224 



Fishery Bulletin 101(2) 



Hampton and Foumier (2001) and Hampton and Fournier^ 

 have shown that growth of smaller-size fish does not con- 

 form to the von Bertalanffy function. They estimated a more 

 linear and an increased growth rate for smaller sizes (<120 

 cm FL for yellowfin and <80 cm FL for bigeye tuna) than 

 could be accounted from the von Bertalanffy function for 

 the entire size range. Although this could, in principal, bias 

 our size-based estimates of F and M, considering the growth 

 variability and the large size ranges we have considered, 

 we feel departure from von Bertalanffy growth is of little 

 importance. 



One of the primary objectives of the HTTP was to im- 

 prove understanding of the dynamics of tuna aggregations 

 at the Cross Seamount and to determine the importance of 

 Cross Seamount associated fish to domestic longline and 

 inshore fisheries. This discussion will therefore focus on the 

 Cross Seamount fishery and its potential interaction with 

 other fisheries. Previous analyses (using fewer data) have 

 estimated gross attrition rates and residence times (Hol- 

 land et al., 1999) and transfer and attrition rates (Sibert 

 et al., 2000). Using the more recent and complete data set 

 and including size specific attrition to improve the tag-at- 

 trition model, we have been able to extend the analysis to 

 provide a more detailed picture of fishery dynamics and 

 interactions. 



The natural mortality rate is a critical parameter in stock 

 assessment models, and size- (or age-) specific estimates 

 would greatly improve stock assessment efforts. Unfor- 

 tunately, natural mortality is not linked to a well-defined 

 in situ process, and M is always estimated indirectly (e.g. 

 Fournier et al.. 1998). In tag attrition models, M is the "re- 

 sidual attrition" that cannot be accounted for by processes 

 specified in the model. In our nuKiel, A/ would also include 

 permanent emigration beyond the model area. Hampton 



(2000) estimated natural mortality rates from tagging data 

 for a large number of size classes from a "single fishery" 

 model. We have shown here that the attrition component 

 can also be partitioned into size classes in a bulk transfer 

 model. The relatively low number of recoveries from most 

 of the sites did not allow us to estimate attrition over a 

 larger number of size classes. However, our estimates ofM 

 are consistent with Hamptons (2000) estimates for both 

 species within the size ranges considered. 



The relatively low transfer rate estimates for both spe- 

 cies from the Cross Seamount to the inshore areas supports 

 earlier findings (Sibert et al., 2000). However, the relatively 

 high transfer rates estimated for both species from the 

 Cross Seamount and the offshore buoys to the longline 

 fishery (and by inference to the Pacific-wide fishery) 



