substrate porosity, configuration of adjoining banks, bank and instream 

 cover, number of channels, pool-riffle frequency, depth, velocity, and 

 wetted perimeters at different flow levels. Additional habitat alterations 

 were noted where appropriate, such as excessive siltation, auf e i s formation 

 and creation of new aquatic habitats. Much of the analysis was subjective 

 because many habitat parameters were difficult to quantify, consequently, 

 the analysis was kept conservative. The results of hydraulic analysis, as 

 described in the EFFECTS OF GRAVEL REMOVAL ON RIVER HYDROLOGY AND 

 HYDRAULICS, allowed for a certain amount of habitat parameter quantification 

 and these results supported the subjective evaluations whenever comparisons 

 were available, indicating that subjectivity was not a major source of 

 error. 



Analysis of changes in fish populations was accomplished by evalu- 

 ating the types of habitat alterations occurring in the mined area relative 

 to the upstream area. Then the measured parameters that appeared to be 

 most important at the particular site were examined to determine if there 

 had been a change in fish distribution, as indicated by a difference in 

 catch rate between the upstream and mined areas. In this manner the combi- 

 nations of habitat alteration could be evaluated for their cumulative effect 

 on the population of fish present during the site visit. Additional effects 

 were postulated based on known life history requirements of the various 

 spec i es. 



The large number of benthic sample replicates obtained at each study 

 site during the field surveys allowed for an analysis of variance to de- 

 termine if significant differences existed in the densities among sample 

 areas within a study reach. All Surber sample data were computer coded and 

 the densities were subjected to an analysis of variance and multiple classi- 

 fication analysis (Nie et al. 1975). A nonpar ameter i c procedure, the Mann- 

 Whitney U-test (Zar 1974), was also used to evaluate differences in density. 

 The results of the two tests were compared and, where the results of the two 

 tests differed, the more conservative nonparametr i c test was used. Addi- 

 tional computer analysis included the calculation of various indices of 

 diversity and similarity, such as the Bray-Curtis and Raabe similarity 



146 



