1Tokyo University of Marine Science and Technology, Minato, Tokyo 108-8477, 2Marine Fisheries Research and Development Center (JAMARC), Fisheries Research Agency, Yokohama, Kanagawa 220-615, Japan
Covered-codend fishing experiments were conducted to assess the selectivity of a 33-mm diamond mesh codend with square mesh window in offshore pair trawlers. Six types of square mesh window were tested: two mesh sizes of 72 mm and 81 mm mesh openings; and three window positions of top, both-sides, and top & both-sides attached at the anterior half of the codend. Logistic parameters to express the selection curve of each haul for blackthroat seaperch Doederleinia berycoides and white-spotted conger Conger myriaster were estimated by the maximum likelihood method with the SELECT process considering sampling fraction. There were large variations in selectivity parameters between hauls in each type of codend, but no effect of catch size in the codend on the selectivity. The square mesh window had a larger 50% selection length in mean selection curves than the 33-mm diamond mesh codend. The top & both-sides window was most effective to allow fish to escape. For white-spotted conger, the both-sides window was equivalent to the top & both-sides window in selectivity. For blackthroat seaperch, the 50% selection length was larger in the top window than in the both-sides one. The effectiveness of the square mesh window to reduce the bycatch of young blackthroat seaperch while retaining marketable white-spotted conger was also discussed.