Laserfiche WebLink
ratios for treatment effects were calculated using the mean square of the block by <br />treatment interaction as the denominator. Normal probability plots showed that the <br />treatment means met the assumption of normality. <br />We also wanted to evaluate the effects of predation on the diversity of <br />copepods and chironomids. The number of samples processed was not always the <br />same, so we used the rarefaction method to obtain the expected number of cyclopoid <br />copepod species and chironomid genera in a sample size of 30. Rarefaction <br />estimates the number of species expected in a random sample of individuals taken <br />from a collection (see Krebs 1989 for methodology). <br />Next, we grouped all taxa into one of four trophic categories: collector/gatherers, <br />herbivores, omnivores, and invertebrate predators. A multivariate analysis of variance <br />(MANOVA) and univariate ANOVA tests were then performed on the log transformed <br />data for both benthic and planktonic organism counts (from weeks 1, 3, and 5) using <br />the same model as in the weighted ANOVA tests. This was to test the effect of the <br />treatments on trophic structure of the whole invertebrate community and on the density <br />of specific trophic levels within the community. <br />Finally, mean densities for control treatments, with 95% confidence intervals, <br />were calculated for the major benthic invertebrate taxa collected on each of the <br />sampling dates. The average densities, with 95% confidence intervals, were <br />calculated for closed and perforated treatments (weeks 1, 3, and 5) for those taxa <br />which had significant p-values (a = 0.05) in weighted ANOVA tests. Means for <br />plankton densities were calculated on log(x+1) transformed count data and then <br />7