902 
Fishery Bulletin 96(4), 1998 
The larvae were classified as Japanese sardine, S. 
melanostictus, Japanese anchovy, E. japonicus, and 
round herring, Etrumeus teres. Sardine larvae were 
preserved in 80% ethanol for otolith examinations. 
Otolith measurement 
Total lengths (TL) of sardine larvae were measured 
to the nearest 0.1 mm with an optical comparator. 
Sagittal otoliths were dissected and cleaned under a 
binocular microscope, mounted on a glass plate 
with enamel resin, and used for measurements and 
counts of daily growth rings. We used the otolith mea- 
surement system (Ratoc System Engineering Inc.) 
composed of a light microscope, a video camera and 
monitor, and an image analyzer controlled by com- 
puter. Because the otolith had not yet developed a 
rostrum and because we were not able to determine 
its orientation at the larval stage, we measured ra- 
dii of all daily rings along the maximum radius of 
the otolith. 
Growth backcalculation 
The relationship between larval TL and maximum 
otolith radius (OR) was considered for larvae 
sampled. Plots of TL against OR can be expressed by 
an allometric relationship (see “Results” section). An 
allometric OR-TL relationship has previously been 
demonstrated in larval S. melanostictus from differ- 
ent waters (Watanabe and Kuroki, 1997). Otolith 
growth rings have been found to be deposited on a 
daily basis in S. melanostictus , with the first ring 
being formed on the third day of hatching (the day of 
first feeding) when larvae are reared at 18°C 
(Hayashi et al., 1989). The relationship of the zth 
otolith ring radius iOR t ) and TL on the day of the ith 
ring formation (TL ; ) is considered to be expressed by 
an allometric formula, TL t -ax OR b , for individual 
larvae. We determined the allometric parameters a 
and b for each larva by using the biological intercept 
method (Campana, 1990; Campana and Jones, 1992) 
and a size of first feeding larvae (first otolith daily 
ring deposition) of 5.0 mm TL (Watanabe and Kuroki, 
1997). The solution of the following two equations 
gives us a and b for each larva: 
TL^ ax OR b and TL capture = a x OR capture b , 
where TL ] = total length (mm) at the first ring 
deposition which was fixed at 5.0 mm; 
OR x = the measured radius of the first daily 
ring; 
TL capture = the measured total length (pm) at cap- 
ture; and 
OR capture = otolith radius (pm) at capture. 
TL t of each larva was thus calculated from the for- 
mula for each larva independently. Mean ±SD (stan- 
dard deviation) of TL at ages from 4 d up to a certain 
age, with at least 10 backcalculated TLs, was calcu- 
lated for the March- and April-hatched cohorts for 
the two years of study. Differences in mean TLs at 
ages 10, 15, 20, and 25 d were examined between 
the same month of hatch cohorts in 1990 and 1991 
by using Student’s t-test, when variances were equal, 
or by Welch’s t-test when they were not equal. 
Sea surface temperature distribution 
Sea surface temperatures (SST) in and around the 
shirasu fishing ground were measured monthly at 
fixed stations off Atsumi Peninsula (Fig. 1). We used 
data from April and May, 1990 and 1991, to describe 
temperature distributions in coastal waters. 
Results 
The SST in coastal waters off Atsumi Peninsula was 
in the range of 16-19°C during 23-25 April and 8-9 
May 1990. From 23 to 25 April, the isotherms ran 
nearly parallel to the coast line but from 8 to 9 May 
offshore waters warmer than 18°C intruded into the 
coastal fishing ground (Fig. 3). In 1991, the SST dur- 
ing 2-3 April and 8-9 May ranged from 11 to 18°C. 
Cold waters extruded from Ise Bay covered part of 
the coastal fishing ground. The SST in the fishing 
grounds were 1.5-2.5°C lower in 8-9 May 1991 than 
in the corresponding season in 1990. 
Local government permits the shirasu fishery to 
operate year round in these fishing grounds, but 
catches were zero in January-February 1990 and 
January-March 1991 (Fig. 4). Japanese sardine lar- 
vae were caught mainly in April in these years. The 
major target species of the shirasu fishery shifted to 
Japanese anchovy after May. A monthly catch of sar- 
dine larvae was 430 t in April 1990, declining to 175 t 
in April 1991. Annual catches of sardine larvae were 
724 t and 298 t in 1990 and 1991, respectively. 
Total length of larval sardines caught in the shirasu 
fishery were 15-27 mm in mid-April 1990 (Fig. 5). 
Frequency of larvae smaller than 15 mm TL in- 
creased after late April in this year, with sizes of 11- 
20 mm. In 1991, the modal size and range of the lar- 
vae remained relatively constant at 22-25 and 20-27 
mm TL, respectively, mid-April to mid-May (Fig. 5). 
Sardine larvae fished on 11 and 16 April 1990 were 
aged from 14 to 39 d after hatching. All of them 
hatched in March (Fig. 6). Larvae caught on 27 April 
