in our study, but Schoener (1982) suggests it is related to the higher 

 precision of total community abundances versus those for individual 

 species. The similarity in niches of the species (e.g. Corophid and 

 Caprellid amphipods, barnacles, and sponges) collected on exposure 

 panels supports her theory. The availability of freespace on panels is 

 an important factor in colonization (Osman 1977) and how quickly this is 

 occupied eventually determines the abundances of individual species from 

 one replicate to the next. 



Although trumpet diagrams describe the tightening of the confidence 

 interval about the mean, they do not indicate how the estimate of the 

 mean changes as the number of replicates is increased. Two species 

 common between replicate panels and exhibiting large differences in 

 abundances were selected from communities collected on asbestos panels 

 at Giants Neck and Effluent in August and May to illustrate how the mean 

 fluctuates as the number of replicates is increased (Figs. 13 and 14). 

 Similarly, the fluctuation in means were also compared for woodborers 

 common to each station (Fig. 15). Even though there were fluctuations 

 in the value of the means as the number of replicates was increased, 

 significant changes (based on overlap of the 95% CI) did not occur in 

 this estimate in any of the eight species considered. In fact, the 

 cumulative mean of four replicates closely approximates that of six. A 

 similar case might be made for three replicates, but the effects of 

 patchiness on species of low abundance make four replicates a better 

 c ho ic e . 



Adequacy of data base 



Three years of qualitative and quantitative data have been collected 

 using the six replicate and six month exposure period method and provide 



27 



