7829 - Laserfiche WebLink

Home Browse Search

644 COPEIA, 1995, NO. 3 have supported the view that G. robusta and G. cypha represent distinct species, though "inter- grades" have been detected in varying frequen- cies. Our results similarly argue for treating G. robusta and G. cypha as separate species; despite the presence of intermediates, the two forms were clearly distinct both in sympatry and in allopatry. In fact, the intermediate specimens in our sample could be assigned to species with confidence, suggesting that morphological dis- tinctiveness of G. robusta and G. cypha is suffi- cient to categorize individuals despite high lev- els of variability. This would not necessarily be expected if the bases for a priori assignment of intermediates were more random than real (McElroy and Kornfield, 1993). Further, this variability may not to be a ubiquitous feature of populations of Gila (Smith et al., 1979; Doug- las et al., 1989; Kaeding et al., 1990): the ma- jority of intermediates in our data set were col- lected from Cataract and Desolation canyons, and these populations may be subject to evo- lutionary forces not operating, or less impor- tant, at other localities (see below). Only one study to date has addressed the issue of intraspecific population divergence in upper basin Gila. Holden and Stalnaker (1970) argued that populations of both G. robusta and G. cypha were homogeneous morphologically. However, their analysis relied on ordinal data and, thus, provided limited resolving power (Smith et al., 1979; Douglas et al., 1989). By contrast, we found subtle but significant differences in mor- phology among all populations in both species. This structure is overlain on patterns of varia- tion at the species level, as a posteriori assign- ments of all individuals to species agreed with a priori classications despite the existence of among-population differences. Our results also indicate that an isolation by distance model of gene flow among populations cannot account for these differences (Fig. 4): the lack of matrix correlation between canonical and river mile distances among populations of both species suggests that local processes must contribute to distinctiveness of groups. The nature of these additional biological forces (e.g., hybridization, selection, gene flow) remains to be elucidated. The most intriguing aspect of our data con- cerns the relationships of populations at the ge- neric level. In particular, although G. robusta and G. cypha are in general distinct morpholog- ically both in sympatry and allopatry, the two species converge morphologically at Cataract and Desolation canyons. At these sites, a locality effect appears to contribute significantly to re- lationships among populations; by contrast, a species effect appears to dominate relationships among other populations (Fig. 7). The Cataract and Desolation population pairs are clearly dis- tinct from all other groups (and from one an- other) in all analyses. This dichotomy between species and locality effects suggests that popu- lation dynamics of G. robusta and G. cypha in Cataract and Desolation canyons may differ qualitatively from those of other populations of Gila in the upper basin. The structure of these populations may be driven by biological forces (see below) whose effects are less pronounced elsewhere. The uniqueness of the G. cypha pop- ulation at Cataract Canyon has been noted pre- viously: specimens from this area tend to be smaller, deeper bodied, and less extreme mor- phologically than are G. cypha from other lo- calities (R. Valdez, Bio/West, Inc., pers. comm.). It has been suggested that these fish represent remnants of a much larger population that in- habited the region prior to closure of Glen Can- yon Dam (Holden and Stalnaker, 1975). Time series data that might provide insight into this hypothesis are currently being analyzed to ex- amine changes in morphology over the past cen- tury U. Lynch, pers. comm.). Potential bases for the locality effect. -We postulate three explanations for the morphological con- vergence of heterospecific populations at Des- olation and Cataract canyons. First, the simi- larity of G. robusta and G. cypha at these localities may reflect extensive introgressive hybridiza- tion, either occurring naturally or as a result of human intervention. Second, sympatric popu- lations may be subject to similar selection pres- sures within each of these localities, i.e., the observed locality effect may reflect parallel local adaptation of the two species in Desolation and Cataract canyons. Third, populations of both species at these sites may retain a high propor- tion of ancestral traits, and their apparent sim- ilarity may be a consequence of shared primitive characteristics. These explanations are not mu- tually exclusive. It is unlikely, however, that our results reflect an inability to consistently char- acterize species, because this would lead to poor classification of individual specimens. Although complementary genetic studies might be useful in evaluating the relative contribution of each alternative, the likelihood of each based on ex- isting data is considered below. Hybridization appears to have played a po- tentially significant role in the evolutionary dy- namics of the G. robusta group (DeMarais et al., 1992; Dowling and DeMarais, 1993). A hybrid origin for G. seminuda, long suspected (Ellis, 1914; Miller, 1946; Smith et al., 1979), has been substantiated by comparison of morphological,