300 
BULLETIN OP THE BUREAU OF FISHERIES 
be small, lowering the average size of the fish in the whole catch. Later, as the fish 
of this year class grow older, the average size of the fish in the commercial catch will 
be gradually raised, until another dominant year class appears and temporarily 
lowers it. Evidence of such changes in size, due to the progression of dominant 
year classes, is shown in Elrington Passage by accurate records kept by one of the 
packers, giving the proportion of the catch pickled each year and the trade categories, 
according to the size of the fish. To make each year comparable the records have 
been compared for the month of July: 
Year 
Number 
barrels 
Per cent 
pickled 
Per cent 
matjes 
Per cent 
medium 
Per cent 
large 
1924 
8, 125 
10.0 
71.5 
14.4 
14.1 
1925 
4, 355 
9.3 
56.6 
25.6 
17.8 
1926 
3, 239 
15.9 
90.9 
9. 1 
1927.. 
2 , 271 
29.1 
13.2 
77.6 
9.2 
Matjes are herring from about 9 % to 10 % inches in total length (197-215 milli- 
meters in body length); mediums from about 10K to 11 % inches in total length (215- 
232 millimeters in body length) ; large 
from about 11 % to 12 % inches or over 
(232-249 millimeters in body length). 
Of course, the proportions of each kind 
pickled will depend to some extent on 
market conditions, and the sizes in- 
cluded under each name will fluctuate 
slightly from year to year, but, in the 
main, the proportion of the catch 
pickled and the percentage of each 
class will depend on the raw material 
available. There are two features so 
pronounced as to appear valid; one 
is the increase in the total amounts 
pickled in 1926 and 1927, the other 
is the shift in sizes pickled. These 
changed from over 70 per cent matjes 
and 14 per cent mediums in 1924 to 
no matjes and over 90 per cent me- 
diums in 1926. In 1927 the percent- 
age of mediums packed is less than in 
1926 owing to the packing of 13 per 
cent of matjes, but the percentage of 
the total catch pickled as mediums in 
1927 is actually greater than in 1926 
and the same is true of the large. 
Comparing these annual changes 
in the sizes and amounts pickled with 
the percentage of herring at each size 
as shown in Figure 35 the existence of 
a rough correlation is at once apparent, lending support to our biological findings, 
and thus demonstrating an important point; namely, that the effect of the progres- 
Figure 38. — Length frequency distributions for the Eshamy Bay and 
McClure Bay fall herring for 1926 and 1927, respectively 
