each collection was resieved under tapwater and all specimens from respective 

 samples were stored in 70-percent isopropanol for final sorting, taxonomic 

 determinations, and species counts. The 0.7 millimeter screen was used 

 instead of a conventional 0.5 millimeter one because the former facilitated 

 sieving operations and retained a percentage of infauna that was shown to be 

 very nearly equivalent to that sampled by the smaller mesh size. 



As in Saloman's (1976) work, biological data presented here include a 

 species checklist and individual station listings that show species occurrence 

 and frequency, together with calculations for number of individuals per square 

 meter and the Shannon-Weaver index of faunal diversity (H'). Also, as a 

 measure of relative species dominance, equitability (J') was computed for each 

 station (Pielou, 1975). Two other statistical procedures were also employed. 

 The first, Morisita's Index (Morisita, 1959; Bloom, 1981), provided a numerical 

 method of comparing faunal similarity between comparable sets of control and 

 experimental samples, and was used to develop similarity matrices and classifi- 

 cation diagrams that graphically show faunal relationships based on station data 

 for diversity and abundance. 



The second procedure, a stability analysis (Bloom, 1980), is a multivariate, 

 nonparametric statistical and geometric procedure that converts biotic data 

 from control and experimental samples into communities that can be represented 

 mathematically. For one representation all base-line and control data were 

 used to define numerical characteristics of a preconstruction community cluster 

 that has a central point, or centroid, and certain specific spatial limits. In 

 the first stability analysis, the distance from the centroid to control and 

 experimental samples was used to determine variability among samples from 

 undredged and dredged bottoms. In the second analysis, community clusters 

 calculated for experimental samples were compared to the preconstruction 

 cluster, in postconstruction sequence. When a boundary or an experimental 

 cluster met the limit of the preconstruction cluster, faunal recovery was 

 accepted. Experimental collections from station 1, where sampling over time was 

 done, were the only borrow pit samples used in this analysis. 



V. RESULTS 



1. General . 



The findings in this section are based on the detailed information given 

 in Appendixes A to F. Appendix A lists abiotic parameters by station. Appendix 

 B is a checklist of all organisms collected at offshore stations from November 

 1974 to November 1977. Appendix C contains all biological station data and 

 indices of diversity (H') and equitability (J'). Appendix D (Similarity 

 Matrices) and E (Classification Analyses and Dendrograms) are both based on 

 Morisita's index of faunal similarity. Appendix F is a graphic representation 

 of the two stability analyses. The first graph shows comparative variability 

 among control and experimental samples when compared with the centroid of a 

 community cluster calculated from all base-line and control samples. The second 



14 



