PACIFIC HERRING 
303 
The racial analysis shows that these two localities do not differ significantly in 
any of the characters compared, their condition factors (Table 29) follow almost 
exactly the same trend, and their rates of growth (Table 28) are comparable, so 
there is no reason for supposing them to be separate stocks. 
CONCLUSIONS 
1. Owing to the short periods during which samples are obtainable from one 
locality, the proof of the validity of sampling must be placed largely upon the 
repeated occurrence of consistent variations. 
2. Dominant year classes are normally present. 
3. The progression of sizes due to the growth of dominant year classes is 
reflected in the commercial catch. 
4. The similarity or difference between the dominant year classes in two adjacent 
localities gives indications for or against the independence of the two populations. 
ANALYSIS OF CATCH STATISTICS 
SOURCES OF STATISTICS 
The statistics which have been used in studying the changes in abundance have 
been derived from a number of fairly reliable sources. For the very early years of 
the fishery (previous to 1904) we have had to rely wholly on published records. Moser 
(1899 and 1902) and the United States Senate (1912) published the best records of 
the Killisnoo plant, and Cobb (1906) summarized ail of the early records obtainable. 
Since 1904 the Bureau of Fisheries has required every individual or company 
fishing in Alaska to make a sworn annual return of the total amounts and kinds of 
fishery products prepared, and of the amounts, kinds, and value of fishing gear, 
boats, and other apparatus used. These sworn returns constitute one of the main 
sources of information. Another major source of information is the annual statistical 
review and the monthly numbers of the Pacific Fisherman, a trade journal published 
in Seattle, Wash. Many of the herring companies have kept careful records of 
their catches for several years, to which we have had access. A fifth source of 
information, available since 1926, is contained in detailed records giving the amount, 
date, and location of every catch made by each individual boat. These records are 
kept by the herring companies on duplicate receipt books issued by the Bureau of 
Fisheries, a system which we patterned after that planned bj^ Will F. Thompson 
and used successfully for several years by the Division of Fish and Game of Cali- 
fornia. These sources have been supplemented by field notes. 
TREATMENT OF DATA 
The records do not give the quantities of raw herring captured (except in the 
case of halibut bait), but give the amounts of various finished products prepared. 
In analyzing the statistics, it was necessary for purposes of comparison that all 
amounts be put on a common basis. The unit selected was the pound of raw herring 
as delivered to the plant. Some of the factors used in converting the weights of 
finished products into the weights of raw herring were more or less empirically 
determined, which may have allowed some small errors to creep in. However, the 
advantages of such a method of treatment are obvious, and such errors as may have 
arisen as a result are too small to have any appreciable effect on the analysis. For 
the conversion of canned herring into raw, 75 pounds of raw herring have been 
allowed for every 48 pounds of canned herring. No statistics are available on the 
