p. eivvifeva was absent; the abundance of C. setosa, anthozoa, and 



P. staminea (bivalve) was higher, and the density of the other bivalves, 



T. modesta and S. pattuxl , was lower. 



In summary. Tables A-3 and A-4 show that the 1.6-kilometer swath area 

 is the least similar of the four areas sampled. The quantitative and 

 qualitative information in Table A-2 supports this conclusion. Samples 

 from the three smaller areas which were closer together are similar. 

 Thus, the fauna appears to be homogeneous on a smaller scale within the 

 20-meter depth contour. Since the permanent stations were restricted to 

 an equally small area, stratification of sampling was not necessary. 



V. SUMMARY 



1. Diver-operated coffee can corers (area, 0.018 square meter; pene- 

 tration, 15 centimeters] permit good replication and precise control of the 

 area sampled. 



2. Eight cores sampled had 76 percent of the species present in 28 

 cores. 



3. The effect of increasing sample size on the means and 95-percent 

 confidence limits was examined for various parameters. In most cases, 



the confidence limits decreased sharply from 4 to 8 cores and continued to 

 decrease gradually to the largest sample size (28 cores). 



4. Cumulative species diversity and evenness values appeared to 

 stabilize between four and eight cores. 



5. A comparison of samples from four progressively larger areas 

 indicated a general similarity among the three smaller areas; the largest 

 deviated from these. 



6. It was concluded that a sample size of eight replicate cores was 

 sufficient to estimate most of the important parameters and that the fauna 

 appeared homogeneous within the 20-meter depth contour for an area equal 

 to that of the permanent stations. 



75 



