2008 SOUTHEASTERN NATURALIST 7(4):665–678 Density Dynamics of a Threatened Species of Darter at Spatial and Temporal Scales Mitchell S. Wine1, Michael R. Weston2, and Ronald L. Johnson2,* Abstract - Etheostoma moorei (Yellowcheek Darter), a candidate species for federal listing, is endemic and obligate to headwater riffles of the upper Little Red River drainage in north central Arkansas. Downstream segments of these tributaries were inundated in 1964 as a result of filling of Greers Ferry Reservoir. We compared riffle densities during drought (1999–2001) and non-drought periods (2003–2004), and to historic data (1979–1980). Upstream sites dried periodically during the drought of 1999–2001, and Yellowcheek Darters occupying those sites were extirpated; even at downstream sites, densities were significantly lower than historical levels. During normal precipitation levels during 2003–2004, densities increased significantly, yet several upstream sites and one complete stream remained extirpated. The loss of downstream refugia as a result of the construction of Greers Ferry Dam in 1964 may exacerbate natural climatic cycles, which include drought, resulting in stream-wide extirpations. Introduction Etheostoma moorei Raney and Suttkus (Yellowcheek Darter), with a maximum standard length of 64 mm (Robison and Buchanan 1988), is one of only two members of the subgenus Nothonotus known to occur west of the Mississippi River (Wood 1996). First collected in 1959 from Devils Fork of the Little Red River of north-central Arkansas (Raney and Suttkus 1964), much of the Yellowcheek Darter’s range was inundated in 1964 when Greers Ferry Dam was completed (Robison and Buchanan 1988). Raney and Suttkus (1964) suggested that the remaining upstream reaches of the four headwater streams (Archey, Middle, South, and Turkey/Beech/Devils forks) would serve as Yellowcheek Darter sanctuaries. Sympatric darter species common to the Little Red River drainage system include E. caeruleum Storer (Rainbow Darter), E. blennoides Rafinesque (Greenside Darter), and E. zonale Cope (Banded Darter) (Wine 2004). A 1979–1980 status survey indicated the Yellowcheek Darter to be the most abundant riffl e fish within its range (McDaniel 1984, Robison and Harp 1981). Robison and Harp (1981) estimated a total population for the four headwater streams of ca. 60,000 individuals. However, during a later study of the genetic structure of Yellowcheek Darter populations in 1998, Mitchell et al. (2002) observed individuals to be rare at several sites where they were 1US Fish and Wildlife Service, Arkansas Ecological Services Field Office, 110 South Amity Road, Suite 300, Conway, AR 72032. 2Arkansas State University, Department of Biological Sciences, State University, AR 72467. *Corresponding author - rlj@ astate.edu. 666 Southeastern Naturalist Vol. 7, No. 4 previously abundant (McDaniel 1984). For example, only one individual was captured during four hours of kick seining in Beech Fork. During sampling by Mitchell et al. (2002), there was a severe drought for several years (NCDC 2007), and several of the sites previously inhabited by Yellowcheek Darters were dry during portions of the year. Currently, the Yellowcheek Darter is a candidate species for listing under the Endangered Species Act of 1973, as amended (Federal Register 2001). Other than the original species description (Raney and Suttkus 1964), published literature regarding this species is limited to genetic analyses (Johnson et al. 2006b, Mitchell et al. 2002, Wood 1996). Reductions in abundance, the prolonged drought during the 1990s and this void of ecological literature led to the present study. The goal of this study was to compare current to historic densities of Yellowcheek Darters in the headwater streams of the Little Red River in the context of variable hydrology. We were particularly interested in recolonization of upstream sites that had recently and periodically dewatered. Methods Study areas The Middle Fork (MF) is the largest Little Red River tributary. Elevation of the streambed declines an average 2.8 m/km (0.28% gradient) along the 65 km of the Middle Fork known historically to support Yellowcheek Darters (McDaniel 1984). The Middle Fork is a third-order stream at the uppermost sample site (MFA) and a fifth-order stream at the lowest sample site (MFD). The South Fork (SF) is the second largest tributary, with a gradient of 0.17% along the 24-km study reach. It is a secondorder stream at the uppermost study site (SFA) and third-order stream at the lowest site (SFC). The Archey Fork is a third-order stream with a gradient of 0.14% along 18 km. Turkey, Beech, and Devils forks are confluent streams, with Turkey Fork representing the uppermost segment and Devils Fork the lowest. This watershed has the steepest stream gradient at 0.30% along 25 km (McDaniel 1984). The upstream site (TF) is a secondorder stream, whereas the downstream site (BF) is a third-order stream. The watersheds of these headwater streams are steep, with relatively impermeable soils that contribute to rapid changes in water levels during and following precipitation events. The 1999–2001 surveys were conducted during an extended drought with reduced stream fl ow, and the surveys during 2003–2004 were performed during a time of normal stream fl ow to determine and compare Yellowcheek Darter densities during differing fl ow regimes. Eleven riffl e sites were chosen due to their historic Yellowcheek Darter populations (Robison and Harp 1981) and availability of access. These sites were surveyed between August 1999 and November 2001, and seven of those eleven sites were surveyed 2008 M.S. Wine, M.R. Weston, and R.L. Johnson 667 from August 2003 through September 2004. The eleven initial sites included four sites on Middle Fork (MFA, B, C, and D), three sites on South Fork (SFA, B, and C), two sites on Archey Fork (AFA and B) and one site each on Turkey (TF) and Beech (BF) forks (Fig. 1). Each site was 0.8 km in length and the number of riffl es surveyed at each site ranged from two to four. These sites coincided with and were directly compared to three sites used in the 1979–1980 survey (MFD, SFA, BF; McDaniel 1984, Robison and Harp 1981). The seven sites sampled during 2003–2004 were MFA–D, SFA, SFC, and TF, with two sites (MFD, SFA) compared to historical data. During the latter study, we focused on the Middle and South Forks (although we had no access to SFB), due to their supporting the greatest numbers of individuals. Uppermost and lowest sites for each stream represent the extreme limits of Yellowcheek Darter distributions (Robison and Harp 1981, McDaniel 1984, Mitchell et al. 2002). Downstream of these riffl e sites, the substrate was primarily composed of bedrock. Additionally, streambanks upstream of MFA were characterized by high erosion. Density estimates Adult Yellowcheek Darters (>30 mm) and sympatric species were collected using a kick-seine method (surface area of 1.5 m2 with 0.48-cm mesh size). Sampling began at the downstream end of riffl es and progressed upstream, with sampling involving ≥50% of the riffl e surface area. The number of sets was recorded and mean fish density for each riffl e was determined by dividing the number of fish collected by the surface area sampled. Each study site was sampled and densities were estimated several times during 1999–2001 and again in 2003–2004 (see Table 1 for sampling frequency). Figure 1. Etheostoma moorei (Yellowcheek Darter) study sites on tributaries of the Little Red River in Arkansas. 668 Southeastern Naturalist Vol. 7, No. 4 Lastly, abundance of this darter relative to other sympatric species was compared for 1999–2001 versus 1979–1980, to determine if abundance differences were a species versus an assemblage phenomenon. Kick seining proved to be the most effective non-lethal method of collecting darters. Although snorkeling has commonly been used in the study of darters (e.g., Chipps et al. 1994, Greenberg 1991, Stauffer et al. 1996), snorkeling proved ineffective at our sites, due partly to the crevice-dwelling behavior of the Yellowcheek Darter and partly to stream fl ow conditions usually being prohibitively low or high for snorkeling. Sampling efficiency of darters in headwater streams has been estimated at between 50–60% (Peterson and Rabeni 2001), which is impacted by the variables of substrate and flow (Fisher 1987, Peterson and Rabeni 2001). Electrofishing (Backpack Model 15 D Smith-Root, Inc.) was therefore conducted to validate the accuracy of the kick seine method and develop conversion factors for estimating densities. For example, one riffle each within MFC and SFC was block-netted at upstream and downstream margins and measured. Fifty percent of the riffle area was sampled using the kick-seine method, and Yellowcheek Darter density was estimated. The entire riffle was then electrofished using the three-pass sum-of-catches depletion method (Riley and Fausch 1992) to determine a second density estimate. During shocking passes, the substrate was agitated to dislodge shocked specimens, which were then collected and put in five-gallon buckets for later identification and enumeration. Dividing the electrofishing estimate by the kickseine estimate produced the respective conversion factor. In each case, kick-seine estimates were multiplied by their respective conversion factors (1.35 for the Middle and Beech forks, 2.88 for the South and Archey forks) to produce a density estimate for each sample site. Table 1. Comparisons of estimates of mean Yellowcheek Darter densities (number/m2), with 95% confidence intervals, for each study site of the Little Red River, AR. 1979–1980 1999–2001 2003–2004 Site Site samples Density Site samples Density Site samples Density Middle Fork A 0 N/A 9 0.05 ± 0.12 5 0.00 ± 0.00 B 0 N/A 1 0.77 4 0.24 ± 0.28 C 0 N/A 7 1.11 ± 0.79 6 1.67 ± 0.65 D 9 2.38 ± 1.13 3 0.66 ± 0.13 5 1.67 ± 0.77 South Fork A 5 1.16 ± 0.62 7 0.22 ± 0.26 6 2.09 ± 1.92 B 0 N/A 3 0.00 ± 0.00 0 N/A C 0 N/A 7 2.06 ± 0.87 7 1.29 ± 0.34 Archey Fork A 0 N/A 2 0.00 ± 0.00 0 N/A B 0 N/A 7 0.06 ± 0.10 0 N/A Turkey Fork 0 N/A 4 0.00 ± 0.00 2 0.00 ± 0.00 Beech Fork 4 0.59 ± 0.66 5 0.47 ± 0.90 0 N/A 2008 M.S. Wine, M.R. Weston, and R.L. Johnson 669 Methods used to derive conversion factors were similar to those used by Robison and Harp (1981), yet the conversion factor for their work (2.71) was derived from the Middle Fork, and then applied to each of the other streams. Although their study was also conducted during a drought, water levels were greater (10% greater mean monthly stage height; USGS 2006), and therefore kick-seining efficiencies were likely lower during their study than at present (G. Harp, Arkansas State University, State University, AR, pers. comm.). This differing application of a correction factor between studies introduces some uncertainty in interstudy analyses. Consistency in technique within and among researches and studies was attempted by having G. Harp, the initial investigator of Yellowcheek Darter distributions in the Little Red River train both M.S. Wine and M.R. Weston on sampling technique. Wine and Weston subsequently performed all sampling for this study, and seined only during low fl ow. Seine and mesh size were also kept constant among sampling periods. Due to the many variables among studies, no attempts were made to compare our density data to those for other darter species. Proximal downstream pool areas (n = 11) were also sampled during zerodischarge riffl e conditions (no surface fl ow) to record pool fish assemblages and to determine if Yellowcheek Darters used pools as refugia during periods when riffl es were dry. Pools were isolated using block nets at upstream and downstream margins and backpack electrofished twice. No Yellowcheek Darters were collected in these downstream pools, nor have they been collected from pools historically despite extensive efforts (Robison and Harp 1981). Some riffl e species can use the hyporheic zone during riffl e drying (Berra and Allen 1989, Gagen et al. 1998, Stegman and Minckley 1959). However, Weston (2006) failed to find any individuals occupying pits dug into the thalweg of MFB during riffl e drying. On the basis of these evidences, it was determined that Yellowcheek Darters were leaving these riffl es during drying (or dying as those riffl es dried), and therefore densities were recorded as zero when sites were dry. We used ANOVA to compare Yellowcheek Darter densities among sites and streams. Downstream sites were compared among streams rather than using all sites, as upstream sites had few individuals and low fl ow. ANOVAs demonstrating significance were followed with Tukey’s multiple comparison a posteriori test to investigate level interactions. Temporal differences within sites among sampling periods (1979–1980, 1999–2001, and 2003–2004) were also compared using ANOVA. All confidence intervals given are 95% group confidence intervals for the difference between the two density means between sites and streams. Alpha levels for all significance tests were set at α = 0.05. Hydrologic analysis The United States Geological Survey (USGS) maintained a gauging station for surface water data at MFD (Gage 07075000 continuously from 1974 670 Southeastern Naturalist Vol. 7, No. 4 to 2002) and SFC (Gage 07075300 continuous through all studies). Stage height was compared to historical data (USGS 2006), as water levels were noticeably lower during the present study as compared to the previous study (G. Harp, Arkansas State University, State University, AR, pers. comm.). During the 1999–2001 sampling period, all upstream sites were periodically dry. Discharge, thalweg velocity, and thalweg depth were measured during 1999–2001 to provide insight into site and stream differences. Velocity was measured by an electronic fl ow probe (Model FP 101, Global Water Instruments, Inc.). Substrate composition and cobble embededness were visually evaluated for each site on the Middle and South forks during summer (August 2004) low-fl ow conditions. Data were collected along five longitudinal transects (upper 20%, 21–40%, 41–60%, etc.) and two to five cross-channel transects (left, ¼, ½, ¾, and right), depending upon stream width. Along each transect, we used a view box to estimate percent silt (<1.0 mm), sand (1.0 mm–0.5 cm), gravel (0.5–8.0 cm), cobble (8.0–30.0 cm), boulder (>30.0 cm), and bedrock. As both species prefer gravel and cobble substrate (Mc- Daniel 1984, Winn 1958), we focused our analyses on these two variables. Cobble embededness was estimated as the percent silt- and/or sand-covered or embedded cobble at these locations. Results Density estimates Yellowcheek Darter densities were highly variable spatially and temporally (Table 1). Upstream sites were subject to drying early in the study. No surface water fl ow was observed in riffl es at: MFA, SFA, TF, AFA or AFB during August 1999; MFA, SFA, or SFB in September 1999; SFB in November 1999; or AFB in September 2000. Densities were lower at MFD during 1999–2001 as compared to both 1979–1980 (80% decline, but no statistical significance, P = 0.08) and 2003–2004 (86% decline, P = 0.04) (Table 2). Densities were also lower in 1999–2000 than in 1979–1980 at SFA (P < 0.05). During 1999–2004, no Yellowcheek Darters were collected at TF, SFB, or AFA, and only intermittently collected at MFA and AFB. Yellowcheek Darter densities were similar among downstream stream sites for all time periods studied (P range of 0.29 to 0.38; Table 1). ANOVA did reveal significant differences in densities among sites within the Middle Fork for both the 1999–2001 (F2,16 = 7.766, P = 0.004) and 2003–2004 periods (F3,16 = 18.073, P < 0.001) (Table 2). A Tukey multiple comparison a posteriori test showed that the MFA Yellowcheek Darter density (where one individual was collected in five sampling years) was less than MFC for 1999–2001 and less than MFC and MFD for 2003–2004 (P < 0.05). MFB Yellowcheek Darter density was also less than densities at MFC and MFD during 2003–2004 (P < 0.05). There were no significant differences in 2008 M.S. Wine, M.R. Weston, and R.L. Johnson 671 Yellowcheek Darter densities among sites on the South Fork (1999–2001: P = 0.06; 2003–2004: P = 0.28). Yellowcheek Darters ranked fifth in abundance of fish collected during 1999–2001, comprising 7% of the overall sample. This represents a substantial drop (20%) in the relative abundance of the Yellowcheek Darter, as during 1979–1980 it was the most abundant species, comprising 27% of that sample (Robison and Harp 1981). Hydrologic analysis Monthly stage height for both MFD and SFC during 1999–2001 sampling was lower during late summer months than during the 1979–1980 study and also lower than the 25-year averages (Fig. 2). Upper sites in all four tributaries, as well as the lower site in the Turkey/Beech fork system (BF), were observed to be completely dry on at least two occasions during the late summer months of 2000. However, during 2001–2004, the upper riffl e sites of Middle and South forks never completely dried, although one site (MFB) was low enough during September 2004 that no fl ow was detectable. Water depth at that time was less than 2.5 cm across one of two MFB riffl es studied. We identified a stage height of 1.53 m at the USGS gage station at MFD and 1.02 m at the USGS gage station at SFC as the minimum currently required to maintain water fl ow at the upper study sites within this system. Stage heights below these minima result in upper riffl e sites experiencing zero discharge. Mean monthly stage height was never below this critical threshold in 2001–2004 (Fig. 2). Flow on the South Fork was less than on the Middle Fork (Fig. 2). Discharge increased downstream on the Middle and Turkey/Beech forks, yet Table 2. Comparisons of densities of Yellowcheek Darters of the Little Red River drainage system spatially and temporally where significance exists among levels. Where a significant difference exists between means, the 95% confidence intervals for the difference between variables is given (column, row). NSD = no significant difference exists for individual comparisons. A. Spatial comparisons among Middle Fork sites; 1999–2000 comparisons on first line and 2003–2004 comparisons on second line. MFA MFB MFC MFB NSD - - NSD - - MFC (-1.761, -0.362) NSD - (-2.486, -0.844) (-2.303, -0.552) - MFD NSD NSD NSD (-2.522, -0.906) (-2.336, -0.517) NSD B. Temporal comparisons within Middle Fork and South Fork sites MFD (2003–4) SFA (1979–81) MFD (1999–2001) (0.071, 1.944) - SFA (1999–2001) - (0.071, 1.944) 672 Southeastern Naturalist Vol. 7, No. 4 was largely constant within the South and Archey forks (Table 3). Stream riffl e velocity ranged from 0.59 to 1.01 m/s, with the MFB data serving as an outlier (MFB sampling was limited to three late summer samples only). Depth generally increased downstream for each stream. Substrate was mostly gravel (50%) and cobble (34%) (Table 3), with minimal boulder (4.0%), silt (<1.0%), and bedrock (<1.0%) within sites. There were significant differences among sites for both gravel and cobble substrate (gravel: F3,164 = 35.936, P < 0.001; cobble: F3,164 = 37.609, P < 0.001). Gravel represented a significantly smaller percent of the substrate composition at MFB than at MFA, MFC, and MFD (Tukey’s: P < 0.01), and MFA had less gravel than MFD (Tukey’s: P < 0.05). Predictably, based upon the predominance of gravel and cobble within sites and therefore an inverse relationship between these two variables, cobble represented a significantly greater percent of the substrate composition at MFB than at MFA, MFC, and Figure 2. Monthly averages for 1 9 7 9 – 1 9 8 0 , 1 9 9 9 – 2 0 0 0 , and 2001–2002 stage heights of MFD (above) and SFC (below) of the Little Red River, AR (Latitude 3 5 ° 3 9 ' 2 5 " N , 9 2 ° 1 7 ' 3 4 " N ; USGS 2006). The horizontal dotted line (at 1.53 m for MFD, and 1.02 m for SFC) represents the stage height at which drying of uppermost sampling sites occurs. 2008 M.S. Wine, M.R. Weston, and R.L. Johnson 673 MFD (Tukey’s: P < 0.01), and MFA and MFC had more cobble than MFD (Tukey’s: P < 0.01). Cobble embeddedness was low at all sites (Table 2). However, there were significant differences among sites (F3,163 = 10.772, P < 0.001), with more embeddedness upstream at MFA and MFB than at MFC and MFD (Tukey’s: P < 0.05). Discussion Fishes of headwater streams, including Yellowcheek Darters, have evolved in systems subject to periodic drying. Recolonization of previously dried and extirpated sites must therefore be a common feature of their life histories. Within the headwaters of the Little Red River, low water conditions during the late 1990s resulted in habitat loss in a system that had previously lost extensive portions of downstream habitat due to reservoir impoundment. When upstream sites of the Little Red River drainage system Table 3. Arithmetic means and standard error of the mean (second line) of physical variables measured for riffl e sites for the Little Red River, AR. All measurements were taken during low fl ow, Spring 2001 to Fall 2003. Discharge, velocity, and depth are therefore not representative of high-fl ow conditions. Thalweg Discharge Velocity Depth Substrate % Site (m3/s) (m/s) (m) Sand Gravel Cobble Embededness Middle Fork A 0.33 0.69 0.14 10.60 58.80 31.30 13.20 (0.05) (0.05) (0.01) (1.51) (4.67) (5.62) (2.91) B 0.11 0.18 0.06 5.30 25.90 60.90 11.20 (0.11) (0.27) (0.04) (1.45) (3.08) (4.08) (1.48) C 0.73 0.80 0.18 6.10 58.40 31.80 2.40 (0.19) (0.04) (0.01) (1.31) (3.89) (3.82) (0.66) D 0.66 0.88 0.21 19.70 57.20 18.50 7.40 (0.12) (0.04) (0.02) (1.89) (3.40) (2.61) (1.03) South Fork A 0.44 0.82 0.13 5.40 37.0 54.0 6.30 (0.07) (0.07) (0.01) (1.29) (5.66) (5.74) (1.48) B 0.37 0.63 0.16 N/A N/A N/A N/A (N/A) (N/A) (N/A) C 0.40 0.99 0.22 11.80 56.40 31.80 12.80 (0.05) (0.09) (0.02) (2.45) (4.01) (4.50) (2.17) Archey Fork A 0.31 0.64 0.15 N/A N/A N/A N/A N/A N/A N/A B 0.47 0.77 0.17 N/A N/A N/A N/A (0.09) (0.15) (0.01) Turkey Fork 0.45 0.78 0.15 N/A N/A N/A N/A (0.07) (<0.01) (<0.01) Beech Fork 1.01 0.59 0.14 N/A N/A N/A N/A (0.13) (0.04) (<0.01) 674 Southeastern Naturalist Vol. 7, No. 4 dry, there are few downstream riffl es to serve as refugia for maintaining populations. For example, Wine (2004) calculated a reduction of available Yellowcheek Darter habitat due to drought in 1999–2000 ranging from 30% for the Middle Fork up to 70% for the South Fork. As riffl es dewater, occupants of the riffl e must move into neighboring pools or, as appears to occur for Yellowcheek Darters, move large distances downstream to a persistent riffl e or perish. Conversely, the sympatric Rainbow Darter and Campostoma anomalum Rafinesque (Central Stoneroller) were frequently found in pools following dewatering of upstream riffl es, and occurred in greater frequencies in riffl es than historically (Robison and Harp 1981, Wine 2004). This apparent obligate occupation of riffl e habitats by Yellowcheek Darters provides insight for the relatively slow recolonization of upper sites compared to other riffl e fishes. Numerous studies have examined downstream faunal changes following stream impoundment (e.g., Johnson and Harp 2004, Johnson et al. 2006a, Shaver et al. 1997). Less frequently studied are upstream faunal effects from impoundment (Erman 1973, Herbert and Gelwick 2003, Winston et al. 1991). Stream impoundment has been associated with the decline of other darter species including the threatened E. nianguae Gilbert and Meek (Niangua Darter; Mattingly and Galat 2002) and endangered E. doration Jordan and Brayton (Bluemask Darter; Layman et al. 1993). Several monthly stage-height averages during 1999–2000 (September and October of both years; USGS 2006) were below that required to maintain surface fl ow in the upper stream sites. As these are monthly averages, they refl ect prolonged periods of riffl e drying. Daily stage height fl uctuations below these monthly averages would represent more extensive dewatering. Water levels during 1999 and 2000 represent 25-year lows (USGS 2006). Fish often recolonize stream reaches rapidly after disturbance events such as drought, provided there are no barriers to their movement (Peterson and Bayley 1993). Distances between suitable habitats can affect the success of dispersal and recruitment (Lonzarich et al. 1998). For example, drying of riffl es during 1999–2000 (and previously during 1998 [Mitchell 1999]) occurred for several km downstream of our upstream study sites, yet intermittent pools remained. This extent represents long distances for upstream recolonization to occur. Data are inconsistent as to the potential for darter migration patterns among species. Some darter species exhibit high site fidelity (e.g., E. radiosum Hubbs and Black [Orangebelly Darter; Scalet 1973], E. cragini Gilbert [Arkansas Darter; Labbe and Fausch 2000]), whereas other species are highly mobile (14 species studied by Winn [1958]). During normal fl ow conditions, Yellowcheek Darters demonstrate high site fidelity (Weston and Johnson 2008). However, migration patterns as these streams dewater remain to be studied. A key niche difference identified among darter species was that the species studied by Winn (1958) were eurytypic and had widespread geographic distribution, whereas darter species in Scalet (1973) 2008 M.S. Wine, M.R. Weston, and R.L. Johnson 675 and this study were obligate riffl e species and endemic to small watersheds. Site fidelity may contribute to limited distribution of some darter species, whereas the ability or propensity to travel may contribute to the widespread geographic distributions of other darter species. The decline of Yellowcheek Darters during 1999–2001 was pervasive throughout the Little Red River drainage system. Comparisons to the Robison and Harp (1981) survey and more recent sampling in 2003–2004 show consistently low densities for comparable sites, other than for site SFC. With the rewatering of riffles since 2000, subpopulations within the upper South Fork have apparently re-established. In contrast, upper sites in the Middle Fork have not effectively recolonized, particularly site MFA. The Middle Fork has a greater stream gradient than the South Fork; therefore, flow and biological responses to that flow should be more variable in the Middle Fork. Archey Fork sites were almost extirpated, with only 23 individuals observed sporadically in the lower site during seven site visits; more recent visits indicate that this site has since been repopulated (M.S. Wine, pers. observ.). Unlike the Turkey Fork, which drains directly to Greers Ferry Reservoir, Archey Fork merges with the South Fork prior to entering the reservoir. Downstream South Fork populations may have assisted in the repopulating of Archey Fork sites. Conversely, the extirpation of the Turkey Fork population has continued through the present (M.S. Wine, pers. observ.). McDaniel (1984) noted a high amount of large cobble and boulders (80%) in the Turkey Fork and recognized that much of the stream at that time represented suboptimal Yellowcheek Darter habitat. Similar results have been identified for other riffle species (e.g., Noturus nocturnus Jordan and Gilbert [Dusky Darter]; Herbert and Gelwick 2003) occupying previously connected streams upstream of reservoirs. Turkey Fork has the smallest watershed of the study streams, and prolonged and frequent drying of this stream could explain localized species extinctions. This population represents a critical loss of genetic diversity for Yellowcheek Darters (Mitchell et al. 2002). Yellowcheek Darter densities of colonized sites presently range from 0.18–1.23 adults/m2, and are comparable to historic levels (Robison and Harp 1981). Trends in the data indicate that Yellowcheek Darter densities are probably closely linked to water levels. Increases in both fl ow and densities occurred with the more recent sampling period. Flow ceased at upstream sites when water levels fell below 1.53 m and 1.02 m at MFD and SFC, respectively. The loss of downstream refugia as a result of the construction of Greers Ferry Dam in 1964 may exacerbate natural climatic cycles, which include drought, resulting in stream-wide extirpations. Acknowledgments Financial support was provided by the US Fish and Wildlife Service. We thank G. Harp, S. Blumenshine, R. Nilius, S. Rogers, S. Shoults, and B. Wagner for technical 676 Southeastern Naturalist Vol. 7, No. 4 support, and R. Mitchell, T. Harmon, S. Fowler, L. Hodgens, J. Gore, N. Bickford, J. Fullington, K. Gillespie, B. Intres, M. Johnson, A. Peck, T. Sanders, M. Trevino, C. Dawes, and C. Weston for assistance with collection. We also thank G. Harp, C. Davidson, R. Mitchell, and additional, anonymous reviewers for improving the quality of this manuscript. Literature Cited Berra, T.M., and G.R. Allen. 1989. Burrowing, emergence, behavior, and functional morphology of the Australian Salamanderfish, Lepidogalaxias salamandroides. Fisheries 14:2–10. Chipps, S.R., W.B. Perry, and S.A. Perry. 1994. Patterns of microhabitat use among four species of darters in three Appalachian streams. American Midland Naturalist 13:175–180. Erman, D.C. 1973. Upstream changes in fish populations following impoundment of Sagehen Creek, California. Transactions of the American Fisheries Society 102:626–628. Federal Register. 2001. The Federal Register main page. Volume 66 p.54811. Available online at http://www.gpoaccess.gov/fr/index.html. Accessed August 7, 2006. Fisher, W.L. 1987. Benthic fish sampler for use in riffl e habitats. Transactions of the American Fisheries Society 116:768–772. Gagen, C.J., R.W. Standage, and J.N. Stoeckel. 1998. Ouachita Madtom (Noturus lachneri) metapopulation dynamics in intermittent Ouachita mountain streams. Copeia 4:874–882. Greenberg, L.A. 1991. Habitat use and feeding behavior of thirteen species of benthic stream fishes. Environmental Biology of Fishes 31:389–401. Herbert, M.E., and F.P. Gelwick. 2003. Spatial variation of headwater fish assemblages explained by hydrologic variability and upstream effects of impoundment. Copeia 2003:273–284. Johnson, R.L., S.C. Blumenshine, and S.M. Coghlan. 2006a. Bioenergetic and foodweb analysis of factors limiting Brown Trout growth in the Little Red River, AR tailwater. Environmental Biology of Fishes 77:121–132. Johnson, R.L., R.M. Mitchell, and G.L. Harp. 2006b. Genetic variation and genetic structuring of a numerically declining species of darter, Etheostoma moorei Raney and Suttkus, endemic to the upper Little Red River, Arkansas. American Midland Naturalist 156:37–44. Johnson, R.L., and G.L. Harp. 2004. Spatio-temporal changes of benthic macroinvertebrates in a cold Arkansas tailwater. Hydrobiologia 537:15–24. Labbe, T.R., and K.D. Fausch. 2000. Dynamics of intermittent stream habitat regulate persistence of a threatened fish at multiple scales. Ecological Applications 10:1714–1791. Layman, S.R., A.M. Simons, and R.M. Wood. 1993. Status of the Dirty Darter, Etheostoma olivaceum, and Bluemask Darter, Etheostoma (Doration) sp., with notes on fishes of the Caney Fork River system, Tennessee. Journal of the Tennessee Academy of Sciences 68:65–70. Lonzarich, D.G., M.R. Warren, Jr., and M.R.E. Lonzarich. 1998. Effects of habitat isolation on the recovery of fish assemblages in experimentally defaunated stream pools in Arkansas. Canadian Journal of Fisheries and Aquatic Sciences 55:2141–2149. 2008 M.S. Wine, M.R. Weston, and R.L. Johnson 677 Mattingly, H.T., and D.L. Galat. 2002. Distributional patterns of the threatened Niangua Darter, Etheostoma nianguae, at three spatial scales, with implications for species conservation. Copeia 2002:573–585. McDaniel, R.E. 1984. Selected aspects in the life history of Etheostoma moorei Raney and Suttkus. M.Sc. Thesis. Arkansas State University, State University, AR. 124 pp. Mitchell, R.M. 1999. Genetic and meristic variations between and within populations of Etheostoma moorei (Yellowcheek Darter) Raney and Suttkus (Perciformes: Percidae). M.Sc. Thesis. Arkansas State University, State University, AR. 77 pp. Mitchell, R., R.L. Johnson, and G.L. Harp. 2002. Population structure of an endemic species of Yellowcheek Darter, Etheostoma moorei (Raney and Suttkus), of the upper Little Red River, Arkansas. American Midland Naturalist 148(1):129– 137. National Climatic Data Center (NCDC). 2007. Climate of 2007 - March: US drought watch. Available online at http://www.ncdc.noaa.gov/oa/ climate/research/2007/ mar/us-drought.html. Accessed February 10, 2007. Peterson, J.T., and P.B. Bayley. 1993. Colonization rates of fishes in experimentally defaunated warmwater streams. Transactions of the American Fisheries Society 122:199–207. Peterson, J.T., and C.F. Rabeni. 2001. Evaluating the efficiency of a one-squaremeter quadrat sampler for riffl e-dwelling fish. North American Journal of Fisheries Management 21:76–85. Raney, E.C., and R. Suttkus. 1964. Etheostoma moorei, a new darter of the subgenus Nothonotus from the White River system. Copeia 1:130–138. Riley, S.C., and K.D. Fausch. 1992. Underestimation of trout population size by maximum likelihood removal estimates in small streams. North American Journal of Fisheries Management 12:768–776. Robison, H.W., and T.M. Buchanan. 1988. The Fishes of Arkansas. The University of Arkansas Press, Fayetteville, AR. 536 pp. Robison, H.W., and G.L. Harp. 1981. A study of four endemic Arkansas threatened fishes. Project E-1-3. Office of Endangered Species, Washington, DC. Unpublished. Report. Scalet, C.G. 1973. Stream movements and population density of the Orangebelly Darter, Etheostoma radiosum cyanorum (Osteichthyes: Percidae). Southwestern Naturalist 17:381–87. Shaver, M.L., J.P. Shannon, K.P. Wilson, P.L. Benenati, and D.W. Blinn. 1997. Effects of suspended sediment and desiccation on the benthic tailwater community in the Colorado River, USA. Hydrobiologia 357:63–72. Stauffer, J.R., Jr., J.M. Boltz, K.A. Kellogg, and E.S. Van Snik. 1996. Microhabitat portioning in a diverse assemblage of darters in the Allegheny River system. Environmental Biology of Fishes 46:37–44. Stegman, J.L., and W.L. Minckley. 1959. Occurrence of three species of fishes in interstices of gravel in an area of subsurface fl ow. Copeia 1959:341. US Geological Service (USGS). 2006. USGS surface-water monthly statistics for Arkansas: Gages 07075000 and 07075300. Available online at http://waterdata. usgs.gov/ar/nwis/monthly. Accessed June 15, 2006. 678 Southeastern Naturalist Vol. 7, No. 4 Weston, M.R. 2006. Density, population size estimates, and movement patterns of Etheostoma moorei Raney and Suttkus. M.Sc. Thesis. Arkansas State University, State University, AR. 101 pp. Weston, M.R., and R.L. Johnson. 2008. Visible implant elastomer as a tool for marking Etheostomine darters (Actinopterygii: Percidae). Southeastern Naturalist 7(1):159–164. Wine, M.S. 2004. Current status of Etheostoma moorei Raney and Suttkus with emphasis on population sizes, distributions and reproductive habitat requirements. M.Sc. Thesis. Arkansas State University, State University, AR. 69 pp. Winn, H.E. 1958. Comparative reproductive behavior and ecology of fourteen species of darters. Ecological Monographs 28:155–191. Winston, M.R., C.M. Taylor, and J. Pigg. 1991. Upstream extirpation of four minnow species due to a damming of a prairie stream. Transactions of the American Fisheries Society 120:98–105. Wood, R.M. 1996. Phylogenetic systematics of the darter subgenus Nothonotus (Teleostei: Percidae). Copeia 2:300–318.