Purves et al.: Distribution, demography, and discard mortality of crabs caugtit as bycatch in the South Atlantic 



877 



mined by plotting the proportion of mature females against 

 size (carapace width) and determining the point of 50% 

 maturity (Sragg). Logistic curves of the form {proportion 

 mature = 1/(1 + expi-ricarapace width-Sm^Q))}] were used 

 to estimate 50% maturity (Sm^g) and its standard error by 

 using the nonhnear fitting function in Splus. 



Three different experiments were conducted between 

 March 2000 and April 2001 to assess crab discard survival 

 rates. During cruise Gl a number of alive and active crabs 

 from one haul were tagged through one of the lateral plates 

 of the abdomen with Hallprint plastic T-bar tags and main- 

 tained in running seawater before they were placed in pots 

 prior to the next setting. Once these pots were rehauled, the 

 vitality of the tagged crabs was assessed by using the four 

 point relative scale shown in Table 1. A control group of crabs 

 were similarly tagged and kept onboard in running seawater 

 to monitor any effect that tagging might have had on sur- 

 vival. Survival experiments conducted on cruises G2 and H 

 differed in that crabs were selected at random and included 

 individuals that were already "limp" prior to re-immersion. 

 Crabs were tagged with thin strips of masking tape around 

 their legs prior to re-immersion. To ensure that the same 

 crabs were assessed for vitality after rehauling, pots were 

 marked and sealed off to prevent any new captures. 



Estimates of the total survival that can be expected 

 after discarding were made in the following manner. By 

 observing crabs on arrival on deck we determined the pro- 

 portion of animals that arrived on deck as lively {lively^), 

 or limp (limp J, or dead. Using the data from the survival 

 experiments we set p(liuely,lively) as the probability that 

 a lively animal that is discarded will recover to a lively 

 condition (this was estimated by calculating the proportion 

 of experimentally re-immersed lively animals that were 

 recovered as lively). We defined p(limp,liuely) similarly. 

 The proportion of discarded animals that were lively and 

 would continue to be lively is lively ^xpdiuely, lively) and the 

 proportion of discarded animals that were limp but would 

 recover to a lively condition is limp ^xpdimpjively). The 

 overall survivorship, S, is then 



S = lively ^xpilivelyMvely) + limp^xpilimp, lively). 



In our experiments, some of the damage may have 

 occurred on rehauling the pots after re-immersion, a 



situation that would not normally occur once crabs have 

 been discarded. p(lively,lively) can be corrected for this by 

 adding to it the proportion of animals that were not lively 

 when first hauled (i.e. 1-livelyJ, termed the rehaul correc- 

 tion. For instance, suppose that 1% of the crabs were not 

 lively on the first hauling, and in the experiment 4% of 

 the re-immersed lively crabs were not lively on recovery. 

 The rehaul correction would indicate that 1% of the re-im- 

 mersed crabs would have been damaged anyway simply by 

 the hauling process, and that therefore the correct damage 

 rate attributed simply to the initial capture and discarding 

 would be 3%. 



Results 



Crab catch 



The majority of the bycatch comprised two species of litho- 

 did crabs (Anomura: Lithodidae), Paralomis spinosissima 

 and P. formosa. Both species have been previously reported 

 in catches around Shag Rocks and South Georgia (Otto and 

 Macintosh, 1996) and have formed a large proportion of the 

 total catch of crab (Table 2). Crab species formed 69.5% of 

 the total weight of all species caught, including toothfish, 

 and 98.2% of the total numbers of individuals caught. 



Three other species of crab were also identified during 

 the pot trials. The most abundant of these was Paralo- 

 mis anamerae (Macpherson, 1988), which like the other 

 Paralomis species, was subject to detailed biological sam- 

 pHng. 12,370 individuals of this species (1 721 kg) were 

 caught. All individuals were discarded because they were 

 smaller than the minimum size limit for the smaller of 

 the two regulated species, P. formosa. Two other species, 

 Neolithodes diomedeae and Lithodes murrayi, were also 

 caught in small numbers. 



Distribution 



Crab distribution was investigated for the areas as defined 

 in Figure 2. There were too few data from South of South 

 Georgia; therefore the analysis was restricted to data from 

 Shag Rocks, North South Georgia, and East South Georgia. 

 A few pot strings had been left for several days because 



