Initial identification of major species associations was based 

 on subjective judgment developed during reconnaissance and permanent 

 transect diving. These preliminary identifications were corroborated 

 by computer analysis of the field data. An R-mode cluster analysis 

 program (unweighted pair-group arithmetic average clustering method 

 (UPGMA) as described by Sneath and Sokal, 1973) was used. Input data 

 consisted of presence-absence designations for all species encoun- 

 tered in each 1 -square meter quadrat from the east and north sides 

 for the sunmier (August) and fall (November) seasonal surveys. 



The program generates a matrix of similarity for all species. 

 A CALCOMP plotter program was used to generate dendrograms showing 

 the aggregate hierarchical classification among species (see App. C) . 

 On the basis of this information, 13 tentative species associations 

 were identified. 



Measurements were made to the boundaries of the various species 

 associations from fixed reference points around the island. Depths 

 (referenced to MLLW) and distances were recorded at transition zones 

 or boundaries, between associations. These measurements were taken 

 along transects located at 10-meter intervals around the island (5- 

 meter intervals were used around the four comers of the island to 

 assure adequate radial coverage) . The starting point for each tran- 

 sect was the upper boundary of the barnacle-limpet zone. In plotting 

 the data, boundaries of associations were extrapolated between transect 

 lines to depict the distributions of the associations. Actual dis- 

 tances were plotted on a base chart of the island. Boundaries of the 

 talus beds, measured during the fourth subtask, were also plotted on 

 this chart. The actual distances were then trigonometrically recti- 

 fied for plan view plotting according to the methodology in Appendix 

 A, 3, 



Areas covered by each species association were determined by 

 cutting out the associations on the base chart (before trigonometric 

 rectification), weighing the pieces from each association on a 

 Mettler analytical balance to a precision of +0.001 gram, and calcu- 

 lating the percent each association represents of the total area of 

 the island bounded by the upper limit of the barnacle-limpet zone and 

 the lower limit of rock on the bottom. 



6. Quantitative Characterization of Species Associations . 



The fifth subtask involved quantitative characterization of the 

 species associations. Biomass and densities of raacrobiota around the 

 island were measured. Analysis of these data provided the rationale 

 for separating or combining associations lying adjacent to one 

 another or on different sides of the island. 



17 



