Egg and Larva Production of Nehu — Tester 
39 
If the eggs classed as agglutinated were 
actually dead at the time of sampling, it would 
seem that this source of natural mortality was 
more pronounced in the winter, the period of 
small spawning, than in the summer, the 
period of large spawning. This suggests a 
further investigation of (1) the assumption 
that agglutinated eggs were dead at the time 
of sampling, and if this is established, (2) the 
causes of mortality of the eggs. 
Reference may again be made to the grand 
geometric means for eggs and larvae (7.01 
and 0.96 per 100 cubic meters) and to the 
great reduction in numbers (86.3 per cent). 
As the eggs hatch within 24 hours whereas 
the larvae persist in the catches for several 
days, the mean count for eggs would be ex- 
pected to be smaller rather than larger than 
that of the larvae. There are several possible 
explanations for the relatively small mean 
larva counts: 
(a) The larvae may be carried away from 
Station 4 by a slow clockwise current drift 
as suggested by Tester (1951: 342) following 
a study of variation in length distribution of 
the larvae. 
(b) The thin filiform larvae, particularly 
after yolk sac absorption, may slip through 
the meshes of the net. 
(c) The newly hatched larvae may sink be- 
low the surface layer which was sampled. 
Yamashita (MS: 14) has shown that in aquaria 
the newly hatched nehu larvae tend to sink. 
Although in simultaneous tows made by 
Tester (1951: 335) there was no significant 
evidence of more larvae at a depth than at the 
surface, the numbers caught were too few to 
warrant a conclusion. 
(d) The newly or recently hatched larvae 
may suffer a high natural mortality. It seems 
idle to speculate further on this or other 
possibilities in view of the probable action 
of (a), (b), and perhaps (c), as outlined above. 
Perusal of Table 2 will disclose that the ratio 
of egg to larva number varies considerably 
Fig. 4. Seasonal variation in the percentage of ag- 
glutinated and normal eggs. 
from month to month. The most striking 
discrepance occurs in the months of Decem- 
ber, January, and February, 1951-52, when 
the mean egg count is less, rather than greater, 
than the mean larva count. Although in these 
months the comparisons are intrinsically less 
reliable than in others because of the small 
numbers involved, nevertheless the discrep- 
ance seems due to factors other than chance. 
The reversal of the egg to larva ratio in the 
winter is contrary to what would be expected 
from the tendency, just discussed, for the 
percentage of agglutinated and presumably 
dead eggs to be higher in the winter than the 
summer: this should increase rather than de- 
crease the winter egg to larva ratio. One 
possible explanation of the reversal is that in 
the winter of 1951-52 spawning of relatively 
small extent took place elsewhere than at 
Station 4, and that the larvae drifted into the 
sampling area. 
