FISHERY BULLETIN: VOL. 73, NO. 3 



numbers, and 5 is in kilograms. The somewhat 

 higher maximum for the selectively fished stock 

 may have resulted from its changed size composi- 

 tion; it included fewer extremely large males than 

 the unselectively fished population. 



Since selection reflected the ability of fish to 

 escape through vertical slots between glass bars, 

 thickness was the controlling dimension. Most of 

 the fish-size data in this report are therefore in 

 thickness. However, to reduce fish handling to a 

 minimum, the growth measurements of live fish 

 during the final part of the experiment were in 

 lengths. Because of this, and because other 

 biologists may wish to compare their length data 

 with my thickness data, I calculated thickness- 

 length relations. 



Measurements for the relation were from the 

 caught fish, for which both thickness and length 

 were recorded. Preliminary analysis showed that 

 data for immature and male fish could be combined 

 into a single rectilinear regression of length on 

 thickness. The regression for females was also 

 rectilinear but had a gentler slope, probably 

 because of the distention of mature fish carrying 

 eggs. It was therefore calculated separately. The 

 regression equations, numbers of pairs of obser- 

 vations in parentheses, and correlation coefficients 

 were (fitting was by least squares): 



Immature and male 



L = 7.027 T (368) r = 0.987 r^ = 0.974. 



Female 



L = 26.89 -I- 5.037 T (207) r = 0.866 

 r^ = 0.750, 



where L is length and T is thickness, both in 

 millimeters. The squared coefficients suggest that 

 97 and 75% of variations in length were associated 

 with variations in thickness. 



RESULTS 

 Exploitation and Response 



Before selective exploitation could be started, it 

 was necessary to determine the selection point. To 

 aid in this the thickness of all of the fish in both 

 populations was measured at month 33.5. At this 

 time the population to be selectively fished (pre- 

 test) consisted of 85 males and 95 females-that to 



30-1 



Female- 



15 20 



THICKNESS (mm) 



T 



25 



30 



35 



Figure 4.-Thickness frequencies at month 33.5. Test population 

 was to be selectively fished; control, unselectively. 



be unselectively fished (precontrol), 77 and 98, 

 respectively. The thickness frequencies (Figure 4) 

 revealed a low point in the pretest population 

 between males and females at about 25 mm. This 

 was used as the initial selection point and it meant, 

 of course, that the early catches were mostly 

 males. It will be shown below, however, that later 

 catches were composed of roughly equal propor- 

 tions of males and females. 



Changes from exploitation, in addition to those 

 described under "Course of Populations," were 

 reflected in the size composition of the catches 

 (Figures 5, 6). In the test population, the catches 

 were roughly the target percentage of the selected 

 group; the percentages in the control population 

 were adjusted to take the same proportion of the 

 entire population as taken in the test population. 

 Percentages were by number at months 39.2 and 

 41.3, but it became evident that this procedure 

 took too large a proportion of the biomass. At 

 months 43.2-75.1, the percentages were by weight 



500 



