Stevens Survival of king and tanner crabs captured incidentally in the Bering Sea 



741 



crabs (4.713), and injured crabs (0.717). The likelihood 

 ratio chi-sciuared value, calculated over all categories, 

 was 2.30 with p = 0.317, indicating a good fit between 

 the observed and expected survival odds. 



In order to compare different methods of crab vitality 

 assessment as tools for predicting future survival, cells 

 of Table 5A were combined. Table 5B shows that using 

 only vitality codes as an indicator of future survival 

 resulted in survival estimates of 0.925 for VIT = 1, and 

 0.347 for VIT = 2, identical to those shown in Table 2. 

 Simulation of the NMFS observer program procedure 

 resulted in survival estimates which were slightly 

 higher for both categories (0.932 and 0.365, respective- 

 ly), but otherwise very similar. Use of only data on 

 presence or absence of injuries did not provide ade- 

 quate information about future survival. 



Loglinear analysis was also conducted using imme- 

 diate mortality (IMORT) as the dependent variable and 

 the interactions of IMORT with body injuries (BODY) 

 and leg injuries (LEGS) as the independent variables 

 (Table 6A). The parameter for mean survival odds was 

 0.966, or about 1:1. Absence of body injuries increased 

 the odds of survival by a factor of 1.407; their presence 

 decreased survival odds by 0.711. Absence of leg in- 

 juries increased survival odds by a factor of 1.290. Coef- 

 ficients for IMORT X BODY and IMORT x LEGS were 

 not significantly different. Expected odds differed 

 slightly but not significantly (x" = 3.084, df = 1, p = 

 0.08) from observed odds (Table 6B). 



Evidence of recent leg autotomy was not significant- 

 ly associated with immediate mortality of king crabs 

 (X- = 2.48; Table 7A). However, leg injuries which oc- 

 curred distal to the autotomy plane were significantly 

 (X- = 34.86) associated with IMORT, as were those 

 which occurred proximal to it (X- = 88.06). Proximal 

 injuries had the greatest effect, increasing deaths 

 above the expected proportion by a factor of 5.47. All 

 three types of injuries were significantly associated 

 with IMORT for Tanner crabs (Table 7B). Immediate 

 survival odds, expressed as the ratio of live/dead crabs, 

 decreased significantly with the number of legs injured 

 for both king (X- = 86.5) and Tanner crabs (X- = 

 225.3; Fig. 8A). Delayed survival odds also decreased 

 significantly with the number of injured legs for king 

 crabs (X- = 17.0), but not Tanner crabs (X- = 4.8), 

 perhaps due to inadequate sample size for the latter. 



Discussion 



The results of this study show that, although the ma- 

 jority of crabs died as a result of capture incidental to 

 commercial trawling operations, a significant propor- 

 tion (>20%) can survive the process. The major fac- 

 tors associated with increased crab mortality were shell 



A IMMEDIATE EFFECTS 



T3 



ni 

 « 

 Q 

 It 



O 



LEGS INJURED 

 B DELAYED EFFECTS 



T3 

 n] 

 V 

 Q 

 It 



I 



O 

 cr 



65 



X Value 



KINQ CRABS 



17.0 •• 



4.78 nS ^ TANNER CRABS 



52'1 



I 



84 7 



LEGS INJURED 



Figure 8 



Effects of multiple leg injuries on survival odds of l<ing and Tanner 

 crabs. *** indicates X" value significant at /)<(). 001. Sample size 

 indicated above bars. 



condition and time in captivity (CAPTIME) prior to 

 recovery. The present study was conducted in August, 

 when both king and Tanner crabs are typically hard- 

 shelled. However, molting for both species usually oc- 

 curs from January through June, coincident with the 

 season in which most sole trawling occurs in the EBS, 

 so mortality due to molting would probably be greater 

 during the normal fishery. Processing times observed 

 in this study ranged from 0.8 to 12.5 hours, which may 

 have been longer than normal due to the sampling 

 operation, so they probably caused higher than normal 

 mortality. If average processing times for the commer- 

 cial fishery are less than those observed here, expected 

 mortality would also be less. The study conditions also 

 differed from normal operations in that crabs were re- 

 tained for several hours after recovery for examina- 

 tion, whereas those recovered during normal opera- 

 tions would have been placed on discard conveyors and 

 returned to sea sooner, probably increasing their sur- 

 vival rates. 



