from 2-year-old jacks to 5-year-old males 

 and females with an occasional fish maturing 

 after its fifth year. In lieu of any specific age 

 data on the returning fish it is important to 

 develop an index of production which will be 

 least affected by this variation. The modal age 

 at return is usually considered to be 4 years, 

 but the numbers returning at other ages could 

 introduce significant bias if it is assumed 

 that production of any brood year may be 

 measured by the return 4 years later. Some 

 variation could be eliminated by omitting the 

 jack returns, but classification of jacks may 

 be somewhat arbitrary and the number of 

 3-year-olds, particularly males, may be ex- 

 cessive. 



In the present study, production for each 

 brood year has been measured by the number 

 of females returning 4 years later.' Holmes* 

 lists the age of returns of marked fish of 

 Spring Creek stock in his study of fingerling 

 Chinook salmon mortalities at Bonneville Dam. 

 Omitting jacks but not older males, approxi- 

 mately 15 percent returned as 3-year-olds, 

 70 percent as 4-year-olds, and 15 percent as 

 5-year-olds. The sex of these fish is not in- 

 dicated, but if no less than 70 percent of the 

 females return as 4-year-olds, any important 

 change in production of a brood year should 

 be reflected by returns of females 4 years 

 later. Scale samples taken by Oregon Fish 

 Commission in the 1959 Columbia River gill 

 net fisnery showed 75 percent of the females 

 were 4-year-olds as compared with only 51 

 percent 4-year-old males. Length dataonl959 

 returns to several hatcheries do not confirm 

 this result, and this anomaly will be discussed 

 in a later paragraph of this section. 



It is unfortunate that more data on age dis- 

 tribution of fish returning to Spring Creek 

 Hatchery are not available, but some length- 

 frequency graphs of Little White Salmon 

 stocks^ are of interest (fig. 1). The mode has 



' The use of female returns lagged 4 years was sug- 

 gested by R, E. Noble for the study of two hatcheries 

 operated by the Washington Department of Fisheries. 



■* H. Holmes. Loss of salmon fingerlings in passing 

 Bonneville Dam as determined by marking experiment?. 

 Manuscript. U.S. Fish and Wildlife Service, 1952. 



^ H. E. Johnson. U.S. Fish and Wildlife Service, 

 unpublished data. 



been placed arbitrarily at 19.5 inches for 

 2-year-olds, 31.5 inches for 3-year-olds, and 

 37,5 inches for 4-year-olds. These values are 

 in reasonable agreement with modal length of 

 marked returns released as fingerlings. The 

 size distribution for adult males returning in 

 1955, 1956, and 1957 illustrates the dominance 

 of the 1953 brood, which returned as 2-year- 

 olds in 1955, 3-year-olds in 1956, and 4-year- 

 olds in 1957. In this respect, with a 4-year 

 lag the total male returns in 1956 would in- 

 correctly classify the 1952 brood as the 

 dominant producer. The size distribution for 

 females, however, shows no such pattern 

 though there is an indication of an increased 

 return of 3-year-olds in 1956. The total num- 

 ber of females returning in 1957 correctly 

 classifies the 1953 brood as the very dominant 

 producer if a 4-year lag is used. 



Recent evidence of an increased return of 

 3-year-old fall chinook to Columbia River 

 hatcheries must be examined with respect to 

 use of a 4-year lag on female returns as an 

 index of adult production for earlier years. In 

 the remainder of this section we will present 

 evidence for this change, consider indications 

 that only returns in recent years are affected 

 by this change, and note the probable effects 

 of this change on the present analysis. 



From the size distributions of both male and 

 female adults returning to Little White Salmon 

 Hatchery in 1958 and 1959 (fig. 1) it appears 

 that the modal age of return has shifted from 

 4 years to 3 years for the 1955 and 1956 brood 

 years. The similarity of the distributions in 

 1958 and 1959 indicates that the peaks at the 

 3-year level in 1958 did not occur due to a 

 dominant 1955 brood, otherwise large peaks 

 should appear as 4-year-olds in 1959. 



It might be postulated that the position of the 

 modes in 1958 and 1959 could have been in- 

 fluenced by release of large numbers of year- 

 ling Chinook of the 1954 and 1955 brood years. 

 Four-year-old fish released as yearlings would 

 be expected to be similar in size to 3-year-old 

 adults from fry or fingerling releases, since 

 effective growing time in salt water would be 

 nearly the same. At Spring Creek Hatchery, 

 however, no yearling releases were made 

 during this period, and the similarity between 



