18 U. S. BUREAU OF FISHERIES 



of so large a number of 3-year-olds gives promise of exceptionally 

 good returns from the experiment. The fish in another sockeye 

 experiment, in which the young fish were liberated during the fall 

 of their first year, should have been 5 years old in 1925, but no 

 returns were received. This experiment seems to have been a total 

 failure, as no 3 or 4 year olds were recovered. These results agree 

 with those from former experiments in showing that when it is not 

 possible to liberate sockeyes in a lake in which they can remain until 

 the spring of their second year (which is their natural time of migra- 

 tion), they should be held in rearing ponds until that time. 



About 50 of the marked chinooks recovered were 3-year-olds that 

 had been marked at the Big White' Salmon River hatchery during 

 May and June, 1923, when they were about 5 months old. Here, 

 also, the recovery of a large number of 3-year-olds gave promise of 

 good returns as 4 and 5 year olds. This one year's returns from 

 this experiment nearly equal the total returns from two former ex- 

 periments, which differed from this one mainly in that the finger- 

 lings were retained for about three months longer before being liber- 

 ated. The greater returns from this experiment would indicate that 

 the best time to liberate the chinooks that spawn in the Little 

 White Salmon and Big White Salmon Rivers is during the spring of 

 their first year. This is to be expected, for a study of the scales of 

 the fish that spawn in those tributaries has shown that normally they 

 migrate to the ocean within a short time after hatching. 



In order to determine the best time to liberate fingerlings of the 

 spring run of chinook salmon, a series of marking experiments 

 was undertaken in 1925 at the McKenzie River hatchery. This 

 series consisted of five markings with liberations ranging from May of 

 the first year to March of the second. 



In addition to conducting the marking experiments on the Colum- 

 bia River, H. B. Holmes has been making a study of the blueback 

 salmon of the Columbia. Representative samples of scales and meas- 

 urements were collected from the commercial catch at intervals 

 throughout the season, and observations were made of the seaward 

 migrants and the spawning adults in the Okanogan River spawning 

 district. 



HERRING OF ALASKA 



In the spring of 1925 a biological investigation of the herring of 

 Alaska was begun by George A. Rounsefell, scientific assistant. 

 During the summer a preliminary survey was made, and as many 

 herring establishments as was possible were visited in the season, 

 touching at Ketchikan, Craig, Killisnoo, and Port Walter in south- 

 eastern Alaska; Sawmill Bay, Prince William Sound, Seldovia, Cook 

 Inlet, and Red Fox Bay; and Kodiak and Three Saints Bay in the 

 Kodiak district. As a necessary preliminary to other work, a study 

 was begun to determine what local races exist and the degree of vari- 

 ation between them. Rticial samples were collected at many of 

 the points visited, and various structural characters were studied. 

 These data show clearly that the Pacific herring is not a homoge- 

 neous population. 



The average vertebral count falls from 52. 78 in Prince William 

 Sound and 52.72 in Shuyak Strait to 52.50 in Cook Inlet, 52.45 in 



