2007 NORTHEASTERN NATURALIST 14(4):519–530 Clinal Variation in Ohio River Basin Populations of the Redfi n Shiner (Lythrurus umbratilis) David J. Eisenhour1,* and Lynn V. Eisenhour2 Abstract - Prior to this study, undocumented morphological variation in Lythrurus umbratilis (Redfi n Shiner) has impaired identifi cation of many samples of eastern populations of the species. Meristics, morphometrics, tuberculation, pigmentation, and nuptial male coloration of over 700 specimens of L. umbratilis were examined in order to assess patterns of geographic variation and species limits in the Ohio River basin. Principle component and spatial autocorrelation analyses of these data demonstrate that morphological variation in L. umbratilis is clinal along most of the length of the Ohio River basin. Specimens from eastern populations have less black in the dorsal fi n of breeding males, have lower mean scale counts, and are more slender than western populations. The cline does not extend into central and northern Ohio, as populations have relatively high meristic counts and more robust males. These analyses suggest that recognition of an additional species in the Ohio River basin is not warranted at this time. The cline may refl ect the infl uence of drainage evolution of the region, or even past gene fl ow with Lythrurus fasciolaris (Scarlet Shiner), a closely related species with a distribution parapatric to that of L. umbratilis. Introduction Lythrurus umbratilis (Girard) (Redfin Shiner) is a small cyprinid that occurs throughout most of the Mississippi River basin, the southern Great Lakes drainages, and some drainages along the West Gulf Slope. It is typically common to abundant in its preferred habitat of pools and raceways of small, warm, moderate-gradient streams. In the Ohio River basin, L. umbratilis has a nearly parapatric distribution with Lythrurus fasciolaris (Gilbert) (Scarlet Shiner) (Burr and Warren 1986, Etnier and Starnes 1993, Stauffer et al. 1995, Trautman 1981). These species usually are ecologically separated, with L. fasciolaris typically occupying streams with clearer water, higher gradients, and coarser substrates than L. umbratilis. At their contact zone, limited to extensive hybridization may occur, and some evidence suggests L. umbratilis may be replacing L. fasciolaris, perhaps as a result of environmental degradation (Burr and Warren 1986, Hopkins 2005, Trautman 1981). Lythrurus umbratilis, L. lirus (Jordan) (Mountain Shiner), and the L. ardens complex, which includes L. ardens (Cope) (Rosefin Shiner), L. fasciolaris, and L. matutinus (Cope) (Pinewoods Shiner) (Dimmick et al. 1996), appear to be closely related, probably forming a monophyletic group (Pramuk et al. 2006, Schmidt et al. 1998). Phylogenetic analysis of 1Department of Biological and Environmental Sciences, Morehead State University, Morehead, KY 40351. 2Rowan County School System, Morehead, KY 40351. *Corresponding author - d.eisenhour@morehead-st.edu. 520 Northeastern Naturalist Vol. 14, No. 4 mtDNA data suggests L. umbratilis is sister to a clade containing L. lirus and L. fasciolaris (Pramuk et al. 2006). Snelson and Pfl ieger (1975) examined systematics of populations west of the Mississippi River and recognized two subspecies, the darkly pigmented Lythrurus umbratilis umbratilis in western Missouri and L. u. cyanocephalus in eastern Missouri. Although they did not examine specimens from the Ohio River basin, L. u. cyanocephalus has been assumed to be the form present in the Ohio River basin (Burr and Warren 1986, Etnier and Starnes 1993). No thorough systematic study of Ohio River basin populations has been published, although published descriptions reveal apparent geographic variation in morphology (Etnier and Starnes 1993, Robison and Buchanan 1988, Stauffer et al. 1995), and L. umbratilis is suspected to be polytypic (W.C. Starnes, North Carolina State Museum of Natural Sciences, Raleigh, NC, pers. comm.). Our preliminary surveys suggested that considerable variation in morphology exists in the Ohio River basin and a thorough systematic study was warranted. Some eastern populations of L. umbratilis are so aberrant that they could be misidentifi ed as L. fasciolaris using characters provided in various guides (Etnier and Starnes 1993, Page and Burr 1991, Stauffer et al. 1995). The objective of this study is to assess geographic patterns of morphological variation in Ohio River basin populations of L. umbratilis and to determine species limits in that region. We also discuss causes of clinal variation of L. umbratilis and other fi shes with a similar clinal pattern in the Ohio River basin. Methods Institutional abbreviations are from Poss and Collette (1995), except MOSU is used as the acronym for the Morehead State University Ichthyology Collection. Morphological data were collected following the methods of Hubbs and Lagler (1964), unless defi ned otherwise, and were analyzed using programs available in SAS 9.1 as modifi ed by David L. Swofford. Eleven counts (lateral-line scales, predorsal scale rows, scales above the lateral line, scales below the lateral line, caudal peduncle scales, circumferential scales, anal fi n rays, pelvic fi n rays, and pectoral fi n rays) were taken from 544 specimens of L. umbratilis collected from 36 localities (Appendix 1) in the Ohio River basin and adjacent watersheds. Sampling from geographic areas with suspected hybridization with L. fasciolaris, such as the Salt River drainage (Hopkins 2005), was avoided (Fig. 1). Vertebrae were counted from 110 specimens cleared and stained using a protocol from Taylor and Van Dyke (1985). Twenty-nine measurements (Eisenhour and Eisenhour 2004) were taken from 123 male specimens and 94 female specimens selected from the pool of specimens used in the meristic analysis. All measurements were made with digital calipers to the nearest 0.1 mm; those under 5 mm were made with the aid of a dissecting microscope. Truss-geometric protocol (Bookstein et al. 1985, Humphries et al. 1981, Strauss and Bookstein 1982) was used, in part, to archive body form. In 2007 D.J. Eisenhour and L.V. Eisenhour 521 addition, tuberculation, pigmentation, and life coloration were examined from numerous specimens from the study area. Multivariate analysis of meristic characters was accomplished using principal component analysis (PCA). Principal components were factored from a correlation matrix of 11 non-transformed meristic variables. Multivariate analyses of morphometric variables were accomplished using sheared PCA (Bookstein et al. 1985, Humphries et al. 1981) to eliminate overall size effects. Principal components were factored from the covariance matrix of 29 log-transformed morphometric characters following recommendations of Bookstein et al. (1985). Spatial autocorrelation analysis tests whether values of a variable are spatially independent (Legendre and Fortin 1989, Sokal and Oden 1978) and has been useful in detecting clines in fishes such as Fundulus zebrinus Jordan and Gilbert (Plains Killifish; Poss and Miller 1983), Aphredoderus sayanus (Gilliams) (Pirate Perch; Boltz and Stauffer 1993), and Macrhybopsis aestivalis (Girard) (Speckled Chub; Eisenhour 2004). Spatial autocorrelation analysis was performed on 544 specimens of L. umbratilis from 33 localities using the R Package, version 3.02 (Legendre and Vaudor 1991). For each locality, latitude and longitude were determined to the nearest second and population means were used as variables. All meristic variables and their PC 1 scores were subjected to spatial autocorrelation analysis. The adjacency matrix used in spatial autocorrelation analysis Figure 1. Localities from which specimens of L. umbratilis were examined. Photos show differences in body shape of nuptial males: Left, 63 mm SL, MOSU 2165, Ferguson Creek, Livingston County, KY, 21 June 2004; Right, 62 mm SL, MOSU 2160, Slate Creek, Montgomery County, KY, 10 June 2004. 522 Northeastern Naturalist Vol. 14, No. 4 characterizes a fully connected graph in which each node (locality) is connected to all others. Edge lengths (distance between localities) were weighted by their geographic distance based on arc lengths. Comparisons were made at 80-km intervals. This grouping provided at least 23 pairs in all but the two largest distance classes, which had 9 pairs each. We originally attempted to estimate melanophore coverage by analyzing digital photographs of dorsal fins with Adobe Photoshop 6.0, similar to the method used by Cox et al. (2005). However, results from this failed to reflect readily apparent variation, perhaps because melanophores exhibited a range of intensity, from pale brown to black. Instead, percent melanophore coverage of dorsal fins of 69 nuptial males from 29 localities was visually estimated by each of the authors and averaged. Correlation between each estimate was high (p < 0.001, r2 = 0.829). Regression analysis was used to examine the relationship of melanophore coverage to standard length (SL). Population means, from raw percent melanophore coverage and from regression residuals, were subjected to spatial autocorrelation analysis to test for the presence of clinal variation. Methods follow those outlined above, except that comparisons were made at 166-km intervals, providing at least 14 pairs in all but the two largest distance classes, which had only 2 or 4 pairs. Results Western populations of L. umbratilis had higher mean scale counts than eastern populations. Frequencies of two counts with the most pronounced geographic variation, lateral-line scales, and circumferential scales, are presented in Tables 1 and 2; other scale counts exhibited a similar pattern. Counts of vertebrae, cephalic pores, and fi n rays exhibited local variation (e.g., several populations have modally 10 anal-fi n rays), but no broad or concordant geographic patterns were noted. Principle component analysis of the meristic data separated mean population scores along an irregular east-west cline (Fig. 2), Figure 2. Mean meristic scores on PC axes 1 and 2 for ten drainage units of 544 L. umbratilis. Percentages indicate variation explained by each axis. 2007 D.J. Eisenhour and L.V. Eisenhour 523 Table 1. Frequency distribution of lateral-line scales for selected drainage units of L. umbratilis. Drainage unit 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 n Mean Lake Erie 1 6 5 3 5 14 3 2 2 2 1 1 45 42.5 Muskingum 2 - 5 1 2 5 4 4 2 25 44.6 West Virginia 2 2 4 8 12 12 7 4 2 4 2 1 2 62 42.1 Little Sandy 2 6 3 4 4 4 4 2 3 32 42.8 Licking 1 6 6 6 7 8 11 8 4 5 5 2 1 2 1 2 75 44.1 Green 1 - 2 5 3 9 8 8 9 8 1 54 44.2 Tradewater 1 - 2 1 4 4 5 6 9 3 6 3 2 - 1 - 1 1 49 45.7 Cumberland-Tennessee 1 2 7 4 5 2 6 10 2 5 3 2 - 3 52 45.1 Mississippi Valley 1 3 3 5 5 6 9 3 14 7 8 8 3 - 1 - 1 77 45.2 Southeast Missouri 1 - 1 3 3 6 4 5 9 10 11 8 3 2 2 1 1 - - 1 73 46.8 Table 2. Frequency distribution of circumferential scales for selected drainage units of L. umbratilis. Drainage unit 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 n Mean Lake Erie 10 12 8 12 2 1 45 34.7 Muskingum 1 2 4 6 5 1 1 25 36.9 West Virginia 2 9 8 25 9 5 4 62 34.9 Little Sandy 2 6 5 6 7 2 2 - 1 32 36.1 Licking 2 14 11 15 11 9 3 3 2 2 2 - 1 75 35.7 Green 3 8 15 13 8 4 1 1 54 35.7 Tradewater 1 - 3 5 7 9 12 4 2 2 3 1 49 36.5 Cumberland-Tennessee 2 10 11 11 9 4 4 1 52 37.9 Mississippi Valley 2 9 10 16 11 12 10 2 - 1 2 - - - 1 - 1 77 37.1 Southeast Missouri 1 - 6 16 9 13 11 8 3 2 3 2 3 1 73 38.0 524 Northeastern Naturalist Vol. 14, No. 4 although extensive overlap among populations was present in a plot of individual scores. Most separation occurred along the PC 1 axis; examination of PC1 loadings indicated that western populations had higher scale counts (Table 3). Although there was considerable local variation in body shape of nuptial males, multivariate analysis of the entire morphometric data set did not clearly reveal clinal variation or other geographic patterns (Fig. 3, top). In general, males from the western Ohio River basin were more robust than those from the eastern Ohio River basin (Fig. 1). Exceptions included relatively robust males from Lake Erie drainages and proximate Ohio River basin populations (Muskingum) and relatively gracile males from Black River populations in southeast Missouri. Removal of these populations from the sheared PCA, all at peripheral ends of the study area, resulted in a graph that shows clinal variation of male body shape from that part of the Ohio River basin (Fig. 3, bottom). Multivariate analysis of female body shape revealed no obvious geographic trends in variation. Despite substantial variation in dorsal-fin pigmentation of nuptial males within populations, there was a general trend showing males from western areas having darker fins than those from eastern areas (Fig. 4). Mean melanophore coverage for western populations was 47% (n = 35), 43% (n = 6) for the Green River, and 18% (n = 28) for eastern populations. Correlation between SL and dorsal-fin pigment was significant (p = 0.018), although the relationship was weak (r2 = 0.284). Substitution of regression residuals for raw melanophore-coverage percentages yielded a nearly identical pattern of geographic variation. Qualitative assessment of additional pigmentation patterns, chromatic coloration, and tuberculation of nuptial males did not reveal meaningful geographic patterns of variation. We judged males from the western Ohio River basin to have less red in median fins, but this appears primarily due to dark melanophores masking red pigment in specimens from that area. Table 3. Principle component loadings for 11 meristic variables for 544 selected specimens of L. umbratilis. Variable PC1 PC2 Anal rays -0.05 0.35 Pelvic rays -0.01 0.42 Pectoral rays 0.20 0.14 Predorsal scales 0.41 -0.16 Lateral-line scales 0.40 0.04 Scales above the lateral line 0.40 0.00 Scales below the lateral line 0.31 -0.10 Caudal peduncle scales 0.33 0.02 Circumferential scales 0.49 -0.09 Infraorbital pores 0.09 0.67 Preoperculomandibular pores 0.12 0.44 Eigenvalue 2.78 1.21 Proportion of variance 25.3% 11.0% 2007 D.J. Eisenhour and L.V. Eisenhour 525 Clinal variation in morphological characters was further corroborated by examination of correlograms (Fig. 5). All presented correlograms were “globally tested” (Legendre and Fortin 1989) by adjusting the α level with a Bonferroni correction and found to be signifi cant. Two variables (lateralline scales and scales around the caudal peduncle) showed a “depression” or bowl cline pattern (Sokal and Oden 1978) in which close populations show signifi cant positive correlation, moderately distant locations show signifi cant negative correlation, and very distant populations are mostly uncorrelated. Figure 3. Mean m o r p h o m e t r i c scores on sheared PC axes 2 and 3 for male L. umbratilis. Top, ten drainage units of 123 males. Bottom, eight drainage units of 84 males. The bottom graph is from a sheared PCA excluding samples from the Muskingum and Lake Erie drainages in Ohio and the Little Black River drainage in Missouri. Figure 4. Photos of dorsal fins of nuptial males of L. umbratilis, showing variation in pigmentation: (A) 53.8 mm SL, MOSU 1968, Dry Creek, Bollinger County, MO, 20 June 2003; (B) 50.6 mm SL, MOSU 1878, Massac Creek, McCracken County, KY, 22 June 2002; (C) 50.2 mm SL, MOSU 1999, Flynn Fk., Caldwell County, KY, 22 June 2003; (D) 52.8 mm SL, MOSU 2184, Carter County, KY, 6 July 2004. 526 Northeastern Naturalist Vol. 14, No. 4 That is, populations at either end of the study area (Lake Erie, Muskingum, and central Ozark populations) had higher counts than those in the middle of the study area (Tables 1 and 2). The remainder of the variables examined exhibited a fairly “smooth” clinal pattern in which close populations were positively correlated and distant populations were negatively correlated (Sokal and Oden 1978). Discussion Analysis of geographic patterns of morphological variation does not support recognition of multiple species of L. umbratilis in the Ohio River basin. Discrete geographic breaks in patterns of variation of morphological characters, suggestive of breaks in gene fl ow and thus evolutionary independence, are not apparent. The most marked shift in character variation, although minor, is above the Green River drainage. However, some of the “steepness” of the cline here may be due to a large geographic area just northeast of the Green River, the Salt River drainage, that was not sampled (Fig. 1). Although L. umbratilis is present in the lower part of the Salt River drainage and nearby small direct tributaries to the Ohio River, morphological and genetic evidence from this area indicates L. umbratilis is hybridizing with and possibly replacing L. fasciolaris (Hopkins 2005). Thus, samples from these areas are inappropriate to use because of probable introgression with L. fasciolaris. Figure 5. Correlograms of 7 variables of L. umbratilis. Signifi cant autocorrelation is indicated by closed symbols. Sample size is 544 for all variables except for dorsal fi n pigment, which is 69. 2007 D.J. Eisenhour and L.V. Eisenhour 527 We recognize that in some cases, morphological intermediacy and clinal variation are not due to genetic introgression, but to retained ancestral polymorphism (Mayden 2002, Mayden and Wood 1995). In such a situation, recognition of additional species-level diversity may be warranted. Because no phylogenetic analysis providing a historical perspective is available, this explanation cannot be completely discounted. However, invoking a retained ancestral polymorphism explanation to recognize species on the ends of clines requires an assumption that populations along the “cline” are isolated from each other. This seems unlikely for L. umbratilis, as there are no natural barriers to gene fl ow among most populations along the Ohio River basin. Extensive hybridization between L. umbratilis and L. fasciolaris across a broad area of the Ohio River basin supports the idea that these species have the capacity for substantial migration and gene fl ow (Hopkins 2005). The geographic pattern of variation in Ohio River basin L. umbratilis could be due to (1) local selective pressures, (2) drainage evolution, (3) gene fl ow with L. fasciolaris, or a combination of these factors. The pattern of clinal variation in L. umbratilis is perhaps most similar to patterns of geographic variation of Etheostoma kennicotti (Putnam) (Stripetail Darter; Page and Smith 1976) and Percina maculata (Girard) (Blackside Darter; Steinberg and Page 1999). Populations of these species in the upper part of the Ohio River basin (Licking River upstream to Big Sandy River) were reported to be morphologically different (although not diagnosable) from populations in western Kentucky and Tennessee. Also, like L. umbratilis, these two species have lower mean scale counts in specimens from the upper part of the Ohio River basin, perhaps suggesting that larger scale size is adaptive in the upper Ohio River basin, or is linked to locally adaptive genes. Perhaps morphological convergence of upper Ohio L. umbratilis to L. fasciolaris may be due to similar selective pressures, as well. In addition, clinal similarity across taxa may refl ect the history of drainage evolution in the Ohio River basin (Burr and Page 1986). Prior to the glacial advances of the Pleistocene, West Virginia, southern Ohio, and eastern Kentucky were drained by the Teays River, which fl owed westward across central Indiana and Illinois to join the Mississippi River. Major modern rivers that likely fed the Teays include the Kentucky, Licking, Big Sandy, and Kanawha rivers. The Pliocene Ohio River was separate from the Teays River and was much smaller than at present, supplied primarily by the Tennessee, Cumberland, Green, and lower Wabash rivers. Thus, morphological differences of L. umbratilis along the length of the Ohio River basin may be due to partial isolation of populations in the Teays and Old Ohio watersheds, followed by gene fl ow subsequent to capture of the Teays watershed by the Ohio River in the Pleistocene. An additional explanation for the pattern of morphological variation present in upper Ohio River basin L. umbratilis is that introgression with L. fasciolaris has occurred. These species have a parapatric or narrowly sympatric distribution (Burr and Warren 1986, Trautman 1981) in the upper Ohio 528 Northeastern Naturalist Vol. 14, No. 4 River basin from the Green River drainage upstream, providing opportunities for genetic exchange. Historical geographic interactions of these species, perhaps involving competitive exclusion, apparently has resulted in their unusual “alternating” distributions in drainages of eastern Kentucky (L. fasciolaris in the Kentucky River, L. umbratilis in the Licking River, L. fasciolaris in Tygart’s Creek, and L. umbratilis in the Little and Big Sandy rivers). Lythrurus umbratilis specimens in the upper Ohio River basin show morphological characteristics, including reduced black pigment on the dorsal fi ns, a more slender body, and lower scale counts, which are similar to those of L. fasciolaris. Evidence of current or historical gene fl ow was documented by Hopkins (2005), who found L. fasciolaris nuclear markers present in putatively “pure” populations of L. umbratilis in southeast Ohio and northeast Kentucky. Despite the similarities of upper Ohio River basin L. umbratilis and L. fasciolaris, they can be diagnosed using morphological characters. All L. fasciolaris lack black pigment in the dorsal fi n (other than the anterior blotch), but breeding male L. umbratilis have at least some dark melanophores scattered throughout, though they may not be visible without magnifi cation. Suborbital tubercles consistently are present in breeding male L. umbratilis, but absent in L. fasciolaris (Hopkins 2005). Although these species differ in some scale and fi n-ray counts and in body shape (Hopkins 2005), pronounced overlap limits their utility as “diagnostic” characters. Acknowledgments We thank R.L. Hopkins, R.R. Hopkins, and J.D. Eisenhour for assistance in fi eld collections and R.L. Hopkins for providing a helpful review of the manuscript. Access to museum specimens was provided by B.M. Burr, T. Cavender, M. Kibbey, and J. Stewart. Partial funding was provided by research grants from the Offi ce of Research, Grants, and Contracts and the Institute for Regional Analysis and Public Policy, Morehead State University. Specimens were removed from wild populations in accordance with permits from Kentucky Department of Fish and Wildlife Resources, Tennessee Wildlife Resources Agency, Missouri Department of Conservation, Ohio Division of Wildlife, West Virginia Division of Natural Resources, and Morehead State University IACUC. Literature Cited Boltz, J.M., and J.R. Stauffer, Jr., 1993. Systematics of Aphredoderus sayanus (Teleostei: Aphredoderidae). Copeia 1993:81–98. Bookstein, F.L., B. Chernoff, R.L. Elder, J.M. Humphries, Jr., G.R. Smith, and R.E. Strauss. 1985. Morphometrics in evolutionary biology. Academy of Natural Sciences, Philadelphia Special Publication 15:1–277. Burr, B.M., and L.M. Page. 1986. Zoogeography of fishes of the lower Ohio-upper Mississippi basin. Pp. 287–324, In C.H. Hocutt and E.O. Wiley (Eds.). The Zoogeography of North American Freshwater Fishes. John Wiley and Sons. New York, NY. 866 pp. Burr, B.M., and M.L. Warren, Jr. 1986. A Distributional Atlas of Kentucky Fishes. Kentucky Nature Preserves Commission, Frankfort, KY. Science and Technical Series 4. 398 pp. 2007 D.J. Eisenhour and L.V. Eisenhour 529 Cox, R.M., S.L. Skelly, A. Leo, and H.B. John-Alder. 2005. Testosterone regulates sexually dimorphic coloration in the Eastern Fence Lizard, Sceloporus undulatus. Copeia 2005:597–608. Dimmick, W.W., K.L. Fiorino, and B.M. Burr. 1996. Reevaluation of the Lythrurus ardens (Cypriniformes: Cyprinidae) complex with recognition of three evolutionary species. Copeia 1996:813–823. Eisenhour, D.J. 2004. Systematics, variation, and speciation of the Macrhybopsis aestivalis complex west of the Mississippi River. Bulletin of the Alabama Museum of Natural History 23:9–47. Eisenhour, D.J., and L.V. Eisenhour. 2004. Morphological variation and systematics of the Notropis rubellus complex (Teleostei: Cyprinidae) in Kentucky and Tennessee. Southeastern Naturalist 3:431–450. Etnier, D.A., and W.C. Starnes. 1993. The Fishes of Tennessee. University of Tennessee Press, Knoxville, TN. 681 pp. Hopkins, R.L. 2005. Hybridization of Lythrurus fasciolaris (Cope) and Lythrurus umbratilis (Girard) (Cypriniformes: Cyprinidae) in the Ohio River Basin. Unpublished M.Sc. Thesis. Morehead State University, Morehead, KY. 95 pp. Hubbs, C.L., and K.F. Lagler. 1964. Fishes of the Great Lakes Region. University of Michigan Press, Ann Arbor, MI. 213 pp. Humphries, J.M., F.L. Bookstein, B. Chernoff, G.R. Smith, R.L. Elder, and S.L. Poss. 1981. Multivariate discrimination by shape in relation to size. Systematic Zoology 30:291–308. Legendre, P., and M.-J. Fortin. 1989. Spatial pattern and ecological analysis. Vegetation 80:107–138. Legendre, P., and A. Vaudor. 1991. The R Package: Multidimensional Analysis, Spatial Analysis. Département de Sciences Biologiques, Université de Montréal, Montréal, QC, Canada. 142 pp. Mayden, R.L. 2002. On biological species, species concepts, and individuals in the natural world. Fish and Fisheries 3:171–196. Mayden, R.L., and R.M. Wood. 1995. Systematics, species concepts, and the evolutionary signifi cant unit in biodiversity and conservation biology. Pp. 58–113, In J.L. Nelson (Ed.). Evolution and the Aquatic Ecosystem: Defi ning Unique Units in Population Conservation. American Fisheries Society, Bethesda, MD. 435 pp. Page, L.M., and B.M. Burr. 1991. A Field Guide to Freshwater Fishes of North America North of Mexico. Houghton Miffl in Co., Boston, MA. 422 pp. Page, L.M., and P.W. Smith. 1976. Variation and systematics of the Stripetail Darter, Etheostoma kennicotti. Copeia 1976:532–541. Poss, S.G., and B.B. Collette. 1995. Second survey of fi sh collections in the United States and Canada. Copeia 1995:48–70. Poss, S.G., and R.R. Miller. 1983. Taxonomic status of the plains killifi sh, Fundulus zebrinus. Copeia 1983:55–66. Pramuk, J.B., M.J. Grose, A.L. Clarke, E. Greenbaum, E. Bonaccorso, J.M. Guayasamin, A.H. Smith-Pardo, B.W. Benz, B.R. Harris, E. Siegfreid, Y.R. Reid, N. Holcroft-Benson, and E.O. Wiley. 2006. Phylogeny of fi nescale shiners of the genus Lythrurus (Cypriniformes: Cyprinidae) inferred from four mitochondrial genes. Molecular Phylogenetics and Evolution 42:287–297. Robison, H.W., and T.M. Buchanan. 1988. Fishes of Arkansas. University of Arkansas Press, Fayetteville, AR. 535 pp. 530 Northeastern Naturalist Vol. 14, No. 4 Schmidt, T.R., J.P. Bielawski, and J.R. Gold. 1998. Molecular phylogenetics and evolution of the cytochrome-b gene in the cyprinid genus Lythrurus (Actinopterygii: Cypriniformes). Copeia 1998:4–22. Snelson, F.F., Jr., and W.L. Pfl ieger. 1975. Redescription of the Redfi n Shiner, Notropis umbratilis, and its subspecies in the central Mississippi River basin. Copeia 1975:231–249. Sokal, R.R., and N.L. Oden. 1978. Spatial autocorrelation in biology 1. Methodology. Biological Journal of the Linnaean Society 10:199–228. Stauffer, J.R., Jr., J.M. Boltz, L.R. White. 1995. The fi shes of West Virginia. Proceedings of the Academy of Natural Sciences, Philadelphia 146:1–389. Strauss, R.E., and F.L. Bookstein. 1982. The truss: Body-form reconstruction in morphometrics. Systematic Zoology 31:113–135. Steinberg, R., and L.M. Page. 1999. Geographic variation in the blackside darter, Percina maculata (Teleostei, Percidae), in the Ohio River drainage of Kentucky. Journal of the Kentucky Academy of Science 60:94–107. Taylor, W.R., and G.C. Van Dyke. 1985. Revised procedures for staining and clearing small fi shes and other vertebrates for bone and cartilage study. Cybium 9:107–119. Trautman, M.B. 1981. The Fishes of Ohio, 2nd Edition. Ohio State University Press, Columbus, OH. 782 pp. Appendix 1. Specimens of Lythrurus umbratilis examined in quantitative analyses. Parenthetical numbers refer to number of specimens used in the meristic, morphometric, and dorsal-fi n pigmentation analyses, respectively. Cleared and stained specimens used for vertebral counts are identifi ed by “CS.” Lake Erie: OSUM 17227 (3,0,0), OSUM 17325 (3,2,0), OSUM 21558 (24,5,0), OSUM 52126 (15,4,0). Muskingum: OSUM 50941 (9,6,0), OSUM 73744 (4,3,0), OSUM 73869 (7,1,0), OSUM 73340 (5,4,0). West Virginia: MOSU 2197 (9,3,0), MOSU 2010 (7,2,0), MOSU 2009 (11,7,1), MOSU 2188 (30,4,2), MOSU 2189 (35 CS), MOSU 2186 (4,1,0). Little Sandy: MOSU 1930 (3,3,2), MOSU 2008 (15,8,2), MOSU 2184 (14,6,3). Licking: MOSU 1832 (27,16,7), MOSU 1962 (18,4,3), MOSU 1963 (30 CS), MOSU 1835 (21,13,3), MOSU 2160 (6,4,1), MOSU 1462 (3,3,0). Green: West MOSU 1839 (19,9,1), MOSU 1792 (8,3,2), MOSU 1967 (17,6,4), MOSU 2167 (10,8,3), MOSU 2168 (13 CS). Tradewater: MOSU 1999 (22,6,2), MOSU 2285 (20,5,1), MOSU 875 (7,1,0). Cumberland-Tennessee: MOSU 1970 (20,6,7), MOSU 1975 (32 CS), MOSU 1343 (18,4,1), MOSU 2165 (14,5,2). Mississippi Valley: SIUC 42837 (6,6,0), SIUC 34798 (25,10,0), MOSU 2208 (2,1,0), MOSU 1978 (17,6,3), MOSU 1878 (13,12,4), MOSU 1987 (15,4,1). Southeast Missouri: MOSU 1968 (20,12,4), MOSU 2280 (25,5,5), MOSU 2284 (28,6,4).