462 



Fishery Bulletin 94(3). 1996 



earlier studies with July releases; and 2) the late- 

 summer decline in abundance resulted in less reduc- 

 tion in release impact in 1992 than in other years. 



Recapture rates were disproportionately higher in 

 this study than after releases in 1990 and 1991 (Table 

 6). Number offish released at Kahaluu Stream in 1992 

 exceeded the number released at that same site in 1990 

 and 1991 by 690% and 180%, respectively Yet the total 

 number of fish recaptured from week 16 on exceeded 

 comparable data (similar sampling effort) from 1990 

 and 1991 by 1,560% and 420%, respectively (Table 6). 

 Increase in the effectiveness of 1992 releases is revealed 

 by comparing recapture frequencies (no. recaptured/ 

 no. released) and by including data from all release 

 sites. From week 16 on, 1.65% of the cultured mullet 

 released in 1992 were recaptured in cast-net samples. 

 This recapture rate was 6 times that seen in the 1990 

 study (0.28% ) and 1.7 times the rate in the 1991 study. 

 Sampling frequency and number of cast-net samples 

 were nearly identical among the three studies. 



These differences among studies were statistically 

 significant. Recapture rates after 1992 releases were 

 significantly greater than expected in all pair-wise 

 comparisons of 1992 data with data from 1990 and 

 1991 (Table 6; G-tests, P<0.001 in all cases). The 1991 

 study yielded significantly greater recapture rates 

 than did the 1990 study, when fish released outside 

 of the Kahaluu site in 1990 are included in the com- 

 parison (Table 6; all sites, x' 2 = 225, P<0.001 ). 



Discussion 



Marine enhancement potential 



Context of this study A rigorous test of marine 

 stock-enhancement involves several phases of re- 

 search. These are considered here to place the cur- 

 rent study in perspective. Key research phases in- 

 clude 1) understanding both wild stock distribution 



