FLUCTUATIONS IN THE SUPPLY OF HERRING 
31 
1921 was a year in which the herring companies operated on a very restricted scale, 
owing to poor economic conditions (Rounsefell, 1930, p. 234). Disregarding 1921, 
June shows a higher level of abundance during the next four years than during the 
last four. July shows minor variations, but the general trend appears to maintain a 
level. August is characterized by a very sudden rise in 1925 and as sudden a drop in 
1927. September fluctuates considerably but shows a very considerable rise, 
especially since 1925. 
In the combined data it will be noted that whereas the curve for all of the months 
does not show a decline, the curve for the first three months, June to August, does 
appear to show a decrease. The reason for the difference between the two curves is 
the rise in September during the 
past few years. Although this Sep- 
tember rise is undoubtedly valid, in 
comparing the various years it is 
better to eliminate the September 
data, for it is only within the last 
few years (possibly owing largely to 
the increase in size and seaworthi- 
ness of the fishing vessels) that the 
September fishing has been very 
successful. Of former years when 
autumn approached and no herring 
were to be had in the more sheltered 
bays in the lee of Cape Ommaney, 
the plants suspended operations for 
the season; but since the fishermen 
have learned to seine herring in the 
dangerous tide rips and high ocean 
swells around Cape Ommaney, the 
plants usually operate with success 
until the end of September. The 
high relative abundance of herring 
during this final run is indicated by 
Figure 16, which gives the standard 
curve for Group I. 
It must be concluded that 
Group I gives evidence of a slight 
decline in abundance during the 
period from June to August. The apparent increase during September is probably 
due in large part, if not wholly, to the changes in the efficiency of the fleet. 
In Group II (figs. 17 and 18) the data cover only the past four years. June shows a 
considerable rise in 1928 and 1929 over the first two years. In both July and August 
the year 1926 is quite high, and the last three years are about equal. In September, 
1929 is at such a high plane as to indicate great abundance, but this is doubtless due 
in part to the absence of data during September in 1926, since the high 1929 year is 
compared to a standard curve considerably lower than would surely be the case were 
1926 included. 
Figube 15. — Showing for Group I the geometric means (expressed as per- 
centages of the 1928 mean) of the percentages that the average weekly- 
catch per boat is of a standard curve (see text) 
