920 
Fishery Bulletin 96(4), 1998 
respective growth curve) constituted the database for 
nonlinear regression estimation of instantaneous 
mortality rate of postyolksac larvae in a simple ex- 
ponential model, 
P t = P 0 exp(-zt), 
where P = the daily production of larvae at age t; 
t = the age in days from the end of the 
yolksac stage; 
P () = the daily larval production at t = 0; and 
z = the daily instantaneous mortality rate. 
The simple exponential mortality model fitted the 
data very well (Watanabe and Lo, 1988). 
Mortality models resulting from the application of 
different methods of correcting (or not correcting) 
catches for light-induced net avoidance were com- 
pared by an analysis of the residual sum of squares 
(Chen et al., 1992). 
Results 
Estimates of mean catch 
No larvae >12-mm were caught in any of the 227 
day-time tows. Estimates of mean catch per length 
group (Table 2) indicated that catches of >10-mm 
larvae were very low and their coefficients of varia- 
tion too great (generally >20%) to be useful in this 
study. Thus, our analysis was restricted to larvae 
<10-mm (see also Lenarz, 1972). 
Retention of anchovy larvae in the 
0.335- and 0.500-mm mesh nets 
Results of the catch analysis indicate that the 0.500: 
0.250-mm mesh catch ratio was less than one for lar- 
vae <7 mm but the 0. 335:0. 250-inm mesh ratio was 
not significantly different from one, at any length 
(Fig. 2). 
Data of maximum head width against standard 
length of the 74 anchovy larvae, along with the mesh 
diagonals of 0.500-, 0.335-, and 0.250-mm, are pre- 
sented in Figure 3. A comparison of head-width mea- 
surements with mesh diagonals indicates that the 
minimum lengths for complete retention in the 0.500- 
and 0.335-mm mesh nets were approximately 7.5 and 
3.5 mm, respectively. The head-width measurements 
further suggest that to ensure full retention of lar- 
vae, it would be necessary to use netting with a mesh 
aperture of 0.250 mm or less. In terms of ontogenetic 
stages, the 0.335-mm mesh net is expected to retain 
all except the yolksac larvae, whereas the 0.500-mm 
mesh net seems to be inefficient for yolksac and 
Figure 2 
Catch ratios of 0.500:0.250 mm (A) and 0.335:0.250 mm 
(B) mesh nets as a function of larval length. Bars indicate 
95% confidence intervals. 
preflexion larvae. It therefore appears that the 
“diagonal rule” is conservative for yolksac larvae of 
the European anchovy, when head width is used as 
the maximum cross-sectional diameter of the larva. 
Differences in day versus night and twilight 
versus night catches 
Results of the catch analysis show a change in the 
daymight and twilightmight catch ratios at L - 6.5 
mm (Fig. 4). In general, both ratios are practically 
one for larvae <6 mm, at which point they drop sub- 
stantially, continuing to decrease more or less lin- 
early thereafter. This decrease is stronger for the 
daymight than for the twilightmight catch ratio, the 
daymight ratio being significantly less than one from 
6 mm onwards and greater than the corresponding 
twilightmight ratio. Thus, for larvae >6 mm, it seems 
