7269 - Laserfiche WebLink

Home Browse Search

TOLERANCE OF DISSOLVED SOLIDS BY ENDANGERED COLORADO RIVER FISHES 597 stituents of the saline water from oil-shale pro- cessing in the drainage (Wagenet et al. 1980). Ions were added to the water in the form of CaC12.21-1201 MgS04.7H2O, NaSO4, NaCl, and KCl (Fisher, Certified American Chemical So- ciety reagent grade). A group of fish (30 Colorado squawfish, 40 humpback chubs, or 40 bonytails) were placed in compartment 1 at the start of an experiment and allowed to migrate throughout the trough for 24 hours. Positions were recorded every 0.5 hour for the first 7 hours (to ensure that all compartments were sampled by the fish) and again at 24 hours to enable us to estimate pref- erence and avoidance. We measured TDS as conductivity with a YSI conductivity meter (Ta- ble 1) before and after each experiment because the concentration in each compartment slowly changed over time. This change was caused by aeration and movement of fish through the dis- continuous layers. Only the last measurement of TDS was used as an indication of preference. Black plastic sheeting around the apparatus re- duced external disturbance. Photoperiod was the same as for the holding tanks. Temperature in the gradient device was not controlled. The fish were tested over a 3-month period (Septem- ber-November 1981); bonytails were tested first at 16-18 C, Colorado squawfish next at 14-16 C, and humpback chubs last at 12 C. There were three replications and three con- trols (all compartments filled with city water) for each species. The fish and water were changed for each experiment. The control ex- periments were run three in a row and the data were pooled to give one control distribution for each species. Each replicate was tested for in- dependence from the pooled control distribu- tion for that species by chi-square. The treat- ment distributions were then pooled and tested along with control distributions for uniformity by the chi-square "goodness of fit" test. The TDS preference was defined as the concentra- tion of the compartment in the gradient occu- pied by the greatest number of fish (that is, the mode of the pooled treatment distribution). Avoidance was defined as the TDS concentra- tions in the gradient avoided by 95% of the pooled fish. Results Inasmuch as all three species exhibited some degree of schooling behavior, independence for the frequency distributions could not be as- sumed. Schooling tendency of the fish changed the effective sample size for chi-square tests, because strongly schooling species have a small- er effective sample size than do weakly school- ing species. Because of the lack of independence and its unknown quantitative effect on sample size, we accepted the results of the chi-square analysis only when the probability of a greater X2 was :50.005. Chi-square values for independence were sig- nificant in two out of three tests for Colorado squawfish and humpback chub, and significant in all three tests for bonytail (Table 2). These results indicated that distribution of fish in the device was affected by the presence of a TDS gradient. The data then were compared with a hypothetical uniform distribution by chi-square "goodness of fit" analysis to see if controls were uniformly distributed and treatments were grouped around a preferred TDS concentra- tion (Table 2). The degree of control uniformi- ty appeared to be inversely related to the school- ing tendency of the fish. The Colorado squawfish appeared to be the weakest schooling fish, and distribution of control fish was the closest to uniform (X2 = 12.8). The humpback chub ex- hibited the strongest schooling tendency and distribution of control fish diverged farthest from uniform (X2 = 48.0). Treatment distribu- tions for all three species had very high chi- square values indicating they diverged greatly from a uniform distribution as one would ex- pect if they were grouping around a preferred concentration. Colorado squawfish had both the lowest pre- ferred TDS concentration (560-1,150 mg/li- ter) and the lowest avoided concentration (more than 4,400 mg/liter) of the three species (Table 3). Its preference was essentially the same as the TDS concentrations found today in the Green River (212-756 mg/liter) and upper main stem of the Colorado River (285-1,330 mg/liter: USGS 1978). Concentrations avoided were higher than the TDS concentrations in most of the major tributaries of the Green and Colo- rado rivers. The TDS preference of bonytails (4,100-4,700 mg/liter) was the highest among the three species and was approximately four times greater than present TDS concentrations of the Green or Colorado rivers. The TDS avoidance of bonytails was also the highest and the fish completely avoided the lowest concen-