7856 - Laserfiche WebLink

Home Browse Search

96 GREAT BASIN NATURALIST [Volume 55 were to (1) describe diets of YOY Colorado squawfish and other small fish in backwaters of the Green River and (2) examine diet overlap and potential for competition with Colorado squawfish. METHODS Samples of small fish were provided by the U.S. Fish and Wildlife Service Field Station at Vernal, UT. These were collected from back- waters of the Green River during summer (30 June-27 August) and autumn (22 September- 10 December) 1987. The study area extends from confluence of the Green and Yampa rivers in Echo Park, Dinosaur National Monu- ment, CO, to Turks Head in Canyonlands National Park, UT-river kilometer (RK) 555 to 35 above confluence with the Colorado River. Upper and lower reaches are divided at Sand Wash (RK 346), UT, a convenient access point just above Desolation Canyon. Each river reach began with a rocky, high-gradient (1.3-2.1 m/km) segment and continued with a sand- and silt-laden, low-gradient (0.2-0.4 m/km) segment known for relatively high catches of YOY Colorado squawfish (Haines and Tyus 1990, Tyus and Haines 1991). The river was further divided into 8-km sections starting from a random location within each reach to help assure an even distribution of collection sites. Backwaters were defined as shallow (typi- cally <0.5 in maximum depth), ephemeral embayments with negligible water velocity. Substrates consisted of silt and sand or silt and mud, sometimes overlaying or interspersed with gravel or cobble. Backwaters had little or no rooted aquatic vegetation, but some had dense mats of algae. Two backwaters were sampled weekly in each 8-km section during daylight (1000-1800 h) using 1-m2 seines (0.8- mm2 mesh) in summer and 1-m X 3-m seines (3.2-mm X 4.8-mm mesh) in autumn. Fish were killed and fixed in 10% formalin immedi- ately after collection. Up to five specimens <- 20 mm total length (TL) and five >20 mm TL of each fish species, representing graded size series, were selected from each sample. Each digestive tract (from esophagus to vent) was removed, opened, and visually assessed for percent fullness. Food items were identified to lowest practical taxon, and a visual estimate was made of percentage contributed by each taxon to total volume of food in each digestive tract (Larimore 1957, Mathur 1977). For diet analyses, food-item taxa (total of 124) were grouped into 20 family, order, or broader-based categories, sometimes divided according to habitat (e.g., aquatic or terrestrial). Data for each fish species were stratified according to length (10-mm TL or larger inter- vals) by season (summer or autumn) within river reach (upper or lower). Only subsets with at least six fish containing food were included in analyses. Diet measures calculated for each subset were (1) mean percentage each food category contributed to total volume of food in each digestive tract (mean of volume percent- ages) and (2) percentage of all digestive tracts in which each food category occurred (per- centage of occurrence). Wallace (1981) evalu- ated several diet measures and concluded that mean of volume percentages is the best mea- sure for calculating overlap. However, per- centage of occurrence is useful for describing general variations in diet (Wallace 1981, Bowen 1983). Similarities in diet by subset between Colo- rado squawfish and other fishes were evaluat- ed by Schoener's (1970) resource-overlap index: n «= 1-0.5(Y,I Pxi-Pyi i=1 where n is the number of food categories, Pxi is the proportion of food category i (expressed as mean of volume percentages) in the diet of species x (Colorado squawfish), and Pyi is the proportion of food category i in the diet of species y (other fishes). Values range from 0.0 (no overlap) to 1.0 (complete overlap). When data on resource availability are absent, Schoener's index is one of the best indices available for calculating resource overlap (Hurlbert 1978, Linton et al. 1981, Wallace 1981). Diet overlap is useful in helping to elu- cidate food relationships among species and has been considered "biologically important" when values exceed 0.60 (Zaret and Rand 1971, Matthews and Hill 1980, Galat and Vucinich 1983). RESULTS Digestive tracts from 2554 fish represent- ing 15 species were examined for food items;