460 



Fishery Bulletin 94(3). 1996 



Effect of release season Although recapture 

 rates among SAR groups >60 mm appeared similar 

 in the aggregate data set, an SAR effect was appar- 

 ent when data were collapsed within release seasons 

 (Table 5). To clarify survival and the interactive ef- 

 fect of release season with SAR, only fish that sur- 

 vived at least 16 weeks are included in Figures 6, 7, 

 and 8. This provides a picture of those cultured fish 

 that remain in the population from 4 months on, 

 having survived initial causes of mortality in the wild, 

 and that have the potential to affect adult abundances. 



All sizes released in spring and summer were 

 prominent in cast-net collections (Table 5). As ex- 

 pected following spring releases (Leber et al. 3 ), no 

 direct relationship was apparent between SAR and 

 recapture rates. However, recapture rates (from week 

 16 on) were significantly less for the smallest size 

 group released than for SAR groups 2, 3, and 5 (Fig. 

 6,P<0.02). 



Summer releases resulted in better recovery of the 

 largest size group released (Fig. 7, 110-130 mm TL; 

 P<0.04; marginally significant). There was also a 

 trend (nonsignificant) towards better recovery of the 



three largest SAR groups when released in summer 

 rather than in spring (Fig. 8). 



Growth offish released in spring was comparable 

 with that offish released in summer (Fig. 9). Data 

 in Figure 9 are for fish from the median SAR group 

 (70-85 mm TL), for which the greatest amount of 

 data was available. This growth comparison is repre- 

 sentative of the other sizes released in both seasons. 



Comparison with 1990 and 1991 release impact 



There was substantially greater impact on juvenile 

 abundances in Kaneohe Bay following 1992 releases 

 than after pilot releases in 1990 (Leber, 1995) and in 

 1991 (Leber et al. 3 ). Proportions of cultured fish at 

 Kahaluu 10 months after releases increased from 

 about 3% following 1990 releases to 10% after 1991 

 releases, to about 50% in the present study (Fig. 10). 

 The general pattern following releases was simi- 

 lar in all years — an initial increase in proportions of 

 cultured fish in samples, followed by a decline over 

 the next year. However, there were two principal dif- 

 ferences in 1992: 1) July releases in 1992 caused a 

 considerably greater increase in abundance than in 



