Potential bias in estimates of the size of maturity of crabs derived from trap samples

The size at the onset of sexual maturity (SOM) of female crustaceans is typically estimated using logistic regression analysis of the proportions of mature females in sequential size classes. The validity of this approach depends on the composition of the samples reflecting accurately that present in the environment. However, catches obtained by traps, a passive fishing method, typically contain disproportionately greater numbers of large crabs, whereas those obtained using active fishing methods, such as seine-netting and otter trawling, will presumably represent far better the size composition of the population. Moreover, we demonstrate that samples of female Portunus pelagicus caught by trapping were predominantly mature, whereas those collected by seining and trawling contained numerous immature as well as mature females. Therefore, the samples of females collected by trap are clearly biased towards mature crabs. Consequently, for any size class, it would be predicted that the proportion of mature females in trap catches will be overestimated, so shifting the logistic curve fitted to the proportions of mature crabs in each size class to the left, and yielding an underestimate of the SOM. This conclusion is substantiated by the fact that the carapace width of female Portunus pelagicus, at which 50% of individuals reach maturity (SOM50), was estimated to be markedly greater when using the proportion of mature females obtained by seine-netting and otter trawling, i.e. 101.1 mm, than by trapping, i.e. 86.1 mm. Data on the size and maturity status of the deep-sea crabs Hypothalassia acerba and Chaceon bicolor are available only from trap catches. From the above data for P. pelagicus, it is considered likely that, through a greater vulnerability of mature females of these species to capture by traps, the respective SOM50s derived for female H. acerba and C. bicolor from trap samples (carapace lengths of 69.7 and 90.5 mm) will be considerable underestimates of the true values