Taxonomic Status of Etheostoma denoncourti Stauffer and van Snik
Abstract
The upper Tennessee River drainage populations of Etheostoma tippecanoe Jordan and Evermann were recently elevated to specific status with the description of the Golden Darter, Etheostoma denoncourti Stauffer and van Snik. In the description, the Duck River system was included as part of the range of E. denoncourti, although no specimens from that system were examined. Evaluation of the taxonomic status of the Duck River system population of E. denoncourti was conducted using meristic, morphometric, and nuptial coloration characters. The Duck River population was compared with other Tennessee River drainage populations of E. denoncourti and with Cumberland and Ohio drainage populations of E. tippecanoe. The Duck River population appears to be unique based strictly on meristics; however, it is similar to other E. denoncourti populations by nuptial coloration, presence of cheek scales behind the eye, and the complete separation of the posterior nasal openings from the interorbital canal. Etheostoma denoncourti is rediagnosed, and current status and distributional information are provided for both E. tippecanoe and E. denoncourti.
The Tippecanoe Darter, Etheostoma tippecanoe Jordan and Evermann, was described from specimens collected in the Tippecanoe River, Ohio River drainage, Indiana (Jordan and Evermann, 1890). Until the early 1960s, Tippecanoe darters were known only from a few widely disjunct populations in the Ohio and Cumberland River drainages. The first Tennessee River drainage specimens were collected in 1963 in the Clinch River (Jenkins and Burkhead, 1994).
Zorach (1969) studied variation in the Tippecanoe darter as part of his review of the darter subgenus Nothonotus. He examined specimens from throughout the known range of E. tippecanoe including a small series from the Clinch River system. He found essentially no differences between Ohio and Cumberland drainage populations of E. tippecanoe; however, specimens from the Clinch River had 1–4 scales on the cheek behind the eye, whereas the Ohio and Cumberland populations were scaleless in this area. He also reported coloration, cephalic lateralis, and possible belly squamation differences between the Clinch River population and populations from the Ohio and Cumberland drainages; yet, he did not think these differences warranted taxonomic recognition.
Subsequent to Zorach's (1969) paper, Tippecanoe darters were discovered in the Duck, Buffalo (a major Duck River tributary), and Sequatchie Rivers of the Tennessee River drainage and Big South Fork and Red Rivers of the Cumberland River drainage. Jenkins and Burkhead (1994) recognized coloration and squamation differences shared by the Clinch and Duck River populations and remarked that they may represent an undescribed species.
Stauffer and van Snik (1997) described the upper Tennessee River populations of E. tippecanoe as Etheostoma denoncourti Stauffer and van Snik, the Golden Darter. They separated the two species by the presence (Tennessee drainage) versus absence (Ohio and Cumberland drainages) of cheek scales and differences in nuptial coloration. They stated that E. denoncourti is endemic to the upper Tennessee River drainage, which is confusing because Tippecanoe darters are known from the Duck and Buffalo Rivers of the lower Tennessee River drainage (Etnier and Starnes, 1993). Further, Stauffer and van Snik (1997) did not examine specimens from the Duck or Buffalo Rivers, although they included those rivers within the range of E. denoncourti. Based on drainage connections, the Duck River system is geographically intermediate between the Clinch and Sequatchie River systems of the upper Tennessee River drainage and the Cumberland and Ohio River drainages. Thus, an analysis of that population is crucial to interpreting variation across the range of E. tippecanoe sensu Zorach (1969).
Because the Duck River system has faunal affinities with both the Cumberland River drainage and the upper Tennessee River drainage (Lee et al., 1980; Etnier and Williams, 1989; Etnier and Starnes, 1993), the initial impetus for this work was to determine whether the Duck River system population represented E. denoncourti, E. tippecanoe, or something intermediate between the two nominal taxa. The study ultimately resulted in our analysis of meristic, morphometric, and fresh nuptial coloration characters for virtually all known populations of the two nominal species. Here, we present those data to allow a thorough evaluation of the status of the recently described E. denoncourti.
Materials and Methods
Counts and measurements generally follow Hubbs and Lagler (1964). All counts (by DAE) and measurements (by CES) were made on the left side except for counts of cheek scales, which were made on both sides of the head. Positions of the posterior nasal pores of the cephalo-lateralis system were also recorded for both sides of the head. Measurements were made with needle-point digital calipers and taken to the nearest 0.01 mm. Shape variation among the Duck/Buffalo (hereafter called Duck) population, the Clinch and Sequatchie populations (E. denoncourti), and seven populations from the Ohio and Cumberland drainages (E. tippecanoe) was assessed using sheared principal components analysis (Humphries et al., 1981; Bookstein et al., 1985) with a program written for SAS by D. Swofford (Statistical Analysis Systems Institute, Inc., Cary, NC). The sheared principal components analysis of 27 morphometric variables taken from 94 individuals showed considerable overlap among all populations examined and was considered uninformative.
Meristic data were analyzed with one-way analysis of variance (ANOVA) among all populations and also among the Duck population and the two nominal taxa. Variables that exhibited significant variation were further evaluated with the Tukey posthoc test (P < 0.05) to determine which populations differed from one another. All statistical procedures were conducted with SAS (vers. 6.12, Statistical Analysis Systems Institute, Inc., Cary, NC, 1996, unpubl.). Color descriptions were made from freshly preserved specimens and photographs. Institutional abbreviations are as listed in Leviton et al. (1985).
Results
Frequency distributions of meristic data are presented in Tables 1–3. For all populations examined (Table 1), number of anal rays is modally 8 (7–9). The Stones River population was significantly different from the East Fork White River, Clinch River, and Little Kanawha River; however, there were no overall differences among nominal E. denoncourti, E. tippecanoe, and the Duck River population. For dorsal spines (Table 2), the Clinch and the Sequatchie River populations (E. denoncourti) were not different from each other but were significantly different from all other populations examined, with modally 13 spines. The Duck River population had modally 12 dorsal spines as did all Cumberland and Ohio River drainage populations (E. tippecanoe). For dorsal soft rays, the Duck River population stands alone with a strong mode of 11 soft rays (Table 2). Dorsal fin rays of all other populations were modally 12, except for the Clinch River (which is bimodal at 11 or 12 rays) and the Red River. These modal differences in dorsal fin soft rays in Duck River specimens also represent statistically significant differences from all other populations except the Red River. In addition, Duck River specimens were unique in often having as few as 10 dorsal fin soft rays (Table 2). Number of pored lateral line scales was highly variable, but the frequency distribution indicates a general increase in pored scales from the Tennessee River to the upper Ohio River (Table 3). The Duck River stands out as having the fewest pored scales (x̄ = 23.3) and is significantly different from all populations except French Creek and the Sequatchie and Stones Rivers. When considering scales in lateral series, the only significant difference was between French Creek and Big Darby Creek, both Ohio River tributaries. This difference may be partially attributable to the limited number of specimens examined from French Creek. A comparison among the Duck River population and the two nominal taxa revealed no significant differences for this count. In contrast, Stauffer and van Snik (1997) reported E. denoncourti to have lower counts of scales in lateral series than did E. tippecanoe. Their Materials section indicates this to be based primarily on upper and middle Ohio River populations versus Clinch River specimens, with no Duck River specimens, nine Sequatchie River specimens, and 10 Cumberland River drainage specimens examined.
Based solely on meristic data, the Duck River population, which differs from other Tennessee River drainage populations (E. denoncourti) in both dorsal fin spines and dorsal fin rays, is nameworthy. Cheek scales were present on all individuals examined from the Clinch, Sequatchie, and Duck Rivers. Specimens from the Cumberland and Ohio drainages rarely had cheek scales. We did, however, find a scale or two on three of 19 specimens from East Fork White River, one of three from French Creek, one of seven from Tippecanoe River (all Ohio River drainage) and one of six from Red River, one of seven from Harpeth River, and one of five from Big South Fork Cumberland River (all Cumberland River drainage). Extent of cheek squamation is a reasonable and nearly diagnostic character for separating the two taxa. Within the subgenus Nothonotus, this character has been shown to be useful in defining the Etheostoma maculatum species group (Etnier and Williams, 1989) and shows very little within-species variation (Page, 1983).
We surveyed the extent of belly squamation but did not quantify our observations. There was a general trend for E. tippecanoe specimens to have fewer belly scales than E. denoncourti and specimens from the Duck River. Specimens from the Sequatchie and Little Kanawha Rivers represented the two extremes. Sequatchie individuals were nearly fully scaled, whereas the Little Kanawha population had virtually naked bellies. Except for the Sequatchie River population, almost all specimens lacked scales along the midline of the belly.
Examination of breeding color indicates some variation between populations that is concordant with the nominal taxa. The first dorsal saddle is almost always more intense on individuals from the Tennessee River drainage (including Duck River). This saddle is followed by a pale area which is the dorsal color of the fish. The next saddle is narrow and is almost always completely beneath the anterior portion of the spinous dorsal fin. In the Ohio and Cumberland River drainages (E. tippecanoe), the first saddle is less intense, and the second saddle almost always extends in front of the spinous dorsal fin.
Examination of the infraorbital canal indicated no differences between E. tippecanoe, E. denoncourti, and specimens from the Duck River system. The infraorbital canal was almost always complete and had a strong mode of 7 (6–8) pores. Zorach (1969) found that the posterior nasal pores of the cephalo-lateralis system were noticeably separated from the posterior nasal openings in Clinch River populations (E. denoncourti). The pores of Clinch River specimens were medial to the posterior nasal openings, whereas the pores of all other Ohio Basin populations adjoined the nasal opening. We found that this character successfully separates E. tippecanoe from E. denoncourti plus specimens from the Duck River system. The posterior nasal pores of E. tippecanoe either adjoin the posterior nasal opening or are completely included in it.
We examined specimens of all other Nothonotus species to ascertain the polarity of this character within the subgenus. The pores were almost always separate from the nasal opening on all Nothonotus species except E. tippecanoe. Occasionally, an individual had the opposite state for this character on one or both sides of the head. This occurred in one of 39 E. denoncourti, three of 46 E. tippecanoe, and two of 10 E. jordani. For the six E. etowahae examined, two had the separate state on one side of the head and the joined state on the other. Among the other species of Nothonotus, the pores come closest to joining the nasal opening on E. rufilineatum and E. acuticeps. Because of the widespread nature of the separate character state within Nothonotus and outgroups (e.g., Etheostoma caeruleum and Etheostoma flabellare), we believe this to be the primitive character state and, thus, consider the joined pore state an autapomorphy for E. tippecanoe.
Discussion
The frequency distributions of the meristic characters indicate that the Duck River population is unique among all of the populations examined. The Duck River is aligned with the Ohio/Cumberland in number of dorsal fin spines and is significantly different from all other populations in number of dorsal fin rays (10–12, mode 11, vs 11–13, mode 12) except for the Red River. The Duck River also has the fewest pored lateral line scales (Table 3). Duck River specimens are aligned with those from the Sequatchie and the Clinch Rivers by the presence of cheek scales behind the eye. Duck River specimens are also similar to Clinch and Sequatchie specimens in having more intense coloration of the first dorsal saddle and by the placement of the second saddle. Finally, the nasal pore character unites specimens from the Duck, Sequatchie, and Clinch systems. We concur with Stauffer and van Snik (1997) and recognize E. denoncourti as a valid species confined to the Clinch, Sequatchie, Duck, and Buffalo Rivers of the Tennessee River drainage.
Diagnosis
Etheostoma denoncourti is a member of the subgenus Nothonotus as defined by Page (1983) and Bailey and Etnier (1988). Its diminutive size and absence of dark horizontal lines separate it from all other species in the subgenus except E. tippecanoe. It differs from E. tippecanoe by the consistent presence of scales on the cheek behind the eye and the complete separation of the posterior nasal pores of the cephalo-lateralis system from the posterior nasal openings. Further, the second dorsal saddle of E. denoncourti is almost always completely beneath the anterior portion of the spinous dorsal fin, whereas this saddle extends anterior to the spinous dorsal fin on E. tippecanoe.
Status and distribution
Etheostoma denoncourti is endemic to the Tennessee River drainage in Tennessee and Virginia (Fig. 1). It has been collected from Copper Creek (Clinch tributary) and the Clinch, Sequatchie, Duck, and Buffalo Rivers (Etnier and Starnes, 1993). In Virginia, E. denoncourti is known only from a few localities in a 50-mile reach of the Clinch River (Jenkins and Burkhead, 1994). Occasional reports of this fish in Copper Creek are likely the result of periodic recruitment from the Clinch (Jenkins and Burkhead, 1994). The Virginia Department of Conservation and Recreation recognizes E. denoncourti as threatened in the state (L. Berlinghoff, Virginia Div. Nat. Heritage, pers. comm.), but because it is restricted to a single, relatively short reach of the Clinch River, we recommend elevating it to a status of endangered.
Citation: Ichthyology & Herpetology 2000, 4; 10.1643/0045-8511(2000)000[1097:TSOEDS]2.0.CO;2
In Tennessee, E. denoncourti is fairly common and widespread in the Clinch and Duck Rivers but is restricted to short reaches in the lower Sequatchie and Buffalo Rivers. We recommend threatened status for E. denoncourti in Tennessee.
Etheostoma tippecanoe is widespread in the Ohio River drainage (Fig. 1) but is considered imperiled over much of its range. It persists in the Tippecanoe River and East Fork White River in Indiana but is considered endangered in that state (B. Fisher, Indiana Fish and Game, pers. comm.). In Kentucky, the Tippecanoe darter has recently been de-listed by the Kentucky State Nature Preserves Commission (R. Cicerello, Kentucky St. Nat. Pres. Comm., pers. comm.) because of discoveries of additional populations. The status of the Tippecanoe darter in Tennessee is Species in Need of Management (P. Shute, Tennessee Valley Authority Nat. Heritage Prog., pers. comm.), but the status is based on its synonymy with E. denoncourti. Elevation of E. denoncourti to species status results in E. tippecanoe becoming restricted to short reaches in the Big South Fork Cumberland, Red, and Harpeth Rivers. It should probably be considered endangered in the state. In West Virginia, the Tippecanoe darter is known from the Elk and Little Kanawha Rivers (Stauffer et al., 1995) and is listed as threatened (D. Cincotta, West Virginia Nat. Heritage Prog., pers. comm.). The Tippecanoe darter is known from the Scioto River system and the lower Muskingum River system in Ohio and is considered threatened by the Ohio Department of Natural Resources, Division of Wildlife (P. Jones, Ohio Dept. of Nat. Res., Div. Nat. Areas and Pres., pers. comm.). In Pennsylvania, the Tippecanoe darter is found in the Allegheny River and French Creek and is considered threatened (J. Stauffer, Penn St. Univ., pers. comm.)
Material Examined
Etheostoma tippecanoe
Ohio River Drainage, Ohio: Big Darby Creek: INHS 74280 (15), INHS 82171 (1), NLU 34867 (10), NLU 56270 (15), NLU 56210 (14), NLU 56229 (1), OSUM 13894 (31), and UT 91.5465 (11). Indiana: Tippecanoe River: uncat., G. Seegert coll. (7); East Fork White River: UGAMNH 216 (19). Kentucky: Green River: INHS 41703 (2), INHS 64017 (7), and TU 19392 (6); Little Barren Fork: UT 91.5498 (8). West Virginia: Little Kanawha River: CU 5868 (12) and CU 5882 (28); Elk River: UT 91.3663 (2) and UT 91.5499 (9). Pennsylvania: French Creek: INHS 87158 (5). Cumberland River Drainage, Tennessee: Red River: UT 91.3742 (6); Harpeth River: TU 32984 (3), UT 91.1122 (5), UT 91.5243 (10), and UT 91.5497 (7); Big South Fork: UT 91.446 (2), UT 91.661 (6), UT 91.1464 (10), and UT 91.4296 (13); Stones River: TU 19481 (7) and TU 32985 (18).
Etheostoma denoncourti
Tennessee River Drainage, Tennessee: Duck River: NLU 52841 (3), NLU 56842 (3), NLU 56640 (13), SIUC 19599 (8), SIUC 20860 (1), SIUC 19911 (3), UT 91.3782 (2), UT 91.2443 (26), UT 91.2461 (20), UT 91.2461 (3), UT 91.5107 (10), UT 91.4780 (5), UT 91.2500 (41), UT 91.4121 (12), and UT 91.5440 (2); Buffalo River: UT 91.4596 (2); Sequatchie River: UT 91.3663 (19), UT 91.5511 (2), and UT 91.2626 (10); Clinch River: UT 91.3903 (50), UT 91.814 (23), UT 91.1676 (13), UT 91.2130 (14), UT 91.3048 (9), and UT 91.1157 (5). Virginia: Copper Creek: CU 2135 (1).
Additional specimens examined
Tennessee River Drainage, Tennessee: Etheostoma acuticeps—Nolichucky River: UT 91.4036 (12); Etheostoma aquali—Duck River: UT 91.5050 (7); Etheostoma chlorobranchium—West Prong Little Pigeon River: UT 91.5436 (11); Etheostoma rufilineatum—Holston River: UT 91.5365 (10); Etheostoma wapiti—Elk River: UT 91.3084 (7). North Carolina: Etheostoma vulneratum—Tuckasegee River: UT 91.5506 (6). Cumberland River Drainage, Tennessee: Etheostoma microlepidum—Red River: UT 91.3739 (19); Etheostoma sanguifluum—New River: UT 91.5203 (15). Ohio River Drainage, Tennessee: Etheostoma bellum—West Fork Drake Creek: UT 91.3772 (20). Ohio: Etheostoma camurum—Big Darby Creek: UT 91.5462 (12). Pennsylvania: Etheostoma maculatum—French Creek: UT 91.3448 (15). Alabama River Drainage, Alabama: Etheostoma chuckwachatte—Cornhouse Creek: UT 91.4157 (20); Etheostoma douglasi—West Fork Sipsey River: UT 91.4171 (3). Coosa River System, Georgia: Etheostoma etowahae—Etowah River: UT 91.1902 (4); Amicalola Creek: UT 91.5157 (2); Etheostoma jordani—Armuchee Creek: UT 91.5161 (10). White River Drainage, Arkansas: Etheostoma juliae—Buffalo River: UT 91.1289 (17); Etheostoma moorei—Middle Fork Little Red River: UT 91.980 (24). Mississippi River Drainage, Mississippi: Etheostoma rubrum—Bayou Pierre: UT 91.311 (10).
Distributions of Etheostoma denoncourti and Etheostoma tippecanoe. The open circle represents an extirpated population of E. tippecanoe