Eggs and Larvae of the Nehu — Tester 
329 
TABLE 5 
Number of Nehu Larvae per 10-Minute Surface Haul with the lOO cm. Net according to Survey, 
Station, and Direction (A and B), and Summary 
station 
SURVEY 1 
SURVEY 2 
SURVEY 3 
SURVEY 4 
mean per cent 
OF SURVEY 
TOTAL 
A 
B 
A 
B 
A 
B 
A 
B 
1 
16 
30 
1 
3 
— 
5 
1 
— 
3.58 
2 
3 
3 
2 
- 
4 
8 
2 
18 
5.50 
3 
14 
9 
4 
- 
- 
3 
2 
6 
2.66 
4 
23 
23 
14 
4 
1 
7 
6 
6 
6.59 
5 
45 
39 
26 
14 
2 
2 
9 
9 
10.09 
6 
90 
50 
25 
24 
3 
2 
25 
21 
14.21 
7. 
57 
115 
15 
16 
2 
15 
6 
3 
14.74 
8 
53 
105 
5 
2 
3 
- 
4 
6 
7.75 
9 
25 
13 
5 
5 
3 
4 
10 
4 
5.40 
10 
2 
7 
8 
— 
5 
13 
17 
5.12 
11 
1 
- 
18 
13 
- 
1 
1 
4 
3.60 
12 
- 
- 
7 
3* 
- 
1 
4 
4 
1.81 
13 
- 
- 
5 
1 
3 
- 
3 
3 
1.91 
14 
2 
- 
3 
2 
- 
' 2 
43 
28 
6.14 
15 
- 
- 
6 
1 
- 
- 
33 
20 
4.38 
16 
- 
- 
- 
2 
- 
2 
6 
7 
1.72 
17 
3 
— 
5 
— 
— 
— 
13 
13 
2.40 
18 
1 
— 
2 
4 
_ 
1 
1 
2 
1.10 
19 
- 
- 
3 
2 
- 
- 
1 
1 
0.61 
20 
21 

— 
: 
: 
— 
1 

1 
0.38 
22 
23 
- 
- 
- 
- 
- 
1 
- 
- 
0.31 
Totals .... 
335 
387 
153 
104 
21 
60 
183 
173 
100.00 
* Part of haul lost. 
of consistency in the relative catches of A and 
B hauls from station to station, due to causes 
other than chance variation. In all surveys, 
the SXN (station by net) interaction is rela- 
tively large, at least larger than the residjiial 
variance SXHXN, but in no case is it sta- 
tistically significant. It probably includes a 
real component of variance, which is of 
small magnitude. For the present, however, 
it may be assumed that the catches of the two 
nets, apart from order of haul, bear a con- 
sistent ratio to each other from station to 
station. The HXN (haul order by net) inter- 
action is small; in fact, it is smaller than the 
SXHXN interaction in all surveys except the 
third. In other words, for the same station the 
catches made in A and B order for each of the 
two nets are somewhat more consistent than 
might be expected on the basis of chance 
As there were obviously large differences in 
egg catches, both between stations and be- 
tween nets, the large mean squares for stations 
and nets were expected. The mean square for 
haul order was relatively large in Surveys 1, 
2, and 4, but includes a large SXH (station by 
haul) interaction component, and cannot be 
regarded as generally significant. Moreover, 
the (geometric) means for A and B hauls 
(converted from logarithms) do not show a 
consistent difference from survey to survey: 
SURVEY A B 
1.... 101.4 67.0 
2 35.2 61.5 
3 36.7 48.6 
4 28.5 20.1 
Greater interest is centered in the inter- 
actions. In all four surveys the SXH inter- 
action is large, and in Surveys 1 and 4 it is 
highly significant. This demonstrates a lack 
