2012 年 48 巻 3 号 p. 173-182
Covered-codend fishing experiments were conducted with using square mesh top window of four different types attached at the anterior half of the codend. Top window of three lengths: 4m, 3.5m, and 3m lengths with square mesh of 90mm nominal mesh size (81mm mesh opening) were tested, compared with 4m-length square mesh panel of 80mm nominal mesh size (72mm mesh opening). Individual haul selection curves and the mean selection curve of each codend type were estimated for blackthroat seaperch Doederleimia berycoides and for white-spotted conger Conger myriaster by maximum likelihood method with SELECT process considering sampling fraction. There was large variation in individual haul selection curve, and no effect of total catch size in codend on the selectivity parameters (50% retention length and selection range) was found. Differences in mean selection curve were found between window-lengths by the AIC model selection. The 50% retention lengths of 4m-, 3.5m- and 3m-window length were 14.1cm, 14.6, and 13.6cm in total length for blackthroat seaperch, and 28.1cm, 29.5cm, and 24.5cm in pre-anal length for white-spotted conger, respectively. Effect of window length on size selectivity of the codend was remarkable for white spotted conger, rather than for balckthroat seaperch. To reduce the bycatch of small-sized blackthroat seaperch while retaining marketable white-spotted conger as much as possible, square mesh top window of 3m length is preferable.