FISHERY BULLETIN: VOL. 86, NO. 4 



are outlined in Beacliam et al. (1985). Lgg was re- 

 named as Tapep and was scored as a four-allele locus 

 (L. Seeb^) (although only three were present in the 

 stocks we surveyed) on both a Tris-boric acid-EDTA 

 buffer described by Markert and Faulhaber (1965) 

 and an amine citrate buffer described by Clayton 

 and Tretiak (1972). Me was renamed as MdhP, 6-Pg 

 as Pgdh, Pmi as Mpi, and Agp as GSpdh. 



We determined allelic frequencies for each locus 

 by summing the numbers of each allele and dividing 

 by the total number of alleles counted. Genotypic 

 frequencies at each polymorphic locus in each popu- 

 lation were tested for departures from Hardy- 

 Weinberg equilibrium by chi-square. We used the 

 log-likelihood ratio statistic (G-test) (Sokal and RohLf 

 1969) to test equality of allelic frequencies between 

 countries, among populations within countries, and 

 between years for samples taken from the same 

 population. An approximate i^-ratio (G-statistic 

 summed over all loci/degrees of freedom) was used 

 to test the relative magnitude of the sources of varia- 

 tion. We calculated genetic distance among popula- 

 tions using Nei's (1978) statistic and the 7 loci in- 

 dicated in Table 7 and constructed a denogram from 

 the matrix of the distances using the unweighted 

 pair group mean method of Sneath and Sokal (1973). 



RESULTS 



Age and Morphology 



Age of Maturity 



The dominant age of maturity for fall chum 

 salmon returning to Canadian rivers during 1984 

 and 1985 was four years, with age 3 and age 5 chum 

 salmon each comprising less than 15% of the total 

 return (Table 1). The proportion of chum salmon 

 returning at three or five years of age varied an- 

 nually within a population, possibly reflecting dif- 

 ferent production from the respective brood years. 

 Yukon River chum salmon matured at a significantly 

 older age than chum salmon in British Columbia 

 (Xz^ = 193.9, P < 0.01), but four years was the 

 dominant age of maturity in both areas. Of 227 chum 

 salmon sampled in 1984 from which age could be 

 determined from both otoliths and scales, the same 

 age was recorded in 184 (83%) of the cases. For the 

 43 other fish, the age estimated from scales was one 

 year older than that estimated from otoliths in 22 

 cases, whereas in the remaining 21 fish, the age 



estimated from otoliths was either one year (19 

 cases) or two years (2 cases) older than that esti- 

 mated from scales. In 1984, age could not be deter- 

 mined from 16% (48 fish) of the scales collected and 

 9% (26 fish) of the otoliths collected. In 1985, age 

 could not be determined from 18% (124 fish) of the 

 scales examined due to scale resorption. Age of 

 maturity was more likely to be determined from 

 otoliths than from scales. 



Table 1 . — Percentage of chum salmon returning at ages 3-5 years 

 and mean age of return for Yukon River populations sampled on 

 thie spawning grounds during 1984-85. The mean for British 

 Columbia chum salmon was derived from Beacham and Murray 

 (1987). 



^University of Idaho, Moscow, ID 83843, pars, commun. 

 December 1985. 



Meristics 



No sexual dimorphism was observed in the num- 

 ber of gill rakers or branchiostegal rays (P > 0.05). 

 No population differences in gill raker number were 

 observed (P > 0.05), but significant differences 

 among populations were observed in branchiostegal 

 ray number (^2,297 = 4.36, P < 0.05). The popula- 

 tion with the greatest number of gill rakers (Fishing 

 Branch) also had the greatest number of branchi- 

 ostegal rays (Table 2). Yukon River chum salmon 

 had more gill rakers and branchiostegal rays than 

 the average British Columbia chum salmon. 



Morphometry 

 Yukon River male chum salmon were longer than 



666 



