468 



Fishery Bulletin 94(3), 1996 



characterizing field data is needed that clearly re- 

 flects 1) recruitment (CPUE) of both cultured and 

 wild fish, 2) hatchery contribution, and 3) standard- 

 ized recovery rates for released fish. One way of il- 

 lustrating these variables together is presented here. 

 All three indicators are used in Figure 11 to com- 

 pare relative hatchery effect at Kahaluu in 1992 with 

 data from three other years of pilot releases (Leber, 

 1995; Leber et al., 1995; Leber et al. 3 ). Overall, CPUE 

 of cultured juveniles was greatest after 1992 releases. 

 However, recovery rate (and presumably survival) 

 was greatest after 1993 releases. Thus, 1992 releases 

 had greatest impact on recruitment, whereas yield 

 per stocked juvenile was greatest from 1993 releases. 

 Recovery rate improved in 1992 and again in 1993. 

 The improvement was largely due to better survival 

 after adjusting SAR protocol to avoid summer re- 

 leases offish <70 mm TL and to avoid release sites 

 outside of habitat preferred by striped mullet. Al- 

 though hatchery contribution was inflated in 1992 be- 

 cause of reduced recruitment of wild fish, overall effect 

 from releasing cultured fish was greatest following 1992 

 releases, and recovery rate improved most in 1992. 



Pilot releases 



Whereas actual survival of cultured fish may be dif- 

 ficult to quantify in open ecosystems, relative sur- 

 vival can be quantified and compared across treat- 

 ment groups in pilot release experiments (Burnham 

 et al., 1987). Results from two previous years of pilot 

 releases were used to identify optimal release strat- 

 egies for this study. Discontinuing releases near the 

 Hawaii Institute of Marine Biology (HIMB) pier af- 

 ter 1990 resulted in a >300% increase in recovery 

 rate in the 1991 study; the increase in recovery rate 

 was compounded in 1992 by modifying SAR protocol 

 as well (Table 6; note the 1990 data in Fig. 11 are 

 only for fish released at Kahaluu). By confining re- 

 leases in this study to the vicinity of Kahaluu Stream 

 and by adjusting minimum SAR upwards to include 

 only fish >70 mm TL in summer, we achieved a 590% 

 increase in recovery rate over the 1990 study and a 

 170% increase over the 1991 study (Table 6: after 16 

 weeks, all sites combined). 



This study provided new information on effects of 

 release site. Choice of release microhabitat at 

 Kahaluu (inlet and upstream lagoon) affected dis- 

 persal north from Kahaluu into other streams but 

 had no apparent impact on survival. The similar 

 survival was surprising, given the poor survival in 

 1990 offish released along the shoreline near HIMB 

 pier, in comparison to fish released at streams ( Leber, 

 1995; Leber and Arce, in press). We hypothesize that 

 refuge from predators, afforded by mangroves and 



UJ 



=1 



CL 



u 



Cultured CPUE 

 Wild CPUE 

 Recovery 



1990 1991 1992 1993 



Year of release 



Figure 1 1 



Stacked bar graph of release impact at Kahaluu Stream 

 following releases conducted there from 1990 to 1993. 

 CPUE is total number of cultured fish (gray bars) and wild 

 fish (open bars) in 30 casts per month averaged over 10- 

 11 monthly collections. Percent cultured fish (hatchery 

 contribution! varied from monthly average of 4.4% (SEM 

 =±1.3) of the striped mullet collected in 1990 to 20.6% ( ±4. 1 ) 

 in 1991, 65.2% (±5.2) in 1992, and 23.5% (±5.0) in 1993. 

 Percent recovery rate of cultured fish (solid line with closed 

 circles) is number recaptured/number released, x lOO'? . 



other shoreline vegetation in the north end of 

 Kaneohe Bay, accounted for the higher survival of 

 mullet that were released at Kahaluu Inlet and that 

 dispersed along the shoreline. 



Release microhabitat affected the extent of en- 

 hancement in Kaneohe Bay in this study by partially 

 controlling colonization of nursery habitats north of 

 Kahaluu. If a management objective for full-scale 

 releases were to have a portion offish from each re- 

 lease lot disperse into adjacent nursery sites in the 

 bay (e.g. in order to maximize use of available nurs- 

 ery habitat), then releases at the inlet to Kahaluu 

 Stream would achieve this. If stronger site fidelity 

 were desired, releases farther upstream would re- 

 sult in lower dispersal during the nursery phase of 

 the life cycle. The ability to affect which nursery habi- 

 tats are selected by released fish, coupled with knowl- 

 edge about recruitment success of wild fish, could be 

 used to help prevent overstocking a particular 

 nursery. 



Implications 



This study shows that information from pilot releases 

 is critical for managing full-scale stock enhancement 



