Percina sciera

Robert H. Robins & Lawrence M. Page, 2007, Taxonomic status of the Guadalupe Darter, Percina apristis (Teleostei: Percidae)., Zootaxa 1618, pp. 51-60 : 51-59

[[ Percina sciera (Swain 1883) ]]

Introduction

The Dusky Darter, Percina sciera (Swain 1883) is a wide-ranging species that occurs in the Mississippi River basin from Ohio and West Virginia west to Illinois and south to Louisiana, and in Gulf drainages from northwest Alabama south to the Guadalupe River, Texas (Fig. 1). Percina sciera is one of four species in the subgenus Hadropterus (Page 1974). All but P. sciera are restricted to Gulf and Atlantic Slope drainages east of the Mississippi River (Page & Burr 1991).

Two subspecies of P. sciera are currently recognized: P. s. sciera (Swain), distributed throughout the majority of the range, and P. s. apristis (Hubbs and Hubbs), restricted to the Guadalupe River system of the San Antonio Bay drainage, Texas (Fig. 1). The San Antonio Bay drainage is the penultimate major drainage on the Gulf Slope between the Mississippi River and the Rio Grande and consists of the Guadalupe and San Antonio River systems. Percina s. apristis is found in the Guadalupe, but not the San Antonio, river system (Hubbs et al. 1991).

Percina s. apristis was diagnosed from P. s. sciera primarily by a reduced number of preopercular serrae (Hubbs & Hubbs 1954). Hubbs (1954) examined specimens from 13 river systems in a north/south gradient from the Wabash River, Indiana, to the Guadalupe River, Texas. He found that counts of preopercular serrae formed a cline, with counts low in the north and highest in the Brazos and Colorado rivers, Texas, with an abrupt shift to much lower counts in the Guadalupe River. He also noted other differences in meristic counts, body proportions and color pattern, as well as the peripheral isolation of apristis , and remarked that “ apristis may be specifically distinct from scierus ” (Hubbs 1954, p. 215). However, he declined to recognize the taxon as a species due to concerns about overlapping variation in analyses that considered all 13 river systems examined.

More recent studies have described additional differences between P. s. apristis and P. s. sciera . Research comparing the egg complements of equal-sized female P. s. apristis and Colorado River P. s. sciera found significant differences in egg size and number (Hubbs & Johnson 1961). Laboratory hybridization crosses between P. s. apristis and Colorado River P. s. sciera produced offspring that exhibited an overall lack of hybrid vigor, prompting speculation that differences in phenotypes of the two subspecies may indicate genetic isolation (Hubbs 1967).

Other than the distinctiveness of P. s. apristis , variation within western P. sciera remains poorly understood. Following the lead of Hubbs and Black (1954), who concluded that P. sciera may be a complex of subspecies, Richards and Knapp (1964) defined P. s. sciera as consisting only of those populations found from eastern tributaries of the Mississippi River basin and along the Gulf coast east to the Pearl River in Mississippi. These authors relegated all populations of P. sciera west of the Mississippi River in Missouri, Arkansas, Oklahoma and Texas, with the exception of P. s. apristis , to uncertain status pending further study.

Reference in the description of P. s. apristis to a yellow to orange band in the distal portion of the spinous dorsal fin (Hubbs & Hubbs 1954) has led some authors to treat the characteristic as diagnostic of the subspecies. However, observations of a yellow band in populations of P. s. sciera by other authors confound this distinction. The present study analyzes variation in meristic features and color patterns in P. sciera from eight western Gulf Slope river drainages and discusses the results relative to the taxonomic status of the population in the Guadalupe River system. We chose not to examine body and fin shapes because of the large ontogenetic and geographic variation found by Hubbs and Hubbs (1954).

Methods

Sixty-five collections totaling 389 specimens of P. sciera were examined for variation in numbers of preopercular serrae, pored lateral-line scales, modified scales (present in males only), transverse scales, circumferential caudal-peduncle scales, dorsal spines, dorsal rays, pectoral rays, and anal rays. Specimens from eight river drainages, encompassing the known range of P. sciera on the western Gulf Slope between the Mississippi River and Rio Grande River, constituted the sample. From southwest to northeast, these are the Guadalupe, Colorado, Brazos, San Jacinto, Trinity, Neches, Sabine, and Calcasieu drainages (Fig. 1).

All counts and measurements were made on the left side using a Leica MZ75 dissecting microscope and follow Hubbs and Lagler (1964) except counts of transverse scales were made from the anal fin origin anterodorsally to the spinous dorsal fin, and counts of modified scales were those in a single midbelly row from the anus to the posterior base of the pelvic symphysis. Counts of pored lateral-line scales include those on the body and caudal fin. Standard length was taken with dial calipers and recorded to the nearest tenth millimeter. Sex was recorded for all specimens. Juveniles of less than 35 mm SL and damaged specimens were excluded from the study. Institutional abbreviations follow Leviton et al. (1985). Pigment patterns on preserved specimens were compared among all drainages, and breeding adults in the Guadalupe and Brazos drainages were observed and photographed in streams and aquaria.

Specimens examined: Catalog numbers are arranged by drainage; numbers of specimens are in parentheses. Guadalupe River system: TNHC 2287 (15), TNHC 2297 (5 of 6), THNC 2321 (12 of 14), TNHC 2609 (6), TNHC 3136 (4), TNHC 7708 (3), TNHC 10387 (2), TNHC 17332 (15), UF 14996 (9), UF 26598 (1), UF 29652 (12), UF 43396 (2 of 3), UF 50605 (4), UF 65778 (9), UF 65788 (15), UF 96807 (5), UF 96809 (8), UF 128463 (2). Colorado River drainage: TNHC 2613 (2) , TNHC 2653 (1) , TNHC 5386 (9) , TNHC 9221 (7) , TNHC 9485 (1) , TNHC 15633 (1) , TNHC 23129 (1) , TNHC 23612 (1) , TNHC 23890 (2) , TNHC 24995 (9 of 10) , TNHC 27756 (2) , UAIC 799 (3) , UAIC 11412.19 (11 of 36) . Brazos River drainage: TNHC 3785 (1) , TNHC 12347 (7) , TNHC 16791 (6) , UF 28104 (3) , UF 50646 (10) , UF 147820 (10) . San Jacinto River drainage: TNHC 2398 (9) , TNHC 2806 (5) , TNHC 2823 (2 of 3) , TNHC 22367 (3 of 4) , UAIC 11417.06 (8) , UF 29467 (8) , UF 29644 (3) . Trinity River drainage: TNHC 21908 (5) , TNHC 22166 (5) , UF 147842 (2) , UF 147868 (1) , UMMZ 162371 (15 of 16) , USNM 160626 (12 of 28) . Neches River drainage: TNHC 2391 (6) , TNHC 2756 (3) , TNHC 2930 (4) , TNHC 3832 (6) , UF 29603 (4) . Sabine River drainage: TNHC 3301 (6) , TNHC 3345 (4) , TNHC 3353 (12) , TNHC 3376 (7) , TNHC 3669 (18) , TNHC 15528 (3) , UAIC 2677 (9) , UF 29589 (5) . Calcasieu River drainage: UAIC 2679 (3) , UF 147870 (5 of 11) .

Results

The number of preopercular serrae varies greatly among western Gulf Slope populations of P. sciera (Table 1). Specimens from the Guadalupe River system have the fewest serrae (-χ= 0.70), and those from the adjacent Colorado and Brazos river drainages have the most (-χ = 12.5 and 14.3, respectively). Other populations are intermediate, with means varying from 7.8 to 11.5.

The number of pored lateral-line scales is higher in the Guadalupe River (-χ= 71.7) than elsewhere. Other populations show little variation, with means ranging from 66.0 to 68.5 (Table 2). The number of modified scales in males is highest in the Guadalupe River (-χ= 27.3) and intermediate in the Colorado River (-χ= 19.8). Other populations have lower values, with means ranging from 13.8 to 17.9 (Table 3).

The number of transverse scales shows a similar pattern to that of modified scales, although variation among drainages is low (Table 4). The Guadalupe and Colorado populations have the highest means and modal ranges (23 or more scales), the Brazos population is intermediate with a mean of 23.1 and a modal range of 23-24 scales, and other populations have the lowest means (-χ= 21.8-22.9) and usually 23 or fewer scales.

The Guadalupe population also has the highest number of caudal-peduncle scales (-χ= 27.5), with the Colorado population having slightly fewer (-χ= 27.2). Both populations usually have 26-28 scales. Other populations have means ranging from 24.8 in the Neches to 26.2 in the Trinity, and modes of 25-27 (Table 5).

The highest mean number of anal rays (9.7) is found in the Guadalupe population where 65% of the individuals have 10 rays; other populations have means of 9.1 to 9.4 and usually 9 rays (60-78%) (Table 6). The number of pectoral rays varies little among drainages (Table 7), although the Guadalupe population has a higher frequency of 13 rays (26% of the individuals) than do other populations (0-16%). Dorsal spine (-χ= 12.1-12.6) and dorsal ray counts (-χ= 12.9-13.3) varied little among populations.

A row of 7-12 dusky to black oval to rectangular blotches is present along the side of the body of all specimens examined. The blotches appear longer and narrower in Guadalupe specimens, as noted by Hubbs (1954).

The spinous dorsal fin of breeding males in clear water in the San Marcos River near the town of San Marcos have a black band proximally and a distinct yellow-orange or orange band distally (Figs. 2 & 3). Two very narrow light yellow bands are visible in the second dorsal fin. Three light yellow bands of similar intensity to those observable in the second dorsal fin are present in the caudal fin but grade from wide to narrow, anteriorly to posteriorly. Anal and pelvic fins are dusky black. The body is dark although the underlying pattern of non-breeding individuals remains visible. A single male collected from murky water in the San Marcos River near Gonzales, Texas, has a yellow, rather than yellow-orange, band distally in the spinous dorsal fin but otherwise differs little in breeding color from individuals observed in the San Marco River near San Marcos. Color in specimens anesthetized with MS-222 (tricaine methane sulfonate) is exaggerated compared to that exhibited prior to treatment; males become much darker once treated with MS-222.

Three males in breeding condition were collected from the Paluxy River (Brazos River drainage) near Bluff Dale, Texas, on 2 March 2005. All have a black band proximally and a clear band distally in the spinous dorsal fin; one specimen has a tinge of yellow in the otherwise clear band. All three specimens are suffused with dusky pigment. One is considerably darker than the others; the underlying non-breeding pattern is nearly obscured, particularly in the area of the caudal peduncle.

Discussion

Percina sciera apristis was described as a subspecies based on the hypothesis that the population in the Guadalupe River system was morphologically distinct from other populations and that populations geographically closest to the Guadalupe River were intermediate in traits, i.e. were intergrades between P. s. sciera and P. s. apristis . Our expanded dataset confirms the decision to recognize the population in the Guadalupe River as taxonomically distinct. However, western Gulf populations closest to the Guadalupe River do not show a consistent pattern of intermediacy in meristic variables, which we consider requisite for continuing to treat apristis as a subspecies. The Guadalupe population of P. sciera has extreme numbers of preopercular serrae, pored lateral-line scales, modified scales, caudal-peduncle scales, and anal rays. The most extreme trait is the number of preopercular serrae; 95% of all individuals from the Guadalupe, but only 6% from other western Gulf drainages, have 0-3 serrae. Sixty-eight percent of Guadalupe individuals, but only 4% from elsewhere, have no serrae. These extreme values clearly demonstrate that the Guadalupe population is genetically isolated.

Variation in counts in western Gulf Slope populations shows one of four patterns: (1) The Guadalupe population is separable from other populations without intermediate counts; this is shown in numbers of pored lateral-line scales and anal rays. (2) Individuals from the Colorado River are intermediate between those from the Guadalupe and other drainages; this is shown in numbers of modified midbelly scales and caudal-peduncle scales. (3) The population in the Brazos River drainage is intermediate between those in the Guadalupe/ Colorado rivers and other populations in numbers of transverse scales. (4) Individuals from all other populations on the western Gulf Slope are intermediate between those of the Guadalupe and Colorado/Brazos in numbers of preopercular serrae.

The first pattern suggests that no population can be recognized as intergrades; the Guadalupe population is isolated and diagnosable from all other populations without intermediates. The second and third patterns suggest that the Colorado and Brazos rivers harbor populations that are intermediate between those in the Guadalupe and elsewhere on the Gulf Slope, and that the Colorado and Brazos populations should be regarded as intergrades. The intermediacy could be due to selection pressures that are intermediate in the Colorado and Brazos to those elsewhere on the western Gulf Slope, or to gene flow between the Colorado and the Guadalupe and between the Colorado and Brazos and other drainages to the north. It also could be the result of genetic drift producing different meristic frequencies in the various drainages. The fourth pattern, in which the Colorado and Brazos populations are the most different from the Guadalupe population, suggests the same as the first pattern, that the Guadalupe population is genetically isolated from nearby streams and is diagnosable from all other populations. It makes little sense to invoke gene flow between distant, and not nearest, neighbors.

Breeding males of P. sciera show a great deal of variation in color depending on environmental cues such as the presence of other males or females, but generally are much darker overall than non-breeding individuals (E.g., Ross 2001, p. 512), and the midlateral row of blotches often is obliterated (Page & Smith 1970, Page 1983). Breeding males from the Guadalupe River system darken but not nearly to the degree seen in other populations, and the midlateral blotches remain apparent (Figs. 2 & 3). Individuals from the San Marcos River, observed in the wild and kept in captivity, courted and spawned in the same manner (Fig. 3) as observed for P. sciera from the Wabash River drainage of Illinois (pers. obs.).

Although the evolution of biological diversity is assumed to be driven by mutation, natural selection, and genetic drift, we rarely understand the relative contribution of each (Hey 2001). The forces driving patterns of variation in morphological traits in western Gulf Slope P. sciera are not easily recognized, especially in view of confounding trends.

Of particular interest in evaluating the taxonomic status of the Guadalupe population is the degree of isolation. Knapp (1953) was the first to attempt to explain the distribution of Texas freshwater fishes in terms of geologic and climatic history. He contended that during the Pleistocene, the major drainage patterns of the region were not greatly different than they are today, but that there existed a much-better developed system of tributaries among which stream piracy was common. To explain what he viewed as a high degree of endemism in southwestern drainages, Knapp (1953) speculated that the Brazos River, in more arid times, with its highly turbid conditions and higher ionic concentrations, grew to represent a barrier to the westward expansion of eastern freshwater fishes.

Conner and Suttkus (1986) concurred with the premise that the Brazos is a barrier to dispersal, but considered other drainage divides also to be important. Calculation of distinction indices for western Gulf Slope drainages revealed that the divides between the Trinity and Brazos rivers, between the Brazos and the Colorado, and between the Colorado and the Guadalupe were significant barriers to dispersal. Conner and Suttkus (1986) considered the divide between the Brazos and the Colorado to be the most significant with an abrupt faunal change between the rivers dividing the region into two broad groups of drainages. The Colorado and Brazos rivers, distinctive in their faunas, constitute poor fits with western and eastern drainages (Conner & Suttkus 1986).

The findings of Conner and Suttkus (1986) suggest a historical explanation for the patterns presented by the results of the current study. There is more isolation between southern, than between northern, river faunas on the western Gulf Slope. Morphological gaps were found between populations in the Brazos and rivers to the north, between the Colorado and the Brazos, and between the Guadalupe and the Colorado.

We consider the definition of subspecies to be: a diagnosable population of a species separated from other subspecies by a geographically and genetically intermediate population. The intermediacy may be caused by intermediate selection pressures or gene flow from neighboring populations. To continue to treat the Guadalupe River population as a subspecies of P. sciera would require the Colorado River population (or Colorado and Brazos populations) to be intermediate between populations in the Guadalupe and rivers to the north. Intermediacy is expressed in the Colorado River in numbers of modified midbelly scales and caudal-peduncle scales. However, the intermediacy in numbers of transverse scales in the Brazos River suggest a different taxonomic scenario, that individuals in the Brazos be recognized as intergrades between the Guadalupe/Colorado river populations and populations to the north. Finally, in numbers of preopercular serrae, individuals from all other populations on the western Gulf Slope are intermediate between those of the Guadalupe and the Colorado /Brazos. The conflicting patterns negate recognition of any population as consisting of intergrades and suggest instead that the uniqueness of the Guadalupe River population be expressed by treating the population as a species, Percina apristis . Although there may be occasional gene flow between the Guadalupe and rivers to the north, meristic data suggest a strong degree of isolation and a unique evolutionary history for the Guadalupe population.

Published descriptions of breeding colors of male P. sciera are limited and contradictory. Some authors (Kuehne & Barbour 1983, Etnier & Starnes 1993) have suggested that a pale orange or yellow-orange band in the spinous dorsal fin of the male is diagnostic of P. s. apristis , other authors recorded the band in various populations of P. s. sciera (Ross 2001; Mettee et al. 1996; Smith 1979; Jenkins & Burkhead 1994, Robison & Buchanan 1988), and Page (1983) considered the band to be a general characteristic of P. sciera . Presence of a yellow or orange band in the spinous dorsal fin appears to be highly variable geographically and not useful for defining taxa. However, overall breeding male body color does appear to be diagnostic for P. apristis . Males of P. apristis darken slightly during the breeding season but do not become dusky black overall as does P. sciera (e.g., Ross 2001, p. 512), and the midlateral blotches remain apparent (Figs. 2 & 3).