30 
BULLETIN OF THE BUREAU OF FISHERIES 
fifth year, with ocean nuclei, foimd both in the ocean and witliin the stream. No 
satisfactory explanation of this can be given, although the possibility of some sort 
of a selective migration naturally suggests itself. Figure 5 shows the distribution 
of egg sizes for this age group dming the season, and it may readily be seen that, 
with the exception of the groups marked by a double circle, the range of egg sizes 
shown by the fish taken outside is quite comparable with the range shown by those 
taken inside the river. The groups indicated by a double circle are those that 
-7 
1 
1 
( 
1 
1 
1 
1 
1 
A 
• 
1 
1 
1 
1 
! M 
A 
1 
1 
1 • 
[•] 
■ 
1 
„ i • 
1 
1 
1 
1 
1 
1 
1 
1 
1 
1 
• 
• 
• 
< 
[°] 
• 
• 
I 
o 
-L. 
Fig. 4. — Sizes of eggs of females in their fifth year, ocean nuclei, from the Columbia River 
and the ocean ofi the mouth of the river. In this and other similar illustrations the average 
size of the eggs is indicated by the circles or dots; the vertical lines e.xtending from these 
points show the limits of twice the standard deviation (± 2 a). Ninety-five per cent of the 
cases will fall within the upper and lov/er limits indicated by these lines. The standard 
deviation has, with a few exceptions, been calculated only for those cases in which at least 
20 individuals of the particular age group in question were contained in the collection 
may be considered as immature, although the interpretation, especially in the case 
of those collections made during May, is somewhat doubtful. 
A correlation table has been prepared (Table 4), giving the sizes of the eggs 
found in the fish of this age group contained in a collection of troll fish taken off 
the mouth of the Columbia River, May 17 and 18, 1920. Reference to this table 
will show that the fish are quite distinctly separated into two groups on the basis 
of egg sizes. The log D of one group is greater than 0.30, averaging 0.4254, and of 
the other group is less, averaging 0.2486. The modes of the two distributions are 
distinct, and it seems probable that two categories are represented — one of mature 
and the other of immature fish. This interpretation is further indicated by the 
