PACIFIC HERRING 
239 
Island purse seining was first carried on in 1922 in Raspberry Strait and Izhut Bay. 
Since 1923 it has been largely confined to Shuyak Strait. 
In western Alaska a very small gill-net fishery was established on Simeonof 
Island, one of the Shumagin group, in 1906. About the same time a small fishery 
was commenced near by at Chignik on the Alaska Peninsula. These two minor 
fisheries have continued up to the present. A small fishery has been carried on at 
Golovin Bay, in the northern part of the Bering Sea, since before 1909. However, no 
extensive fishery existed in western Alaska until 1928, when about half of the central 
Alaska purse-seine fleet fished at Unalaska (Dutch Harbor) in the Aleutian Islands. 
(Figs. 10 and 11.) 
COLLECTION AND DESCRIPTION OF SAMPLES 
SELECTION OF THE SAMPLE 
In the collection of samples an effort was made to obtain a truly representative 
sample from each load. Care was taken not to select the sample in such a manner as 
to influence the size of the fish that were to be contained therein. The usual procedure 
was to hold a bucket under the fish elevator in such a manner that the fish dropped 
into it without any voluntary selection. (Fig. 12.) This was always done after the 
fishermen had shoveled off the top of the load since there is a tendency for the larger 
fish to rise to the top of the load and for the smaller fish to sink to the bottom. This 
tendency appears to act only upon the upper and lower few inches of the load, so that 
by taking samples from the middle of the load one does not incur the danger of under- 
representing the extreme sizes. In localities where the fish in the same load tend to 
cover a large range of sizes, samples of about 100 fish were usually taken. Where the 
range of sizes in the same load tended to be small, but the average sizes of fish of 
different loads varied considerably, then smaller samples of about 25 to 50 fish were 
taken, and an endeavor was made to sample more loads. 
DATA TAKEN ON EACH SAMPLE 
Having collected a random sample from a load of herring, the next step was to 
lay the herring in a row. Then a scale sample was taken from each fish from the 
middle of the side below the dorsal insertion. The scales were preserved in serially 
numbered coin envelopes and the corresponding number given each fish when its 
measurements were entered on the data sheet. 
The fish were next measured, the measurements all being taken parallel to the 
body axis from a plane at right angles to the body axis at the tip of the mandible 
with the mouth closed. These were read in millimeters on an improved model of the 
fish-measuring board employed by Thompson (1917) in his investigation of the herring 
of British Columbia and later modified by Thompson (1926, p. 60), Elmer Higgins, 
and the author for taking sardine measurements at the California State Fisheries 
Laboratory (see fig. 13). 
In making the measurements the wire on the cross arm of the measuring board 
was invariably aligned with its own reflection in the mirror. This always insured the 
eye being held vertically above the wire. As soon as each fish had been measured it 
was weighed on' a spring balance graduated to 2 grams, with a capacity of 500 grams. 
After the measurements were all completed, the rays in the dorsal and anal fins were 
counted and the sex was then determined. The fish was reweighed with the entire 
