^1 ^2 



with n^ = n_ 2 df. When significant F ratios were found, indicat- 

 ing disparate variances, an approximate t test was used (Sokal and 

 Rohlf, 1969): , 



(Y^ - YJ CMi - M.) 



t' 

 s 



Summary data for all 54 species selected for seasonal analysis 

 are presented in Table B-1. For each species, this table presents 

 transformed and untransformed means, standard deviations, transformed 

 variances, transformed range data, and an indication of whether the F 

 and t tests are significant at the 95-percent confidence level. 



These values are tabulated for each of the four seasons with 

 data combined for all four sides, and for each of the four sides with 

 data combined for all four seasons. Side differences were not tested 

 for significance. 



Note that the values in Table B-1 of Appendix B for mean densities 

 for each species refer to their abundance only over the parts of the 

 island wherein the species may reasonably be expected to occur--not 

 over the entire extent of the island revetments. 



Because of the lack of data during two seasons for the west-side 

 macrophytic algae, Mytilus-Pollicipes , and barnacle- limpet zones, 

 special consideration was required for the species that occurred in 

 these zones. These included most of the algae species and the follow- 

 ing invertebrates: Anthopleura sp. , Lottia gigantea, Mytilus 

 calif ornianus , and Pisaster ochraceus. For these species, means for 

 seasons 1 and 4 were compared since data from seasons 2 and 3 were 

 questionable. A rerun of the entire analysis for all these species 

 resulted in changes from significant to nonsignificant (at the 95- 

 percent confidence level) for only four species: Laurencia pacifica, 

 Rhodoglossum affine, Lottia gigantea, and Pisaster ochraceus. No 

 species changed from nonsignificant to significant with the 

 re analysis. 



72 



