180 



Fishery Bulletin 97(1), 1999 



However, for a combination of a 50-mni-long and 22- 

 mm-wide internal tag (L-tag) with a gray Petersen disc 

 (external) tag (G-tag) (rows 7-10, Table 3), A(i,B.t(i))= 

 0.7347 (±0.10r2)/yr if data are pooled for both sexes, 

 with a -log-likelihood of 45.4817. For the sex-specific 

 model, ?i(i,B,t(i)) = l.l439 (±0.2534)/yr for males; 

 ?i(i,B,t(i))=0.5202 (±0.1016)/yr for females, with a (male 

 and female) combined -log-likelihood of 41.7940 

 (=14.3301+27.4639). The increase in value of the -log- 

 likelihood function for an extra parameter is statisti- 



cally significant (x'l oo()66=2x<45. 4817-41. 7940)= 

 7.3754), suggesting significant differences in tag shed- 

 ding rates between sexes for gray Petersen discs. Simi- 

 larly, for a combination of a 35-mm-long and 10-mm- 

 wide internal tag (S-tag) with a gray Petersen disc (ex- 

 ternal) tag (rows 15-18, Table 3), A(;,B,^(/;i=3.0653 

 (+0.4739)/vr if data are pooled for both sexes, with a 

 -log-likelihood of 47.9105. For the sex-specific model, 

 Xd.Bjd ))^4.5992 (±1.0705)/yr for males; k(i,B,tii>)= 

 2.3509 (±0.4955 )/yr for females, with a (male and fe- 



