Somerton and Otto Net efficiency of a survey trawl for Chionoecetes optlto and C bairdi 



621 



skimmed over the bottom at an 

 estimated height of 1-2 cm. No 

 difference in the proximity of the 

 footrope to the bottom could be 

 visually detected among the 

 three trawl configurations. Be- 

 cause the range in wing spread 

 among the three configurations 

 was greater than the difference 

 between the survey average and 

 the experiment average, it is 

 likely that the addition of the 

 auxiliary net did not alter the 

 contact of the footrope in the cen- 

 ter of the trawl where most of the 

 escapement usually occurs (Walsh, 

 1992). 



Description of the data 



Tanner crab 



Mixed sexes 



Snow crab 



Mixed sexes 



n = 3487 



Ol^ 



ISO- 



IOC 



50 



0\ 



n = 3458 



;• A 



i % 



• r\ ... 



50 



100 



150 



50 



100 



150 



J3 



E 



Mature females 



30 



20 



10 



Thirty-one hauls of the experi- 

 mental trawl were completed, 

 but the data from the first seven 

 were not used in subsequent 

 analysis because of unusual 

 trawl performance (the narrow- 

 ing of wing spread associated 

 with large catches of epibenthic 

 fauna in the auxiliary net I. For 

 snow crab, width measurements 

 were collected from 3458 indi- 

 viduals from the mixed sexes cat- 

 egory, but from only three indi- 

 viduals from the mature female 

 category. The low sample size for 

 mature females prevented further analysis. Cara- 

 pace-width frequency distributions indicated that a 

 broad range of carapace widths was sampled, but the 

 observations were concentrated in three carapace 

 width modes (Fig. 2). For Tanner crab, width mea- 

 surements were collected from 3487 individuals from 

 the mixed sexes category, and from 579 individuals 

 from the mature female category. Sampling for both 

 categories covered a broad range of carapace widths, 

 but for the mixed sexes category the observations 

 were concentrated at smaller carapace widths. Plots 

 of net efficiency against carapace width (Fig. 3) for 

 the mixed sexes groups of both species showed an 

 initial decline with increasing width until a mini- 

 mum in net efficiency was reached in the range of 

 40-50 mm, then a steady increase as carapace width 

 increased. For mature female Tanner crab, the nar- 

 row range of carapace widths and the large variance 

 in net efficiency obscured any obvious change in net 

 efficiency with carapace width (Fig. 3). 



50 



100 



150 



Carapace width (mm) 



Figure 2 



Carapace-width frequency distributions for the mixed sexes of snow crab 

 [Chionoecetes opilio) and for the mixed sexes and mature females of Tanner crab 

 iC. bairdi \. 



The fitted models 



For the mixed-sexes category of both species, the 

 models that included both the P^and P^ terms fitted 

 the data significantly better than the models with 

 only the P. term (Tanner crab, likelihood ratio= 189.7, 

 P<6.001; snow crab, likelihood ratio=60.5, P<0.001 ). 

 This indicates that the apparent increase in net effi- 

 ciency at small sizes is statistically detectable. Like- 

 wise, the separate-species model fitted the net effi- 

 ciency data significantly better than did the combined 

 species model (likelihood ratio=85.5, P<0.001). This 

 indicates that the change in net efficiency with width 

 differs between species. Plots of the fitted models 

 (Fig. 3) indicated that the predicted values of P^ de- 

 cline rapidly with increasing carapace width and be- 

 come negligible at about 50 mm for Tanner crab and 

 60 mm for snow crab. For mature female Tanner crab, 

 the model considering only the P.. term fitted the data 

 significantly better than did the sample mean (likeli- 



