Leber et al.: Influence of release season on size-dependent survival of Mugil cephalus 
275 
to coincide with peak abundances of young-of-the- 
year recruits in these nursery habitats, the modal 
size of cultured fish led that of the wild recruitment 
pulse by around 30 mm at one month after spring 
releases. In contrast, the size structures of wild 
young-of-the-year and cultured fish were nearly iden- 
tical 1 month after summer releases (Fig. 9). 
Discussion 
Recapture rates and release impact 
Release impact on striped mullet abundance was 
comparable to contributions from experimental re- 
leases of cod in Norway (e.g. Kristiansen and 
Svasand, 1990; Nordeide et al., 1994), red drum in 
Florida (Willis et al., 1995), and to proportions of 
cultured flounder in commercial landings in Japan 
(Kitada et al., 1992). Cultured striped mullet 
amounted to no less than 7% of the fish in monthly 
samples throughout the 11-month study period at 
both release sites. By the end of this study, cultured 
fish represented about 12% of the striped mullet 
sampled at Kahaluu Stream, over 40% of those 
sampled in Kaneohe Stream, and 13.6% of the total 
collected in Kaneohe Bay. 
There was clearly an improvement in this study 
in recapture frequencies compared with initial re- 
leases into the Bay in 1990 (Leber, 1995). The im- 
provement was largely due to adjusting release strat- 
egy in this study to avoid releases outside of streams, 
the nursery habitats preferred by striped mullet. 
Recovery rates (number recaptured/number released) 
x 100%) of fish released at Kahaluu Stream in 1991 
(this study) were similar to rates following a release 
of 10,000 fish at the same site in 1990; whereas, con- 
siderably fewer striped mullet were recovered follow- 
ing 1990 releases of 30,000 fish into more marine 
conditions near Coconut Island in the southern por- 
tion of the bay (authors’ unpubl. data for juveniles; 
and see Leber and Arce, 1996, for data on adults). 
Temporal changes in abundance of released 
fish 
Reduction in abundance of cultured fish over time at 
release sites was likely a result of 1) mortality, 2) 
emigration from nursery habitats into adjacent reef 
habitats in the bay, and 3 ) sampling bias as fish grew 
to larger sizes and moved out of shallow water. Mortal- 
ity appeared to be more important than emigration as 
the cause for reduction over time in recapture rates. 
Juvenile striped mullet have a relatively strong 
affinity for brackish water during the nursery stage 
of their life cycle (Major, 1978; Blaber, 1987). After 
earlier pilot releases, when cultured striped mullet 
were released into more marine conditions (surface 
salinities >25 ppt), they schooled in both directions 
along the shore and rapidly occupied nearly all 
striped mullet nursery habitats (streams and tribu- 
taries) in those bay systems (Leber, 1995). In con- 
trast, when striped mullet were released into habi- 
tats with lower surface salinities, as in this study, 
the majority of individuals recaptured were caught 
at or near the release site (Leber et al., 1995, 1996). 
Had emigration out of release habitats remained 
high in this study after fish had had time to accli- 
<D 
o 
c 
(0 
"O 
c 
_Q 
CO 
0 
> 
OJ 
CD 
a: 
Total length (mm) 
Figure 8 
Size structures of wild and cultured Mugil cephalus col- 
lected in samples made about 1 month following spring 
releases into Kaneohe Bay. 
ai 
cr 
- 
Wild fish in August 
r 
n=453 
..all 
- 
20 40 60 80 
10C 
120 140 160 180 200 220 240 260 
Cultured fish in August 
i 
n=184 
. ,_jiS 
illl __ 
20 40 60 80 100 120 140 160 180 200 220 240 260 
Total length (mm) 
Figure 9 
Size structures of wild and cultured Mugil cephalus col- 
lected in samples made about 1 month following summer 
releases into Kaneohe Bay. 
