A-12, 



PS = 100(1. 0-0. 5^ 



Pia"Pib ^ = 100Smin(p.^,p.j^) 



where Pj^^ is the proportion of sample "a" composed by species 

 "i" and p^, is the same for sample for sample "b" . These 

 values were then used to assess faunal similarities between 

 depths and various locations. This method basically measures 

 the degree of faunal overlap, in terms of species composition, 

 between various samples. 



To identify patterns of community similarity and species 

 distribution we subjected the data to rotated principal com- 

 ponents analysis. Each dive was divided into 50-meter depth 

 intervals. All species seen, and their abundance, within that 

 depth interval were then combined into a composite sample. 

 After elimination of 50-meter depth intervals in which fewer 

 than 50 m2 were viewed, we retained 169 composite samples. Of 

 133 recognizable taxa possibly occurring in our study area, 30 

 were not encountered. This resulted in a data matrix of 169 

 samples and 103 taxa. To facilitate computation, while still 

 analyzing all 169 samples at the same time, we needed to reduce 

 the niomber of taxa to 60. The 43 species that were dropped 

 were taxa that were represented by fewer than 23 individuals in 

 the entire data set. This should not substantially alter the 

 results of the analysis, because generally each sample was 

 dominated by only a few species. 



