7829 - Laserfiche WebLink

Home Browse Search

McELROY AND DOUGLAS-MORPHOLOGICAL VARIATION IN GILA 637 neous and represented distinct evolutionary lin- eages, although they considered the status of G. cypha somewhat ambiguous. Using a combina- tion of morphometric and meristic characters, Smith et al. (1979) were able to differentiate museum specimens of G. robusta, G. cypha, and G. elegans in multivariate space. Three studies have been carried out using nondestructive sampling methods. Douglas et al. (1989) dem- onstrated the utility of a qualitative approach to classify individual G. robusta and G. cypha col- lected from the Yampa River, as well as the difficulty in separating these specimens on the basis of eight quantitative characters scored in the field. Using a different set of morphometric characters, Kaeding et al. (1990) found good concordance between taxonomic assignments made in the field and results of quantitative analyses of G. robusta and G. cypha from the upper Colorado River at Black Rocks. Douglas (1993) applied video image acquisition tech- niques to test sexual dimorphism in G. cypha and rejected the hypothesis (Holden, 1991) that the pronounced nuchal hump in this species rep- resents a male secondary sexual characteristic. Here we extend the use of video imaging to a consideration of morphological variation within and among populations of G. robusta ro- busta (hereafter designated G. robusta) and G. cypha in the upper Colorado River basin and the Grand Canyon. Using canonical discriminant and cluster analyses, we examine patterns of variation at three levels of biological organi- zation. Specifically, we ask three questions. First, is there evidence of significant divergence among populations of G. robusta or G. cypha, and is there a geographic component to any patterns that may exist? Second, do specimens identified as G. robusta and G. cypha represent distinct mor- phologies, or does the pattern of variation rep- resent a continuum across putative species boundaries? Third, what are the phenetic re- lationships among geographic populations of G. robusta and G. cypha.; that is, do conspecific pop- ulations cluster together, or is there evidence of a locality effect on the similarity of groups? We discuss our findings in light of both conser- vation and evolutionary implications and iden- tify avenues for future research. MATERIALS AND METHODS Field sampling.-Three hundred sixty-three adult G. robusta (n = 215) and G. cypha (n = 148) were collected with hoop and trammel nets and/ or electroshocking from eight localities in the upper Colorado River basin and the Grand Can- yon between May 1991 and Oct. 1992 (Fig. 1; WYOMING Yampa River Green River\ Y -C o,ado Fiver D QR UTAH B COLORADO W C lake P-11 \Gien Cary- D- G /UWe Colorado River ARIZONA NEW MEXICO Fig. 1. Collection localities for samples of Gila ro- busta and G. cypher taken from the upper Colorado River basin. Population labels are identified in Table 1. Table 1). Specimens ranged in size from 187- 433 mm TL. Fish were sexed according to de- velopment of the urogenital papillus (Suttkus and Clemmer, 1977) and assigned to species by MED based on overall appearance, body pro- portions, fin-ray counts, and squamation (Minckley, 1973; Douglas, 1993; Douglas et al., 1989). Sagittal views of each individual were col- lected on videotape following the procedures of Douglas (1993). Briefly, specimens were an- aesthetized in MS-222 and placed against a neu- tral background with a 10-cm rule to provide scale. Background material contained a shallow depression; positioning fish in this depression minimized error associated with two-dimen- sional projection of a three-dimensional object (Schaefer, 1991). Dorsal and anal fins were spread, and several critical anatomical land- marks that are difficult to locate from video images were identified with insect pins. Fish were videotaped perpendicular to the midsagittal plane for approximately 10 sec using either a Sony CCD-V701 8mm or General Electric 9-9808 SE 16mm camcorder. To balance the need for image resolution against the potential for image distortion through spherical aberra- tion, the working distance of the camera was adjusted such that each specimen filled the cen- ter two-thirds of a frame. Following videore- cording, fish were placed in 19 liters of fresh river water, allowed to fully revive, and re- leased.