MICROHABITAT USE AND MOVEMENTS OF GILT DARTERS (PERCINA EVIDES) IN

TWO SOUTHEASTERN STREAMS

by

JESSICA PAIGE SKYFIELD

(Under the Direction of Gary David Grossman)

ABSTRACT

The patterns of abundance and resource use for most stream fishes are largely unknown.

Quantifying resource use of these organisms is a prerequisite for understanding a species’ ecology for theoretical and practical applications. I examined microhabitat use by gilt darters

(Percina evides) in two Southeastern streams. I used Principal Component Analysis to examine patterns of use. Darters used microhabitats with higher percent cobble and average velocities than randomly available with male darters deviating from random more than females. In the size- based analyses, larger gilt darters tended to use microhabitats with more heterogeneous substrata

and larger amounts of boulder than the smaller size classes. I also conducted a short-term

movement study and population estimates. Darters did not seem to conform to the Restricted

Movement Paradigm. Nonetheless, 40% of movements were within 5 meters of initial capture.

The population density of 0.31 darters/m2 is a point estimate; nonetheless, it provides novel information.

INDEX WORDS: microhabitat, Percina, darters, principle component analysis (PCA), benthic, movement, population estimates

MICROHABITAT USE AND MOVEMENTS OF GILT DARTERS (PERCINA EVIDES) IN

TWO SOUTHEASTERN STREAMS

by

JESSICA PAIGE SKYFIELD

B.A., Rhodes College, 2002

A Thesis Submitted to the Graduate Faculty of The University of Georgia in Partial Fulfillment

of the Requirements for the Degree

MASTER OF SCIENCE

ATHENS, GEORGIA

2006

© 2006

Jessica Paige Skyfield

All Rights Reserved

MICROHABITAT USE AND MOVEMENTS OF GILT DARTERS (PERCINA EVIDES) IN

TWO SOUTHEASTERN STREAMS

by

JESSICA PAIGE SKYFIELD

Major Professor: Gary D. Grossman

Committee: Brett W. Albanese John P. Wares

Electronic Version Approved:

Maureen Grasso Dean of the Graduate School The University of Georgia May 2006

ACKNOWLEDGEMENTS

I would like to acknowledge all that have helped me throughout my academic career, especially during my thesis work at the University of Georgia. This includes, but is not limited to, my major professor, Gary Grossman, committee members Brett Albanese and John Wares, labmates and friends Megan Hill, Bob Rataczjak, Duncan Elkins, and Peter Hazelton, and last but not least my family, Mom, Dad, Gretchen, Walter, Parker, Linton, Isabelle, and my grandparents.

iv

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS...... iv

LIST OF TABLES...... vi

LIST OF FIGURES ...... vii

INTRODUCTION AND LITERATURE REVIEW ...... 1

CHAPTER

1 MICROHABITAT USE AND MOVEMENT OF GILT DARTERS (PERCINA

EVIDES) IN TWO SOUTHEASTERN STREAMS...... 12

SUMMARY AND CONCLUSIONS ...... 64

v

LIST OF TABLES

Page

Table 1: Univariate mean and standard deviation (SD) of microhabitat availability and overall

darter use data for all sites and seasons...... 36

Table 2: Number and percent of darters by season, sex, and size ...... 37

Table 3: Univariate mean and standard deviation (SD) of microhabitat availability and male and

female darter (51-59 mm) use for all sites and seasons ...... 38

Table 4: Univariate mean and standard deviation (SD) of microhabitat availability and size

sorted darter use for all sites and seasons...... 39

Table 5: Summary of capture and recapture history for positively identifiable darters in Tellico

Creek, Autumn 2005 ...... 40

Table 6: Number and percentage of overall darters moved, by sex, and size class in Tellico

Creek, Autumn 2005 ...... 40

Table 7: Summary statistics (mean, standard deviation) of darter movement in Tellico Creek,

Autumn 2005 (downstream movements are absolute value) ...... 40

Table 8: Results of univariate statistical comparisons of darter movement distances...... 41

Table 9: AIC calculations using log likelihood for testing models of varying capture-recapture

probabilities to estimate population size in Tellico Creek, Autumn 2005. Values and

models in bold are ones in the confidence set ...... 41

vi

LIST OF FIGURES

Page

Figure 1: Principle Component Analysis (PCA) of constrained random habitat comparisons

across sites and seasons: factors 1 and 2. Ellipses represent mean values and 95%

confidence intervals. Only loadings ≥ 0.40 are presented...... 42

Figure 2: Principle Component Analysis (PCA) of constrained random habitat comparisons

across sites and seasons: factors 1 and 3 ...... 43

Figure 3: Principle Component Analysis (PCA) of constrained random habitat comparisons

across sites and seasons: factors 2 and 3 ...... 44

Figure 4: PCA of Tellico Summer and Autumn unconstrained microhabitat availability ...... 45

Figure 5: PCA of Tellico Summer constrained and unconstrained microhabitat availability ...... 46

Figure 6: PCA of Tellico Autumn constrained and unconstrained microhabitat availability...... 47

Figure 7: PCA of Coweeta Spring microhabitat use for all darters compared to constrained

random microhabitat availability...... 48

Figure 8: PCA of Tellico Summer microhabitat use for all darters compared to constrained

random microhabitat availability...... 49

Figure 9: PCA of Tellico Summer microhabitat use for all darters compared to unconstrained

random microhabitat availability, factors 1 and 2 ...... 50

Figure 10: PCA of Tellico Summer microhabitat use for all darters compared to unconstrained

random microhabitat availability, factors 1 and 3...... 51

vii

Figure 11: PCA of Tellico Summer microhabitat use for all darters compared to unconstrained

random microhabitat availability, factors 2 and 3...... 52

Figure 12: PCA of Tellico Autumn microhabitat use for all darters compared to constrained

random microhabitat availability...... 53

Figure 13: PCA of Tellico Autumn microhabitat use for all darters compared to unconstrained

random microhabitat availability...... 54

Figure 14: PCA of Coweeta Spring sex linked microhabitat use ...... 55

Figure 15: PCA of Tellico Summer sex linked microhabitat use ...... 56

Figure 16: PCA of Tellico Autumn sex linked microhabitat use...... 57

Figure 17: PCA of Coweeta Spring size linked microhabitat use ...... 58

Figure 18: PCA of Tellico Summer size linked microhabitat use ...... 59

Figure 19: PCA of Tellico Autumn size linked microhabitat use...... 60

Figure 20: Movement frequency of darters in Tellico Creek Autumn 2005 ...... 61

Figure 21: Sex linked movement frequency of darters in Tellico Creek Autumn 2005...... 62

Figure 22: Size linked movement frequency of darters on Tellico Creek Autumn 2005...... 63

viii

INTRODUCTION AND LITERATURE REVIEW

Temperate lotic systems are spatially and temporally heterogeneous. This heterogeneity

creates patches, or microhabitats of varying qualities and types (e.g., Grossman et al. 1995;

Wiens 2002) and affects a variety of biological characteristics of lotic species including: 1) distribution (Levins 1963; Grossman et al. 1998), 2) movements (Mundahl and Ingersoll 1983;

Railsback et al. 1999; Petty and Grossman 2004), and 3) habitat choice (Freeman and Grossman

1993; Wu and Loucks 1995; Poff and Allan 1995; Grossman et al. 2002, Resitarits, Jr 2005).

Quantifying habitat use and movement within this patchy landscape is crucial to understand a species’ ecology (sensu Pulliam 1988; Hanski 1994; Wu and Loucks 1995). Despite this fact, the mechanisms affecting patterns of abundance and resource use for most stream fishes are largely unknown (Grossman et al. 2002). Nonetheless, quantifying resource use patterns of these organisms is a prerequisite for scientifically based management of these species, provides information on habitat variation and quality, and can aid in the construction of ecological models.

In lotic systems, fish habitat can be quantified on a variety of scales including macro

(river), meso (reach) or micro (position of fish) levels. In this study I focus on habitat use of gilt darters at the microhabitat scale.

Microhabitat

Microhabitat is defined as the suite of habitat characteristics present at the position of an individual fish at a given time (Baltz 1990). Fish perceive habitat as a constellation of physical

1

and biological factors and a number of environmental characteristics have been shown to affect microhabitat use: 1) substratum composition (Greenberg 1991), 2) current velocity (Chipps et al.

1993), 3) prey availability (Petty and Grossman 1996; Thompson et al. 2001), 4) predator avoidance (Schlosser 1987; Persson and Eklöv 1995), 5) season (Grossman et al. 2002; Carter et al 2004) or 6) combinations of both these and other variables (e.g. Thompson et al. 2001). The variables of importance to particular species determine their microhabitat use in streams, and this also may be affected by fish size or sex. Quantification of the importance of physical and biological factors to microhabitat use by stream fishes frequently is accomplished via direct, in situ observations (Grossman and Freeman 1987; Baltz 1990; Stauffer et al. 1996).

For some stream fishes, microhabitat (or patch) use patterns have been shown to be based on decisions that appear to result in increased fitness (Hill and Grossman 1993; Grossman et al.

2002; Rosenfeld 2003; Wildhaber and Lamberson 2004), such as higher prey availability or reduced energy expenditure. Microhabitat use data for several benthic stream fishes, i.e., Cottus bairdi, Rhinichthys cataractae, and Etheostoma nianguae, show that these species select microhabitats that put them in contact with more invertebrate prey (Petty and Grossman 1996;

Thompson et al. 2001; Mattingly and Galat 2004). By contrast, rainbow darters (E. caeruleum) appear to increase fitness by choosing microhabitats that decrease energy expenditure (Harding et al. 1998).

Many fish microhabitat studies focus either on both physical and chemical (abiotic) habitat characteristics (e.g., Strange et al. 2002; Schaefer et al. 2003) or biotic habitat characteristics (Matthews 1998). Studies focusing on both abiotic and biotic variables are not common and are needed for a more comprehensive understanding of fish microhabitat selection.

2

Movement

Spatial and temporal heterogeneity in habitat availability not only influence microhabitat use but also fish movement in lotic systems. Physical factors such as floods may increase both upstream and downstream movement (Albanese et al. 2004) in stream fishes, whereas droughts have equivocal effects (Matthews 1998; Albanese et al. 2004). Physiochemical changes such as temperature shifts also may influence fish movement (Schaefer et al. 2003). Other temporal or spatial changes in stream characteristics, such as presence of cover or water depth may decrease stream fish movement (Aparicio and Sostoa 1999). Changes in mesohabitat type (i.e, from pool to riffle) may decrease movement for both water column and mid-water column guild (sensu

Grossman et al. 2002) stream fish movement (Schaefer 2001).

Life history traits and differential physiological conditions may affect movement patterns and dispersal of stream fishes (Turner and Trexler 1998; Matthews 1998; Fraser et al. 2001).

Roberts and Angermeier (2005) found few correlations between life-history characteristics, movement and dispersal patterns of three darters (E. flabellare, E. podostemone, and Percina roanoka). However, life-history characteristics such as rheotactic movements to seasonal spawning areas precipitates stream fish movement for some species (Mathews 1998). Nocomis leptocephalus (Skalski and Gilliam 2000) and Cottus bairdi juveniles (Petty and Grossman 2004) have higher growth rates when they move (i.e., switch microhabitats) although adult C. bairdi display the opposite trend (Petty and Grossman 2004). Albanese et al. (2004) found that body size was not correlated with distance moved for Rhinichthys atratulus, Semotilus corporalis, or

Thoburnia rhothoeca, although it was weakly correlated with Nocomis leptocephalus movements. Life history requirements (e.g., Turner et al. 1996; Turner and Trexler 1998;

3

Matthews 2002) and individual fitness (e.g., Railsback et al. 1999; Petty and Grossman 2004) all may influence stream fish movements.

Many stream fishes appear to display short movements over significant portions of their lifespan. This has lead to what is called “The Restricted Movement Paradigm” (RMP, Gowan et al. 1994; Rodriguez 2002). Barbus haasi in a Mediterranean stream demonstrate restricted movement, with 55.6% of 573 fish moving less than 10 m between captures and 38.8% remaining within 100 m of initial capture (Aparicio and Sostoa 1999). Petty and Grossman

(2004) found that Cottus bairdi also display restricted movement over a 45 day period, with most individuals moving less than 3 m during this time and 79% of the fish not moving out of the 1 km site between spring 1994 and autumn 1996. Mundahl and Ingersoll (1983) observed that

>96% of Etheostoma nigrum and >87% of E. flabellare did not move out of their pool or riffle of initial capture. The RMP fits the observed data for many stream fish species, however biases supporting the RMP exist if fish have moved out of the study site and this area is not sampled for recaptures (Gowan et al. 1994; Albanese et al. 2003).

Some stream fish populations exhibit leptokurtotic movement distributions (e.g., Nocomis leptocephalus and Semotilus atromaculatus: Skalski and Gilliam 2000; Rivulus hartii: Fraser and

Gilliam 2001). These leptokurtotic distributions model populations in which many individuals move relatively short distances but some individuals move rather long distances, hence what

Skalski and Gilliam refer to as ‘movers’ and ‘stayers.’ Petty and Grossman (2004) found

‘movers’ and ‘stayers’ within populations of Cottus, respectively, with some individuals moving often and some moving seldom if at all. Currently, whether a fish is a ‘mover’ or ‘stayer’ is thought to be caused by an individual internal motivation (Skalski and Gilliam 2003), and while

4

quantifiable at the population level, individual predictions of whether a fish will be a ‘mover’

versus ‘stayer’ is unknown.

In conclusion, basic movement data is lacking for a majority of stream fish species.

Studies on stream fish movement patterns may be useful from both a basic ecological and

applied perspective.

The Gilt Darter (Percina evides)

As land development continues within the United States, many stream fishes are

becoming imperiled, endangered, or extirpated due to habitat degradation (Marguiles et al. 1980;

Labbe and Fausch 2000). This trend is particularly evident within the highly speciose and diverse

fish fauna of the Southeastern United States (Scott and Helfman 2001), with Etheostomatinae

Percids (darters) having a disproportionately high imperilment rate (Warren et al. 1997). Many of the declines in darter abundance can be attributed to declines in stream habitat quality such as siltation, damming, or diversion. Habitat degradation is a particular problem for darters with small ranges (e.g., Conasauga logperch, Percina jekinsi),

(http://cars.er.usgs.gov/Southeastern_Aquatic_Fauna/Freshwater_Fishes/Logperch/logperch.html

), the snubnose darters (Etheostoma barrenense and E. rafinesquei,

http://www.inhs.uiuc.edu/annualreports/95_96/CBD.html) or the Niangua darter (E. nianguae,

Mattingly and Galat 2002). Gilt darters, in contrast, are relatively widespread with three disjunct

regional populations: Eastern, Missouri/Upper Mississippi River, and White River (Ozarks; Near

et al. 2001). Despite the relatively widespread distribution of gilt darters within the United

States, both their range and general abundance have been shrinking (Page 1983; Hatch 1985;

NatureServe Explorer 2006). As Hatch (1982) and Margulies et al. (1980) recognized over two

5

decades ago, stream impoundment and degradation are important factors in the reduction of both

gilt darter numbers and habitat. Anthropogenic threats in combination with ecological attributes

such as small body size and use of the benthos (for feeding and reproducing) put darters at a high

risk of imperilment (sensu Angermeier 1995). The Georgia Department of Natural Resources

lists gilt darters as a species of special concern and the United States Forest Service lists darters

as sensitive (at risk) in Region 9 (Midwestern and Northeastern United States)

(http://georgiawildlife.dnr.state.ga.us/content/specialconcernanimals.asp,

http://www.fs.fed.us/r9/wildlife/).

Given the shrinking range and decreasing abundance of gilt darters, I began a study of the microhabitat requirements and short-term movements of this species in two relatively undisturbed North Carolina streams. Quantification of gilt darter microhabitat use, movement, and population estimates provides data for ecologically based management decisions and a baseline for future studies.

Objectives/Hypotheses

My primary objective was to determine gilt darter microhabitat use over multiple seasons and sites. Specifically, I examined the following hypothesis: microhabitat use by darters would not differ from what was randomly available in the habitat. In addition to my primary hypothesis,

I also conducted a short-term movement study to gain insights into this species’ movements.

Finally, I examined whether microhabitat use by darters differs by sex or size and also conducted a short-term movement study to determine whether darters exhibit restricted movements (Mundahl and Ingersoll 1983; Aparicio and Sostoa 1999; Fraser at al. 2001; Petty and Grossman 2004).

6

LITERATURE REVIEW

Albanse, B., Angermeier, P.L., Gowan, C. 2003. Designing mark-recapture studies to reduce effects of distance weighting on movement distance distributions of stream fishes. Transactions of the American Fisheries Society 132: 925-939.

Albanese, B., Angermeier, P.L., Dorai-Raj, S. 2004. Ecological correlates of fish movement in a network of Virginia streams. Canadian Journal of Fisheries and Aquatic Sciences 61(6): 857-869.

Angermeier, P.L. 1995. Ecological attributes of extinction prone species: loss of freshwater fishes of Virginia. Conservation Biology 9(11): 143-158.

Aparicio, E., de Sostoa, A. 1999. Pattern of movements of adult Barbus haasi in a small Mediterranean stream. Journal of Fish Biology 55: 1086-1095.

Baltz, D.M. 1990. Autoecology. Methods for Fish Biology. C.B. Schreck and P.B. Moyle, eds. American Fisheries Society, Bethesda, MD. pgs. 585-600.

Carter, M.G., Copp, G.H., Szomlai, V. 2004. Seasonal abundance and microhabitat use of bullhead Cottus gobio and accompanying fish species in the River Avon (Hampshire), and implications for conservation. Aquatic Conservation: Marine and Freshwater Ecosystems 14: 395-412.

Chipps, S.R., Perry, W.B., Perry, S.A. 1993. Patterns of microhabitat use among four species of darters in three Appalachian streams. American Midland Naturalist 131: 175-180.

Fraser, D.F., Gilliam, J.F., Daley, M.J., Le, A.N., Skalski, G.T. 2001. Explaining leptokurtotic movement distribution: intrapopulation variation in boldness and exploration. The American Naturalist 158 (2): 124-135.

Freeman, M.C., Grossman, G.D. 1993. Effects of habitat availability on dispersion of a stream cyprinid. Environmental Biology of Fishes 37 (2): 121-130.

Georgia Department of Natural Resources, Wildlife Resources Division, Georgia Natural Heritage Program. Krakow, G., data manager. Special Concern Animal Species in Georgia. Updated [10/22/2004] [Accessed 13 January 2005]

Gowan, C., Young, M.K., Fausch, K.D., Riley, S.C. 1994. Restricted movement in resident stream salmonids-a paradigm lost. Canadian Journal of Fisheries and Aquatic Sciences 51 (11): 2626-2637.

Greenberg, L.A. 1991. Habitat use and feeding behavior of thirteen species of benthic stream fishes. Environmental Biology of Fishes 31 (389-401).

7

Grossman, G.D., Freeman, M.C. 1987. Microhabitat use in a stream fish assemblage. Journal of Zoology 212: 151-176.

Grossman, G.D., Ratajczak Jr, R.E., Crawford, M., Freeman, M.C. 1998. Assemblage organization in stream fishes: effects of environmental variation and interspecific interactions. Ecological Monographs 68 (3): 395-420.

Grossman, G.D., Rincon, P.A., Farr, M.D., Ratajczak Jr, R.E. 2002. A new optimal foraging model predicts habitat use by drift feeding stream minnows. Ecology of Freshwater Fish 11: 2-10.

Hanski, I. 1994. A practical model of metapopulation dynamics. Journal of Animal Ecology 63 (1): 151-162.

Harding, J.M., Burky A.J., Way, C.M. 1998. Habitat preferences of the rainbow darter, Etheostoma caeruleum, with regard to microhabitat velocity shelters. Copeia 4: 988-997.

Hatch, J. T. 1985. Distribution, habitat, and status of the gilt darter (Percina evides) in Minnesota. Journal of the Minnesota Academy of Science 51(2): 11-16.

Hill. J., and Grossman, G.D. 1993. An energetic model of microhabitat use for rainbow trout and rosyside dace. Ecology 74: 685-698.

Labbe, T.R., Fausch, K.D. 2000. Dynamics of intermittent stream habitat regulate persistence of a threatened fish at multiple scales. Ecological Applications 10 (6): 1774-1791.

Levins, R. 1963. Theory of fitness in a heterogeneous environment. II. Developmental flexibility and niche selection. The American Naturalist 97 (893): 75-90.

Margulies, D., Burch, O.S., Clark, B.F. 1980. Rediscovery of the gilt dater (Percina evides) in the White River, Indiana. American Midland Naturalist 104 (1): 207-208.

Matthews, W.J. 1998. Patterns in Freshwater Fish Ecology. Kluwer Academic Publishers. Norwell, Massachusetts.

Mattingly, H.T., Galat, D.L. 2004. Predictive performance of a summer microhabitat model for the threatened Niangua darter, Etheostoma nianguae. Journal of Freshwater Ecology 19 (1): 109-114.

Mattingly, H.T.Galat, D.L 2002. Distributional patterns of the threatened Niangua darter, Etheostoma nianguae, at three spatial scales, with implications fo species conservation. Copeia 3: 573-585.

8

Mundahl, N.D., Ingersoll, C.G. 1983. Early autumn movements and densities of Johnny (Etheostoma nigrum) and fantail (E. flabellare) darters in a southwestern Ohio stream. Ohio Journal of Science 83 (3):103-108.

Nature Serve Explorer, version 4.7. 2006. Percina evides Status report. http://www.natureserve.org/explorer/servlet/NatureServe?init=Species [Accessed 21 April 2006]

Near, T. J., Page, L.M., and Mayden, R.L. 2001. Intraspecific phylogeography of Percina evides (Percidae: Etheostomatinae): an additional test of the Central Highlands pre-Pleistocene vicariance hypothesis. Molecular Ecology 10: 2235-2240.

Page, L.M. 1983. Handbook of darters. TFH Publications. Neptune City, NJ.

Persson, L., Eklöv, P. 1995. Prey refuges affecting interactions between piscivorous perch and juvenile perch and roach. Ecology 76 (1): 70-81.

Petty, J.T. Grossman, G.D. 1996. Patch selection by mottled sculpin (Pisces: Cottidae) in a southern Appalachian stream. Freshwater Biology 35: 261-276.

Petty, J.T, Grossman, G.D. 2004. Restricted movement by mottled sculpin (pisces: cottidae) in a southern Appalachian stream. Freshwater Biology 49: 631-645.

Poff, N.L., Allan, J.D. 1995. Functional organization of stream fish assemblages in relation to hydrological variability. Ecology 76 (2): 606-627.

Pulliam, H.R. 1988. Sources, sinks. and population regulation. The American Naturalist 132 (5): 652-661.

Railsback, S.F., Lamberson, R.H., Harvey, B.C., Duffy, W.E. 1999. Movement rules for individual-based models of stream fish. Ecological Modeling 123 (2-3): 73-89.

Resetarits, Jr. W.J. 2005. Habitat selection behavior links local and regional scales in aquatic systems. Ecology letters 8: 480-486.

Roberts, J.H. and Angermeier, P.L. 2005. Relationships of environmental an species attributes to the dispersal patterns of stream fishes. Master’s of Science manuscript, Virginia Tech.

Rodriguez, M.A. 2002. Restricted movement in stream fish: the paradigm is incomplete, not lost. Ecology 83 (1): 1-13.

Rosenfeld, J. 2003. Assessing the habitat requirements of stream fishes: an overview and evaluation of different approaches. Trans. of the Am. Fisheries Soc. 132 (5): 953- 968.

9

Schaefer, J. 2001. Riffles as barriers to interpool movement by three cyprinids (Notropis boops, Campostoma anomalum and Cyprinella venusta). Freshwater Biology 46: 379-388.

Schaefer, J.F., Marsh-Matthews, E., Sponner, D.E., Gido, K.B., Matthews, W.J. 2003. Effects of barriers and thermal refugia on local movement of the threatened leopard darter, Percina pantherina. Environmental Biology of Fishes 66: 391-400.

Schlosser, I.J. 1987. The role of predation in age- and size-related habitat use by stream fishes. Ecology 66: 651-659.

Scott, M.C., Helfman, G.S. 2001. Native invasions, homogenization, and the mismeasure of integrity of fish assemblages. Fisheries 26 (11): 6-15.

Skalski, G.T., Gilliam, J.F. 2000. Modeling diffusive spread in a heterogeneous population: a movement study with stream fish. Ecology81 (6): 1685-1700.

Stauffer, Jr., J.R., Boltz, J.M. Kellogg, K.A., van Snik, E.S. 1996. Microhabitat partitioning in a diverse assemblage of darters in the Allegheny River System. Environmental Biology of Fishes 46: 37-44.

Strange, K.T., Vokoun, J.C, Noltie, D.B. 2002. Thermal tolerance and growth differences in orangethroat darter (Etheostoma spectabile) from thermally contrasting adjoining streams. American Midland Naturalist 148: 120-128.

Thompson, A.R., Petty, J.T., Grossman, G.D. 2001. Multi-scale effects of resource patchiness on foraging behavior and habitat use by longnose dace, Rhinichthys cataractae. Freshwater Biology 46: 145-160.

Turner, T. F., Trexler, J.C., Kuhn, D.N., Robison, H.W. 1996. Life history variation and comparative phylogeography of darters (Pisces: Percidae) from the North American Central Highlands. Evolution 50(5): 2023-2036.

Turner, T. F., Trexler, J.C. 1998. Ecological and historical association of gene flow in darters (Teleostei: Percidae). Evolution 52(6): 1781-1801.

United States Fish and Wildlife: Regional Forester Region 9. created 29 February 2000, Updated 20 October 2003 http://www.fs.fed.us/r9/wildlife/tes/docs/rfss_animals.pdf [accessed 13 January 2005]

Warren, Jr. M.L., Angermeier, P.L., Burr, B.M., Haag, W.R. 1997. Decline of a diverse fish fauna: patterns of imperilment and protection in the Southeastern United States. In: Aquatic Fauna in Peril: The Southeastern Perspective. Eds. G.W. Benz and D.E. Collins. Southeast Aquatic Research Institute Special publication I. Lenz Design and Communication, Decatur, GA.

10

Wiens, J.A. 2002. Riverine landscapes: taking landscape ecology into the water. Freshwater Biology 47: 501-515.

Wildhaber, M.L., Lamberson, P.J. 2004. Importance of the habitat choice behavior assumed when modeling the effects of food and temperature on fish populations. Ecological Modeling 175 (4): 395-409.

Wu, J., Loucks, O.L. 1995. From balance of nature to hierarchical patch dynamics: a paradigm shift in ecology. Quarterly Review of Biology 70 (4): 439-466.

11

MICROHABITAT USE AND MOVEMENTS OF GILT DARTERS (PERCINA EVIDES) IN

TWO SOUTHEASTERN STREAMS

______

Skyfield, J.P. and Grossman, G.D. To be submitted to Ecology of Freshwater Fish

12

ABSTRACT

Quantification of many stream fishes’ habitat use, movement, and population size is

lacking. We examined habitat use by gilt darters (Percina evides) in two Southeastern streams at

the microhabitat scale, comparing overall darter use to randomly available microhabitat,

differences in male and female use, and differences in size-based use. Using Principal

Component Analysis (PCA), we found that darters preferred more erosional areas with higher

percent cobble and average velocities than randomly available microhabitat. Male darters tended

to use microhabitats that deviated more from the randomly available microhabitat than females.

In the size-based analyses, larger (≥60 mm) gilt darters tended to use microhabitats with more

heterogeneous substrata and larger amounts of boulder than the other two size classes (≤50 mm,

51-59 mm). We also conducted a short-term movement study and calculated population

estimates based on mark-recapture data in Autumn 2005. Darters moved longer distances than

expected if they conformed to the Restricted Movement Paradigm. Nonetheless, 40% of

movements were within 5 meters of initial capture. Using Program MARK and model averaged

parameter estimates, the population size of 225 darters in 730 m 2 (density = 0.31 darters/m2) provides novel information regarding gilt darter populations. Our data provide quantification of habitat use, movement, and population size of gilt darters in two streams in the Southeastern

U.S., necessary information for management of this declining species.

13

INTRODUCTION

Temperate lotic systems are spatially and temporally heterogeneous. This heterogeneity creates patches, or microhabitats of varying qualities and types (e.g., Grossman et al. 1995;

Wiens 2002) and affects a variety of biological characteristics of stream fishes including: 1) distribution (Grossman et al. 1998; Rieman et al. 2006), 2) movements (Mundahl and Ingersoll

1983; Railsback et al. 1999; Petty and Grossman 2004), and 3) habitat use (Freeman and

Grossman 1993; Poff and Allan 1995; Grossman et al. 2002; Resitarits, Jr 2005). Quantifying habitat use and movement within any patchy landscape is crucial for our understanding and management of these species and systems (sensu Pulliam 1988; Hanski 1994; Wu and Loucks

1995). Despite this fact, the mechanisms affecting patterns of abundance and resource use for most stream fishes are largely unknown (Grossman et al. 2002).

The paucity of information on habitat requirements of stream fishes is a significant problem for fishes residing in the high diversity streams of the southeastern United States. In fact, concomitant with increased land development in this region, a large number of southeastern stream fishes have decreased in abundance; with many being classified as imperiled (Marguiles et al. 1980; Labbe and Fausch 2000; Warren et al. 2000). This trend is particularly evident within the highly speciose darters (Etheostomatinae: Warren et al. 1997, Scott and Helfman 2001).

Given these conditions, we conducted a study of microhabitat use and short-term movements of the gilt darter (Percina evides). Gilt darters are relatively widespread with three disjunct regional populations: Eastern, Missouri/Upper Mississippi River, and White River

(Ozarks; Near et al. 2001). Despite their relatively widespread distribution, the range and general abundance of gilt darters has been declining for decades (Page 1983; Hatch 1985; NatureServe

Explorer 2006). Consequently, we quantified microhabitat use patterns of gilt darters in two

14

sites over several seasons. We addressed whether microhabitat use varied seasonally or by fish

sex or size. In addition, in one season we conducted a short-term movement study to determine

whether darters displayed restricted movements.

METHODS

Study Sites

Our study sites were Coweeta Creek and Tellico Creek, located in Macon County, North

Carolina. These streams are both tributaries of the Little Tennessee River (35°11’ N; 83°23’ W)

and are relatively undisturbed (<3% non-forested land cover) (Sutherland et al. 2002). Coweeta

Creek is a fifth order stream with an average discharge of 1.26 m3/s (Sutherland et al. 2002). Our study site was a 100 m section of the creek approximately 4 km downstream of the U.S.D.A.

Coweeta Hydrologic lab. The general physiognomy of the Coweeta Creek drainage has been described by Grossman et al. (1998). Tellico Creek is a third order stream, with an average

discharge of 0.63 m3/s (Sutherland et al. 2002). The Tellico study site was 110 m in length. Both

study sites have morphologies typical of streams in the Blue Ridge Province of the southern

Appalachians (i.e., riffle, run and pool sequences).

Microhabitat availability

We quantified microhabitat availability in two ways. We used a stratified random

sampling plan and collected five random availability samples per 10 m of bank for a total of at

least 50 samples per site per seasonal sample. Random locations (meter mark and percent

distance across the stream) were chosen using a random number table. We also collected

constrained random microhabitat availability samples between 20 cm and 200 cm from each

15

darter’s location. The direction of the constrained microhabitat availability sample was randomly chosen from 0-24, corresponding to 30 minute intervals on a clock face; the distance to the constrained sample was randomly chosen from numbers 20-200. Having constrained random samples allowed us to compare microhabitat use to a random patch that a gilt darter could potentially have occupied (Petty and Grossman 1996).

Microhabitat measurements

We made fish microhabitat measurements in Coweeta Creek between 7 May and 7 June

2005 (Spring sample) and between 9-28 August 2005 (Summer sample) and 21 October to 5

November (Autumn sample) in Tellico creek. During the Spring sample some gilt darters exhibited breeding behavior, thus habitat use patterns during this period represent a mixture of breeding and non-breeding patterns. We made fish observations during daylight hours (from

~0900 to ~1700) by entering the study site at a randomly determined location and snorkeling slowly upstream while searching for gilt darters. We covered the entire channel width by slowly moving laterally until the far bank was reached and then moving slowly upstream and repeating this procedure. Upon sighting an undisturbed fish (see Grossman and Ratajczak 1998), we placed a painted lead weight at its location.

We measured a series of microhabitat parameters at the location of each fish.. These measurements included mean water velocity (velocity at 0.6 depth), focal point velocity (i.e., at the fish’s nose), water depth, distance from the substratum (gilt darters were always on the bottom) and substratum composition (Grossman and Freeman 1987). We quantified current velocity using a Marsh-McBirney Model 201 electronic velocity meter. We measured depth to the nearest 0.5 cm with a meter stick. We visually estimated substratum composition within a

16

400cm2 quadrat centered on the fish’s position. Substrata were divided into seven categories

(bedrock, boulder, cobble, gravel, sand, silt, and debris).

Movements and Population Size

During the Autumn 2005 microhabitat sampling period, we also conducted a short term movement study and a mark-recapture population estimate for gilt darters in Tellico Creek. We captured gilt darters with dip nets and anesthetized them with clove oil (25 ppm). Gilt darters were then injected subcutaneously with acrylic paint (magenta, orange, light pink, and turquoise) using a unique combination of three marks at four possible body locations (left and right anal fin, left and right caudal peduncle; Hill and Grossman 1987, Roberts and Angermeier 2004). We weighed each fish with an electronic balance (nearest 0.1 g) and measured them with a straightedge (nearest mm, SL). After marking, gilt darters were kept in a live well to monitor for marking injury or mortality. If fish did not display evidence of physical damage or abnormal behavior they were replaced in their exact capture location in the stream.

We made three sequential passes through the Tellico Creek site on 21 October to 22

October, 29 October to 30 October, and 3 November to 5 November. We maintained equal sampling effort during each pass, spending three to four hours per 15 meters of the site, or until all observed gilt darters were captured. All gilt darters were examined for marks, and marked if they had not been marked previously. Exact locations were recorded and we measured length and weight at each capture. At the end of the third pass, we captured all gilt darters within 25 m zones above and below the study site to check for marked individuals who may have moved out of the study site (Albanese et al. 2003). Gilt darters also were marked during summer microhabitat measurements but logistical problems prevented us from obtaining accurate

17

movement data during this season. We established permanent benchmarks at 5 meter intervals

along the stream bank to triangulate (+ 1 cm) exact darter locations. We then compared recapture locations to derive an estimate of linear movement between captures (Petty and

Grossman 2004).

DATA ANALYSIS

Microhabitat availability and use

We tested for differences in habitat availability within and among sites using the

Principle Component Analysis (PCA) technique of Grossman and Freeman (1987) and

Grossman and deSostoa (1994). In addition, we used Wilcoxon ANOVA and Tukey’s Honestly

Significant Difference (HSD) tests on the ranked data to detect differences in univariate data among sites.

Non-random microhabitat use by gilt darters was identified using the procedures of

Grossman and Freeman (1987) and Grossman and deSostoa (1994). Because microhabitat represents a constellation of correlated factors, we performed a Principle Component Analysis

(PCA) on both habitat availability and darter microhabitat use data. Linear data were natural log transformed and percentage data were arcsine-square root transformed. We only reported ecologically interpretable components with eigenvalues greater than one (Grossman & Freeman

1987). We then plotted means and 95% confidence ellipses for both gilt darter and habitat availability data for each pair of components (Grossman and deSostoa 1994) or in the case of sex and size comparisons among darters, we plotted the varying ellipses for darter microhabitat use.

If the 95% confidence ellipses did not overlap another mean, darter microhabitat use differed significantly from microhabitat availability or the other sex/size class. We only collected

18

constrained microhabitat availability data for both sites, hence microhabitat comparisons generally were conducted with these data. Nonetheless, we also tested for non-random microhabitat use using unconstrained microhabitat availability data from the two Tellico Creek samples and compared microhabitat availability estimates derived from both constrained and unconstrained samples.

We used the PCA microhabitat analysis previously described to test for sex and size- linked differences in microhabitat use within seasonal samples. Gilt darters have sexually dimorphic coloration and sex is correlated with size (Skyfield and Grossman, unpublished data).

Juvenile gilt darters (~< 45 mm SL) all appear to have female coloration, although by ~50mm SL sexually dimorphic coloration is present (Hatch 1982). Although all darters between 45mm and

50mm had female coloration, almost all gilt darters greater than 60mm displayed male coloration. Consequently, sex-based comparisons only were made on fish between 50 and 59 mm where both sexes are represented. We also assumed that external coloration represented the true sex of the individual (i.e., there were no males that mimicked female coloration).We also performed size-linked analyses but recognize that size and sex are confounded in these analyses.

Movement

We quantified movements of darters with individually recognizable marks (i.e., those with three readable marks) during Autumn microhabitat sampling in Tellico Creek. We obtained linear movement distances using ARCView 3.2. We tested for significant differences in the number of darters making upstream versus downstream movement as well as mean upstream and downstream distances moved using Wilcoxon rank sum tests. We used a Wilcoxon ANOVA with a posteriori Tukey’s HSD tests on ranked data to test for differences in movement distances

19

between gilt darters recaptured at varying sampling intervals (i.e., pass 1 - pass 2, pass 2 - pass 3, and pass 1 - pass 3). Finally, we used a Wilcoxon rank sum test to examine whether there were differences in mean movement distances between males and females. We also used a Wilcoxon

ANOVA with a posteriori Tukey’s HSD tests on the ranked data to test for differences in mean movement distances for darters in different size classes (≤50mm, 51-59 mm and ≥60 mm).

Population size

We used Program Mark (White and Burnham 1999; Program MARK) to derive a population size estimate and 95% confidence interval from mark-recapture data from Tellico

Creek in Autumn 2005. We used a closed population model with a Huggins estimator to calculate capture, recapture, and abundance separately. To evaluate the relative fit of the candidate models, we used Akaike’s Information Criteria with sample size correction (AICc).

The relative fits of the candidate models were determined by calculating the Akaike weights, with the best fitting model having the highest weight. We ran the analysis using 111 marked and

53 recaptured gilt darters.

RESULTS

Microhabitat Availability

Inter-Site and Seasonal Variation in Habitat Availability

Approximately half (49%) of the variance in the constrained microhabitat availability data set was explained by the first three factors from the PCA. Summer microhabitats for both

Coweeta and Tellico samples displayed faster average and bottom velocities with more cobble than Autumn Tellico microhabitats (Table 1, Fig. 1). In addition, microhabitats in Tellico Creek

20

possessed greater amounts of boulder than were available in Coweeta Creek (Table 1, Fig. 3).

Because we do not have microhabitat availability data for both sites in the same season, it was impossible to determine whether differences were attributable to inter-site or inter-seasonal differences. Microhabitat availability also differed between constrained Summer and Autumn microhabitat availability samples in Tellico Creek (Table 1, Figs. 1-3). Summer constrained microhabitat availability samples at Tellico had higher percentages of cobble, higher average and focal velocities and less sand and debris than Autumn samples.

Seasonal microhabitat availability comparisons for Tellico Creek using unconstrained microhabitat availability data produced three interpretable components (54% variance explained,

Figs. 4-6). Samples differed only on the first two components, and in Summer, Tellico Creek had faster average and bottom velocities than in Autumn (substratum differences were not ecologically interpretable) (Table 1, Fig. 4).

Comparisons of Tellico constrained versus unconstrained microhabitat availability in

Summer and Autumn showed that the Summer samples were different (Fig. 5), but the Autumn availability was not significantly different (Fig. 6). In Tellico Creek in Summer, the constrained microhabitat had higher average velocities and less sand than the unconstrained data set (Table 1,

Fig. 5).

Our single univariate comparison indicated that Coweeta Creek (Spring) was significantly deeper (F = 26.89, df = 2,336, p = <0.0001) than Tellico Creek in either Summer or

Autumn. It is unclear whether this difference is a seasonal or site effect.

21

Microhabitat use

Coweeta Spring

During Spring 2005, gilt darters in Coweeta Creek showed non-random microhabitat use

(Fig. 7) using constrained microhabitat availability samples. The PCA extracted three

interpretable components (51 % variance explained); however, only components 1 and 2 showed

differences in among-group values. Gilt darters occupied microhabitats with slightly lower focal

point velocities, greater amounts of cobble and lower quantities of boulders and depositional

substrata, than randomly available (Table 1, Fig. 7). Gilt darters also occupied microhabitats that were slightly deeper than those randomly available (Table 1).

Tellico Summer

During Summer 2005, gilt darters in Tellico Creek also occupied microhabitats that differed significantly from those available (constrained microhabitat availability data) (Fig. 8).

The PCA extracted three interpretable components (53% variance explained). Only components

1 and 2 showed differences in the among group values, and these components indicated that gilt darters occurred in microhabitats with higher percentages of cobble and gravel and less boulder than randomly available (Table 1).

Gilt darters also displayed non-random microhabitat use in analyses using unconstrained microhabitat availability data. Unconstrained comparisons typically illustrate more general patterns of non-random microhabitat use because they are not constrained by the position of the fish. The PCA extracted three ecologically interpretable components that accounted for 52 % of the variance in the data sets. All three components yielded differences and gilt darters in Tellico

22

Creek occupied deeper microhabitats with faster average velocities, greater amounts of cobble and gravel and lower quantities of both large and depositional substrata (Table 1, Figs. 9-11).

Tellico Autumn

Gilt darters in Autumn 2005 in Tellico Creek also showed non-random microhabitat use

(constrained & unconstrained microhabitat availability data) (Figs. 12-13). Three components were extracted by the PCA (49% variance explained, constrained data), although among-group differences only were present for PC’s 1 and 2. Gilt darters were over-represented in higher velocity microhabitats with greater amounts of cobble and less boulder, sand, and debris than randomly available (constrained microhabitat availability data, Table 1, Fig. 12). Analyses based on unconstrained microhabitat availability data yielded similar results with three PC’s (among- group differences on two) explaining 52% of the variance in the data set. Gilt darters also preferentially occurred in deeper microhabitats with more cobble and less boulder and depositional substrata than were randomly available (Table 1, Fig. 13).

Sex-linked differences in microhabitat use

Our sex-based microhabitat comparisons were based on fish 51-59 mm SL because this was the only size class in which the sexes overlapped considerably for all seasons (Table 2).

Coweeta Spring

Male and female gilt darters were found in different microhabitats during Spring in

Coweeta creek. The PCA extracted three components (66% variance explained) from the data set and male gilt darters occupied microhabitats with higher average and focal velocities, greater amounts of cobble and less sand than females (Table 3, Fig. 14).

23

Tellico Summer

Male and female darters occurred in different microhabitats during Summer in Tellico

Creek. The PCA delimited three components (69 % variance explained), and male gilt darters

utilized microhabitats with more cobble, higher average velocities, and less boulder than females

(Table 3, Fig. 15).

Autumn Tellico

Similar to Summer, male and female gilt darters in Tellico Creek occupied distinct

microhabitats. The PCA identified three components (61 % variance explained) from the

constrained data set, and male gilt darters used deeper microhabitats with lower average and

bottom velocities and greater amounts of cobble than females (Table 3, Fig. 16).

Size-linked differences in microhabitat use

Coweeta Spring

We observed size-related differences in microhabitat use by gilt darters in Coweeta

Creek. The PCA extracted three factors (61 % variance explained) in the data set and small (45-

50 mm -- immature and female fish) darters occupied microhabitats with lower average and

focal-point velocities and more depositional substrata then larger (50-60+mm) fish (Table 4, Fig.

17). In addition medium gilt darters (males & females, 51-59 mm) were found in microhabitats

with higher bottom velocities, deeper water, and greater amounts of cobble than both larger and

smaller fish (Table 4, Fig. 17). Finally, large gilt darters (> 60 mm) used microhabitats with greater amounts of boulder than smaller fish (Table 4, Fig. 17).

Tellico Summer

24

Gilt darters in Tellico Creek in Summer also displayed size-related differences in microhabitat use (Fig. 18). Fifty-five percent of the variance was explained by the first three factors of the PCA. Small (45-50 mm) gilt darters occupied microhabitats with greater amounts of cobble than larger gilt darters (Fig 18) and the same relationship was observed between medium sized (50-59 mm, more cobble) and large gilt darters (Table 4, Fig. 18). Microhabitats used by medium gilt darters (51-59 mm) had higher average and bottom velocities than microhabitats used by either large or small gilt darters (Table 4). Finally, large gilt darters were found over more gravel and large erosional substrata than either small or medium fish (Table 4,

Fig. 18).

Tellico Autumn

In contrast to the previous seasons, microhabitat use by medium and large gilt darters did not significantly differ. The PCA (3 PC’s, 57 % variance explained) indicated that small gilt darters occupied microhabitats with lower average velocities and more boulder and sand than those used by medium and large gilt darters (Table 4, Fig. 19). Velocity differences in microhabitat use among the three size classes were not apparent (Table 4, Fig. 19).

Movement

We marked 164 darters and recaptured 35 positively identifiable individuals (Table 5; see

Autumn sampling dates for capture intervals). Net movement for these individually identifiable gilt darters ranged from ~31 meters downstream to ~58 meters upstream, with the majority of movements occurring within 8 meters of the capture site (Table 6, Fig. 20) and 40% staying within 5 meters of the initial capture site. During our study, gilt darters did not exhibit significant differences in: 1) frequency of upstream or downstream movement, (20 upstream versus 21

25

downstream), 2) mean distances moved upstream and downstream, 3) distances moved while at

large for differing time periods (e.g., recaptures on pass 2 versus pass 3), 4) movements by males

and females or 5) movement distances of small, medium or large gilt darters (Tables 7-8, Figs

21-22).

Both male and female gilt darters underwent both long and short distance movements

(Tables 6-7, Fig. 21). Wilcoxon rank sum tests indicated that there was no difference between

male and female movement (Table 8) in Autumn 2005 in Tellico Creek. In the size analysis of

movement, all size classes of gilt darters seemed to move long and short distances as well, but

darters greater than 60 mm did not move more than 21 meters up or downstream, and both the

long distance movements were made by the smallest size class of darters (Table 6, Fig. 22).

Nevertheless, both the Wilcoxon ANOVA and the a posteriori tests of difference indicated that

there were no differences in movement distances among size classes (Table 8).

Population Size

Because our AICc weights were low, and the best two models did not provide good fits to

the data, we used model averaging from the remaining models in the confidence set to obtain our

parameter estimates (Table 9). We used the 1/8 minimum cutoff suggested by Royall (1997) to

define models in the confidence set. Using our model averaged parameter estimates, we obtained

a mean population estimate of 225 (95% CI 167-282) individuals within an area of 730.35 m2.

Density of gilt darters within our site was 0.31 darters/m2.

26

DISCUSSION

The gilt darters of both Coweeta and Tellico Creeks exhibited non-random microhabitat

use in all sites and seasons regardless of the microhabitat availability data set (constrained or

unconstrained). Our lack of an unconstrained data set for Coweeta lessens our ability to make

more inter-site comparisons. However, the comparisons of the Tellico constrained and unconstrained microhabitat availability measurements showed no significant difference in

Autumn and only slight differences in average velocity and percent sand composition in

Summer. This similarity suggests that even within a fine spatial scale of 2 m, darters select microhabitats different than what is randomly available and that, at least in our sites and seasons, the differences between the constrained and unconstrained data set was not great.

In general gilt darters selected higher velocity microhabitats with greater amounts of cobble and to a lesser extent, gravel. Gilt darters always were found on the substratum (benthic guild, sensu Grossman and Freeman 1987) and generally avoided microhabitats with substantial amounts of depositional substrata (sand, silt, or debris). In addition, there was some evidence that gilt darters selected deeper microhabitats in Coweeta Creek, however, this may be due to either seasonal, site-specific, or reproductive requirements of this species. Although we have no direct evidence, it is possible that gilt darters are over-represented in high velocity microhabitats dominated by cobble, because cobble substrata may provide excellent habitat for invertebrate colonization (Mattingly and Galat 2004), and benthic stream fishes may select microhabitats that have higher prey abundances (see Petty and Grossman 1996; Thompson et al. 2001). In addition, cobbles provide extensive interstitial spaces which may serve as refuges from both current and predators (Harding et al. 1998).

27

Gilt darters displayed relatively consistent microhabitat use patterns despite differences in

microhabitat availability and the abundances of benthic species in the two streams. For example,

mottled sculpin (Cottus bairdi) were much more abundant in Coweeta Creek than in Tellico

Creek (J. Skyfield, personal observation). This suggests that interspecific interactions with mottled sculpin have little effect on microhabitat use by gilt darters. Similar results have been found in microhabitat use studies for other species in Coweeta Creek (i.e., few interspecific effects, Grossman et al. 1998).

Given the between-stream and seasonal consistency of microhabitat use patterns for gilt

darters, it is possible that our results are extrapolable to other gilt darter populations using type

III (use/availability) habitat suitability curves (HSCs) (sensu Glozier et al. 1997). Our data also

are consistent with studies that have identified important habitat types for darters (Bowen et al.

1998; Freeman et al. 2001) such as shallow-coarse (<35 cm deep, gravel or larger substrate) and

shallow-slow (<35 cm deep, <35 cm/s) microhabitats. Nonetheless, without replication of

seasonal and site-related patterns the potential transferability of our microhabitat results are

limited (sensu Rosenfeld 2003; Rosenfeld et al. 2005).

In addition to the microhabitat use patterns observed in all darters, we detected significant

size-related and sex-related differences in microhabitat use, although these differences were

confounded by a sex-size interaction. Sexual dimorphism in coloration or easily recognizable

external morphology are not common in stream fishes, thus, there are few studies documenting

sexual differences in microhabitat use. Nonetheless, male gilt occupied microhabitats with

greater amounts of cobble over all sites and seasons, and also tended to occur in higher velocity

microhabitats over less boulder substratum than female darters. During Autumn in Tellico

Creek, however, male gilt darters shifted to slower, deeper microhabitats than females.

28

Size-related analyses indicated that large gilt darters selected microhabitats with greater

amounts of boulder and more heterogeneous substrata than medium and small darters in both

Coweeta Creek and Tellico Summer samples. Size-related differences were less pronounced for

Tellico Autumn samples and medium and large gilt darters did not differ in microhabitat use during this season. It is possible that the greater amounts of depositional microhabitats (more debris, shallower, slower water) during this season reduced the amount of preferred microhabitat for large darters. Based on segregated microhabitat use by size, other researchers have found that larger individuals are more likely to inhabit areas that either maximize energy gain or minimize energy expenditure (Hill and Grossman 1993; Petty and Grossman 1996; Page 2000;

Thompson et al. 2001; Grossman et al. 2002; Rosenfeld 2003; Wildhaber and Lamberson 2004;

Mattingly and Galat 2004). This may explain the differences observed in size-related analyses, although we have no data to directly test this hypothesis.

Our findings are similar to those of other investigators who have studied microhabitat use in darters. The use of cobble substrata by gilt darters is well documented (Hatch 1985;

Greenberg 1991) and other species of darters are known to exhibit selection for cobble-sized substrata (Chipps et al. 1994; van Snik Gray and Stauffer, Jr. 2001). Similar to our results, Hatch

(1985) also found that gilt darters avoided depositional substrata (Hatch 1985).

Movement data and population estimates for gilt darters were based on a single season and hence, must be viewed as tentative. Nevertheless, little is known about darter movements, population sizes or densities. We observed both long and short movements by gilt darters with

40% of fish being recaptured within 5m of their original capture location. The Tellico Creek gilt darter population appeared to include both movers and stayers (Gowan et al 1994; Fraser et al.

2001; Rodriguez 2002; Petty and Grossman 2004) with little evidence of different movement

29

patterns for either sexes or size classes of fish, with the exception that large fish never moved

more than 21m from their site of capture. We found no fish that had moved into the 25 m buffer

zones, suggesting that fish may not have escaped the study area; however, our study design

cannot detect movements greater than 25 meters out of the study area. Our population densities

are on the low end of what Hatch (1982) found for gilt darters (0.2 – 1.55 darters/m2), however the temporal gap and the fact that Southeastern gilt darters represent a different monophyletic

group than Minnesota gilt darters (Near et al. 2001) may not permit these densities to be

compared.

Gilt darters are a species of special concern within Georgia and their shrinking numbers

(Margulies et al. 1980; Hatch 1982) may be due to a variety of factors contributing to habitat

degradation. Declines in habitat quality in streams such as sedimentation may negatively affect

stream fishes by reducing fecundity or prey abundance or eliminating preferred spawning or non-

spawning habitats (Hatch 1985; Sutherland et al. 2002; Tabit and Johnson 2002; Bolliet et al.

2005). Given that gilt darters avoided depositional substrata, anthropogenic sedimentation is

likely to negatively influence gilt darter fitness and population persistence.

In conclusion, we have shown that gilt darters are over-represented in faster

microhabitats with greater amounts of erosional substrata. Microhabitat use by gilt darters

differed by sex (cobble and boulder) and size (substratum variables) although there was little

apparent seasonal variation. Gilt darters showed both restricted and relatively long-distance

movements although 40% of marked individuals were recaptured within 5 m of their capture site.

Densities of gilt darters in this relatively natural stream were 0.31 darters/m2. Given the paucity of information on this species of special concern, our data should be of use to both state and federal resource management agencies.

30

LITERATURE CITED

Albanese, B., Angermeier, P.L., Gowan, C. 2003. Designing mark-recapture studies to reduce effects of distance weighting on movement distance distributions of stream fishes. Transactions of the American Fisheries Society 132: 925-939.

Bolliet, V., Bardonnet, A., Jarry, M., Vignes, J.C., Gaudin, P. 2005. Does embeddedness affect growth performance in juvenile salmonids? An experimental study in brown trout Salmo trutta L. Ecology of Freshwater Fish 14: 289-295.

Bowen, Z.H., Freeman, M.C., Bovee, K.D. 1998. Evaluation of generalized habitat criteria for assessing impacts of altered flow regimes on warmwater fishes. Transactions of the American Fisheries Society 127: 455-486.

Fraser, D.F., Gilliam, J.F., Daley, M.J., Le, A.N., Skalski, G.T. 2001. Explaining leptokurtotic movement distribution: intrapopulation variation in boldness and exploration. The American Naturalist 158 (2): 124-135.

Freeman, M.C., Grossman, G.D. 1993. Effects of habitat availability on dispersion of a stream cyprinid. Environmental Biology of Fishes 37 (2): 121-130.

Freeman, M.C., Bowen, Z.H., Bovee, K.D., Irwin, E.R. 2001. Flow and habitat effects on juvenile fish abundance in natural and altered flow regimes. Ecological Applications 11(1): 179-190.

Glozier, N.E., Culp, J.M., Scrimgeour, G.J. 1997. Transferability of habitat suitability curves for a benthic minnow, Rhinichthys cataractae. Journal of Freshwater Ecology 12 (3): 379- 391.

Gowan, C., Young, M.K., Fausch, K.D., Riley, S.C. 1994. Restricted movement in resident stream salmonids-a paradigm lost. Canadian Journal of Fisheries and Aquatic Sciences 51 (11): 2626-2637.

Greenberg, L.A. 1991. Habitat use and feeding behavior of thirteen species of benthic stream fishes. Environmental Biology of Fishes 31 (389-401).

Grossman, G.D., Freeman, M.C. 1987. Microhabitat use in a stream fish assemblage. Journal of Zoology 212: 151-176.

Grossman, G.D., and de Sostoa, A. 1994. Microhabitat use by fish in the uppoer Rio Matarrana, Spain, 1984-1987. Ecology of Freshwater Fish 3: 141-152.

Grossman, G.D., Ratajczak Jr, R.E., Crawford, M., Freeman, M.C. 1998. Assemblage organization in stream fishes: effects of environmental variation and interspecific interactions. Ecological Monographs 68 (3): 395-420.

31

Grossman, G.D., Ratajczak Jr, R.E. 1998. Long-term patterns of microhabitat use by fish in a southern Appalachian stream from 1983 to 1992: effects of hyrologic period, season, and fish length. Ecology of Freshwater Fish 7: 108-131.

Grossman, G.D., Rincon, P.A., Farr, M.D., Ratajczak Jr, R.E. 2002. A new optimal foraging model predicts habitat use by drift feeding stream minnows. Ecology of Freshwater Fish 11: 2-10.

Hanski, I. 1994. A practical model of metapopulation dynamics. Journal of Animal Ecology 63 (1): 151-162.

Harding, J.M., Burky A.J., Way, C.M. 1998. Habitat preferences of the rainbow darter, Etheostoma caeruleum, with regard to microhabitat velocity shelters. Copeia 4: 988-997.

Hatch, J.T. 1982. Life history of the gilt darter, Percina evides (Jordan and Copeland), in the Sunrise River, Minnesota. PhD dissertation. University of Minnesota.

Hatch, J. T. 1985. Distribution, habitat, and stauts of the gilt darter (Percina evides) in Minnesota. Journal of the Minnesota Academy of Science 51(2): 11-16.

Hill, J., and Grossman, G.D. 1987. Effects of subcutaneous marking on stream fishes. Copeia 1987 (495-499).

Hill. J., and Grossman, G.D. 1993. An energetic model of microhabitat use for rainbow trout and rosyside dace. Ecology 74: 685-698.

Labbe, T.R., Fausch, K.D. 2000. Dynamics of intermittent stream habitat regulate persistence of a threatened fish at multiple scales. Ecological Applications 10 (6): 1774-1791.

Margulies, D., Burch, O.S., Clark, B.F. 1980. Rediscovery of the gilt dater (Percina evides) in the White River, Indiana. American Midland Naturalist 104 (1): 207-208.

Mattingly, H.T., Galat, D.L. 2004. Predictive performance of a summer microhabitat model for the threatened Niangua darter, Etheostoma nianguae. Journal of Freshwater Ecology 19 (1): 109-114.

Mundahl, N.D., Ingersoll, C.G. 1983. Early autumn movements and densities of Johnny (Etheostoma nigrum) and fantail (E. flabellare) darters in a southwestern Ohio stream. Ohio Journal of Science 83 (3):103-108.

NatureServe Explorer. 2006. Distribution report for Percina evides. http://www.natureserve.org/explorer/servlet/NatureServe?init=Species [accessed 21 April 2006].

32

Near, T. J., Page, L.M., and Mayden, R.L. 2001. Intraspecific phylogeography of Percina evides (Percidae: Etheostomatinae): an additional test of the Central Highlands pre-Pleistocene vicariance hypothesis. Molecular Ecology 10: 2235-2240.

Page, L.M. 1983. Handbook of darters. TFH Publications. Neptune City, NJ.

Page, L.M. 2000. Etheostomatinae. in Percid Fishes Systematics, Ecology, and Exploitation . John F. Craig, ed. volume 3 of Fish and Aquatic Resources Series. Tony J. Pitcher, ed. Blackwell Science, Oxford, UK.

Petty, J.T. Grossman, G.D. 1996. Patch selection by mottled sculpin (Pisces: Cottidae) in a southern Appalachian stream. Freshwater Biology 35: 261-276.

Petty, J.T, Grossman, G.D. 2004. Restricted movement by mottled sculpin (pisces: cottidae) in a southern Appalachian stream. Freshwater Biology 49: 631-645.

Poff, N.L., Allan, J.D. 1995. Functional organization of stream fish assemblages in relation to hydrological variability. Ecology 76 (2): 606-627.

Pulliam, H.R. 1988. Sources, sinks. and population regulation. The American Naturalist 132 (5): 652-661.

Railsback, S.F., Lamberson, R.H., Harvey, B.C., Duffy, W.E. 1999. Movement rules for individual-based models of stream fish. Ecological Modeling 123 (2-3): 73-89.

Resetarits, Jr. W.J. 2005. Habitat selection behavior links local and regional scales in aquatic systems. Ecology letters 8: 480-486.

Rieman, B.E., Peterson, J.T., Myers, D.L. 2006. Have brook trout (Salvelinus fontinalis) displaced bull trout (Salvelinus confluentus) along latitudinal gradients in central Idaho streams? Canadian Journal of Fisheries and Aquatic Science 63(1): 63-78.

Roberts, J.H., Angermeier, P.L. 2004. A comparison on injectable fluorescent marks in two genera of darters: effects of survival and retention rates. North American Journal of Fisheries Management 24: 1017-1024.

Rodriguez, M.A. 2002. Restricted movement in stream fish: the paradigm is incomplete, not lost. Ecology 83 (1): 1-13.

Rosenfeld, J. 2003. Assessing the habitat requirements of stream fishes: an overview and evaluation of different approaches. Trans. of the Am. Fisheries Soc. 132 (5): 953- 968.

Rosenfeld, J.S., Leiter, T., Lindner, G., Rothman, L. 2005. Food abundance and fish density alters habitat selection, growth, and habitat suitability curves for juvenile coho salmon (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Science 62: 1691-1701.

33

Royall, R.M. 1997. Statistical evidence: a likelihood paradigm. Chapman and Hall, New York.

Scott, M.C., Helfman, G.S. 2001. Native invasions, homogenization, and the mismeasure of integrity of fish assemblages. Fisheries 26 (11): 6-15.

Sutherland, A. B., Meyer, J.L., Gardiner, E.P. 2002. Effects of land cover on sediment regime and fish assemblage structure in four southern Appalachian streams Freshwater Biology 47( 9): 1791-1805. Tabit, C.R., Johnson, G.M. 2002. Influence of urbanization on the distribution of fishes in a southeastern upper Piedmont drainage. Southeastern Naturalist 1(3): 253-268.

Thompson, A.R., Petty, J.T., Grossman, G.D. 2001. Multi-scale effects of resource patchiness on foraging behavior and habitat use by longnose dace, Rhinichthys cataractae. Freshwater Biology 46: 145-160.

Thompson, W.L. 2003. Hankin and Reeves’ approach to estimating fish abundance in small streams: limitations and alternatives. Transactions of the American Fisheries Society 132 (1): 69-75.

van Snik Gray, E. Stauffer, J.R., Jr. 2001. Substrate choice by three species of darters (Teleostei: Percidae) in an artificial stream: effects of a nonnative species. Copeia 1: 254-261.

Warren, Jr. M.L., Angermeier, P.L., Burr, B.M., Haag, W.R. 1997. Decline of a diverse fish fauna: patterns of imperilment and protection in the Southeastern United States. In: Aquatic Fauna in Peril: The Southeastern Perspective. Eds. G.W. Benz and D.E. Collins. Southeast Aquatic Research Institute Special publication I. Lenz Design and Communication, Decatur, GA.

Warren, Jr. M.L., Burr, B.M, Walsh, S.J., Bart, Jr., H.L., Cashner, R.C., Etnier, D.A., Freeman, B.J., Kuhajda, B.R., Mayden, R.L., Robison, H.W., Ross, S.T., Starnes, W.C. 2000. Diversity, distribution, and conservation status of the native freshwater fishes of the Southern United States. Fisheries 25 (10): 7-29.

White, G.C. and K. P. Burnham. 1999. Program MARK: Survival estimation from populations of marked animals. Bird Study 46 Supplement, 120-138: http://www.warnercnr.colostate.edu/~gwhite/mark/mark.htm

Wiens, J.A. 2002. Riverine landscapes: taking landscape ecology into the water. Freshwater Biology 47: 501-515.

Wildhaber, M.L., Lamberson, P.J. 2004. Importance of the habitat choice behavior assumed when modeling the effects of food and temperature on fish populations. Ecological Modeling 175 (4): 395-409.

34

Wu, J., Loucks, O.L. 1995. From balance of nature to hierarchical patch dynamics: a paradigm shift in ecology. Quarterly Review of Biology 70 (4): 439-466.

35

TABLES AND FIGURES

Table 1. Univariate mean and standard deviation (SD) of microhabitat availability and overall darter use data for all sites and seasons Site, season, habitat N depth av. bot. vel. (m/s) % % % % % % % (cm) vel. bedrock boulder cobble gravel sand silt debris (m/s) Coweeta Spring 80 47 (18) 0.41 0.15 (0.14) 1 (7) 10 (22) 48 (33) 10 (14) 26 2 3 (10) constrained (0.19) (28) (8) Coweeta Spring use 80 50 (16) 0.42 0.13 (0.10) 0 (0) 6 (15) 67 (27) 12 (11) 14 0 1 (3) (0.16) (16) (1) Tellico Summer 55 36 (11) 0.39 0.17 (0.12) 3 (15) 22 (31) 56 (33) 14 (17) 5 0 0 (1) constrained (0.18) (13) (0) Tellico Summer 59 31 (18) 0.35 0.17 (0.20) 6 (21) 23 (32) 46 (33) 12 (20) 10 1 2 (7) unconstrained (0.32) (21) (4) Tellico Summer use 58 39 (9) 0.40 0.14 (0.13) 1 (6) 6 (11) 71 (22) 20 (19) 2 (5) 0 0 (0) (0.16) (0) Tellico Autumn 88 29 (13) 0.27 0.11 (0.11) 2 (9) 20 (29) 47 (30) 14 (14) 11 1 5 (14) constrained (.19) (17) (1) Tellico Autumn 57 25 (11) 0.27 0.12 (0.14) 5 (20) 24 (33) 38 (29) 17 (17) 11 0 5 (9) unconstrained (0.23) (15) (1) Tellico Autumn use 90 33 (10) 0.32 0.10 (0.10) 0 (2) 8 (13) 69 (22) 17 (15) 5 (8) 0 1 (3) (0.18) (0)

36

Table 2. Number and percent of darters by season, sex, and size

Sex and season Number of individuals Percent of total Coweeta Spring males 1 5% 45-50 mm 51-59 mm 8 38% ≥ 60 mm 12 57% Coweeta Spring females 17 31% 45-50 mm 51-59 mm 37 69% ≥ 60 mm 0 0% Tellico Summer males 0 0% 45-50 mm 51-59 mm 6 38% ≥ 60 mm 10 62% Tellico Summer females 16 39% 45-50 mm 51-59 mm 17 41% ≥ 60 mm 8 20% Tellico Autumn males 1 4% 45-50 mm 51-59 mm 5 21% ≥ 60 mm 18 75% Tellico Autumn females 24 38% 45-50 mm 51-59 mm 39 62% ≥ 60 mm 0 0%

37

Table 3. Univariate mean and standard deviation (SD) of microhabitat availability and male and female darter (51-59 mm) use for all sites and seasons

Site, season, habitat N depth av. bot. vel. % % % % % % % (cm) vel. (m/s) bedrock boulder cobble gravel sand silt debris (m/s) Coweeta Spring constrained 80 47 0.41 0.15 1 10 48 10 26 2 3 microhabitat availability data Coweeta Spring 7 46 (7) 0.50 0.17 0 (0) 0 (0) 80 (9) 13 (9) 7 (6) 0 0 (0) 51-59 mm male darter use (0.12) (0.12) (0) Coweeta Spring 36 56 0.41 0.13 0 (0) 6 (18) 68 (30) 14 (14) 12 0 0 (1) 51-59 mm female darter use (17) (0.19) (0.11) (14) (0) Tellico Summer constrained 55 36 0.39 0.17 3 22 56 14 5 0 0 microhabitat availability data Tellico Summer unconstrained 59 31 0.35 0.17 6 23 46 12 10 1 2 microhabitat availability data Tellico Summer 6 35 (7) 0.54 0.15 0 (0) 0 (0) 83 (19) 15 (16) 2 (3) 0 0 (0) 51-59 mm male darter use (0.07) (0.15) (0) Tellico Summer 17 37 (9) 0.38 0.20 0 (0) 3 (7) 74 (17) 20 (17) 3 (4) 0 0 (0) 51-59 mm female darter use (0.13) (.16) (0) Tellico Autumn constrained 88 29 0.27 0.11 2 20 47 14 11 1 5 microhabitat availability data Tellico Autumn unconstrained 57 25 0.27 0.12 5 24 38 17 11 0 5 microhabitat data Tellico Autumn 5 38 0.28 0.01 (0) 0 (0) 0 (0) 78 (18) 18 (19) 4 (4) 0 0 (0) 51-59 mm male darter use (13) (0.13) (0) Tellico Autumn 40 30 (9) 0.34 0.12 0 (3) 8 (11) 70 (20) 17 (15) 4 (6) 0 1 (2) 51-59 mm female darter use (0.18) (0.10) (0)

38

Table 4. Univariate mean and standard deviation (SD) of microhabitat availability and size sorted darter use for all sites and seasons

Site, season, habitat N depth av vel. bot. vel. % % % % % % % (cm) (m/s) (m/s) bedrock boulder cobble gravel sand silt debris Coweeta Spring constrained 80 47 0.41 0.15 1 10 48 10 26 2 3 microhabitat availability Coweeta Spring 25 48 (16) 0.37 0.10 0 (0) 5 (11) 63 10 (9) 19 1 2 (4) x ≤ 50 darter use (0.13) (0.07) (27) (21) (2) Coweeta Spring 43 55 (16) 0.43 0.14 0 (0) 5 (16) 70 13(13) 12 0 0 (1) 51 < x < 59 darter use (0.18) (0.11) (28) (13) (0) Coweeta Spring 11 39 (12) 0.42 0.11 0 (0) 10 (16) 62 12 (9) 15 0 1 (2) x ≥ 60 darter use (0.15) (0.08) (24) (13) (0) Tellico Summer constrained 55 36 0.39 0.17 3 22 56 14 5 0 0 microhabitat availability Tellico Summer unconstrained 59 31 0.35 0.17 6 23 46 12 10 1 2 microhabitat availability Tellico Summer 17 42 (11) 0.37 0.10 0 (0) 3 (6) 81 15 1 (1) 0 0 (0) x ≤ 50 darter use (0.16) (0.08) (16) (14) (0) Tellico Summer 23 37 (9) 0.43 0.19 0 (0) 2 (6) 76 19 3 (4) 0 0 (0) 51 < x < 59 darter use (0.14) (0.16) (18) (16) (0) Tellico Summer 18 40 (8) 0.39 0.13 4 (10) 12 (16) 54(23) 27(24) 3 (9) 0 0 (1) x ≥ 60 darter use (0.18) (0.11) (0) Tellico Autumn constrained 88 29 0.27 0.11 2 20 47 14 11 1 5 microhabitat availability Tellico Autumn unconstrained 57 25 0.27 0.12 5 24 38 17 11 0 5 microhabitat availability Tellico Autumn 26 34 (12) 0.26 0.10 0 (0) 12 (17) 65 16 6 (9) 0 1 (2) x ≤ 50 darter use (0.18) (0.10) (27) (17) (0) Tellico Autumn 46 31 (10) 0.32 0.11 0 (3) 7 (10) 71 17 4 (7) 0 1 (2) 51 < x < 59 darter use (0.18) (0.10) (19) (15) (0) Tellico Autumn 18 34 (6) 0.39 0.07 0 (0) 6 (10) 71 17 4 (7) 0 2 (5) x ≥ 60 darter use (0.16) (0.08) (20) (14) (0)

39

Table 5. Summary of capture and recapture history for positively identifiable darters in Tellico Creek, Autumn 2005

originally marked recaptured pass 2 recaptured pass 3 pass 1 46 17 16 pass 2 26 x 8 pass 3 39 x x There were 12 darters that were marked in pass 1 and recaptured in pass 2 and 3.

Table 6. Number and percentage of overall darters moved, by sex, and size class in Tellico Creek, Autumn 2005 meters % of % of darters % of darters % of darters % of % of moved darters 45-50 mm 51-59 mm ≥ 60 mm males females > 5 40 42 38 45 37 46 5 ≥ 10 28 34 33 9 37 8 10 ≥ 15 12 0 10 27 7 23 15 ≥ 20 5 0 5 9 3 8 20 ≥ 30 9 8 10 9 10 8 30 ≥ 40 2 8 0 0 0 8 40 ≥ 50 2 0 5 0 3 0 50 ≥ 60 2 8 0 0 3 0

Table 7. Summary statistics (mean, standard deviation) of darter movement in Tellico Creek Autumn 2005 (downstream mean movements are absolute values)

upstream movement downstream movement

N mean standard deviation N mean standard deviation all gilt darters 22 13.59 15.97 21 7.65 6.98 males 8 9.11 10.44 5 9.75 8.42 females 14 16.14 18.27 16 7.00 6.65 45-50 mm 7 16.34 20.93 5 8.28 11.20 51-59 mm 9 16.21 16.24 12 6.07 3.86 ≥ 60 mm 6 6.43 6.26 4 11.62 8.44

40

Table 8. Results of univariate statistical comparisons of darter movement distances

Wilcoxon rank sum comparisons p-value d.f. upstream versus downstream 0.4877 1,41 males versus females 0.9581 1,41 Wilcoxon ANOVA tests size class comparisons 0.7134 2,40 45-50 mm versus 51-59 mm ns na 45-50 mm versus ≥ 60 mm ns na 51-59 mm versus ≥ 60 mm ns na Capture history comparison 0.0532 2,39 pass 1 versus pass 2 ns na pass 2 versus pass 3 ns na pass 1 versus pass 3 ns na

Table 9. AIC calculations using log likelihood for testing models of varying capture-recapture probabilities to estimate population size in Tellico Creek, Autumn 2005. Values and models in bold are ones in the confidence set.

lower upper Model K LogL AICc ∆i wi mean CI CI population size = capture probability varies, recapture probability is constant 4 1.00 598.543 0.00 0.608 239 164 54488 population size = capture and recapture probability vary 5 0.405 600.349 1.806 0.247 201 164 24914 population size = capture probability and recapture probability are equal and constant 1 0.113 602.897 4.354 0.069 240 213 284 population size = capture probability and recapture probability are constant but not equal 2 0.089 603.382 4.839 0.054 212 185 271 population size = capture probability is constant and recapture probability varies 3 0.036 605.171 6.628 0.022 212 185 271

41

Factor 2 1.00 % Sand 0.43

Coweeta Summer

0.50 Factor 1 Average Velocity 0.75 % Debris -0.46 % Cobble 0.73 % Sand -0.45 Bottom Velocity 0.58 0.00 -1.00 -0.50 0.00 0.50 1.00

Tellico Autumn

-0.50 Tellico Summer

% Boulder -0.86 -1.00

Figure 1: Principal Component Analysis (PCA) of constrained random habitat comparisons across sites and seasons: factors 1 and 2. Ellipses represent mean values and 95% confidence intervals. Only loadings > 0.40 are presented.

42

Factor 3

1.00 Depth 0.65 % Sand 0.50 Average Velocity 0.75 % Cobble 0.73 Bottom Velocity 0.58 0.50 Coweeta Summer Factor 1 % Debris -0.46 % Sand -0.45 0.00 -1.00 -0.50 0.00 0.50 1.00

Tellico Summer

-0.50

Tellico Autumn

-1.00 % Gravel -0.47 Average velocity -0.40

Figure 2: PCA of constrained random habitat comparisons across sites and seasons: factors 1 and 3

43

Factor 3 1.00 Depth 0.65 % Sand 0.50

0.50 Coweeta Summer

Factor 2 % Boulder -0.86 % Sand 0.43 0.00 -1.00 -0.50 0.00 0.50 1.00 Tellico Summer

-0.50

Tellico Autumn

-1.00 % Gravel -0.47 Average velocity -0.40

Figure 3. PCA of constrained random habitat comparisons across sites and seasons: factors 2 and 3

44

Factor 2 1.00 % Boulder 0.74

0.50

Factor 1 Average velocity % Sand -0.64 0.85 % Silt -0.53 0.00 -1.00 -0.50 0.00 0.50 1.00 Tellico Autumn Tellico Summer

-0.50

-1.00 % Cobble -0.84

Figure 4: PCA of Tellico Summer and Autumn unconstrained microhabitat availability

45

Factor 2 1.00 % Boulder 0.55 % Bedrock 0.48 Depth 0.47 Factor 1

% Sand -0.57 0.50 % Silt -0.47 Average velocity 0.77 Bottom velocity 0.73 % Boulder 0.54

0.00 -1.00 -0.50 0.00 0.50 1.00 Unconstrained Constrained

-0.50

-1.00 % Cobble -0.90

Figure 5. PCA of Tellico Summer constrained and unconstrained microhabitat availability

46

Factor 2 1.00 % Sand 0.71 % Silt 0.43

0.50 Factor 1 Average velocity 0.78 % Boulder -0.55 Constrained Bottom velocity 0.58 % Debris -0.50 % Cobble 0.69 0.00 -1.00 -0.50 0.00 0.50 1.00

Unconstrained

-0.50

% Boulder -0.66 -1.00 Bottom velocity -0.44

Figure 6. PCA of Tellico Autumn constrained versus unconstrained microhabitat availability

47

Factor 2 1.00 % Boulder 0.73 Bottom velocity 0.47

0.50 Factor 1 Habitat Average Velocity 0.78 % Sand -0.76 % Cobble 0.70 % Debris -0.43 Bottom Velocity 0.54 0.00 -1.00 -0.50 0.00 0.50 1.00

Darter

-0.50

% Cobble -0.54 -1.00 % Silt -0.40

Figure 7: PCA of Coweeta Spring microhabitat use for all darters compared to constrained random microhabitat availability

48

Factor 2 1.00

% Debris 0.57 Average Velocity 0.54 % Sand 0.51 0.50 Bottom velocity 0.48

Factor 1 Habitat 0.00 -1.00 -0.50 0.00 0.50 1.00 % Cobble -0.91 % Boulder 0.63 % Bedrock 0.57 Darter -0.50

Gilt

-1.00 % Gravel – 0.47

Figure 8: PCA of Summer Tellico microhabitat use for all darters compared to constrained random microhabitat availability

49

Factor 2 1.00 % Boulder 0.63 % Bedrock 0.48

Habitat 0.50

Factor 1 % Sand -0.70 Average velocity 0.75 % Silt -0.57 0.00 Bottom velocity 0.66 -1.00 -0.50 G0.ilt00 0.50 1.00

Darter

-0.50

% Cobble -0.86 -1.00

Figure 9: PCA of Summer Tellico microhabitat use for all darters compared to unconstrained random microhabitat availability, factors 1 and 2

50

Factor 3

1.00 % Gravel 0.74 Depth 0.70

0.50 Factor 1

% Sand -0.70 Darter Average velocity 0.75 % Silt -0.57 Bottom velocity 0.66 0.00 -1.00 -0.50 0.00 0.50 1.00

-0.50 Habitat

-1.00

Figure 10: PCA of Summer Tellico microhabitat use for all darters compared to unconstrained random microhabitat availability, factors 1 and 2

51

Factor 3

1.00 % Gravel 0.74 Depth 0.70

0.50

Factor 2 % Boulder 0.63 % Cobble -0.86 Darter % Bedrock 0.48 0.00 -1.00 -0.50 0.00 0.50 1.00

-0.50 Habitat

-1.00

Figure 11: PCA of Summer Tellico microhabitat use for all darters compared to unconstrained random microhabitat availability, factors 2 and 3

52

Factor 2 1.00 % Boulder 0.56 Focal velocity 0.49

0.50 Factor 1 Habitat % Boulder -0.56 % Cobble 0.74 % Sand -0.48 Average velocity 0.67 % Debris -0.46 Bottom velocity 0.43 0.00 -1.00 -0.50 0.00 0.50 1.00

Darter -0.50

Depth -0.54 -1.00 % Cobble -0.49 %Sand -0.40

Figure 12: PCA of Tellico Autumn microhabitat use for all darters compared to constrained random microhabitat availability

53

Factor 2 1.00 Bottom velocity 0.62 % Boulder 0.50 Average velocity 0.47 Habitat Factor 1 0.50 Average velocity 0.71 % Cobble 0.66 Depth 0.41

0.00 -1.00 -0.50 0.00 Darter 0.50 1.00 % Debris -0.58 % Boulder -0.54 % Silt -0.49 -0.50 % Sand -0.47

% Cobble -0.58 -1.00

Figure 13: PCA of Tellico Autumn microhabitat use for all darters compared to unconstrained random microhabitat availability

54

Factor 2 Average velocity 0.70 % Gravel 0.67 1.00 Bottom velocity 0.64

males Factor 1 % Cobble -0.88 0.50 % Sand 0.79 Average velocity -0.48 % Boulder 0.75 Depth -0.47 % Debris 0.54 Bottom velocity -0.44

0.00 -1.00 -0.50 0.00 0.50 1.00

females -0.50

% Silt -0.40

-1.00

Figure 14. PCA of Coweeta Spring sex linked microhabitat use

55

Factor 2 Average velocity 0.87 Bottom velocity 0.55 males 1.00

Factor 1 0.50 % Gravel 0.81 Depth 0.62 % Cobble -0.96 % Boulder 0.49 % Sand 0.45 0.00 -1.00 -0.50 0.00 0.50 1.00

-0.50 females

% Debris -0.58 -1.00 % Sand -0.49

Figure 15. PCA of Tellico Summer sex linked microhabitat use

56

Factor 2

1.00 Bottom velocity 0.64 % Cobble 0.57 Average velocity 0.52

0.50 females Factor 1 % Gravel 0.80 % Boulder 0.54 % Cobble -0.78 males Average velocity 0.51 0.00 -1.15 -0.65 -0.15 0.35 0.85

-0.50

% Sand -0.72 % Boulder -0.42 -1.00 Depth -0.42

Figure 16. PCA of Tellico Autumn sex linked microhabitat use

57

Factor 2

1.00 % Gravel 0.64 Average velocity 0.63 Bottom velocity 0.42 Darters ≥ 60 Factor 1 0.50 Darters 51 < x < 59 % Cobble 0.86 % Sand -0.76 Average velocity 0.54 % Boulder -0.71 Bottom velocity 0.51 % Debris -0.48 Depth 0.44 0.00 -1.00 -0.50 0.00 0.50 1.00

Darters 45 < x < 50 -0.50

-1.00 % Silt -0.41 % Cobble -0.40

Figure 17. PCA of Coweeta Spring size linked darter microhabitat use

58

Factor 2

1.00 % Sand 0.44 Depth 0.42 % Bedrock 0.41

Factor 1 Darters 45 < x < 50 0.50 Darters ≥ 60

% Cobble -0.94

0.00 -1.00 -0.50 0.00 0.50 1.00

% Gravel 0.58 -0.50 Depth 0.49 Darters 51 < x < 59 % Bedrock 0.48 % Boulder 0.40

% Gravel -0.64 -1.00 Bottom velocity -0.54

Figure 18. PCA of Tellico Summer size linked darter microhabitat use

59

Factor 2 1.00 % Gravel 0.66 Average velocity 0.64 Bottom velocity 0.55

Factor 1 0.50 Darters 51 < x < 59 Darters ≥ 60 % Boulder 0.73 % Cobble -0.93 % Sand 0.64 % Gravel 0.43 0.00 -1.00 -0.50 0.00 0.50 1.00

-0.50 Darters 45 < x < 50

Depth -0.54 -1.00 % Sand -0.52

Figure 19. PCA of Tellico Autumn size linked analysis

60

7

6

5

y 4 nc que e

r 3 f

2

1

0 -30 -26 -22 -18 -14 -10 -6 -2 2610 14 18 22 26 30 34 38 42 46 50 54 58

movement distance) (

Figure 20. Movement frequency of darters in Tellico Creek Autumn 2005

61

7

6

5

females

y 4 males nc que e

r 3 f

2

1

0 -30 -26 -22 -18 -14 -10 -6 -2 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58

movement di t

Figure 21. Sex linked movement frequency of darters in Tellico Creek Autumn 2005

62

7

6

5 > 60 mm 51-59 mm

y 4 45-50 mm nc que e

r 3 f

2

1

0 -30 -26 -22 -18 -14 -10 -6 -2 2 6 10 14 18 22 26 30 34 38 42 46 50 54 58

movement distance Figure 22. Size linked movement frequency of darters on Tellico Creek Autumn 2005

63

SUMMARY AND CONCLUSIONS

Quantifying a species’ habitat use is a prerequisite for scientifically based management of the species. Understanding how a species relates to its environment can aid in the construction of ecological models; potentially be extrapolated to other systems; and, more simply, provide information on habitat quality and variation. I quantified microhabitat use, movement patterns, and population estimates for gilt darters within the Southeastern United States to address the paucity of known information for this species.

I examined the microhabitat use of gilt darters in three seasons at two sites. I looked at whether darters differed in microhabitat use from randomly available microhabitat, whether male and female gilt darters utilized different microhabitats, and whether microhabitat selection by gilt darters differed based on fish size. Gilt darters consistently used microhabitats that were different than what were randomly available, specifically selecting more erosional type microhabitats

(more cobble and larger substrata size, higher average velocities). The applicability of this microhabitat selection to other streams is not yet quantified, but our results are similar to conclusions made for similar darter species in similar systems (Bowen et al. 1998; Freeman et al.

2001). These results suggest that continued stream degradation in the form of siltation, sedimentation and impoundment would reduce the type of microhabitats and water velocities used by gilt darters in this study.

Both the sex-based and size-based analyses of gilt darter microhabitat use showed partitioning of microhabitat by sex and size, but the intersite and interseason comparisons are

64

somewhat equivocal. Nevertheless, large gilt darters used more heterogeneous substrata than the other two size classes of darters, and males deviated farther from the microhabitat availability than females, selecting areas with more cobble and higher average velocities. Whether these microhabitat use patterns by large darters and male darters confers a fitness advantage cannot be answered by these data, but previous research indicates that habitat preferences can result in increased fitness (Hill and Grossman 1993; Petty and Grossman 1996; Harding et al. 1998; Page

2000; Thompson et al. 2001; Grossman et al. 2002; Rosenfeld 2003; Wildhaber and Lamberson

2004; Mattingly and Galat 2004).

The movement and population estimates in this study are from only one site and season.

Nevertheless, they represent novel information regarding gilt darter populations and movement patterns. More data is needed for extrapolation, however the long distance movements we observed indicate that gilt darters do not adhere strictly to the Restricted Movement Paradigm

(Gowan et al. 1994; Rodriguez 2002). Whether the trends I observed are indicative of ‘movers’ and ‘stayers,’ as has been indicated in previous studies (Fraser et al. 2001; Petty and Grossman

2004), will have to be answered over larger spatial and temporal scales.

In conclusion, based on the use of more erosional type microhabitats by gilt darters, stream degradation will likely negatively influence population persistence as it has for other species (Sutherland et al. 2002; Tabit and Jonson 2002; Bolliet et al. 2005).

Given gilt darter declines in abundance (Hatch 1985; NatureServe Explorer 2006) and concern with population status (http://georgiawildlife.dnr.state.ga.us/content/specialconcernanimals.asp, http://www.fs.fed.us/r9/wildlife/, http://www.dec.state.ny.us/website/dfwmr/wildlife/endspec/giltdart.html),

65

the quantification of habitat use, movement, and population size of gilt darters provides imperative information for future conservation and management of this species, especially within the ichthyofaunally speciose Southeastern United States.

LITERATURE CITED

Bolliet, V., Bardonnet, A., Jarry, M., Vignes, J.C., Gaudin, P. 2005. Does embeddedness affect growth performance in juvenile salmonids? An experimental study in brown trout Salmo trutta L. Ecology of Freshwater Fish 14: 289-295.

Bowen, Z.H., Freeman, M.C., Bovee, K.D. 1998. Evaluation of generalized habitat criteria for assessing impacts of altered flow regimes on warmwater fishes. Transactions of the American Fisheries Society 127: 455-486.

Freeman, M.C., Bowen, Z.H., Bovee, K.D., Irwin, E.R. 2001. Flow and habitat effects on juvenile fish abundance in natural and altered flow regimes. Ecological Applications 11(1): 179-190.

Georgia Department of Natural Resources, Wildlife Resources Division, Georgia Natural Heritage Program. Krakow, G., data manager. Special Concern Animal Species in Georgia. http://georgiawildlife.dnr.state.ga.us/content/specialconcernanimals.asp Updated [10/22/2004] [Accessed 13 January 2005]

Gowan, C., Young, M.K., Fausch, K.D., Riley, S.C. 1994. Restricted movement in resident sream salmonids-a paradigm lost. Canadian Journal of Fisheries and Aquatic Sciences 51 (11): 2626-2637.

Grossman, G.D., Rincon, P.A., Farr, M.D., Ratajczak Jr, R.E. 2002. A new optimal foraging model predicts habitat use by drift feeding stream minnows. Ecology of Freshwater Fish 11: 2-10.

Harding, J.M., Burky A.J., Way, C.M. 1998. Habitat preferences of the rainbow darter, Etheostoma caeruleum, with regard to microhabitat velocity shelters. Copeia 4: 988-997.

Hatch, J. T. 1985. Distribution, habitat, and status of the gilt darter (Percina evides) in Minnesota. Journal of the Minnesota Academy of Science 51(2): 11-16.

Hill. J., and Grossman, G.D. 1993. An energetic model of microhabitat use for rainbow trout and rosyside dace. Ecology 74: 685-698.

66

Mattingly, H.T., Galat, D.L. 2004. Predictive performance of a summer microhabitat model for the threatened Niangua darter, Etheostoma nianguae. Journal of Freshwater Ecology 19 (1): 109-114.

Nature Serve Explorer, version 4.7. 2006. Percina evides status report. http://www.natureserve.org/explorer/servlet/NatureServe?init=Species [Accessed 21 April 2006].

New York State Department of Environmental Conservation. Gilt darter fact sheet. 21 December 2004. http://www.dec.state.ny.us/website/dfwmr/wildlife/endspec/giltdart.html [accessed 8 April 2005]

Page, L.M. 2000. Etheostomatinae. in Percid Fishes Systematics, Ecology, and Exploitation . John F. Craig, ed. volume 3 of Fish and Aquatic Resources Series. Tony J. Pitcher, ed. Blackwell Science, Oxford, UK.

Petty, J.T. Grossman, G.D. 1996. Patch selection by mottled sculpin (Pisces: Cottidae) in a southern Appalachian stream. Freshwater Biology 35: 261-276.

Rodriguez, M.A. 2002. Restricted movement in stream fish: the paradigm is incomplete, not lost. Ecology 83 (1): 1-13.

Rosenfeld, J. 2003. Assessing the habitat requirements of stream fishes: an overview and evaluation of different approaches. Trans. of the Am. Fisheries Soc. 132 (5): 953- 968. Sutherland, A. B., Meyer, J.L., Gardiner, E.P. 2002. Effects of land cover on sediment regime and fish assemblage structure in four southern Appalachian streams Freshwater Biology 47( 9): 1791-1805. Tabit, C.R., Johnson, G.M. 2002. Influence of urbanization on the distribution of fishes in a southeastern upper Piedmont drainage. Southeastern Naturalist 1(3): 253-268. Thompson, A.R., Petty, J.T., Grossman, G.D. 2001. Multi-scale effects of resource patchiness on foraging behavior and habitat use by longnose dace, Rhinichthys cataractae. Freshwater Biology 46: 145-160.

Wildhaber, M.L., Lamberson, P.J. 2004. Importance of the habitat choice behavior assumed when modeling the effects of food and temperature on fish populations. Ecological Modeling 175 (4): 395-409.

67