792 



Fishery Bulletin 101(4) 



ranged from 3 to 3.5 hours (Table 1), and depths ranged 

 from 29 to 41 m. 



Sample collection 



On the research vessel, catches were spilled from the 

 codend onto the flat deck (equivalent to the sorting tray 

 on commercial vessels). The entire catch of each trawl was 

 progressively partitioned by shovelling the catch into con- 

 secutively numbered boxes (subsample replicates), each of 

 about 10 kg (according to the methods described by Heales 

 et al., 2000). Partitioning of the catch began at the outer 

 edges and continued in a clockwise direction, and sub- 

 samples were taken from each of four major compass direc- 

 tions: north, east, south, and west until the entire catch had 

 been collected in successively numbered boxes. The direc- 

 tion of the ship's bow was always designated as north. 



Samples on the research vessel were sorted immediately. 

 Fish and invertebrates were identified to the lowest taxo- 

 nomic level possible (mostly species). Where this was not 

 possible, the data were grouped to genus and in a few cases 

 to family. Total numbers and weights were recorded for 

 each species in each subsample and entered directly into a 

 relational database. 



On the commercial vessels, the catches were spilled onto 

 the sorting tray and the commercial-size prawns were 

 removed. The bycatch would then normally move down 



a trash chute and spill overboard. However, to sample a 

 catch, the trash chute was diverted so that all the bycatch 

 was collected in consecutively numbered boxes (subsample 

 replicates) each of approximately 10 kg. 



All samples collected from commercial vessels were 

 frozen on board and transported to the laboratory for sub- 

 sequent sorting, identification, and data entry according to 

 the methods described above. 



Although most bycatch species were identified to spe- 

 cies level, some could only be identified to genus and in a 

 few cases to family. In order to be consistent in terms used 

 throughout this study, we use the term species (plural 

 form) even when referring to multispecies groups. 



The methods used to collect subsamples on both the re- 

 search and commercial vessels differed only in the position 

 from which subsamples were taken (see earlier "Material 

 and methods" section). However, a previously published 

 study (Heales et al., 2000) showed that the majority of 

 the "abundant" bycatch species were evenly distributed 

 throughout the catch. Consequently, for all analyses we 

 combined the 14 catches collected from the research vessel 

 with the six catches from commercial vessels. 



Data analysis 



Abundance groupings Within each catch, there was a large 

 range of species (both fish and invertebrates) and they 



