8097 - Laserfiche WebLink

Home Browse Search

The influence of migratory spawning runs of trout was avoided by exclud- ing all streams which appeared to contain spawning trout from analysis. The assumption that the second sample is random was addressed by Seber (1973) such that if all animals have the same catchability, a random sample can be approximated if each area in the population has the same probability of being sampled. All electrofishing was done with the intention of shocking all available trout habitat on each run. In addition, the methods for col- lecting trout on the second set of electrofishing runs matched as closely as possible methods on the first. The consistent sampling methods eliminated any systematic differences in within and between week population estimates. The assumptions of no tag loss and all tags reported could have been violated and population estimates may have been biased but the bias should have been consistent in within and between week estimates, i.e., both estim- ates should have been similarly affected so that although biased, both estimates should still have been the same. The loss of jaw tags could have caused the between week estimates to be higher than within week estimates, however, on the Big Horn River, tag loss over a 6 month period (May - Novem- ber) was measured at only 3.4%. Also, batch marks were often used in place of jaw tags. Although some regeneration of batch marks was evident, this WaS easily detectable. Violation of these assumptions did not appear to account for the difference between within and between week population estim- ates. Lagler (1968) reported that small fish were generally less vulnerable to capture using electrofishing gear than were larger fish. This problem will create bias in population estimates when data for all sizes of fish are pooled. However, pooling creates a consistent bias between within and be- tween week population estimates as long as the size of fish collected does not differ between runs. On the North Platte River at Bessemer Bend, the population estimates were stratified into three size groups to eliminate some of the size selective bias and results showed between week estimates to be larger than within week estimates. Although violation of this assumption does cause biased estimates, the bias should be consistent for within and between week estimates, yielding similar estimates. Differences in the probability of capture of marked and unmarked trout appears to be responsible for most of the difference between within and be- tween week population estimates. Schreck, et al. (1976) and Whaley (1975) reported on physiological changes in rainbow following electroshock and concluded that more than six hours were required for a fish to return physio- logically to "normal" preshock conditions. Cross and Stott (1975) working with roach (Rutilus rutilus (L.)) and gudgeon (Gobio gobio (L.)) in ponds showed population estimates varied depending on the length of time between mark and recapture runs. They attributed differences to a learned avoidance behavior of fish that were shocked, but escaped during the mark run and were less catchable during the recapture run. Observations made while electrofishing small Wyoming streams where both mark and recapture runs were done several hours apart on the same day have supported the contention that marked trout swim to the electrode much less vigorously, and remain at the positive electrode for shorter periods of time than unmarked trout. 26