De Lestang et al.: Reproductive biology of Portunus pe/agicus 



747 



Sex of small crabs, i.e. with a CW < about 30 mm, was 

 determined with a dissecting microscope to ascertain 

 whether their pleopods bore setae and thus the crabs were 

 females. At CWs > about 30 mm, the female crabs could 

 readily be distinguished from male crabs by their posses- 

 sion of a far wider abdominal flap (Van Engel, 1958; Warner 

 1977). During the pubertal molt of female portunids, the 

 abdominal flap changes from a triangular to oval shape 

 and from being tightly to loosely fixed to the cephalothorax 

 (Ryan, 1967b; Fielder and Eales, 1972; Ingles and Braum, 

 1980; Fisher, 1999). 



The size and time of occurrence of all ovigerous females 

 were recorded. The ovary of each crab was assigned to one 

 of four stages by using macroscopic characters similar 

 to those described for the development of the ovaries of 

 P. pelagicus and other portunids (Ryan, 1967b; Meagher, 

 1971; Krol et al., 1992; Kumar et al.^). The assignment of 

 these stages was augmented by examining the characteris- 

 tics of a subset of 200 of these ovaries in 6-/.im histological 

 sections that had been stained with Mallory's trichrome. 

 For 5-10 ovaries of each macroscopic stage, the diameters 

 of 30 randomly selected oocytes that had been sectioned 

 through the nucleus were measured to the nearest 5 pm. 

 Two measurements (the longest diameter and shortest di- 

 ameter) for each oocyte were then averaged to provide an 

 estimate of each oocyte diameter. 



Male crabs were designated as either morphometrically 

 immature or mature by using differences in the regression 

 equations for the relationships between the natural loga- 

 rithms of the length of the dorsal propodus of their largest 

 cheliped and carapace width in what were clearly either 

 juvenile (small and gonadally immature) or adult crabs 

 (large and gonadally mature). For full description of the 

 method see Somerton (1980). 



On the basis of their macroscopic appearance, the vas 

 deferentia of each male crab were assigned to one of three 

 stages by using criteria derived from the description of go- 

 nadal development for P. pelagicus by Meagher ( 1971) and 

 for P. sanguinolentus by Ryan (1967a). Aquarium studies 

 by Meagher (1971) showed that male crabs with gonads at 

 stages I and II did not copulate and are thus considered im- 

 mature, whereas those with gonads at stage III copulated 

 successfully with females and thus have mature gonads. 



Ovaries and vas deferentia from a wide size range of 

 at least 20 females and 20 males, respectively, from each 

 sampling occasion in each of the five bodies of water were 

 weighed to the nearest 0.01 g. The mean gonad weight 

 at a constant carapace width for each sex in each month 

 in each water body was determined by using analysis of 

 covariance (ANCOVA) of the natural logarithm of the go- 

 nad weight as the dependent variable, month as a fixed 

 factor, and the natural logarithm of the carapace width 

 as a covariate. The common constant carapace width of 



'^ Kumar, M., Y. Xiao, H. Williams, G. Ferguson, G. Hooper, and S. 

 Venema. 1999. Blue crab fishery. South Australian Fisher- 

 ies Assessment Series. 99/02, 64 p. South Australian Research 

 Development Institute, Grenfell Centre Level 14. 2.5 Grenfell 

 Street Adelaide 5000, Australia. 



crabs in all bodies of water was a default value calculated 

 by the ANCOVA. 



Size frequency and reproductive data for the corre- 

 sponding months in the different years in each water body 

 were pooled for describing intra-annual trends in these 

 variables. 



Size at maturity 



The percentages of female crabs of different carapace 

 widths which, in each water body, had undergone a pubertal 

 molt, were subjected to logistic regression to determine the 

 size at which 50% of the female crabs would have become 

 mature sensu Hartnoll (1974). Data for each assemblage 

 were randomly resampled and analyzed to create 1000 

 sets of bootstrap estimates of the parameters of the logis- 

 tic regression and estimates of the probability of maturity 

 within the range of recorded carapace widths. The 95% con- 

 fidence intervals of the CWgg's were derived by using this 

 resampling technique, which produced slightly more con- 

 servative estimates than those obtained from the Hessian 

 matrix of the logistic regression and thus reflected better 

 the uncertainty of the parameter that was associated with 

 the data. The 95% confidence intervals of the probability 

 of maturity at each specified carapace width were taken 

 as the 2.5 and 97.5 percentiles of the corresponding pre- 

 dicted values resulting from this resampling analysis. The 

 point estimate of each parameter and of each probability 

 of maturity at the specified carapace width were taken as 

 the medians of the bootstrap estimates. 



The percentages of mature male crabs at different cara- 

 pace widths in each of the five bodies of water, with matu- 

 rity being assigned by using firstly morphometric and then 

 gonadal criteria (see earlier), were subjected to logistic re- 

 gressions to determine the CWgg's for these variables. The 

 percentages of male crabs in Cockburn Sound and Shark 

 Bay, which possessed an abdomen that was loosely fixed 

 to the cephalothorax, were likewise subjected to logistic 

 regression analysis. The logistic regressions relating ma- 

 turity and carapace width for both the females and males 

 in the different assemblages were compared by using a 

 likelihood ratio test, as described by Cerrato (1990) and 

 emplo3ring a Bonferroni correction. 



Fecundity 



The total wet weight of eggs in each batch of eggs of 40 

 early-stage ovigerous females, i.e. with yellow eggs, from 

 Cockburn Sound and which covered a wide size range, 

 was weighed to the nearest 0.001 g. The number of eggs 

 in each of four replicate subsamples from each batch 

 were recorded, after which each of those subsamples was 

 weighed to the nearest 0.001 g. These data were then used 

 to estimate the total number of eggs in each batch of eggs 

 of each female. The relationship between batch fecundity 

 (BF) and carapace width (CW) was described by using the 

 equation \nBFj=m\nCWj+b. 



The number of batches of eggs produced by a full size 

 range of mature females during the spawning period 

 was estimated by determining the spawning period (SP), 



