Copeia,2001(1), pp. 254-261

Substrate Choice by Three of Darters (Teleostei: ) in an Artificial Stream: Effects of a Nonnative Species

ELLEN VAN SNIK GRAYAND JAY R. STAUFFERJR.

Etheostomazonale, the banded darter, was introduced to the SusquehannaRiver basin, Pennsylvania, through an interbasin transfer. We examined the effects of darter density and the presence of E. zonaleon the substrate choice of two native darter species (Etheostomaolmstedi, the , and Percinapeltata, the shield darter) in an artificial stream. In single species trials, E. olmstedi,E. zonale, and P. peltata exhibited nonrandom substrate selection and occupied patches of large substrate significantly(P < 0.05) more often than expected. No intraspecific density effects were observed. The presence of E. zonaledid not affect the sub- strate choice of P. peltata but did induce a shift of E. olmstedifrom large to small substrate in high-densitytrials (P < 0.05). Etheostomazonale was most frequently the aggressor in behavioral interactions. Neither E. olmstedinor P. peltatadirected any agonistic behavior toward E. zonale;however 40% of aggressive acts initiated by E. zonalewere directed toward E. olmstediand P. peltata. Displacement of E. olmstediby the aggressor E. zonale suggests potential deleterious effects on the native species.

THE introduction, establishment, and prolif- throughout the Susquehanna basin (Raes- eration of nonnative fishes has become ly, 1991). Etheostomazonale has hybridized with a one of the primary factors contributing to the native darter species, Etheostomaolmstedi, the tes- endangerment of native fishes, secondary only sellated darter, in the basin to habitat alteration (Taylor et al., 1984; Miller (Raesly et al., 1990). In a companion study, Gray et al., 1989; Lassuy, 1995). In North America, 70 (1998) compared the habitat use of E. olmstedi fish species of foreign origin (exotic) have be- in sympatry and allopatry with the nonnative E. come established (Courtenay et al., 1986; Cour- zonale. olmstedi exhibited a habitat tenay, 1995; Courtenay and Moyle, 1996). Most shift when sympatric with E. zonale; E. olmstedi nonnative fishes in the (150 spe- was found in shallower, slower velocity habitats cies), however, are the result of interbasin rath- with smaller substrates in sites sympatric with E. er than intercontinental transfers (Courtenay zonale compared to sites in which E. zonale was and Moyle, 1996). Interbasin transfers include absent (Gray, 1998). Thus, it appeared that E. fishes that have been intentionally or uninten- zonale excluded E. olmstedifrom preferred riffle tionally transplanted beyond their native range and run habitats in sympatry, restricting E. olms- in North America (Moyle et al., 1986; Ross, tedi to pool and stream margin habitats. 1991; Courtenay and Moyle, 1996). Although The purpose of this study was to examine the many studies have focused on impacts of exotic substrate choice of E. zonale, E. olmstedi,and Per- fish on native fauna, comparatively few studies cina peltata, the shield darter, a second native have examined effects of interbasin transfers on species, in an artificial stream, thus allowing native fishes despite their numerical dominance control of abiotic and biotic factors not possible (Ross, 1991). Nonnative fishes introduced by in- in the field. We chose to examine substrate terbasin transfers have affected native fishes choice because previous studies have shown through hybridization (Hocutt and Hambrick, substrate size to be important in darter habitat 1973), predation (Lemly, 1985), displacement selection (Kessler et al., 1995; Stauffer et al., from preferred habitats (Brown and Moyle, 1996; Gray, 1998), and segregation by habitat is 1991; Douglas et al., 1994), and suppression of predicted to be important among fishes (Wer- population size or biomass (Moore et al., 1983; ner, 1977). Our objectives were to examine the Lemly, 1985; Nakano et al., 1998). relative effects of intraspecific and interspecific Etheostomazonale, the banded darter, was most biotic interactions on the substrate selection of likely introduced to the Susquehanna River ba- E. olmstedi, E. zonale, and P peltata, determine sin in the 1960s via a bait-bucket introduction whether E. zonale induced a habitat shift of E. from an Allegheny River drainage population olmstedior P peltata, and determine whether ag- (Kneib, 1972; Gray, 1998) and is currently the gressive interactions occurred among the three most abundant darter in many localities species.

? 2001 by the American Society of Ichthyologists and Herpetologists GRAYAND STAUFFER--DARTER SUBSTRATE CHOICE 255

sectionI secti~on 2 CD ChillerJeturnits tubes that emptied at the stream bottom at the lnozzlefrorpurnp upstream end of each substrate patch in each Channel wall 0 --- Permanentsclreell barrier of the four sections. Sbsthvate divider (PVC The native were collected section 4 section 3 withattached seine) species by seining in the River Directi of i i : Susquehanna basin, Pennsylvania, flow=;i including Conococheague Creek, Emma Creek, Little River, Marsh Creek, Creek, 0O Juniata Sugar and Swatara Creek. was col- 10 I 0.91Mlected from Little Pine Cl"K~0iro-jW'o ~r Creek, Pennsylvania. 2.04m0.41 m--- Fish were kept in holding tanks by species, fed as during the experiments, and kept for a min- __0"41 ___.04 imum of four weeks before use in experiments. 1. of the artificial viewed Fig. Diagram stream, Only adult darters were used in experiments, from above (not to scale). and the fish were not sexed. Size ranges of dart- ers used were 4-9 mm for E. olmstedi (mean = MATERIALSAND METHODS 6.25 mm) and P peltata (mean = 6.32 mm) and 4-6 mm for E. zonale (mean = 5.35 mm). The experiments were conducted in an in- We designed substrate choice trials to exam- door, recirculating, acrylic artificial stream with ine the effect of interspecific versus intraspecific four 0.91 m x 2.04 m experimental sections effects upon the three species' substrate choice. separated by 0.7 mm mesh screens (Fig. 1). Densities in single species trials included 6, 12, Growlights above the stream were on a Lutron'1 and 24 fish, corresponding to 3, 6, and 12 fish/ lighting system that allowed for a 14:10 L:D m2, which did not exceed natural fish densities summer photoperiod including a 2 h sunrise (Grady and Bart, 1987; Ross et al., 1990; Gray, and sunset period in which light intensity slowly 1998). Results of three trials were compared to changed. Two chiller units maintained water determine whether E. zonale induced a habitat temperature at 20.6 C +- 1 C. Water depth was shift of each native species: (1) E. olmstedior P maintained at 25.0 ? 0.8 cm. Two water pumps peltata held alone; (2) E. olmstedi + P peltata; discharging upstream of each section main- and (3) E. olmstedi+ P peltata + E. zonale. It was tained water velocity at 0.20 ? 0.04 m/sec on necessary to include two experimental designs the inside of the stream and 0.14 ? 0.03 m/sec to have uncomfounded conclusions regarding on the outside of the stream. The pH was 7.6, interspecific interactions. We also included two and ammonia was below detectable amounts. A density levels for each experimental design. The curtain enclosed the artificial stream to mini- additive experimental design tested for the ex- mize external disturbances during experiments. istence of interspecific biotic interactions upon The entire bottom of each section was cov- substrate choice (Underwood, 1986; Fausch, ered with 2.6 cm of sand. Within each of the 1988). In this design, the total density of fish in four sections, five patches of differing substrate the three trials varies; however the number of sizes (diameters) were installed: 2-4 cm, 4-6 conspecifics is fixed across trials. The low-den- cm, 6-10 cm, 10-25 cm, and a mixture of the sity trials for the additive design were (1) 6o, four substrate sizes (Fig. 1). All substrate but the (2) 6o + 6p, and (3) 6o + 6p + 6z (o = E. largest size class was obtained from a local quar- olmstedi,p = P. peltata, and z = E. zonale); and ry. Substrate 10-25 cm in diameter was collect- the high-density trials were (1) 12o, (2) 12o + ed from a local stream, cleaned, dried indoors 12p, and (3) 12o + 12p + 12z. The substitutive for five months, and scrubbed a second time design tests the relative strength of interspecific before use. The order of the substrates was ran- relative to intraspecific biotic interactions for domly assigned, and substrate was placed on the each species (Underwood, 1986; Fausch, 1988). sand in each patch. Substrate patches were sep- In this design, the total fish density in the three arated by 4.45 cm PVC pipes that contained a experiments is fixed, and the number of con- rolled seine. All four seines within a section specifics varies across trials. The low-density tri- were tied to a bar on the outside of the stream, als for the substitutive design were (1) 12o, (2) such that the seines could be lifted vertically in 60 + 6p, and (3) 40 + 4p + 4z, and the high- unison, thus simultaneously isolating the five density trials were (1) 240, (2) 120 + 12p, and substrate patches and preventing darter move- (3) 80 + 8p + 8z. ment from patch to patch during data collec- The order of experiments as well as which tion. We fed fish a combination of live Califor- section was used for each experiment was de- nia blackworms and ground trout pellets three termined randomly. All first replicate experi- times daily at 0800, 1200, and 1600 through ments were completed before beginning the 256 COPEIA, 2001, NO. 1 second replicates. Fish were selected randomly To examine the substrate preference of each from holding tanks, measured to the nearest 0.5 species, a X2 test was used. For each species and cm standard length (SL), and placed in the each density (6, 12, and 24 individuals), we stream. Each experiment included 72 h of ac- compared the frequency of use of small and climation in the artificial stream, and observa- large substrate patches with the expected fre- tions were made on days 3, 5, and 7. Observa- quency of use based on an equal distribution of tions were conducted at three different times: fish across the two substrate size classes. If no 0900, 1300, and 1500 (hereafter referred to as substrate selection occurred, we expected 40% morning, afternoon, and evening, respectively), of individuals to be in small substrate patches one hour after feeding, and the order of these and 60% of individuals to be in large substrate observation times was determined randomly. patches because there were two small substrate First, the fish in each section were observed for patches and three large substrate patches. For a period of 15 min, and behavioral data were the experiments testing a density of six fish, collected. Each visible fish was observed for 1 Fisher's exact test was used because the expect- min and aggressive interactions were recorded. ed frequency of at least 20% of the cells was less Aggressive behaviors were classified into cate- than five (Glantz, 1992). Fisher's exact test was gories used by Hartman (1965): nips, chases, also used to determine whether intraspecific ef- and displays (only lateral displays were ob- fects on substrate choice occurred. Each spe- served). In addition, the position of the fish cies' substrate choice was compared across the with respect to the substrate was recorded at three densities to determine whether substrate first sighting: in the water column, under, or selection changed with density of conspecifics. above (on the top surface of) substrate. Follow- We used X2tests of homogeneity to test for ing the behavioral data, the seines were lifted habitat shifts resulting from interspecific effects. simultaneously, and the water pumps were Again, Fisher's exact test was used in experi- turned off. The number and species of fish in ments with small sample sizes. The results of each substrate patch were recorded. A day of three trials were compared to determine wheth- reacclimation was allowed between observations er E. zonale induced a habitat shift of either of because disturbance of substrate was necessary the native species as outlined previously: (1) E. to count all the fish in each section. All five sub- olmstedior P peltata; (2) E. olmstedi + P peltata; strate patches were disturbed during data col- and (3) E. olmstedi+ P peltata + E. zonale. When lection. significance was found, further X2 tests were We used individual fish in more than one ex- used to determine which experiments differed. periment because of facility restraints. In select- Reduced alpha levels (ot/n, where n is the num- ing fish for the experiments, all fish of a given ber of pairwise comparisons) were used to cor- species were used before using any of the fish a rect for experimentwise error. In addition, the second time (Tyler and Clapp, 1995). On aver- substrate choice of E. zonale in single species tri- age, each individual E. olmstediwas used in three als was compared to the substrate choice of E. experiments, each E. zonale was used twice, and zonale when held with the two native species to each P peltata was used in four experiments. determine whether the substrate choice of E. We used a two-factor analysis of variance to zonale was affected by the native species. determine the effect of time of day and average SL on substrate selection by the three darter RESULTS species. P-values were considered significant if less than or equal to 0.05 for all statistical tests. In single species trials at all three densities, E. Neither time of day nor average SL influenced olmstedi, E. zonale, and P peltata exhibited non- darter substrate selection (F = 1.35, df = 149, random substrate selection and selected large P = 0.26). Therefore, data from the three time substrate significantly more often than expected periods were pooled for subsequent analyses. (P < 0.05; Table 1). None of the three species Examination of histograms indicated that dart- showed intraspecific effects on substrate choice ers did not discriminate among all five substrate (P > 0.05; Table 1). patches. Instead, darters discriminated among Mixed species trials examined interspecific ef- patches that allowed use of interstitial spaces for fects on the three species' substrate use. At low cover. Therefore, the data from patches with densities, there were no significant differences substrate 2-4 cm and 4-6 cm were pooled and in substrate choice of E. olmstediin the additive referred to as "small substrate," and the data (60/60 + 6p/6o + 6p + 6z) and substitutive from patches with substrate 6-10 cm, 10-25 cm, (120/6o0 + 6p/40 + 4p + 4z) analyses (P > and a mixture of all substrate sizes were pooled 0.05; Table 1). At high densities, however, the and referred to as "large substrate." substrate use by E. olmstediwas significantly dif- GRAYAND STAUFFER-DARTER SUBSTRATE CHOICE 257

TABLE1. AVERAGESIZE (mm SL) ANDPERCENT OF INDIVIDUALSPER SPECIES OCCURRING IN SMALLAND LARGE SUBSTRATEPATCHES IN EXPERIMENTALTRIALS. Abbreviations include o = Etheostomaolmstedi, p = Percinapeltata, and z = Etheostomazonale.

E. olmstedi P peltata E. zonale Average Average Average size Small Large size Small Large size Small Large Trial Time of day (mm) substrate substrate (mm) substrate substrate (mm) substrate substrate 6 morning 6.2 0.0% 100.0% 5.6 16.7% 83.3% 5.2 0.0% 100.0% afternoon 33.3% 66.7% 16.7% 83.3% 0.0% 100.0% evening 16.7% 83.3% 0.0% 100.0% 0.0% 100.0% morning 6.3 16.7% 83.3% 6.7 0.0% 100.0% 5.5 0.0% 100.0% afternoon 0.0% 100.0% 0.0% 100.0% 0.0% 100.0% evening 0.0% 100.0% 16.7% 83.3% 16.7% 83.3% 12 morning 6.3 20.0% 80.0% 6.1 0.0% 100.0% 5.0 0.0% 100.0% afternoon 0.0% 100.0% 18.2% 81.8% 0.0% 100.0% evening 0.0% 100.0% 0.0% 100.0% 0.0% 100.0% morning 6.3 0.0% 100.0% 6.5 25.0% 75.0% 5.7 16.7% 83.3% afternoon 0.0% 100.0% 8.3% 91.7% 16.7% 83.3% evening 0.0% 100.0% 16.7% 83.3% 8.3% 91.7% 24 morning 6.3 25.0% 75.0% 6.5 16.7% 83.3% 5.0 12.5% 87.5% afternoon 8.3% 91.7% 8.3% 91.7% 8.3% 91.7% evening 16.7% 83.3% 8.3% 91.7% 12.5% 87.5% morning 6.1 8.3% 91.7% 6.6 4.2% 95.8% 5.3 12.5% 87.5% afternoon 8.3% 91.7% 12.5% 87.5% 4.2% 95.8% evening 4.2% 95.8% 16.7% 83.3% 8.3% 91.7% 60+6p morning 6.3 0.0% 100.0% 5.7 33.3% 66.7% - - - afternoon 0.0% 100.0% 33.3% 66.7% - - evening 16.7% 83.3% 33.3% 66.7% - - morning 6.3 16.7% 83.3% 5.9 16.7% 83.3% - - - afternoon 16.7% 83.3% 33.3% 66.7% - - evening 0.0% 100.0% 0.0% 100.0% - - 12o0+12p morning 5.7 10.0% 90.0% 5.9 0.0% 100.0% - - - afternoon 10.0% 90.0% 0.0% 100.0% evening 0.0% 100.0% 0.0% 100.0% - - morning 6.6 25.0% 75.0% 6.8 16.7% 83.3% - - - afternoon 33.3% 66.7% 8.3% 91.7% - - evening 0.0% 100.0% 8.3% 91.7% - - 4o+4p+4z morning 5.6 25.0% 75.0% 6.5 0.0% 100.0% 5.4 0.0% 100.0% afternoon 0.0% 100.0% 25.0% 75.0% 0.0% 100.0% evening 0.0% 100.0% 25.0% 75.0% 0.0% 100.0% morning 7.5 0.0% 100.0% 7.1 0.0% 100.0% 5.9 0.0% 100.0% afternoon 0.0% 100.0% 0.0% 100.0% 0.0% 100.0% evening 25.0% 75.0% 50.0% 50.0% 0.0% 100.0% 6o+6p+6z morning 6.5 16.7% 83.3% 5.8 0.0% 100.0% 4.8 16.7% 83.3% afternoon 50.0% 50.0% 16.7% 83.3% 16.7% 83.3% evening 0.0% 100.0% 0.0% 100.0% 0.0% 100.0% morning 5.9 50.0% 50.0% 6.6 16.7% 83.3% 5.2 0.0% 100.0% afternoon 16.7% 83.3% 0.0% 100.0% 16.7% 83.3% evening 16.7% 83.3% 16.7% 83.3% 16.7% 83.3% 8o+8p+8z morning 6.6 0.0% 100.0% 6.7 12.5% 87.5% 5.6 0.0% 100.0% afternoon 50.0% 50.0% 0.0% 100.0% 12.5% 87.5% evening 37.5% 62.5% 25.0% 75.0% 12.5% 87.5% morning 6.1 50.0% 50.0% 6.3 12.5% 87.5% 5.9 0.0% 100.0% afternoon 12.5% 87.5% 0.0% 100.0% 0.0% 100.0% evening 75.0% 25.0% 0.0% 100.0% 0.0% 100.0% 12o+12p+12z morning 5.5 41.7% 58.3% 6.0 41.7% 58.3% 5.1 0.0% 100.0% afternoon 75.0% 25.0% 25.0% 75.0% 25.0% 75.0% evening 18.2% 81.8% 0.0% 100.0% 0.0% 100.0% morning 6.4 33.3% 66.7% 6.6 25.0% 75.0% 5.5 8.3% 91.7% afternoon 66.7% 33.3% 25.0% 75.0% 8.3% 91.7% evening 41.7% 58.3% 8.3% 91.7% 0.0% 100.0% 258 COPEIA, 2001, NO. 1

80 0.06 E. olmstedi OSmall substrate 70 U 0.05 iDE. olmstedi (2) Large substrate SP peltata (19) ,60 0.04 E. zonale (15) -50 0.03 40 S0.02 a 30 20 10 Nips Chases Displays Behavior of the aggressor 0 3. Number and of behaviors 12o 12o+12p 12o+12p+12z Fig. aggressive per minute initiated by Etheostomaolmstedi, pel- 70 tata,and Etheostomazonale. Numbers in parenthesesof P. peltata 0 Smallsubstrate indicate total number of acts initi- 60 legend aggressive N Largesubstrate ated per species. 45 c?40 not direct any agonistic behavior toward P pel- 0l30 tata or E. zonale (Figs. 3-4). Of 245 P peltata F-20 observed, two nips, 12 chases, and five displays were observed 3). (16 10 (Fig. Eighty-four percent acts) of the agonistic acts initiated by P peltata 0 were directed at conspecifics, and 16% (3 acts) 12o 12o+12p 12o+12p+12z were directed at E. olmstedi;neither E. olmstedi nor P directed behavior to- 2. Number of Etheostomaolmstedi and Percina peltata any agonistic Fig. ward E. zonale Fifteen acts observed in small and substrate (Fig. 4). agonistic peltata large patches were observed in which E. zonale was the in single and mixed species trials (including both rep- ag- five two and licates), illustratingthe shift of E. olmstedito smallsub- gressor, including nips, displays, stratesin the presence of the nonnativeEtheostoma zon- eight chases (183 E. zonale were observed; Fig. ale. Percinapeltata did not exhibit a habitat shift. Ab- 3). Etheostomazonale directed nine agonistic acts breviationsinclude o = E. olmstedi,p = P peltata,and intraspecifically (60%) and three acts (20% = z E. zonale. each) each toward E. olmstediand P peltata (Fig. 4). The number of agonistic acts performed per minute was therefore 0.009, 0.078, and 0.082, ferent when E. zonale was present than when E. for E. olmstedi, P peltata, and E. zonale, respec- < olmstedi occurred alone or with P peltata (P tively. Sixty-four percent of the aggressive acts 0.05; Table 1; Fig. 2). This habitat shift from by the three species occurred during mixed spe- large to small substrate (Fig. 2) was evident in cies experiments, including 44% in trials includ- both the substitutive (24o/12o + 12p/8o + 8p ing all three species. + 8z) and additive (12o/12o + 12p/12o + 12p The three species differed in their position + 12z) analyses (P < 0.05; Table 1, Fig. 2). with respect to the substrate. Percina peltata ex- There were no significant differences among hibited vertical segregation from E. olmstediand the substrate choice of R peltata multispecies tri- als at low or high densities in the additive or substitutive analyses (P > 0.05; Fig. 2; Table 1). 16 BE. olmstedi (2) Percina peltata occupied large substrate habitats 14 OP peltata (19) zonale regardless of the presence of other species. In 12- ME. (15) addition, the presence of E. olmstediand P pel- C 10 tata did not affect the substrate choice of E. zon- I- ale. There were no significant differences in the 6 substrate use by E. zonale at low or high densities in the additive or substitutive analyses (P > 0.05;

Table 1). Conspecific E. olmstedi P. peltata E. zonale In 657 min of behavioral observations, E. zon- Species to which aggression directed ale was most the in frequently aggressor agonis- Fig. 4. Frequencyof aggressivebehavior directed and chases were the most com- tic interactions, toward conspecifics and heterospecificsby Etheostoma mon aggressive behavior observed (Fig. 3). Of olmstedi,Percina peltata, and Etheostomazonale. Numbers 229 E. olmstediobserved, one nip and one dis- in parenthesesof legend indicate total number of ag- play were directed conspecifically; E. olmstedidid gressive acts initiated per species. GRAYAND STAUFFER--DARTER SUBSTRATE CHOICE 259

0.9 IMEoolnmstedi Brown and et 0.8 1994; Brasher, 1995; Greenberg [UP. peltata al., 1997), or exclusion from 0.7 ME. zonale primary spawning 0.70,6 habitat causing lowered reproductive success re- sulting from suboptimal oxygenation of eggs 0.4 (Gray, 1998). The habitat shift exhibited by E. olmstedican be attributed to acts E. zonale. Al- o.1 agonistic by though 36 aggressive interactions were ob- Above Under Water Column served, the probability of observing agonistic Position with respect to the substrate acts in three 15-min periods per week is low (Di- amond, 1978). More acts minute 5. of Ethoestomaolmstedi, Percina agonistic per Fig. Proportion were initiated E. zonale than R peltata, and Etheostomazonale individuals observed by (0.082) by pel- above substrate,under substrate,and in the watercol- tata (0.078) or E. olmstedi (0.009). In addition, umn. the percentage of aggressive interactions direct- ed at heterospecifics varies by species: 0%, 16%, and 40%, respectively, for E. olmstedi,R peltata, E. zonale by occurring in the water column in and E. zonale. Although neither E. olmstedinor >75% of the observations (Fig. 5). Both E. olms- P peltata directed aggressive behavior toward E. tedi and E. zonale occurred on top of the sub- zonale, 40% of aggressive acts initiated by E. zon- strate in >75% of the observations; however E. ale were directed toward the two native species, olmstediwas found under rocks (21.4% of the supporting interference competition for space observations) more frequently than E. zonale as the mechanism in the habitat shift exhibited (14.8% of the observations, Fig. 5). by E. olmstedi. This scenario is similar to that found by Hindar et al. (1988): cutthroat trout DISCUSSION (Salmo clarki) excluded dolly varden (Salvelinus malma) from littoral and near-surface habitats in Habitat shifts of native species are one of the sympatry through aggressive dominance. Etheos- most likely immediate effects of introduction toma zonale may also influence the habitat use events (Werner, 1984). In this study, substrate of other darters in its native range (White and choice by E. olmstediwas influenced by the pres- Aspinwall, 1984). White and Aspinwall (1984) ence of a nonnative species. Etheostomaolmstedi found that the habitat use of Etheostoma blen- exhibited a habitat shift from preferred large nioides, the greenside darter, and Etheostomate- substrate to small substrates when E. zonale was trazonum, the Missouri saddled darter, differed present at high densities in both the additive in sympatry and allopatry with E. zonale. Etheos- and substitutive experimental designs. Shifts toma zonale occurred frequently in submergent were not observed at low densities. However as vegetation (Potamogeton),and E. blennioidesand space became limiting at higher densities, E. E. tetrazonum exhibited a marked increase in zonale displaced E. olmstedifrom preferred larger their use of Potamogetonwhere E. zonale was ab- substrates through aggressive interactions. The sent (White and Aspinwall, 1984). greatest effect on the substrate choice of E. olms- Percina peltata occupied large substrates re- tedi was not the density of conspecifics but the gardless of intraspecific or interspecific darter presence of E. zonale. The interaction was asym- densities. Many ecological studies have high- metric, with neither native species affecting the lighted ecological differences between the gen- habitat use of E. zonale. The shift observed in era Etheostoma and Percina (Greenberg, 1991; this study supports the results of the field study Gray, 1998; Welsh and Perry, 1998). Percina pel- by Gray (1998), in which E. olmstedioccurred in tata exhibits vertical segregation from E. olmstedi shallower, slower velocity habitats with smaller and E. zonale, in addition to occupying water ve- substrates in sites sympatric with E. zonale com- locities in between E. zonale (fastest water veloc- pared to allopatric sites. Further studies on the ities) and E. olmstedi (slowest water velocities, effect of E. zonale on the growth, survival, and Gray, 1998). Because R peltata uses substrate for reproduction of E. olmstediare required to de- cover less frequently than the Etheostomaspecies, termine the implications of the habitat shift for it is not surprising that neither E. olmstedinor populations of E. olmstediin the Susquehanna E. zonale affects its substrate selection. River basin. However, restriction of E. olmstedito There is a need for more studies of interbasin shallow pool and marginal habitats with smaller transfers. As Ross (1991) pointed out, although substrates may result in a reduced food base there are at least double the number of inter- (Schlosser, 1987; Vogt and Coon, 1990), in- basin transfers than exotic species introductions creased susceptibility to predation (Greenberg, in North America, exotic species are studied dis- 260 COPEIA, 2001, NO. 1 proportionally. Understanding the mechanisms 1994. Indigenous fishes of western North America of species interactions among unnaturally sym- and the hypothesis of competitive displacement: Meda as a case fish species may us into com- fulgida (Cyprinidae) study. Copeia patric give insight 1994:9-19. 1978; Fausch, munity organization (Diamond, K. D. 1988. Tests of between as well as remind us of the of FAUSCH, competition 1988) importance native and introduced salmonids in streams: What conservatism when new fish using introducing have we learned? Can. J. Fish. Aquat. Sci. 45:2238- species beyond their natural range. 2246. GLANTZ,S. A. 1992. Primer of biostatistics. McGraw- ACKNOWLEDGMENTS Hill, Inc., . GRADY,J. M., AND H. L. BARTJR. 1987. Life history of We are indebted to T. Stecko and K. Kellogg Etheostoma caeruleum (Pisces: Percidae) in Bayou for designing the artificial stream, and we thank Sara, Louisiana and , p. 71-81. In: Envi- D. K. Kellogg, D. Parise, T. Stecko, Z. Whisel, and ronmental biology of darters. G. Linquist and L. B. for with its construction. We M. Page (ed.s). Dr. W.Junk Publishers,The Hague, Zachesky help The Netherlands. appreciate J. Yarnell's faithful care for the dart- GRAY, E. S. 1998. Effects of species packing and in- ers. We thank A. Balendy, B. Bealer, M. B. Gray, troduction events on resource use by darters (Te- P. Nest, and D. Sullivan for assistance in Lees, leostei: Percidae). Unpubl. Ph.D. diss., Pennsylva- the data collection and A. M. K. Balendy, Gray, nia State Univ., University Park. Kellogg, D. Parise, J. Polaski, J. Reber, B. Row- GREENBERG,L. A. 1991. Habitat use and feeding be- don, R. Ruffing, K. Walter, and M. Youndt for havior of thirteen species of benthic fish species. assistance in fish collection. All experiments Environ. Biol. Fish. 31:389-401. were conducted under and approved by the In- . 1994. Effects of predation, trout density and stitutional Care and Use Committee of discharge on habitat use by brown trout, Salmo trut- the State ta, in artificial streams. Freshwater Biol. 32:1-11. Pennsylvania University (protocol E. AND A. G. 1997. Effects number The Fish , BERGMAN, EKLOV. 91R1904C095). Pennsylvania of and interactions on hab- and Boat Commission the predation intraspecific provided necessary itat use and foraging by brown trout in artificial collecting permit 018. This study was supported streams. Ecol. Freshwater Fish 6:16-26. by the National Biological Survey RW042. HARTMAN, G. F. 1965. The role of behavior in the ecology and interaction of underyearling coho LITERATURECITED salmon (Oncoryhynchuskisutch) and steelhead trout (Salmo gairdneri). J. Fish. Res. Brd., Can. 22:1035- BROWN,L. R., ANDA. M. BRASHER.1995. Effects of 1081. predation by Sacramento squawfish (Ptychocheilus HINDAR, K., B. HONSSON, J. H. ANDREW, AND T. G. grandis) on habitat choice of California roach (La- NORTHCOTE. 1988. Resource utilization of sympat- vinia symmetricus)and rainbow trout (Oncorhnchus ric and experimentally allopatric cutthroat trout mykiss)in artificialstreams. Can. J. Fish.Aquat. Sci. and dolly varden charr. Oecologia 74:481-491. 52:1639-1646. HOCUTT,C. H., ANDP. S. HAMBRICK.1973. Hybridiza- , AND P. B. MOYLE.1991. Changes in habitat tion between the darters Percina crassa roanoka and and microhabitat partitioning within an assemblage Percina oxyrhyncha(Percidae: Etheostomatini), with of stream fishes in response to predation by Sacra- comments on the distribution of Percina crassa roan- mento squawfish (Ptychocheilus grandis). Ibid. 48: oka in New River. Am. Midl. Nat. 90:397-405. 849-856. KESSLER,R. K., A. F. CASPER,AND G. K. WEDDLE.1995. COURTENAYJR.,W. R.. 1995. The case for caution with Temporal variation in microhabitat use and spatial fish introductions, p. 413-424. In: Uses and effects relations in the benthic fish community of a stream. of cultured fishes in aquatic ecosystems. H. L. Ibid. 134:361-370. Schramm Jr. and R. G. Piper (eds.). American Fish- KNEIB, R. T. 1972. The effect of man's activity on the eries Society, Bethesda, MD. distribution of five stream fishes in Little Pine Penn. Acad. Sci. 46:49-51. I, AND P. B. MOYLE.1996. Biodiversity, fishes, Creek, Pennsylvania. J. and the introduction paradigm, p. 239-252. In: Bio- LASSUY,D. R. 1995. Introduced species as a factor in diversity in managed landscapes. R. C. Szaro and D. extinction and endangerment of native fish species, W. Johnston (eds.). Oxford Univ. Press, New York. p. 391-396. In: Uses and effects of cultured fishes in H. L. Schramm and R. G. , D. A. HENSLEY,J. N. TAYLOR,AND J. A. MC- aquatic ecosystems. Jr. CANN. 1986. Distribution of exotic fishes in North Piper (eds.). American Fisheries Society Sympo- America, p. 675-698. In: The zoogeography of sium 15, Bethesda, MD. North American freshwater fishes. C. H. Hocutt LEMLY,A. D. 1985. Suppression of native fish popu- and J. R. Stauffer Jr. (eds.). Wiley-Interscience, New lations by green sunfish in first-order streams of York. Piedmont . Trans. Am. Fish. Soc. DIAMOND,J. M. 1978. Niche shifts and the rediscov- 114:705-712. ery of interspecific competition. Am. Sci. 66:322- MILLER, R. R., J. D. WILLIAMS, AND J. E. WILLIAMS. 331. 1989. Extinctions of North American fishes during DOUGLAS,M. E., P. C. MARSH,AND W. L. MINCKLEY. the last century. Fisheries 14:22-38. GRAYAND STAUFFER-DARTER SUBSTRATE CHOICE 261

MOORE,S. E., B. RIDLEY,AND G. L. LARSON.1983. biology, and management of exotic fishes. W. R. Standing crops of brook trout concurrent with re- CourtenayJr. and J. R. StaufferJr. (eds.). Johns moval of rainbow trout from selected streams in Hopkins Univ. Press, Baltimore,MD. Great Smoky Mountains National Park. N. Am. J. TYLER,J. A., AND D. P. CLAPP.1995. Perceptual con- Fish. Manage. 3:72-80. straints on stream fish habitat selection: effects of MOYLE,P. B., H. W. LI, ANDB. A. BARTON.1986. The food availabilityand watervelocity. Ecol. Freshwater Frankenstein effect: impact of introduced fishes on Fish 4:9-16. native fishes in North America, p. 415-426. In: Fish UNDERWOOD,T. 1986. The analysisof competition by culture in fisheries management. R. H. Stroud field experiments,p. 240-268. In: Communityecol- (ed.). American FisheriesSociety, Bethesda, MD. ogy: pattern and process.J. Kikkawaand D. J. An- NAKANO,S., S. KITANO, K. NAKAI,AND K. D. FAUSCH. derson (eds.). Blackwell Scientific Publications, 1998. Competitiveinteractions for foraging micro- Melbourne, Victoria, Australia. habitat among introduced brook charr, Salvelinus VOGTJR., G. F., ANDT. G. COON. 1990. A comparison fontinalis, and native bull charr, S. confluentus, and of the foraging behavior of two darter (Etheostoma) westslope cutthroat trout, Oncorhynchusclarki lewisi, species. Copeia 1990:41-49. in a Montana stream. Environ. Biol. Fish. 52:345- WELSH, S. A., AND S. A. PERRY. 1998. Habitat parti- 355. tioning in a community of darters in the Elk River, RAESLY,R. L. 1991. Genetic and morphological con- West . Environ. Biol. Fish. 51:411-419. sequences of a founder event in an introduced pop- WERNER,E. E. 1977. Species packing and niche com- ulation of the banded darter, Etheostomazonale plementarityin three sunfishes. Am. Nat. 111:553- (Cope) (Teleostei: Percidae). Unpubl. Ph.D. diss., 577. Pennsylvania State Univ., University Park. 1984. The mechanisms of species interac- , J. R. STAUFFERJR., AND R. F. DENONCOURT. tions and community organization in fish, p. 360- 1990. Hybridization between Etheostomazonale and 382. In: Ecological communities: conceptual issues Etheostoma olmstedi (Teleostei: Percidae) following and the evidence. D. R. Strong, D. Simberloff, L. an introduction event. Copeia 1990:584-588. G. Abele, and A. B. Thistle (eds.). Princeton Univ. Ross, S. T. 1991. Mechanisms structuring stream fish Press, Princeton, NJ. assemblages: Are there lessons from introduced WHITE,M. M., ANDN. ASPINWALL. 1984. Habitat par- species? Environ. Biol. Fish. 30:359-368. titioning among five species of darters (Percidae: , J. G. KNIGHT,AND S. D. WILKINS. 1990. Lon- Etheostoma),p. 55-60. In: Environmental biology of gitudinal occurrence of the bayou darter (Percidae: darters. D. G. Lindquist and L. M. Page (eds.). Dr Etheostomarubrum) in Bayou Pierre-a response to W. Junk Publishers, Boston, MA. stream order or habitat availability? Polish Arch. Hydrobiol. 37:221-233. PENNSYLVANIASTATE UNIVERSITY, SCHOOL OF SCHLOSSER,I. J. 1987. The role of predation in age- FOREST RESOURCES, UNIVERSITYPARK, PENN- and size-related habitat use stream fishes. Ecol- by SYLVANIA 16802. PRESENT ADDRESS: (ESG) 68:651-659. ogy UNITED STATES NATIONAL PARK SERVICE,NA- STAUFFER R., M. K. A. KELLOGG,AND JR., J. J. BOLTZ, TIONAL CAPITAL REGION, 4598 MACARTHUR E. S. VAN SNIK. 1996. Microhabitat partitioning in BOULEVARD, NORTHWEST, WASHINGTON, DC a diverse assemblage of darters in the Allegheny E-mail: Riversystem. Environ.Biol. Fish. 46:37-44. 20007. (ESG) [email protected]. Send to ESG. Submitted: 20 TAYLOR,J. N., W. R. COURTENAYJR.,ANDJ. A. MCCANN. reprint requests 1984. Known impacts of exotic fishes in the conti- May 1999. Accepted: 28 May 2000. Section ed- nental United States, p. 322-373. In: Distribution, itor: W. L. Montgomery.