=== T [rset wal i The data for the two regions were 
compared by the analysis of covariance 
30 = (Snedecor 1946, pp. 318-321). The results 
(table 9) showed that the differences between 
them were so small that the data could be 
considered together. Combining the data 
from both regions resulted in the following 
e quation of regression: 
OVARY NO.65 
Y= 006935 x - 0.13607 
m 
(o) 
In order to determine at what level 
of relative ovary weight the eggs first show 
signs of maturing, i. e. begin to turn opaque, 
all central and western equatorial Pacific 
ovaries with relative ovary weights smaller 
than 7 were sampled and investigated for 
opaque eggs. Randomly picked ovaries from 
the first year of collection in the Hawaiian 
area were also investigated. These are noted 
in the tables in the appendix, The few ovaries 
with a relative ovary weight of 2 were all 
NUMBER OF EGGS 
fo) 
immature. Only 11 percent of the ovaries 
with a relative ovary weight of 3 had eggs 
EGG DIAMETER (MN.) in the early maturing stage, whereas 65 
percent of those with a relative ovary weight 
of 4 had such eggs. 
io) 0.I 0.2 03 04 0.5 06 
Figure 3, --Method of determining position 
of mode, 
The possibility of using the relative 
ovary weight as a tool for measuring the ripeness of a fish easily and objectively now arises. 
The size of the egg at the different stages of ripeness is known (page 2), This size may be pre- 
dicted from the relative ovary weight. Thus, by using the regression of modal diameter on 
relative ovary size, relative ovary weight values were assigned to the different classes of ripe- 
ness as follows: 0 to 3 are considered immature, 4 to 18 early maturing, 19 to 28 late maturing, 
and larger than 28 mature. Spent fish are not recognized by this method, being classified in the 
stage of maturity which represents the most advanced of the remaining groups of eggs. 
A certain amount of erroris involved in classifying the degree of ripeness using relative 
ovary weights, This error was estimated by: (1) calculating the probabilities of sampling a fish 
12 in the immature stage and classifying it as 
immature, early maturing, late maturing, 
and mature; of sampling a fish in the early 
maturing stage and classifying it as immature, 
early maturing, late maturing, and mature; 
etc. (Pearson 1931, tables VIII and IX); (2) 
calculating a percentage of error for each 
category based on these probabilities (see 
below). The probabilities (table 10) are based 
on a theoretical population of which the 33 
ovaries used for the regression line for the 
central equatorial Pacific (page 10) constitute 
a sample. The ovaries involved in the regres- 
sion computations of the western equatorial 
Pacific were not used as these had been 
RELATIVE OVARY WEIGHT X 103 selected not randomly but according to their 
relative ovary weights. 
fe} 
08 
MODAL DIAMETER (MM.) 
is} 5 10 15 20 46.25 30 35 40 45 50 
Figure 4,--Relation of modal diameter to rela- 
Pye cv dnyawcsgit If the probability of an individual fish 
actually being in the stage of ripeness in 
13 
