8029 - Laserfiche WebLink

Home Browse Search

5 appropriate use of significance tests and balancing biological and statistical significance in stating conclusions. In trial one, fifteen shortnoee sucker larvae (15-21 mm standard length) were stocked in each tank with one fathead minnow from the appropriate gape width class. Trial two was similar to trial one but used three fathead minnows per tank. No predation occurred during trials one or two (Table 1). In trial three, a wider array of prey sizes was presented to fathead minnows by adding 5 host River sucker larvae (13.5-14.0 mm) to the 15 shortenee sucker larvae remaining firm earlier trials, bringing the total number of sucker larvae to 20 per tank. Predation occurred in the medium and large gape width treatments, although no significant differences were found in predation rates among fathead minnow gape width classes (Table 1). However, the low power of the test (-0.08) precludes concluding with any certainty that predation rate was unrelated to gape width. Kolmogorov-Smirnov tests comparing length frequency distributions of larvae in individual tanks before and after the trial showed no significant differences in length frequencies, suggesting that fathead minnows were not selecting particular larval sizes. In trial four, six fathead minnows from the appropriate gape width classes and ten Lost River sucker larvae (14-18 mm) were randomly assigned to each tank. At the end of the trial larvae remained in only three of the nine tanks, and no significant differences in predation rates were found among fathead minnow gape width classes (Table 1). Again, statistical power of the ANOVA was low (-0.09). Only two tanks had large numbers of larvae remaining, and upon close inspection I noticed the standpipes (which controlled water depth) in these tanks were slightly taller than in other tanks. As a result, a ledge (a reinforcing structure for the tank) about 1 cm wide was 1 cm below the water surface, and the surviving larvae were clearly using these ledges to avoid predators. These ledges likely influenced predation rates in these two tanks in trial three as well, because they were the only tanks in which no predation occurred in the medium and large gape width treatments. Both standpipes were shortened before subsequent trials. Trial five was similar to trial four (six fathead minnows and ten Lost River sucker larvae in each tank), but fathead minnows were provided with alternate prey in the form of live Daphnia and dry feed. Large numbers of Daphnia were concentrated in a small water volume, and 150 ml of this mixture was added to each tank, along with 1 g of dry food. At the trial's end, predation on sucker larvae had occurred in five of the nine tanks (Table 1). Again, there were no significant differences in feeding rates on sucker larvae among treatments, and power of the ANOVA was low (-0.17). Lack of an appropriate control treatment precluded statistical comparison of predation rates with and without