Résumé
Conventional “spaghetti” tags and tags originally designed for sport fishing, called Betyp tags, were used during a tuna tagging program conducted on board Dakar bait boats in 1999. With the aim of quantifying a difference in tag losses between the conventional and the Betyp tags, we integrated the tag type into the tag-attrition model used to fit the recoveries per unit of time. To account for possible over-dispersion in mark-recapture data, the model selection was based on a quasi-likelihood information criterion. The main findings indicate that the immediate mortality on bigeye tuna due to the Betyp tags is around 23%. In contrast, because an additional type-2 tag losses effect was not supported by the data, we cannot conclude there is a difference in continuous tag shedding and continuous mortality between the two types of tags. The attrition rate for juveniles of bigeye tuna (average FL = 57.7
cm) provided by this model (0.58 per year) is consistent with previous estimates of natural mortality rate. One interest of combining different data sets, which have processes in common, is to provide information for the shared parameters, increasing the precision of their estimates. In addition, advantages and disadvantages of each type of tags are discussed. Based on these results and the difficulties to implement Betyp tags, we highlight the fact that Betyp tags are not well suited for intensive tropical tuna tagging programs.