466 



Fishery Bulletin 104(3) 



170'E 



180 



65'N 



60 N 



Russia 



1 70 W 



H- 



55'N 



50 N 



45"N 



Bering Sea 



4D 



.4C 



160 W 



. . y 







150 W 



Alaska 



140W 



130 W 



N I 



.^i.^" Closed -rV-' S 





-^>- 



4B 



Aleutian Is- 



4B 





4A 



^Y^""^ 3B 



.--.J!^ 



4A 



Gulf of Alaska 





Queen Charlolte Is. xt -X^ *<* 



28 4 ^^ 



Vancouver Is %. J" ^^' ^ ' ^ 



2A  



60 N 



55'N 



50 N 



45N 



170°E 



180 



uo^w 



130*W 



120"W 



170'W 160W 150'W 



Figure 1 



International Pacific Halibut Commission regulatory areas. The area marked "closed" is permanently closed 

 to directed halibut fishing. 



our study was to define selectivity to be 1.0 at 120 cm. 

 Estimated selectivity could therefore exceed 1.0 at other 

 lengths. 



Results 



Figure 2 shows the estimates of commercial length-spe- 

 cific selectivity in areas 2B, 2C, 3A, and 3B obtained by 

 the method of Myers and Hoenig (1997) with the use of 

 all available data in each area. There were insufficient 

 data in area 4 to calculate useful estimates. The esti- 

 mates in Figure 2 were calculated by using all recoveries 

 from each release during the first two years at liberty, 

 including recoveries from outside the release area and 

 recoveries from unknown locations. Estimates computed 

 by using only recoveries from the area of release were 

 no different from those obtained by using all of the 

 recoveries. 



In all areas, commercial selectivity in the period of 

 1960-90 anpears to increase with length up to a maxi- 

 mum and then decline. In area 2B, the peak occurs at 

 about 110 cm and there is a substantial decline thereaf- 

 ter, to around half the peak value. In Alaska (areas 2C, 

 3A, 3B), selectivity peaks at a much larger size (about 

 150 cm). Thereafter the decline is about as steep as in 

 area 2B, but not as large because so little of the length 

 composition remains beyond 150 cm. 



Recoveries from releases at spot fishing locations 

 show a selectivity pattern similar to that for the entire 



dataset. The same is true of survey releases, except in 

 area 2B where the selectivity pattern does not show a 

 decline among larger fish. But this impression depends 

 on a small number of recoveries, and therefore it may 

 be false. 



Discussion 



In previous modeling of length-specific selectivity, the 

 IPHC staff generally assumed some kind of asymp- 

 totic function, with full selection occurring at 110-130 

 cm. A function of this form is consistent with video 

 observations of halibut behavior when they are hooked 

 (Kaimmer, 1999), and it produces satisfactory fits to the 

 observed length compositions of survey and commercial 

 setline catches in the annual stock assessment. It is 

 also consistent with the conventional view that hook 

 selectivity varies little with size among fish large enough 

 to take the bait (Lokkeborg and Bjordal, 1992). But the 

 large body of mark-recapture data shows a different 

 pattern: selectivity declining after the peak at 110 cm 

 in area 2B, and not reaching a peak until 150 cm or so 

 in Alaska. These patterns are, in fact, quite similar to 

 those reported by Myhre (1969). 



Commercial fishing selectivity reflects ground selec- 

 tion by the fleet, as well as size selection by the gear. 

 It is therefore possible that the selectivity of commer- 

 cial setline gear on a given ground has the expected 

 asymptotic form, but ground selection has the effect of 



