286 



Fishery Bulletin 94(2). 1996 



mately 50% of the rockfish were distributed above 

 the 1.8-m height sampled by the net. The similar 

 density estimates imply that rockfish dove to within 

 1.8 m of the seafloor or were herded horizontally into 

 the net opening by the bridles and doors. Herding of 

 fish by trawl doors and bridles has been documented 

 for several semidemersal species by using acoustic 

 tags (Harden-Jones et al., 1977), direct observations 

 from a towed submersible (Main and Sangster, 1981 ), 

 and underwater video (Glass and Wardle, 1989). For 

 Pacific ocean perch, herding and diving in response 

 to trawl gear was indicated in studies using sonar 

 and variable bridle lengths. Sonar recordings of Pa- 

 cific ocean perch indicated that the fish were diving 

 as a fishing vessel passed over them (Kieser et al., 

 1992) and were diving in response to trawl gear 

 (Nunnallee 3 ). Herding was indicated in two studies 

 where increased bridle lengths resulted in increased 

 catch rates of Pacific ocean perch: catch rates in- 

 creased 2.79- and 3.24-fold when 36.6-m bridles were 

 increased 27.4 m and 54.9 m (Harling and Daven- 

 port, 1977), and catch rates were about 2.5 times 



3 Nunnallee, E. P. 1991. An investigation of the avoidance re- 

 actions of Pacific whiting iMerluccius productus) to demersal 

 and midwater trawl gear. ICES Council Meeting 1991, paper/ 

 B:5, Sess. U., Fish Capture Committee, 17 p. 



higher for a net with 33% longer bridles, even though 

 its vertical and horizontal opening was smaller 

 (Wilkins and Weinberg, 1987). 



This study demonstrates that estimates of fish 

 density are feasible from a submersible and that 

 these estimates can be used to determine bottom- 

 trawl catchability coefficients. Rockfish were ideal 

 for counting within the illuminated area because they 

 were brightly colored, solitary or loosely grouped, not 

 hidden by rugged habitat, usually motionless, and 

 generally distributed within 7 m of the seafloor 

 (Krieger, 1993). Most rockfish remained oriented into 

 the current while the submersible passed, indicat- 

 ing minimal response to the submersible. The only 

 movements observed were those of individual fish 

 moving out of the direct path of the submersible and 

 those of a few groups moving toward the submers- 

 ible. Counts offish within these groups were not hin- 

 dered, because the fish swam slowly and maintained 

 their spacing and orientation. The pilot observed 

 similar behavior on all sides of the submersible. In 

 addition to rockfish, density estimates appear fea- 

 sible for numerous invertebrate species and ground- 

 fish species, including shortspine thornyhead, 

 Sebastolobus alascanus, and several flatfish species. 



The value of comparing in situ observations of 

 groundfish with trawl catches has been documented 



