268 
Fishery Bulletin 95(2), 1997 
and Kristiansen, 1990; Stoner, 1994; Leber, 1995; 
Willis et al., 1995). Pilot releases have shown that 
survival rates following hatchery releases of striped 
mullet, Mugil cephalus, in Hawaii (Leber and Arce, 
1996; Leber et al. 1 ) and of queen conch, Strombus 
gigas, in the Caribbean (Stoner, 1994) were strongly 
affected by release habitat. Pilot releases with M. 
cephalus have also shown differential survival based 
on size-at-release. Pilot releases conducted during 
summer and fall in Maunalua Bay, Hawaii, (south- 
ern exposure) and during summer in Kaneohe Bay 
(eastern, windward exposure) have shown poor sur- 
vival of cultured M. cephalus smaller than 70 mm 
total length (TL) at the time of release, compared 
with survival of larger-size individuals (e.g. 70 to 130 
mm TL, Leber, 1995). In this study, we document a 
substantial effect of the seasonal timing of releases 
upon size-at-release-dependent recapture rates 
(number recaptured /number released) of cultured M. 
cephalus. 
Materials and methods 
Hatchery releases 
Striped mullet were spawned at The Oceanic Insti- 
tute in 1991 and reared to fingerling size. Batches of 
striped mullet eggs were hatched approximately ev- 
ery 5-6 weeks over a 5-month period and reared 
through three stages in cylindrical tanks. Larvae 
from each batch were hatched and cultured in 5,000- 
L conical-bottom tanks for 45 days. Stage- 1 juveniles 
(i.e. postlarvae 45 days old, 20 mm total length [TL]) 
were transferred to 8,000-L tanks and reared for 40 
days to stage-2 juveniles (i.e. the age and size at 
which we typically transfer fish out of nursery tanks 
into larger growout tanks, 85 days old, around 40 
mm TL). Stage-2 juveniles were transferred to 
30,000-L tanks and reared to tagging size (45 to 130 
mm TL). 
A factorial-design release-recapture experiment 
was performed to compare interactive effects of re- 
lease season and fish size-at-release upon growth and 
survival of about 90,000 cultured striped mullet in 
the wild. During the period 5 May through 17 May 
1991, and again from 12 July through 26 July 1991, 
1 Leber, K. M., D. A. Sterritt, R. N. Cantrell, and R. T. 
Nishimoto. In press. Contribution of hatchery-released 
striped mullet, Mugil cephalus , to the recreational fishery in 
Hilo Bay, Hawaii. In K. Lowe (ed.), Proceedings of the first 
biennial symposium for the Main Hawaiian Islands Marine 
Resources Investigation. Technical Rep. 96-01. Hawaii Depart- 
ment of Land and Natural Resources, Division of Aquatic Re- 
sources, Honolulu, HI. 
juvenile striped mullet, ranging in size from 45 to 
130 mm TL, were harvested from culture tanks and 
transferred to 40,000-L holding tanks. These fish 
were graded into five size groups, tagged, then re- 
leased into Kaneohe Bay; half were released in May, 
the other half in July. 
To identify experimental treatment conditions, all 
released fish were tagged with binary coded- wire tags 
(Jefferts et al., 1963). Tags identified release season, 
release site, size-at-release (SAR), release lot (date), 
and number of fish per treatment condition. Fish 
were tagged in batches, with a different code for each 
season-site and SAR-lot combination (2x2x5x3=60 
batch codes). The five size groups released were 45- 
60 mm; 60-70 mm; 70-85 mm; 85-110 mm; and 110- 
130 mm TL. 
Tags were implanted in the snout area with an 
automatic injector with head molds designed specifi- 
cally for striped mullet. Previous studies have shown 
a coded- wire tag retention rate of 97% for striped 
mullet over a 6-month period (Leber, 1995). To verify 
tag-retention rates in this study, at least 5% of the 
fish tagged for each release lot were randomly 
subsampled prior to each release. The subsamples 
were retained in tanks for up to 6 months to check 
tag retention. Subsampled fish were not released. 
Release statistics 
During May and July 1991, 90,817 juvenile striped 
mullet were tagged and released into Kaneohe Bay. 
Numbers of fish released varied among size groups 
but were held nearly constant among release lots and 
between release sites and seasons (Table 1). At least 
7,500 tagged fish were released in each of 12 release 
lots. There was size variation in all batches of mul- 
let reared for this study. However, the primary dif- 
ference among size-at-release groups was fish age. 
For each season and SAR treatment combination, 
the experiment was replicated at two sites in 
Kaneohe Bay, and within each site, three replicate 
release lots were made (Table 1). The release lots 
were introduced into the bay over a 3-week period 
during both seasons (spring and summer). In each 
season, releases were made simultaneously at the 
inlets of two primary striped mullet nursery habi- 
tats, Kahaluu Stream and Kaneohe Stream. Kahaluu 
Stream is located in the north end of Kaneohe Bay 
(Fig. 1). This tributary is fed by several stream sys- 
tems that originate in the Ko’olau mountain range. 
Kahaluu Stream expands into a lagoon about 300 m 
upstream. The mouth of Kaneohe Stream is 11.6 km 
southeast of Kahaluu Stream. Kaneohe Stream is 
also a Ko’olau mountain drainage system. Selection 
of release habitats in the vicinity of fresh-water tribu- 
