time t (adapted from Elliott and Persson 1978; 1 is added

to allow logarithmic transformation of zeros). The same equation

applies for determining evacuation in tank-held fish. When evacuation

data were missing for a period, evacuation rate was presumed to be an

average of rates before and after that period. Finally, the various rates

during the day were weighted by time to produce an average evacuation

rate (Table 2).

Evacuation rates for each of the time periods were used in calculat-

ing consumption over each of these periods. Each such evacuation rate

was combined with average foregut content before and after the period

94

J. Douglas Oliver

Gambusia affinis

AUGUST

xlOO

Ll.

O

a 80

o

u_

1 60

z

o

a!40-

I '■» I i i i I ■ i i I i i i I i i » I — i i i i i i i ■ i 9 i ■

NOON 4PM 8PM MIDNT 4AM 8AM NOON 4PM 8PM

I i i i I i i i

Fig. 1. Foregut content of and food consumption by Gambusia affinis in

summer. In the upper panel, filled circles and solid lines show diel content

trends in field fish; open squares and dashed lines show changes in content of

fish held in food-free tanks. In the lower panel, filled circles and solid lines show

trends in consumption during each time period; the open circle and dotted lines

show the presumed trend based on difference between the final (1630 hours) and

initial (2050 hours) consumption values.

Pygmy Killifish and Mosquitofish

95

Leptolucania ommata

AUGUST

x 50!

CO

o

q40H

o

u.

!§30

20-

=> 10

CD

LU

tr

o

"■ 0

is100

e>

q80

o

2 60

z

o

co

§20

i — | i i — i — | — i i i \ i i ^i — | — » r^ | i i »i | i i i — f— i — i 'r | i »i \

NOON 4PM 8PM MIDNT 4AM 8AM NOON 4PM 8PM

OLr

I i « » i « > •

TT P

I ' ' ' \

Fig. 2. Foregut content and food consumption of Leptolucania ommata. (See

Fig. 1 legend for details.)

96

"° o ?

m 3 c

< O £

.2 £

"O o >>

r-2 3 W — -.

a §

C

. o

O <-> to ^

<N OO

OS v©

CM Tt

© ©

— CO

GO

c

6

CO

u

C/3

^t r4 oo ^-

m 2° aj on ro oo

M ^ 3 — oo O

^ """ fe >> °^ ~~

« cd w 2

fc 2 3 3>

.!•» 3 <

C/2

I |

OX) 3

< o

ON <?

<N Ov

>,_ < o

c SO —

</->

X)

>

O

u

00

4

tf

Pygmy Killifish and Mosquitofish 97

to yield consumption during the period (adapted from Elliott and

Persson 1978):

Ct=((St+l)-(S0-H)e-rt)rt

l-e~rt

Daily food consumption, £Ct, is the sum of consumption rates over

each period. For G. affinis in summer, the graph of consumption versus

time (bottom panel of Fig. 1) indicates that peak consumption occurred

in the afternoon.

Gambusia affinis in winter samples showed feeding trends similar

to those in summer but at lower levels. Evacuation rate and daily food

consumption were lower in February than in August (Table 2). There

was a single feeding peak in late afternoon, as in summer. Low daily

food consumption combined with low fish biomass to produce very low

area-based consumption.

Leptolucania ommata. Leptolucania ommata showed diel feeding

patterns similar to those of G. affinis. Foregut content in summer

peaked in late afternoon (Fig. 2). Consumption showed one daily peak,

in the afternoon. In winter (March 1984), L. ommata also consumed

maximally in the afternoon, as other species-season combinations had

done.

Leptolucania ommata had seasonal feeding dynamics similar to

those of G. affinis. The evacuation rate in summer was 0.279/ hour,

similar to the 0.262/ hour calculated for G. affinis in summer (Table 2).

The evacuation rate in winter was also similar in the two species, 0.143

and 0.1 57/ hour, respectively. In both species, daily food consumption

increased from winter to summer by a factor of about 4 (3.85 for L.

ommata, 4.62 for G. affinis). This increase might be largely a result of

temperature-dependent feeding: If the "Q10" for food consumption were

about 2, the approximately 20° C difference from winter to summer

would give an increase of about 4 times. Such a doubling of feeding for

every I0°C is consistent with the results of Salvatore et al. (1987), who

found that the feeding rate of Dorosoma cepedianum Lesueur approxi-

mately doubled when laboratory water temperature increased from 10

to 20° C. Thus such differences in feeding rates may be largely attributed

to the environmental temperature.

DISCUSSION

The two fishes fed on similar kinds of foods. This should not be

surprising, because I captured them in the same areas and observed

them feeding at about the same depth (near the surface). Diets,

evacuation, and consumption in this study were similar to values

98 J. Douglas Oliver

reported for fishes in other types of habitats (see comparisons in the

following discussion).

The diet of L. ommata in the Okefenokee marshland was mainly

chironomid larvae and other insects, as well as Cladocera. Similarly, in

the St. Johns River system of north Florida, L. ommata ate mainly

chironomids, Cladocera, and Copepoda (McLane 1955). In the

Okefenokee, Gambusia affinis consumed mainly chironomids, odonates,

other insects, and Cladocera (Table 1). Similarly, G. affinis in shallow

areas of Wheeler Reservoir, Ala., ate mostly Entomostraca (presumably

Cladocera or Copepoda), Chironomidae, and other juvenile insects

(Anopheles, Hess and Tarzwell 1942). In California rice fields, G. affinis

ate mostly Cladocera and immature Chironomidae (Washino and

Hokama 1967) and some Ostracoda (Farley 1980). Those in California

ponds ate mostly Cladocera (Miura et al. 1979). Thus, in Okefenokee

marshes, the fishes fed primarily on the same kinds of foods that they

ate in other localities and in different habitats. The two fishes in the

Okefenokee freshwater marsh ate oribated mites, which are often

associated with an algal, detrital system of this kind (Pennak 1978), but

Oribatei were a minor component of their diets (Table 1).

Both fishes showed peak consumption rates at about the same time

of day, the afternoon. The afternoon had bright sunlight and the highest

temperatures of the day, and these factors may have contributed to

increased consumption by fish (Reddy 1975, Mann 1978, Smagula and

Adelman 1982, Garcia and Adelman 1985), and high illumination may

make foods more conspicuous, particularly in a vegetated blackwater

environment.

Evacuation rates of the two fishes were alike, and similar to rates

reported for other fishes. Foregut evacuation rates were 0.143 and

0.279/ hour for L. ommata in winter and summer, respectively. Similarly,

rates for G. affinis were 0.157 and 0.262/ hour. Doble and Eggers (1978)

reported rates of 0.109 and 0.267/hour for small juvenile Oncorhynchus

nerka Walbaum in winter and summer. Ruggerone (1989) measured

evacuation of 0.274 to 0.329 for Oncorhynchus kisutch Walbaum at

13° C. Persson (1982) found rates of 0.129 and 0.499/hour for Rutilus

rutilus L. held at 12 and 24° C in the laboratory. Amundsen and

Klemetsen (1988) reported rates of 0.08 to 0.16/ hour for Salvelinus

alpinus L. at 13°C. Thus, evacuation rates for L. ommata and G. affinis

appear to be well within the range of values reported for various fishes.

Because evacuation rates are similar to rates of other fishes,

consumption measures of Gambusia and Leptolucania in the Okefenokee

Swamp are in the same range as estimates for other species. Daily food

consumption by L. ommata was 24.2 and 93.1 mg/g/day (dry weights),

in winter and summer, respectively. Values for G. affinis were 32.1 and

Pygmy Killifish and Mosquitofish 99

148.3 mg/g/day. Doble and Eggers (1978) found that Onchorhynchus

nerka juveniles ate 15.3 and 44.1. mg/g/day in Lake Washington in

winter and summer, respectively. Garcia and Adelman (1985) reported

that Cyprinus carpio L. in the Mississippi River consumed 204 mg/g/day

in summer (assuming a fish dry to wet ratio of 20%). Cech et al. (1981)

stated that in the laboratory, newborn G. affinis ate from 70 to 820

mg/g/day at 10-35° C. Thorpe (1977, in Elliott and Persson 1978)

reported summer consumption by Perca fluviatilis L. in Loch Leven to

be 54 mg/g/day (dry weights, assuming fish dry to wet ratio of 20% and

prey wet to dry ratio of 6; Freeman and Freeman 1985), but this was an

underestimate according to Elliott and Persson (1978). Basimi and

Grove (1985) reported that summer consumption by small Pleuronectes

platessa L. off the coast of Wales was 43 mg/g/day (assuming the same

ratios). Sagar and Glova (1988) found that juvenile Oncorhynchus

tshawytscha Walbaum in the Rakaia River, New Zealand, ate 83

mg/g/day at a mean temperature of about 15°C. Consumption by O.

kisutch ranged from 21 to 44 mg/g/day at 5.8-7.5° C (Ruggerone 1989).

Food consumption rates of L. ommata and G. affinis from the

Okefenokee wetland clearly fall within the range offish in other types of

environments.

As expected, area-based consumption by the two fishes was low in

winter and higher in summer. Leptolucania ommata in the Okefenokee

marsh consumed 0.71 and 22.99 mg/m /day in winter and summer,

respectively. Gambusia affinis ate less, presumably because of their

lower biomass; they consumed 0.33 and 3.32 mg/m /day in winter and

summer, respectively. In comparison, in a small New Zealand lake with

only one fish species, Staples (1975) reported that Philypnodon breviceps

Stokell in summer consumed 203 mg/m /day (assuming a wet to dry

ratio of 6); this value was an underestimate according to Elliott and

Persson (1978). In a New Zealand stream where trout and eels were also

present (Hopkins 1970, in Staples 1975), the maximum reported area-

based consumption by P. breviceps was equivalent to 74 mg/m2/day.

Consumption data are consistent with a hypothesis that invertebrate

prey production in Okefenokee blackwater marshes is substantial.

Consumption values in spring and fall are usually between winter and

summer values, rising in a nonlinear manner as a function of temperature

(see Feeding Dynamics above; Staples 1975, Doble and Eggers 1978).

Thus, the geometric mean of winter and summer consumption values

may give a reasonable estimate of mean daily food consumption for the

whole year. Calculating the geometric mean of consumption values

(from Table 2) yields estimates of 47.5 and 69.0 mg/g/day for L.

ommata and G. affinis, respectively. When each of these values is

multiplied by average dry biomass per m2 (from Oliver and Schoenberg

100 J. Douglas Oliver

1989) for each of these fishes, consumption by both L. ommata and G.

affinis is calculated to be 5.88 mg/m /day. Assuming a wet to dry

conversion of 6, invertebrate prey production would have to be at least

130 kg/ ha/ year (wet weight), just to meet consumption needs of these

fish. At a recently abandoned bird rookery on the west side of the

Okefenokee, guano fertilization apparently increased standing stocks of

several trophic levels, including fish (Oliver and Schoenberg 1989):

Average annual biomass estimates of L. ommata and G. affinis were

elevated to 4.5 times the levels of the present study. Invertebrate

production may have been about 580 kg/ ha/ year, just to meet con-

sumption by these fish. At the same site, I measured fish production

directly by the size-frequency method (Hynes and Coleman 1968,

Freeman and Freeman 1985) and made a second estimate of prey

production: Based on a combined production for L. ommata and G.

affinis of 89.8 kg/ ha/ year (wet weight, unpublished data) and a gross

conversion efficiency of <25% for fish (Brett and Groves 1979),

production of invertebrate prey would have to be at least 360 kg/ ha/ year,

just to meet consumption requirements of the two fishes. These values

are within the range of estimates of production for the total zoobenthos

in non-blackwater systems given by Waters (1977), and they are con-

sistent with the findings of Freeman and Freeman (1985) that indicated

substantial productivity of another Okefenokee blackwater ecosystem.

ACKNOWLEDGMENTS.— I thank Robert E. Reinert and Steven

A. Schoenberg, of the University of Georgia, and Tom Richards for

their valuable help and ideas. This research was supported by National

Science Foundation grants BSR 81 14823 and BSR 82155587. Paper No.

63, Okefenokee Ecosystem Investigations.

LITERATURE CITED

Amundsen, P. A., and A. Klemetse-n. 1988. Diet, gastric evacuation rates and

food consumption in a stunted population of Arctic charr, Salvelinus

alpinus L., in Takuatn, northern Norway. J. Fish Biol. 33:697-709.

Araujo-Lima, C. A. R. M., L. P. S. Portugal, and E. G. Ferreira. 1986. Fish-

macrophyte relationship in the Anavilhanas Archipelago, a black water

system in the Central Amazon. J. Fish Biol. 29:1-1 1.

Barber, M. C. 1983. Nutrient dynamics of orb weaving spiders (Araneae:

Araneidae) in Okefenokee shrub swamps. Ph.D. dissertation, Univ. of

Georgia, Athens.

Basimi, R. A., and D. J. Grove. 1985. Estimates of daily food intake by an

inshore population of Pleuronectes platessa L. off eastern Anglesey, North

Wales. J. Fish Biol. 27:505-520.

Brett, J. R., and T. D. Groves. 1979. Physiological energetics. Pages 279-352 in

Pygmy Killifish and Mosquitofish 101

Fish Physiology, Vol. 8 (W. S. Hoar et al., editors). Academic Press, New

York.

Cech, J. J., W. A. Wurtsbaugh, and B. C. Vondracek. 1981. Effect of

temperature and ration size on the food consumption and growth rates of

mosquitofish, Gambusia affinis. Proc. Calif. Mosq. Vector Control Assoc.

49:37.

Doble, B. D., and D. M. Eggers. 1978. Diel feeding chronology, rate of gastric

evacuation, daily ration, and prey selectivity in Lake Washington juvenile

sockeye salmon (Oncorhynchus nerka). Trans. Am. Fish. Soc. 107:36-45.

Dumont, H. J., I. Van de Velde, and S. Dumont. 1975. The dry weight

estimate of biomass in a selection of Cladocera, Copepoda and Rotifera

from the plankton, periphyton and benthos of continental waters. Oecologia

(Bed.) 19:75-97.

Edgar, W. D. 1971. Aspects of the ecological energetics of the wolf spider

Pardosa (Lycosa) lugubris (Walckenaer). Oecologia (Berl.) 7:136-154.

Elliott, J. M., and L. Persson. 1978. The estimation of daily rates of food

consumption for fish. J. Anim. Ecol. 47:977-991.

Farley, D. G. 1980. Prey selection by the mosquitofish Gambusia affinis in

Fresno County rice fields. Proc. Calif. Mosq. Vector Control Assoc.

48:51-55.

Fittkau, E. J., U. Irmler, W. J. Junk, F. Reiss, and G. W. Schmidt. 1975.

Productivity, biomass, and population dynamics in Amazonian water bodies.

Pages 289-31 1 in Tropical Ecological Systems, Ecol. Stud. 1 1 (F. B. Golley

and E. Medina, editors). Springer Verlag, New York.

Freeman, B. J., and M. C. Freeman. 1985. Production of fishes in a

subtropical blackwater ecosystem: The Okefenokee Swamp. Limnol.

Oceanogr. 30:686-692.

Freeman, B. J., H. S. Greening, and J. D. Oliver. 1984. Comparison of three

methods for sampling fishes and macroinvertebrates in a vegetated fresh-

water wetland. J. Freshwater Ecol. 2:603-609.

Garcia, L. M., and I. R. Adelman. 1985. An in situ estimate of daily food

consumption and alimentary canal evacuation rates of common carp,

Cyprinus carpio L. J. Fish Biol. 27:487-493.

Hall, D. J., W. E. Cooper, and E. E. Werner. 1970. Dynamics and structure of

freshwater animal communities. Limnol. Oceanogr. 15:839-928.

Harrington, R. W., Jr., and E. S. Harrington. 1961. Food selection among

fishes invading a high subtropical salt marsh: From onset of flooding

through the progress of a mosquito brood. Ecology 42:646-666.

Hess, A. D., and C. M. Tarzwell. 1942. The feeding habits of Gambusia affinis

affinis, with special reference to the malaria mosquito, Anopheles

quadrimaculatus . Am. J. Hyg. 35:142-151.

Hurlbert, S. H., and M. S. Mulla. 1981. Impacts of mosquitofish {Gambusia

affinis) predation on plankton communites. Hydrobiologia 83:125-151.

Hynes, H. B., and M. J. Coleman. 1968. A simple method of assessing the

annual production of stream benthos. Limnol. Oceanogr. 13:569-573.

Janzen, D. H. 1974. Tropical blackwater rivers, animals, and mast fruiting by

Dipterocarpaceae. Biotropica 6:69-103.

102 J. Douglas Oliver

Kjelson, M. A., D. S. Peters, G. W. Thayer, and G. N. Johnson. 1975. The

general feeding ecology of postlarval fishes in the Newport River estuary.

Fish. Bull. 73:137-144.

Krumholz, L. A. 1948. Reproduction in the western mosquitofish, Gambusia

affinis affinis (Baird & Girard), and its use in mosquito control. Ecol.

Monogr. 18:1-43.

Laerm, J., B. J. Freeman, L. J. Vitt, and L. Logan. 1980. Vertebrates of the

Okefenokee Swamp. Brimleyana 4:47-73.

Lagler, K. F., J. E. Bardach, R. R. Miller, and D. R. M. Passino. 1977.

Ichthyology. 2nd ed. Wiley, New York.

Mann, K. H. 1978. Estimating the food consumption offish in nature. Pages

250-273 in Ecology of Freshwater Fish Production (S. D. Gerking, editor).

Wiley, New York.

McLane, W. M. 1955. The fishes of the St. Johns River system. Ph.D.

dissertation, Univ. of Florida, Gainesville.

Miura, T., R. M. Takahashi, and R. J. Stewart. 1979. Habitat and food

selection by the mosquitofish, Gambusia affinis. Proc. Calif. Mosq. Vector

Control Assoc. 47:46-50.

Oliver, J. D., and S. A. Schoenberg. 1989. Residual influence of macronutrient

enrichment on the aquatic food web of an Okefenokee Swamp abandoned

bird rookery. Oikos 55: 175-182.

Pennak, R. W. 1978. Fresh-water Invertebrates of the United States. 2nd ed.

Wiley-Interscience, New York.

Persson, L. 1982. Rate of food evacuation in roach (Rutilus rutilus) in relation

to temperature, and the application of evacuation rate estimates for studies

on the rate of food consumption. Freshwater Biol. 12:203-210.

Persson, T., and U. Lohm. 1977. Energetical Significance of the Annelids and

Arthropods in a Swedish Grassland Soil. Swedish Natural Science Research

Council, Stockholm.

Reddy, S. R. 1975. Effect of water temperature on the predatory efficiency of

Gambusia affinis. Experientia 31:801-802.

Ruggerone, G. T. 1989. Gastric evacuation rates and daily ration of piscivorous

coho salmon, Oncorhynchus kisutch Walbaum. J. Fish Biol. 34:451-463.

Ruttner-Kolisko, A. 1977. Suggestions for biomass calculations of plankton

rotifers. Arch. Hydrobiol. Beih. Ergen. Limnol. 8:71-76.

Sagar, P. M., and G. J. Glova. 1988. Diel feeding periodicity, daily ration and

prey selection of a riverine population of juvenile chinook salmon,

Oncorhynchus tshawytscha (Walbaum). J. Fish Biol. 33:643-653.

Sainsbury, K. J. 1986. Estimation of food consumption from field observations

of fish feeding cycles. J. Fish Biol. 29:23-36.

Salvatore, S. R., N. D. Mundahl, and T. E. Wissing. 1987. Effect of water

temperature on food evacuation rate and feeding activity of age-0 gizzard

shad. Trans. Am. Fish. Soc. 1 16:67-70.

Smagula, C. M., and I. R. Adelman. 1982. Day-to-day variation in food

consumption by largemouth bass. Trans. Am. Fish. Soc. 111:543-548.

Smock, L. A. 1980. Relationships between body size and biomass of aquatic

insects. Freshwater Ecol. 10:375-383.

Pygmy Killifish and Mosquitofish 103

Staples, D. J. 1975. Production biology of the upland bully Philypnodon

breviceps Stokell in a small New Zealand lake. III. Production, food

consumption and efficiency of food utilization. J. Fish Biol. 7:47-69.

Stinner, D. H. 1983. Colonial wading birds and nutrient cycling in the

Okefenokee Swamp ecosystem. Ph.D. dissertation, Univ. of Georgia, Athens.

Washino, R. K., and Y. Hokama. 1967. Preliminary report on the feeding

pattern of two species of fish in a rice field habitat. Proc. Calif. Mosq.

Control Assoc. 35:84-87.

Waters, T. F. 1977. Secondary production in inland waters. Pages 91-164 in

Advances in Ecological Research, Vol. 10 (A. MacFadyen, editor). Academic

Press, New York.

Weller, M. W. 1981. Freshwater Marshes. Univ. Minnesota Press, Minneapolis.

Wetzel, R. L. 1971. Analysis of cohabitation by Gambusia affinis and Poecilia

latipinna (Pisces: Poeciliidae) in a salt-marsh canal in Florida. M.S. thesis,

Univ. of West Florida, Pensacola.

Accepted June 1991

104

ENDANGERED, THREATENED, AND

RARE FAUNA OF NORTH CAROLINA

PART II.

A RE-EVALUATION OF THE MARINE AND

ESTUARINE FISHES

by

Steve W. Ross, Fred C. Rohde, and David G. Lindquist

This is the second in a series of reports by committees appointed in

1985 by the North Carolina State Museum of Natural Sciences to re-

evaluate the faunal lists presented in Endangered and Threatened Plants

and Animals of North Carolina (John E. Cooper, Sarah S. Robinson,

and John B. Funderburg, editors. N.C. State Mus. Nat. Hist., Raleigh,

1977), which is now out of print. The report on marine and estuarine

fishes by Ross, Rohde, and Lindquist treats one Endangered species, six

Vulnerable species, and four anadromous fishes that, while not formally

listed, are of some concern. Five species listed as being of Special

Concern in 1977 no longer warrant formal status. The publication

includes six original drawings by Renaldo Kuhler.

1988 20 pages Softbound ISBN 0-917134-17-6

Price: $3 postpaid. North Carolina residents add 5% sales tax. Please make checks

payable in U.S. currency to NCDA Museum Extension Fund.

Send order to: ETR MARINE FISHES, N.C. State Museum of Natural Sciences,

P.O. Box 27647, Raleigh, NC 2761 1.

Morphological Variation in Turtles

of the Genus Pseudemys (Testudines: Emydidae)

From Central Atlantic Drainages

Michael E. Seidel

Department of Biological Sciences

Marshall University, Huntington, West Virginia 25701

AND

William M. Palmer

North Carolina State Museum of Natural Sciences

P.O. Box 27647, Raleigh, North Carolina 2761 1

ABSTRACT. — Thirty morphometric and 15 qualitative characters

were analyzed to compare Pseudemys rubriventris, P. floridana

floridana, and P. concinna concinna in the eastern United States.

Taxonomic characters that have been employed to define these species

are reexamined. Principal components and discriminant analyses

indicate that P. rubriventris is morphologically distinct from the other

two Pseudemys. Several additional useful taxonomic characters were

found, but some character convergence or hybridization between P.

rubriventris and congeners was detected. No morphometric divergence

was found between P. f. floridana and P. c. concinna, and only

markings appear to separate the two forms. As reported in previous

works, P. floridana inhabits the coastal plain and P. concinna inhabits

the piedmont. Populations occurring in a relatively broad area

overlapping the Fall Line of North Carolina have morphological

character states that are variable and somewhat intermediate between

these two species.

Cooter and redbelly turtles are aquatic species of emydids that

inhabit the eastern and south-central United States. They are relatively

large (up to 420 mm carapace length) basking species with striped head

markings and primarily herbivorous feeding habits. Following Seidel

and Smith (1986) and Ward (1984), current classification places these

turtles in the genus Pseudemys, separate from sliders (Trachemys) and

painted turtles (Chrysemys) (Collins 1990, Ernst and Barbour 1989,

King and Burke 1989). The genus Pseudemys includes three redbelly

species [P. alabamensis Baur, P. nelsoni Carr, P. rubriventris (LeConte)]

and three cooter species [P. concinna (LeConte), P. floridana (LeConte),

P. texana Baur].

Taxonomic relationships in the genus Pseudemys are problematic

as indicated by an extensive history of species-subspecies revisions (see

Smith and Smith 1980 for a review). Frequently, in areas of sympatry,

Brimleyana 17:105-135, December 1991 105

106 Michael E. Seidel and William M. Palmer

evidence of hybridization has been reported. Some populations with

intermediate (hybrid?) characters are geographically broad, which

suggests subspecific relationships. These interactions have been examined

in Florida (Crenshaw 1955) and Louisiana (Fahey 1980). Part of the

problem has arisen from the absence of clearly defined quantifiable

characters that separate species of Pseudemys. Another problem has

been the relatively small number of specimens examined, especially

from northern populations. The most recent taxonomic analysis with

species diagnoses of Pseudemys relies heavily on cranial musculature

and osteology (Ward 1984). Unfortunately those characters are of little

use in field identification or in evaluation of fluid-preserved museum

material.

In spite of the taxonomic attention Pseudemys has received, we

have found that species of the eastern United States (Atlantic slope)

remain very difficult to identify using available diagnostic characters.

Nearly all key characteristics are qualitative and based on highly

variable markings and shell shapes. The problem of identification is

especially acute in the coastal plain of Virginia and North Carolina

where the ranges of P. rubriventris, P. concinna concinna, and P.

floridana floridana overlap or come in contact. In that area Crenshaw

(1965) noted putative hybridization between P. rubriventris and P.

floridana, and Martof et al. (1980) reported frequent hybridization

between P. concinna and P. floridana, commenting that some specimens

defy classification at the species level. The objectives of the present

study were: (1) to identify external characters that more reliably

distinguish these turtles in Virginia and North Carolina, (2) to identify

individual turtles from this region that appear to be morphologically

intermediate, and (3) to characterize patterns of Pseudemys distribution

in the central Atlantic coastal plain.

METHODS

For morphometric analysis, 76 fluid-preserved P. rubriventris (New

Jersey, Pennsylvania, West Virginia, Virginia, North Carolina), 57 P. c.

concinna (Virginia, North Carolina, South Carolina), and 59 P. f

floridana (Virginia, North Carolina, South Carolina) were analyzed

(Fig. 1). Specimens included freshly collected individuals with typical

coloration as well as museum, specimens (see Specimens Examined).

Abbreviations for museums follow Leviton et al. (1985), and MES =

reference collection of the senior author. All turtles were tentatively

identified to species a priori using traditional qualitative characters

(mostly markings, see Table 1) that have been applied to distinguish P.

concinna, P. floridana, and P. rubriventris (Ernst and Barbour 1972,

1989, Conant and Collins 1991). If assignment was questionable, that

Morphological Variation in Pseudemys

107

Vo.

—#-*

T3-

A >\

A a

OA

•A

J

M.*' A

km

Fig. 1. Localities of adult Pseudemys specimens examined for morph

analysis. The Fall Line is indicated by a broken line.

ometric

108 Michael E. Seidel and William M. Palmer

was noted. These identifications often agreed with specific assignments

in museum collections, although many specimens from coastal Virginia

and North Carolina are catalogued as "Pseudemys sp?" or only

tentatively assigned to species.

Character states for 15 qualitative characters were recorded for

each turtle. Twenty-six shell characters were measured on each specimen

with calipers (Helios) or a goniometer (Jamar). Two head-neck stripes,

head width, and maxillary cusp length were also measured. These 30

measurements (Fig. 2-7) included all quantifiable external characters

that have been used to diagnose P. concinna, P. floridana, or P.

rubriventris, as well as additional characters suspected to have taxonomic

value: carapace length along midline (CL), carapace width at sulcus

between marginals V-VI (CW), carapace width at sulcus between

marginals VII-VIII (SW), plastron length along midline (PL), shell

height at sulcus between vertebrals II-III (CH), shell height at sulcus

between vertebrals III-IV (SH), cervical scute dorsal length (CS), cervical

scute dorsal posterior width (CD), cervical scute ventral length (CU),

cervical scute ventral posterior width (CV), marginal XII length (MH),

marginal XII anterior dorsal width (MA), marginal XII posterior

(ventral) width (MP), lateral angle of carapace formed by dorsal and

ventral surfaces of marginal VI (SA), posterior angle of carapace

formed by midline slope of vertebral V and midline sulcus of anal scutes

(PG), anal notch depth (AN), length of interfemoral sulcus (IL), shortest

distance between inguinal scute and pectoral-abdominal sulcus (IE),

anterior plastral lobe width (PW), posterior plastral lobe width (XW),

taper of anal scutes measured as the angle formed by posterior extension

of lines along the lateral edge of the anal scutes (AA), epiplastron

thickness at mid-humeral scute (ET), depth of epiplastral lip measured

as the distance between the anterior tip of the intergular sulcus and a

line formed by resting a straightedge across the dorsal epiplastral lip

(EP), cervical scute recession measured from anterior tip of cervical

scute to a straight line along the anterior tip of first pair of marginals

(NR), ventral extension of posterior carapace measured from posterior

tip of interanal sulcus to posterior edge of vertebral V (AV) and to

posterior tip of intermarginal XII sulcus (AP), head width at anterior

margin of tympanum (HW), length of cusps on upper tomium (LC),

greatest width of supratemporal stripe (SS), and width of post-

symphyseal (ventral) stripe at level of tympanum (GS).

For multivariate analysis, only turtles with a midline carapace

length > 120 mm were included, and males and females were analyzed

separately. That reduced the effects of ontogenetic and sexually

dimorphic character variation, which may be pronounced in Pseudemys

(Iverson and Graham 1990). Principal components analysis (PCA-SAS;

Morphological Variation in Pseudemys 109

Table 1. Qualitative characters used for initial identification of Pseudemys

species.

P. rubriventris

1. Upper jaw with a prominent notch bordered on each side by tooth-like

(tomiodont) cusps (Carr 1952, Crenshaw 1955, Ernst and Barbour 1989).

2. Second pleural scute without C-shaped mark.

3. Plastron red or faded pink with central dark figure extending along seams

(Carr 1952, Crenshaw 1955, Ernst and Barbour 1989).

4. Carapace of large individuals with numerous lateral rugosities but flat or

concave along the vertebrals (Ernst and Barbour 1989, Weaver and Rose

1967).

5. Dark markings on bridge, including inguinal scute.

6. Posterior (after bridge) inframarginal spots or circles do not overlap

intermarginal seams (Ward 1984).

7. Posterior margin of carapace weakly serrated and marginal notches weak.

P. concinna

1. Upper jaw with a very weak notch, not bordered by tooth-like cusps.

2. Second pleural scute with C-shaped mark (Carr 1952, Crenshaw 1955, Ernst

and Barbour 1989).

3. Plastron yellow or orange with central dark figure extending along seams.

4. Carapace of large individuals slightly keeled along vertebrals; carapace not

finely rugose.

5. Dark markings on bridge, including inguinal scute.

6. Posterior inframarginal circles overlap intermarginal seams.

7. Posterior margin of carapace serrated and marginals prominently notched

(Weaver and Rose 1967, Ward 1984).

P. floridana

1. Upper jaw entirely smooth, no notch or cusps.

2. Second pleural scute without C-shaped mark.

3. Plastron pale yellow without any dark markings.

4. Carapace of large individuals rounded or flat (not keeled) along vertebrals.

5. Dark markings usually absent from bridge and inguinal scute (Ward 1984).

6. Markings usually very faint or absent on posterior inframarginals.

7. Posterior margin of carapace weakly serrated and marginal notches weak.

110

Michael E. Seidel and William M. Palmer

MP

-*/

Fig. 2. Illustration of morphometric characters measured on carapace: CS =

cervical scute dorsal length, CU = cervical scute ventral length, CD = cervical

scute dorsal width, CV = cervical scute ventral width, CW = anterior carapace

width, SW = posterior carapace width, CL = carapace length, MA = anterior

marginal XII width, MP = posterior marginal XII width, MH = marginal XII

length. All measurements are between the appropriate dots.

Morphological Variation in Pseudemys

1 1 1

IE

Fig. 3. Illustration of morphometric characters measured on plastron: PW =

anterior plastral lobe width, PL = plastron length, XW = posterior plastral lobe

width, IE = distance between inguinal scute and pectoral-abdominal sulcus, IL =

length of interfermoral sulcus, AN = anal notch depth, AA = lateral angle of

anal scutes.

112

Michael E. Seidel and William M. Palmer

1/ SA

Fig. 4. Anterior view of shell illustrating lateral angle of the carapace (SA) and

depth of epiplastral lip (EP).

Fig. 5. Dorsal anterior view of the carapace illustrating recession of the cervical

scute (NR).

Morphological Variation in Pseudemys

13

5. CL

114

Michael E. Seidel and William M. Palmer

HW

LC

Fig. 7A. Anterior view of the head illustrating length of the tomial cusp (LC)

and head width (HW).

Barr et al. 1976) was initially applied, thus avoiding assignment of

individuals to groups (species). Morphological similarity or divergence

was examined by observing clustering of individuals on bivariate plots

of their principal component scores. That provided a test to determine if

a priori species identifications based on qualitative characters could be

corroborated by mensural characters. It also provided a more objective

means to determine morphological overlap between species and possible

cases of hybridization or intergradation. If the a priori assignment of a

specimen had been noted as questionable (based on qualitative

characters) and its PCA plot was clearly outside its species cluster but

within the range of another species, it was reidentified. Otherwise,

taxonomic reassignment was avoided. Principal components analysis

was followed by stepwise discriminant analysis (BMDP7M, Dixon

1977). Discriminant analysis was applied to test for significant

morphometric differences between P. rubriventris, P. concinna, and P.

floridana. Sexes were again examined separately and the influence of

size (age) was reduced by linear regression. Size-adjusted residuals were

obtained from the 30 shell and head-stripe measurements by regressing

each character on carapace length.

Multivariate analysis of variance (MANOVA-SAS) followed by

Fisher's protected least significant difference (/-tests) were used to test

for utilitarian taxonomic characters that might provide a more objective

Morphological Variation in Pseudemys

115

Fig. 7B. Lateral view of the head and neck i!

supratemporal stripe (SS).

lustrating maximum width of the

Fig. 7C. Ventral view of the head and neck illustrating width of the post-

symphyseal stripe at level of tympanum (GS).

116

Michael E. Seidel and William M. Palmer

(quantitative) means for identifying species of Pseudemys. Thirty

character ratios were constructed from 27 of the original 30 characters.

These included character ratios that have been reported to be useful in

discriminating between P. concinna, P. floridana, and P. rubriventris.

Sexes were again treated separately. Although several of these ratios are

somewhat redundant and therefore strongly correlated, each was

examined to allow direct comparisons with previously reported values

(e.g. Ward 1984). Despite the theoretical problems with using ratios in

statistical analyses, their effectiveness in taxonomic studies of turtles has

been clearly demonstrated (Iverson and Graham 1990).

We also examined large series of hatchling and juvenile P.

rubriventris, P. floridana, and P. concinna. Young individuals were very

difficult to identify. Characters that we found to be diagnostic in adults

of these species were either impossible to resolve in young turtles or

extremely variable, even within a single brood of hatchlings.

oo

° * .8 o

O

PC II

Fig. 8. Plot of individual adult male Pseudemys based on principal components

analysis (PC II) and discriminant function analysis (DF 1) of morphometric

characters (see text). Open circles represent P. rubriventris, closed circles

represent P. concinna, and triangles represent P. floridana.

Morphological Variation in Pseudemys

117

STATISTICAL RESULTS

The first factor (PC I) extracted by principal components analysis

was size-related, as expected (Wiley 1981). It accounted for more than

50% of the total variance in male and female turtles and all loading

coefficients (eigenvectors) were high and positive (except angle of anal

scute, AA). PC II accounted for 24 and 26% of the remaining variance,

respectively, by sex. Among the 30 components extracted, only PC II

showed evidence of clustering by species. When individuals were plotted

according to their PC II scores (Fig. 8 and 9), P. rubriventris showed

distinct separation from P. concinna and P. floridana, which clustered

together and did not appear morphologically distinct. The most

influential mensural characters loaded on PC II are identified in Table

2. Two male specimens (NCSM 11365 and 13812) from extreme

northeastern North Carolina (Gates Co.) had been tentatively identified

as P. rubriventris. Because these two individuals plotted well outside the

0 _

■3 _

7£u

y,°

PC II

Fig. 9. Plot of individual adult female Pseudemys based on principal components

analysis (PC II) and discriminant function analysis (DF 1) of morphometric

characters. Open circles represent P. rubriventris, closed circles represent P.

concinna, and triangles represent P. floridana.

1 18 Michael E. Seidel and William M. Palmer

range of P. rubriventris on PC II and within the range of P. floridana,

they were reidentified as P. floridana. Another male specimen (NCSM

28704) from the same general area (Beaufort Co., N.C.) had been

tentatively identified as P. floridana, but because it plotted within the

range of P. rubriventris on PC II, it was reassigned to the latter species

for further analysis.

Discriminant analysis of male and female Pseudemys also revealed

marked separation between P. rubriventris and the other two congeners

(Fig. 8 and 9). Eighty-eight percent (females) and ninety-four percent

(males) of the variance was explained by the first discriminant function

(DF 1). Coefficients for the most influential characters are listed in

Table 2. Although differences (P<0.05) were found between all three

species, significance values were much larger comparing P. rubriventris

with P. concinna and P. floridana {F - 18-42) than comparing P.

concinna with P. floridana (F '= 3.0 and 6.9). For males, 100% of the P.

rubriventris were classified (by group discriminant function) correctly,

whereas there was a 17-18% classification error between P. concinna

and P. floridana. For females, 87% of the P. rubriventris, 78% of the P.

concinna, and 80% of the P. floridana were correctly grouped.

Table 2. Coefficients and factor loadings for the most influential morphometric

characters of discriminant analysis (DF 1) and principal components analysis

(PC II).

Morphological Variation in Pseudemys 1 19

Multivariate analysis of variance (Wilks' criterion) for 30 character

ratios and 3 unadjusted characters indicated significant differences

(P<0.01) between species. Differences (P<0.01) were found in 15

characters for males and 15 characters for females. Fisher's test indicated

that most of these characters separate P. rubriventris from P. concinna

and P. floridana (Table 3). The only characters that separate P. concinna

and P. floridana are based on head and neck markings and shell height

(P<0.05). The following characters and character ratios showed no

significant difference (7>>0.05) between species: PG, CW/CL, CD/CL,

CD/CW, CD/SW, CV/CD, MA/MD, MP/MH, MH/MA, IL/PL,

IE/CL, PX/XW, XW/PL, ET/PL, EP/PL, NR/CL.

CHARACTER ANALYSIS

Morphometric analysis of P. rubriventris, P. concinna, and P.

floridana revealed several measurements that distinguish those species in

northern and middle Atlantic slope drainages. In all Pseudemys, females

have considerably deeper shells than males. However, interspecific

comparisons of the same sex indicated a deep shell (SH and CH) in P.

rubriventris, an intermediate depth in P. floridana, and a shallow shell

in P. concinna (Table 3). These results are similar to differences in shell

height reported by Seidel (1981), Ward (1984), and Weaver and Rose

(1967). Head-striping patterns also distinguish these three forms. The

broadest gular (post-symphyseal) stripes (GS) and supratemporal stripes

(SS) are seen in P. concinna, moderate stripes are found in P. floridana,

and the narrowest stripes occur in P. rubriventris (Table 3, Fig. 10).

Several additional mensural characters distinguish P. rubriventris from

P. concinna and P. floridana, but do not separate those two from each

other. Compared with P. concinna and P. floridana, P. rubriventris has

a longer cervical scute (CU, CS), shallower anal notch (AN), broader

anal scute angle (AA), wider lateral angle (slope) of carapace (SA),

greater ventral extension of the posterior carapace (AV, AP), greater

head width (HW), and longer tomial cusps (LC) (Fig. 11 and 12). A

long cervical scute and prominent tomial cusps have frequently been

cited as diagnostic characteristics of the redbelly turtles, P. rubriventris,

P. nelsoni, and P. alabamensis (Carr 1952, Weaver and Rose 1967,

Ward 1984). Weaver and Rose (1967) noted ventral projection of the

carapace (pygal bone) in P. rubriventris but not in P. concinna and P.

floridana, and Ward (1984) reported a deeper anal notch in P. concinna

and P. floridana compared with redbelly turtles. Angle of the anal scute

(xiphiplastron) is a characteristic that shows pronounced sexual

dimorphism (Fig. 12).

Ward (1984) described the following scute and shell characters that

reportedly distinguish P. floridana from P. concinna: wider cervical

o

I

o

x

o,

O

CO

_<L>

X)

Rl

H

<3

2 rM

5, in

^r o^

<*> So

Si

£! oo r» o

ro w y> i

J^ r^ <n «o

^ ^. - ^

-,; <n on o

2 "H -H -H

fLJ on m no

■^- m m

NO NO Tjf

<N — ! — I

-H -H -H

esj "^ no

r-»" «ri cnJ

tt m m

no o ~Z —

J. V J rn

r» on

oo O

-H -H

O NO

wS r-'

^i^~

en. en CM ;_J

/-N NO O 3

^67

ON

<N 2 O ^ V

So - £ ~ S? - ">

9. 2 2 d 9

+122*+! +i

-h t: s S -h 2 2

^ £ i oo ^ 5 jj

* 5 -• « ° «

Tf ON ~~

2 ON'

en °? *?

M- CM OO

r- r- ^ »r>

*- — * cm r--

"J © "7 cm

r^ — cm v->

OO N— ' W N_^

^ NO oo en

^ - - vb

-H-H-H+j^-H-H-H

Z^NOcnH ilr-cMCM

no en

cm'

o en

1 ^

oo ?

"t — _• CM

— 1* ON «Ti

oo r~-

cs oq

rn — '

-H -H

SO r^-i

rn o

"0 o

S5

-H +1

«0 OO ^_J <N

~z? oo 3 >r

+1 T; 3 +1

^+^

^ O /-s * *

<n — i r- >v ?-s

r- — > -^ ^ m no

1 i ' i rN —

o r- — A i i

m »o — g, oo no

^f oq ^T °^ ^ p

On — ! — ' 9 <^i CN

-H r, -h ij -H -H

- -H ^

CM

ON

V~)

Cn|

On OO

^■' d

(S (N M rr

Tt oo r- o

<N U-> On i S

m no rn Nn c-

— rn o

ON

— NO ^-N

— uo

r-N ON ^

8 2?~

— CT «o en

O <N

yQ W> W« ■ ' "/I OO

^ ^ r^ O «0 "O

j w. n w " " -; w

oo SO — — i en ^

«N rf On O

NO en <N

NO CM*

m cm

co oq

CM — -

■H -H

Tfr OO

no *ri

/-S NO ^

tj- ^-s m m

r^ — i — rM

oo

CM

on i oo o

m r^ on — ^3

rt r^ CM »?T ^

On' O On' CM* 9

■H-H-H-H^

r-# «n on vq n

•n oo' on m' ***

^-n CM

O — S ^

oo .1 i NO

«n \0 *_J o

>n

m

oC sp

wn —

o

oo o SoOooSoo

Qu^Uu^Uu<n-)^

o o o o

0050

S S x S

V "r- 1 »T<

" - - —*■' J c^ O

C J/N

.2 -a

£ c

2 *

.S U

»-. ^-^

0. ja

o "§

U f

x x:

(U on

g s

<u

-8 s

* CO

55 J=

^ o ?

S o *

50 ^

3 g

t_ o

«6^

o y c

C NV <D

c ^ ^

CO o <u

. C X)

"3 CJ j_,

CJ t- —

CO u 7i

2 S S

J- •- u

c "° iS

SS'-a

a § >>

e Ha c

— B at

C/3 00 O

OX)

c

cd

"O

c

cd

C

O

RJ X

> o

T3 o

■3 «

Cu *-

00 aJ

O 00

^ c

S

cd

<L>

5 3 S

5 co B

1/3 'vT U

O S» j-j

-H

X S

_ ^ B

^ <~ <

O ° .

.2 3

si I

2 £^

Morphological Variation in Pseudemys

121

X

\

if)

150.

140_

130_

120_

110.

100_

090 _

080

070

.060 _

050.

C

c

F

070

T~"

090

GS/HW

Fig. 10. Graph of post-symphyseal stripe width/head width (GS/HW) versus

supratemporal stripe width/head width (SS/HW). Open rectangles (females)

and shaded rectangles (males) are formed by lines two standard errors above

and two standard errors below means. C = P. concinna, F = P.floridana, R = P.

rubrivenths.

scute (CD/CL, CD/CW, CD/SW), longer cervical scute underlap-

ventral length (CU/CS), greater anterior extension of cervical scute

(NR/CL), deeper curve of epiplastron (EP/PS), greater ratio of

anterior/ posterior width of marginal XII (MA/ MP), broader angle of

anal scutes (AA), and greater ratio of anterior/ posterior plastral lobe

width (PW/XW). In our sample that compares P. concinna concinna

with P. floridana floridana, we found no differences in these characters

and thus conclude that they have no taxonomic value in separating

species in the central Atlantic drainages. Unfortunately, Ward (1984)

did not provide a list of the specimens he examined. Most of Ward's

characters do appear to separate the Florida subspecies, P. floridana

peninsularis, from P. concinna (personal observation). Therefore, we

assume that Ward's interspecific comparisons of P. floridana were based

primarily or exclusively on character analysis of P.f peninsularis.

Many of the qualitative characters that have been used to define

species of Pseudemys show considerable intraspecific variation in Atlantic

slope populations. Among 15 characters recorded, only five were found

to have taxonomic value. Of the P. concinna examined, 63% showed

122

Michael E. Seidel and William M. Palmer

Dl

<

CS/CL

Fig. 1 1. Graph of cervical scute length/ carapace length (CS/CL) versus ventral

extension of posterior carapace (AV/ AP). Symbols are defined in Fig. 10.

evidence of a C-shaped mark on the second pleural scute, whereas only

6% of the P. floridana and 4% of the P. rubriventris had this marking.

As reported in earlier literature, the lateral yellow lines on the pleural

scutes of P. floridana (Fig. 14) form irregular bands or bars, whereas

these markings are more circular, forming ocelli, in P. concinna (Fig.

13). A dark figure on the plastron was detected in 61% of the P.

rubriventris and 35% of the P. concinna (Fig. 15), whereas 96% of the

P. floridana showed no evidence of plastral markings (Fig. 16). The

submarginal circles anterior to the bridge were solid (spots) in 42% of

the P. rubriventris, 28% of the P. floridana, and 6% of the P. concinna.

Ward (1984) reported that the anterior submarginal spots are solid

blotches in P. floridana, whereas in our sample, 72% of the P. floridana

had open circles. The apex of the lower jaw (viewed ventrally) was

rounded, not angled, in 51% of the P. rubriventris, 2% of the P.

floridana, and in none of the P. concinna. There were more than 1 1

prominent head stripes (at level of the posterior margin of the tympanum)

in 94% of the P. concinna, 10% of the P. rubriventris, and 6% of the P.

floridana examined.

As in the mensural characters, several of the qualitative characters

that reportedly distinguish these species do not effectively separate them

in the central Atlantic coast drainages. Ward (1984) stated in his

Morphological Variation in Pseudemys

123

(f)

95 0_

900_

F

C

50 0

65 0

75 0

AA

Fig. 12. Graph of lateral angle of carapace (SA) versus angle of anal scutes

(A A). Symbols are defined in Fig. 10.

definition of the subgenus Ptychemys that redbelly turtles (including P.

rubriventris) have posterior marginals without a notch. Of the P.

rubriventris we examined, 62% had posterior marginals that were

serrated (offset at seam) and clearly notched. Ward (1984) and Weaver

and Rose (1967) reported that the redbelly turtles (in contrast to P.

concinna and P. floridana) have a strongly rugose carapace (Fig. 17),

occasionally even as juveniles. In our sample, we found pronounced

carapacial rugosity in 22% of the P. concinna, 45% of the P. floridana,

and 49% of the P. rubriventris. These rugosities were observed exclusively

in large (old) adults. Ward (1984) reported that the inguinal scute of P.

floridana (in contrast to P. rubriventris and P. concinna) lacks any

black markings. We found that 74% of the P. floridana in our sample

had black markings on the inguinal scute. However, these markings are

usually absent in P. f peninsularis (personal observation). Ward also

reported that inframarginal spots posterior to the bridge are mostly

confined anterior to the seam in P. floridana. In 40% of the P. floridana

we examined, these inframarginal spots broadly overlap the seams.

In old male Pseudemys from the Atlantic drainage areas, melanism

was detected in 2% of the P. floridana, 6% of the P. concinna, and 29%

of the P. rubriventris. Melanism in P. rubriventris was not only more

frequent, but also more complete compared with the other two species.

124 Michael E. Seidel and William M. Palmer

The pattern of male melanism observed in all species was a loss of

yellow lines on the soft parts and carapace and development of a

reticulated (worm-like) pattern of dark speckled markings on the head,

carapace, and plastron (Fig. 18). This pattern is quite different from the

melanism that develops in populations of Trachemys scripta (Lovich et

al. 1990), but similar to that of the Cuban slider, Trachemys decussata

decussata (Seidel 1988). A different form of melanism is found in adult

female and young male P. rubriventris in northern portions of their

range (Pennsylvania and New Jersey). In those areas, the soft parts and

carapace become nearly solid black, but the plastron remains bright

coral or red. It is interesting that darkening to this extent and loss of

yellow lines apparently do not occur in the southern populations in

Virginia and North Carolina.

DISCUSSION

Morphological divergence between P. concinna and P. floridana is

much less than their collective divergence from P. rubriventris. Neither

principal components analysis nor discriminant analysis clearly separated

the two former taxa from each other (Fig. 8 and 9). Male specimens of

P. rubriventris showed no overlap with P. concinna and P. floridana,

but some overlap was observed for females. Six female P. rubriventris

(NCSM 20166, 28753, 28897, 29278; AMNH 90644; USNM Field Series

159366) and two female P. floridana (NCSM 14783; CM Field Series

24447) plotted intermediately between species clusters (Fig. 9). Specific

identification of all these individuals was originally noted as questionable,

and all were collected from the relatively small area of southeastern

Virginia and northeastern North Carolina where the ranges of these

species overlap. There is little doubt that P. rubriventris in the southern

portion of its range is somewhat morphologically convergent with P.

floridana. One possible explanation for this is that reproductive isolation

is not complete and a limited amount of gene flow occurs between these

species. That would support Crenshaw's (1965) proposal of hybridization

in the region. Another explanation is that selection pressures are similar

in this area of sympatry, resulting in homoplastic (convergent) character

states (as suggested for other geographic regions by Carr 1952 and

Ward 1984).

Two of the morphologically intermediate specimens from north-

eastern North Carolina (NCSM 29278 from Dare Co. and 14783 from

Gates Co.) strongly suggest hybridization of P. rubriventris with other

Pseudemys . Skulls were prepared from these two turtles to examine

osteological characters that have been used to distinguish P. rubriventris

from P. concinna and P. floridana (McDowell 1964, Ward 1984). One

Morphological Variation in Pseudemys 125

specimen (NCSM 29278) is an adult female (315 mm carapace length)

with a "C" on the second pleural scute and broad gular and

supratemporal head stripes (13 and 15% of head width) similar to P.

concinna or P. floridana. However, it has a relatively long cervical scute

(8% of carapace length), well-defined premaxillary notch, and broad

xiphiplastron angle (75°) as in P. rubriventris. The skull of that turtle is

also clearly intermediate. The vomer marginally contributes to the

triturating (alveolar) surface, and the lateral edge of the dentary is

weakly serrated. Alveolar width on the dentary surface is 15% of the

condylobasal length, and maxillary alveolar width is 21%. In eight P.

rubriventris skulls that we examined, the dentary width ranged from 17

to 21%, and the maxillary width ranged from 20 to 25%. In 11 P.

concinna and P. floridana examined, the dentary width ranged from 13

to 16%, and the maxillary ranged from 16 to 20%. The other intermediate

specimen (NCSM 14783) is an adult female (254 mm carapace length)

without a "C" mark on the second pleural scute and with fairly narrow

gular and supratemporal stripes (7-8% of head width) as in P.

rubriventris. However, similar to P. concinna and P. floridana, the

cervical scute is less than 8% of the carapace length, the lateral angle of

the carapace is 95°, and the xiphiplastron angle is 55°. The skull of

NCSM 14783 is also somewhat intermediate. Although the vomer does

not project to the alveolar surface, the lateral edge of the dentary is

weakly serrated. Alveolar width on the dentary surface is 17% of the

condylobasal length, and the maxillary width is 20%. Both NCSM

29278 and 14783 are thus morphological intermediates (presumed

hybrids). The former more closely resembles P. rubriventris, whereas

the latter is more similar to P. floridana.

Ward (1984) indicated that markings and coloration are too variable

to be reliable for diagnosing cooter species. If, as Ward suggests, only

osteological characters reliably separate P. concinna from P. floridana,

those species should be considered cryptic (sibling) based on their

external morphology. That would also imply that there are two clearly

recognizable osteomorphs, with little intergradation or polymorphism.

Because large series of skeletal material taken throughout the range of

Pseudemys have not been examined, there is no basis for that conclusion.

Hedges (1990) appropriately stated that "speciation is a dynamic process

and we should expect borderline cases." Results from the present study

suggest that the relationship between P. concinna and P. floridana in

the Atlantic drainages of North Carolina is more characteristic of

subspecies than species. Nearly all of the typical examples of P. concinna

occur in the piedmont, whereas individuals easily identified as P.

floridana are in the coastal plain (Fig. 1). That is similar to Carr's (1952)

observations that led him to consider the two forms subspecies of P.

126

Michael E. Seidel and William M. Palmer

Fig. 13. Dorsal view of a young male P. concinna from Guilford Co., N.C.

(NCSM 30278).

Morphological Variation in Pseudemys

127

j~

Fig. 14. Dorsal view of an adult male P. floridana from Pender Co.

(NCSM 30475).

N.C.

128

Michael E. Seidel and William M. Palmer

Fig. 15. Plastral view of a young male P. concinna from Guilford Co., N.C.

(NCSM 30278).

Morphological Variation in Pseudemys

129

Fig. 16. Plastral view of an adult male P. floridana from Pender Co.

(NCSM 30475).

N.C.

130

Michael E. Seidel and William M. Palmer

Fig. 17. Dorsal view of an adult melanistic male P. rubriventris from Dare Co.

N.C. (NCSM 22818).

Morphological Variation in Pseudemys

131

Fig. 18. Plastral view of an adult melanistic male P. rubriventris from Dare

Co., N.C. (NCSM 22818).

132 Michael E. Seidel and William M. Palmer

floridana. Many of the specimens in the area of the Fall Line appear to

be intermediate between P. concinna and P. floridana, forming an

apparent zone of intergradation. In a few instances typical P. floridana

and P. concinna appear to be in geographic proximity, but they are

rarely, if ever, observed in microsympatry. Pseudemys concinna is

found in rivers or impoundments, and P. floridana frequents more

lentic habitats, which include backwaters of coastal-plain rivers. Typical

examples of these turtles may be distinguished readily by their markings,

but there is no consistent external difference in cranial or shell

morphology, except perhaps shell depth. Furthermore, some of the

diagnostic characters that separate these species elsewhere in their

ranges apparently have little diagnostic value for P. concinna and P.

floridana on the Atlantic slope. In spite of these observations, we feel it

is premature to propose a conspecific relationship for the two taxa. The

senior author (MES) is currently examining morphological variation in

P. floridana and P. concinna throughout their entire ranges. Results

from that analysis, particularly an evaluation of P. f peninsularis,

should provide critical data for taxonomic decisions.

KEY TO ADULT PSEUDEMYS

IN ATLANTIC COAST DRAINAGES

1. Upper jaw with prominent notch bordered on each side by tooth-

like cusps (length of cusp 3-7% of head width). Gular and

supratemporal stripes narrow, 5-8% of head width. Cervical scute

long, 8-9% of carapace length. Lateral angle (slope) of carapace

steep, 1 10-1 17° in females and 100-106° in males. Angle formed by

lateral edges of xiphiplastron (anal scutes) broad, 68-75° in females

and 64-72° in males. Plastral ground color in living specimens pink

or coral P. rubriventris

Upper jaw with only a very shallow notch or notch entirely absent.

Cusps either absent or very small (length less than 3% of head

width). Gular and supratemporal stripes broad, 8-12% of head

width. Cervical scute short, 7-8% of carapace length. Lateral angle

(slope) of carapace moderate, 97-109° in females and 87-95° in

males. Angle formed by lateral edges of xiphiplastron (anal scutes)

sharp, 58-67° in females and 44-55° in males. Plastral ground color

in living specimens pale yellow to orange 2

2. Head and neck stripes numerous, at posterior edge of tympanum

more than 11. C-shaped mark on second pleural scute and/ or a

dark figure on the plastron. Posterior shell depth (at vertebral III-

IV sulcus) 33-36% of carapace length in females and 30-32% in

Morphological Variation in Pseudemys 133

males. Plastral ground color in living specimens yellow to orange.

P. concinna

Head and neck stripes few, less than 11. No C-shaped mark on

second pleural scute and figure absent from plastron. Posterior shell

depth (at vertebral III-IV sulcus) 36-38% of carapace length in

females and 32-33% in males. Plastral ground color in living

specimens pale yellow P. floridana

SPECIMENS EXAMINED

Pseudemys rubriventris

New Jersey: MES 132, 1751. North Carolina: MES 1896. NCSM 9360,

16669-70, 16672, 17910, 20116, 20166, 22818, 25080, 28658-59, 28704,

28752-53, 28871, 28897, 28899, 29275-76, 29278, 30034. CM 53026.

AMNH 72746, 80218-19, 81869, 90640-44. Pennsylvania: CM 27420,

28969, 29400, 29457, 29502, 31244, 32651. Virginia: AMNH 129302,

129312. CM 13262, 23136. CM(field series) 53632, 53634, 53636, 53695.

USNM (field series) 114462, 140441-47, 140759, 157085-88, 157853,

158685, 158881, 159362, 159366, 159568-69, 159572-73. MES 188-90.

West Virginia: CM 26630. MES 1902.

Pseudemys concinna

North Carolina: NCSM 6184, 8518, 11364, 11373, 13759, 13810, 13966,

15030, 15135, 17045, 17276, 17339, 17938, 19169, 19356, 19432,20128,

20236, 20240, 20253, 22966, 24030, 24182, 25044, 25281, 25205-06,

25234, 25265, 26061, 26225, 26525-27, 28688, 29279, 29595, 29968-69,

30038, 30280-81, 30431-34. USNM(field series) 158447. South Carolina:

MES 1790, 1875. SREL 2137, 2229. Virginia: USNM(field series)

141102, 141105, 141364, 158811. MES 489. West Virginia: MES 863.

Pseudemys floridana

North Carolina: NCSM 5881, 5883-85, 5927-30, 10330, 1 1365, 13812-16,

14783, 16476, 17046, 17302, 17581-82, 19353-54, 20190, 20678, 20928,

21001-02, 21603, 23405, 24334, 25737, 25739, 25833, 26344, 26528,

28657, 28705, 28740, 28898, 29301, 29617, 30001, 30291-92, 30423-30.

USNM(field series) 158448. South Carolina: AMNH 50985. SREL

0117. Virginia: CM 24447. MES 1900. USNM(field series) 159365.

ACKNOWLEDGMENTS.— We are indebted to the following

persons and institutions for loan of museum material: W. Auffenburg

and D. Auth, Florida Museum of Natural History; J. Cole and M.

Klemens, American Museum of Natural History; J. W. Gibbons,

Savannah River Ecology Laboratory; C. J. McCoy, Carnegie Museum

134 Michael E. Seidel and William M. Palmer

of Natural History; and G. R. Zug of the National Museum of Natural

History (Smithsonian Institution). We acknowledge Marshall University

for providing computer time and sabbatical leave to M.E.S. We thank

Lu Ann South for typing the manuscript, Steve Gotte and Tom Jones

for assistance in measuring specimens, and R. Kuhler of the North

Carolina State Museum of Natural Sciences for the illustrations of

specimens.

LITERATURE CITED

Barr, A. J., J. H. Goodnight, J. P. Sail, and J. T. Helwig. 1976. SAS User's

Guide: Statistics. SAS Institute Inc., Cary, N.C.

Carr, A. 1952. Handbook of Turtles. Cornell Univ. Press.. Ithaca, N.Y.

Collins, J. T. 1990. Standard common and current scientific names for North

American amphibians and reptiles. 3rd ed. Herpetol. Circ. Soc. Study

Amphib. Reptiles (19): 1-4.

Conant, R., and J. T. Collins. 1991. A Field Guide to Reptiles and Amphibians.

Eastern and Central North America. Houghton Mifflin Co., Boston.

Crenshaw, J. W. 1955. The ecological geography of the Pseudemys floridana

complex in the southeastern United States. Ph.D. dissertation, Univ.

of Florida, Gainesville.

Crenshaw, J. W. 1965. Serum protein variation in an interspecies hybrid

swarm of turtles of the genus Pseudemys. Evolution 19:1-15.

Dixon, W. J. 1977. BMD Biomedical Computer Programs. Univ. California

Press, Berkeley.

Ernst, C. H., and R. W. Barbour. 1972. Turtles of the United States. Univ.

Press, Kentucky, Lexington.

Ernst, C. H., and R. W. Barbour. 1989. Turtles of The World. Smithsonian

Inst. Press, Washington, D.C.

Fahey, K. M. 1980. A taxonomic study of the cooter turtles, Pseudemys

floridana (Le Conte) and Pseudemys concinna (Le Conte), in the lower Red

River, Atchafalaya River, and Mississippi River basins. Tulane Stud. Zool.

22:49-66.

Hedges, S. B. 1990. Species concepts and speciation. Syst. Zool. 39:95-96.

Iverson, J. B., and T. E. Graham. 1990. Geographic variation in the redbelly

turtle, Pseudemys rubriventris (Reptilia: Testudines). Ann. Carnegie Mus.

59:1-13.

King, F. W., and R. L. Burke. 1989. Crocodilian, Tuatara, and Turtle Species

of the World. Assoc. Syst. Collections, Washington, D.C.

Leviton, A. E., R. H. Gibbs, E. Heal, and C. E. Dawson. 1985. Standards in

herpetology and ichthyology: Part 1. Standard symbolic codes for institu-

tional resource collections in herpetology and ichthyology. Copeia

1985:802-832.

Lovich, J. E., C. J. McCoy, and W. R. Garstka. 1990. The development and

significance of melanism in the slider turtle. Pages 233-254 in Life History

and Ecology of the Slider Turtle (J. W. Gibbons, editor). Smithsonian

Inst., Washington, D.C.

Morphological Variation in Pseudemys 135

Martof, B. S., W. M. Palmer, J. R. Bailey, and J. R. Harrison III. 1980.

Amphibians and Reptiles of the Carolinas and Virginia. Univ. North

Carolina Press, Chapel Hill.

McDowell, S. B. 1964. Partition of the genus Clemmys and related problems

in the taxonomy of the aquatic Testudinidae. Proc. Zool. Soc. London

143:239-279.

Seidel, M. E. 1981. A taxonomic analysis of pseudemyd turtles (Testudines:

Emydidae) from the New River, and phenetic relationships in the subgenus

Pseudemys. Brimleyana 6:25-44.

Seidel, M. E. 1988. Revision of the West Indian emydid turtles (Testudines).

Am. Mus. No vitates 2918:1-41.

Seidel, M. E., and H. M. Smith. 1986. Chrysemys, Pseudemys, Trachemys

(Testudines: Emydidae): Did Agassiz have it right? Herpetologica 42:242-248.

Smith, H. M., and R. B. Smith. 1980. Synopsis of the Herpetofauna of

Mexico, Vol. VI: Guide to Mexican Turtles. John Johnson, North

Bennington, Vt.

Ward, J. P. 1984. Relationships of chrysemyd turtles of North America

(Testudines: Emydidae). Spec. Publ. Texas Tech. Mus. 21:1-50.

Weaver, W. G., Jr., and F. L. Rose. 1967. Systematics and fossil history of the

genus Chrysemys. Tulane Stud. Zool. 14(2):63-73.

Wiley, E. O. 1981. Phylogenetics: The Theory and Practice of Phylogenetic

Systematics. John Wiley Sons, New York.

Accepted July 1991

136

STATE LIBRARY OF NORTH CAROLINA

3 3091 00748 4827

NEW RELEASE

THE FRESHWATER FISHES

OF

NORTH CAROLINA

by

Edward F. Menhinick

North Carolina Wildlife Resources Commission, 1991. i-vi +

227 pages. Index and glossary. Available by mail from

Larkin Distributors, 8924 Nottoway Drive, Charlotte, NC

28213. Single copies $34.70 plus $3 for postage and handling.

INFORMATION FOR CONTRIBUTORS

Submit original and two copies of manuscripts to Editor, Brimleyana, North Carolina

State Museum of Natural Sciences, P.O. Box 27647, Raleigh, NC 27611. In the case of

multiple authorship, indicate correspondent. Manuscripts submitted for publication in this

journal should not currently be submitted elsewhere.

Preparation of manuscript. Adhere generally to the Council of Biology Editors Style

Manual, Sixth Edition. Use medium-weight bond paper, 8V2 x 1 1 inches, and leave at least

an inch margin on all sides. Double-space all typewritten material, including tables and

literature cited.

The first page will contain only the title, the name(s) and address(es) of the author(s),

and necessary footnotes to those. Where appropriate, the title will indicate at least two

higher categories to which taxa belong. Example: Studies of the genus Hobbseus Fitzpa-

trick and Payne (Decapoda: Cambaridae).

A brief informative abstract on a separate sheet follows the title page. Indicative

abstracts are not acceptable. Footnotes, other than those in tables (see below), will be used

only when absolutely necessary, and will be numbered consecutively throughout the

paper.

Individuality of writing style and text organization are encouraged, but most papers are

best presented in the form of Introduction, Materials and Methods, Results, Discussion,

and Literature Cited, with those headings centered and all capitals. Headings plus sub-

headings must be kept to a total of three levels.

Scientific names in taxonomic papers will include the author when first used in the text.

Descriptions of new taxa must be in accordance with the requirements of established

international codes. Presentations of etymologies are desirable.

Form for literature cited:

Woodall, W. R., Jr., and J. B. Wallace. 1972. The benthic fauna in four small southern

Appalachian streams. Am. Midi. Nat. 88:393-407.

Crocker, D. W., and D. W. Barr. 1968. Handbook of the Crayfishes of Ontario. Univ.

Ontario Press, Toronto.

Authors, not the editor, are responsible for verifying references.

Form for citing references in text: parenthetical (Woodall and Wallace 1972). For more

than two authors use "et al." (not italicized).

All tables go on separate sheets at the end of the manuscript. Do not use vertical lines in

tables. Footnotes in each table will be indicated by consecutive superscript lower-case

letters. Indicate lightly in pencil in the margin of the original manuscript where tables and

illustrations would best fit.

Preparation of illustrations. Illustrations, including maps, graphs, charts, drawings, and

photographs, should be numbered consecutively as figures. They should not be larger than

21.5 x 28 cm (8V2 x \\ inches). They should be proportioned to make optimal use of the

page format of the journal. Plates must be prepared and presented as they are to appear,

and should not be presented as collections of individual items for arrangement by the

editors. Do not mount individual photographs. The author's name, title of the manuscript,

figure number, and the notation "Top," should be penciled lightly on the back of each

illustration. Lettering on original drawings and maps should be of adequate size to permit

proper reduction when needed. Do not type on illustrations; legends should be typed,

double-spaced, on separate sheets.

Page charges, reprints, and proofs. A charge of $30 per page is expected from authors

who have funds available from institutions, grants, or other sources. Those without such

funds should so indicate in their correspondence with the Editor at time of submittal. This

will not affect acceptance for normal publication. Contributors who pay full page costs

will be furnished 100 free reprints. Reprint order forms will be sent with page proofs and

are to be returned to the Managing Editor. On papers with more than one author, it is the

responsibility of the correspondent to assure that other authors have an opportunity to

obtain reprints. Both galley and page proofs are to be corrected, signed, and returned to

the Managing Editor within seven days. Please notify the Managing Editor if additional

time is required for proofreading.

BRIMLEYANA No. 17, DECEMBER 1991

CONTENTS

Recent Catastrophic Decline of Mussels (Bivalvia: Unionidae) in the Little South Fork

Cumberland River, Kentucky. Robert M. Anderson, James B. Layzer, and

Mark E. Gordon 1

Ants and Cockroaches Trapped Outside Suburban Houses in the Area of Raleigh, Wake County,

North Carolina. C. G. Wright, T P. Nuhn, and H. E. Dupree, Jr 9

Bats (Chiroptera: Vespertilionidae) of the Great Dismal Swamp of Virginia and North Carolina.

Thomas M. Padgett and Robert K. Rose 17

Occurrence of an Introduced African Cichlid, the Blue Tilapia, Tilapia aurea, in a Tidal Creek of

the Skidaway River, Georgia. L. Stanton Hales, Jr 27

Home Range and Substrate Use by Two Family Groups of Red-cockaded Woodpeckers in the

North Carolina Sandhills. Richard R. Repasky and Phillip D. Doerr 37

Mating and First-season Births in Interstate Transplanted River Otters, Lutra canadensis

(Carnivora: Mustelidae). Peter J. Tango and Edwin D. Michael , 53

Habitat Associated With Home Ranges of Female Odocoileus virginianus (Mammalia: Cervidae)

in Eastern Kentucky. Richard C. Pais, William C. McComb, and John Phillips 57

Cation Concentrations and Acidity in Breeding Ponds of the Spotted Salamander, Ambystoma

maculatum (Shaw) (Amphibia: Ambystomatidae), in Virginia. Charles R. Blem and

Leann B. Blem 67

Spawning Activities of Notropis chlorocephalus, Notropis chiliticus, and Hybopsis hypsinotus,

Nest Associates of Nocomis leptocephalus in the Southeastern United States, With Comments

on Nest Association (Cypriniformes: Cyprinidae). Carol E. Johnston 77

Consumption Rates, Evacuation Rates, and Diets of Pygmy Killifish, Leptolucania ommata, and

Mosquitofish, Gambusia affinis (Osteichthyes: Atheriniformes) in the Okefenokee Swamp.

J. Douglas Oliver 89

Morphological Variation in Turtles of the Genus Pseudemys (Testudines: Emydidae) From

Central Atlantic Drainages. Michael E. Seidel and William M. Palmer 105