Eggs and Larvae of the Nehu — Tester 
between stations are due primarily to erratic 
spawning at a station from day to day. This 
possibility cannot be denied, but such evi- 
dence as is presently available lends support 
to the interpretation given previously. 
DISCUSSION 
One of the objects of the present investi- 
gation was to formulate a satnpling program 
for the estimation of egg and larva production 
throughout the year. Therefore it is in order 
to consider the design of such a program in 
the light of the present results. Its scope, 
however, must be limited by available funds. 
The 100 cm. net is too large to be handled 
conveniently from a small vessel. The 22 cm. 
net is too small to give numbers of eggs and, 
particularly, larvae of sufficient magnitude 
for precise statistical analysis in a haul of 
reasonable time duration. Therefore a net of 
intermediate size, perhaps 50 cm. mouth 
diameter, is recommended. The net should 
be constructed of No. 40 grit gauze, and 
should be equipped with a current or flow 
meter and a detachable cod end. In design, it 
might follow that of the Hensen egg net. 
Because of the small variability between 
catches of nets hauled simultaneously and the 
large variability between individual hauls, it 
would be more efficient to make several hauls 
with one net than to make one haul with 
several nets. For the same amount of work 
(chiefly involved in sorting plankton) four 
short hauls at one station, each of 5 minutes’ 
duration, each spaced in regular fashion about 
the station, and each following a close, cir- 
cular course, would be more informative than 
the sampling design followed in the present 
investigation. 
Because of the slight but apparently con- 
sistent difference in egg abundance between 
surface and deep samples, horizontal tows at 
the surface, rather than either simultaneous 
surface and deep tows or oblique tows, could 
be used. However, the vertical distribution of 
the eggs should be subject to further investi- 
gation to determine the nature of the abun- 
343 
dance gradient and its variation in time and 
space. 
For a study of variation in abundance of 
eggs in the southern sector, one station lo- 
cated in the vicinity of Station 4 might suffice, 
provided it were sampled according to the 
above plan. Samples should be taken at close 
intervals of time throughout the year, say 
two per week. With a program such as this it 
should be possible to obtain information on 
both short-term and long-term fluctuations 
in the abundance of eggs. If the horizontal 
and vertical gradients of abundance, and their 
variation, could be established with precision, 
it should be possible to calculate the absolute 
abundance of eggs with reasonable accuracy. 
Interpolation in time and integration of the 
results would then lead to an estimate of total 
egg production for the southern sector which, 
according to this preliminary work, includes 
about 75 per cent of the total eggs spawned 
in Kaneohe Bay. 
The sampling of eggs in the middle sector 
of the bay presents greater difficulty because 
of the lack of consistent variation from station 
to station with time. At least two, and prefer- 
ably more than two stations should be in- 
cluded. With limited funds, it would be more 
efficient to make fewer hauls per station and 
to include more stations. A possible compro- 
mise might be to make two hauls at each of 
Stations 14 to 17 once a week. 
Sampling for eggs in the northern sector 
might be neglected without any great loss of 
information, for the eggs occurring there 
form only about 4 per cent of the total. 
However, if included, it might be sampled at 
Station 20, two hauls, once a week. 
If carried out as outlined, the above pro- 
gram would produce 936 samples per year. 
In addition, it would be desirable to check 
horizontal distribution at all stations and 
vertical distribution at certain key stations, 
e.g.. Stations 4 and 16, at least at quarterly 
intervals, thus producing an additional two 
to three hundred samples. The total, about 
1,200 samples, is regarded as the minimal 
