A strict quality assurance/quality control protocol (QA/QC) was followed by SAIC and 

 MEC to insure a degree of sorting thoroughness and efficiency that resulted in >95 percent 

 effectiveness. A minimum 10 percent re-sort was conducted on every sample. A statistical 

 procedure, which sets an upper limit to the number of organisms that can be round in a re-sort 

 of a fraction of the sample, was used to determine whether a sample had passed the 

 QA/QC check. The pass criterion was that, at the 95 percent probability level, 95 percent 

 of the organisms in a sample were removed during the original sort. In the first fraction re- 

 sort, 10 percent of the original sample was re-sorted and checked for missed organisms. A 

 sample passed if the number of organisms found did not exceed a predetermined number (this 

 predetermined number is part of a specially developed statistical program designed for this 

 purpose). If the sample failed, a second 10 percent of the sample was re-sorted, and a third 

 if needed. Failure of the third 10 percent necessitated complete re-sorting of the entire 

 sample. The complexity of the present samples, which contained large quantities of byssal 

 threads, amphipod tubes, polychaete tubes, and extensive shell hash, resulted in several 

 complete sample re-sorts. Sample re-sorts were checked daily by the laboratory supervisor. 



Following satisfactory completion of QA/QC procedures, the specimens from each 

 sample were distributed to taxonomists. Specialists in the taxonomy of each phyletic group 

 enumerated and identified specimens to the lowest practical taxonomic level. To insure 

 taxonomic consistency and provide QA, 10 percent of all species identifications were checked 

 against reference collection organisms. In addition, 1 percent of the identified specimens 

 were sent out, unlabeled, for taxonomic QA checks. Taxonomic data were entered directly 

 onto computer keypunch sheets, thus precluding transcription errors. Unique National 

 Oceanographic Data Center (NODC) codes were used to identify each taxon. 



Biomass was determined for each major taxonomic group. Specimens were placed on a 

 0.3-mm screen and aspirated for 10 seconds, then placed in tared containers and weighed to 

 0.01 g on an electronic Sartorius balance. Data were coded directly on keypunch data sheets 

 for entry into the data base. 



Data base development. The quality of the data was checked during various stages of 

 handling. Keypunch sheets were examined by laboratory supervisors prior to submission to 

 insure that data sheets were completed and all fields correctly entered. After the data 

 were entered into the computer data base, an error-checking program was used to validate 

 the data entries. All data fields were checked for alphabetic, alphanumeric, or numeric 

 characters, and for acceptable ranges and characteristics. In addition, 10 percent of all data 

 entries were checked manually by data management personnel. Additional 10 percent 

 increments were checked if errors were found, and this interactive process continued until no 

 errors were found or until the entire data set was reviewed. After the QA checks, the data 

 base was considered complete and ready for analysis. 



Data analyses. Data were organized for presentation at three different levels and reported 

 in detail in a final contractor report available from NOAA (Barnett el ah, 1987). Level 1 

 data analyses included determinations of the name and number of individuals (abundance) 

 for each species in each replicate grab, and the total number of organisms and total biomass 

 in each replicate grab {i.e., per 0.1 nvO for the major taxonomic groups. Level 2, the station 

 summary, included the means and standard deviations of the sample data from each of 

 three stations at a site. Level 3, the site summary, included the means and standard 

 deviations calculated for the three stations at each site. Levels 2 and 3 included the means 

 and standard deviations for the following parameters: 1) abundance of each species, 2) 

 biomass of each major taxonomic group, 3) total abundance and total biomass of all biota, 

 and 4) numerical proportion of the most abundant species and taxonomic groups to the total 

 abundance of all biota. Level 3 also included the following community descriptive 

 parameters: 1) the Shannon-Weiner diversity index (H), 2) Pielou's measure of equitability 

 (evenness, J), 3) dominance (D) measured as the complement of equitability (i.e., D=l-J), and 

 4) species richness (R). 



17 



