NOTE Somarakis et a I.: Catchability and retention of larval Engraulis encrasicolus 
923 
Table 3 
Estimates of length-specific larval production for the 1994 and 1995 cruises (and their standard errors in parentheses) with 
different methods of adjusting data for net avoidance. Method 1 is based on daymight catch ratio values available for Engraulis 
mordax, and methods 2 and 3 are based on daymight and twilightmight values calculated in the present study (see text for 
details). 
Cruise Correction 
Length (mm) 
4.5 
5.5 
6.5 
7.5 
8.5 
9.5 
1994 none 
1.984 
1.655 
1.169 
0.893 
0.518 
0.413 
(0.198) 
(0.198) 
(0.139) 
(0.127) 
(0.110) 
(0.093) 
method 1 
2.142 
1.793 
1.281 
0.972 
0.569 
0.450 
(0.214) 
(0.212) 
(0.152) 
(0.138) 
(0.121) 
(0.102) 
method 2 
1.984 
1.655 
1.231 
0.944 
0.553 
0.445 
(0.198) 
(0.198) 
(0.146) 
(0.134) 
(0.118) 
(0.101) 
method 3 
1.984 
1.655 
1.258 
0.979 
0.573 
0.475 
(0.198) 
(0.198) 
(0.150) 
(0.140) 
(0.122) 
(0.109) 
1995 none 
1.537 
1.235 
0.928 
0.690 
0.472 
0.341 
(0.111) 
(0.090) 
(0.080) 
(0.071) 
(0.052) 
(0.046) 
method 1 
1.721 
1.409 
1.076 
0.722 
0.559 
0.399 
(0.029) 
(0.015) 
(0.058) 
(0.045) 
(0.046) 
(0.047) 
method 2 
1.537 
1.235 
1.009 
0.681 
0.531 
0.390 
(0.111) 
(0.090) 
(0.086) 
(0.062) 
(0.057) 
(0.051) 
method 3 
1.537 
1.235 
1.026 
0.706 
0.549 
0.412 
(0.111) 
(0.090) 
(0.087) 
(0.064) 
(0.059) 
(0.054) 
vae, the efficiency of the towed net decreased with 
larval length relative to the seine. 
The change in catchability at flexion, observed in 
our study, can be attributed to a change in the swim- 
ming ability of larvae, which is associated with flex- 
ion. Batty ( 1984) found that Atlantic herring larvae, 
as they grow, change from an anguilliform mode of 
swimming to a subcarangiform mode of swimming. 
This change of swimming mode occurs as the caudal 
fin develops (at a body length of about 22 mm). Sub- 
sequently, Heath and Dunn (1990) using the large 
(5 m 2 ), high-speed LOCHNESS sampler, found that 
the daymight catch differential increased with body 
length for larval herring in the North Sea, with a 
maximum fivefold difference between day and night 
catches for larvae >25mm. In another study, Osse 
and van den Boogaart (1995) reported results similar 
to those of Batty (1984) for common carp ( Cyprinus 
carpio). The morphological differentiation of the cau- 
dal fin, which begins at flexion and is accompanied 
by ossification of the caudal-fin rays and the caudal 
part of the notochord, closely parallels a change in 
the swimming mode from anguilliform to carangi- 
form. Burst-speed capability, which is believed to 
determine, in part, the ability of a larva to avoid 
plankton nets (Hunter, 1976), might also change with 
the development of the tail. 
Table 4 
Estimated parameters of the exponential mortality mod- 
els (P,=P 0 exp(-zf )) for the 1994 and 1995 cruises. A differ- 
ent model was fitted for each different method of adjust- 
ing data for net avoidance (see text for details), r 2 = coeffi- 
cient of determination. 
Cruise 
Correction 
Po 
z 
r 2 
1994 
none 
2.931 
0.143 
0.980 
method 1 
3.156 
0.142 
0.979 
method 2 
2.872 
0.135 
0.979 
method 3 
2.832 
0.131 
0.978 
1995 
none 
2.195 
0.136 
0.992 
method 1 
2.463 
0.135 
0.989 
method 2 
2.139 
0.126 
0.989 
method 3 
2.109 
0.121 
0.989 
Alternatively, changes in catchability may have a 
behavioral component. A substantial change in 
daymight catches of northern anchovy ( Engraulis 
tnordax ) with bongo nets occurs at approximately 11 
mm (Fig. 2 in Hewitt and Methot, 1982). Flexion in 
this species occurs at around 11 mm (Watson and 
Sandknop, 1996). Laboratory and field studies 
(Hunter and Sanchez, 1976; Hewitt, 1981; Hunter 
