510 



Fishery Bulletin 94(3), 1996 



ery) reported in Utter et al. ( 1989) were also included 

 1 ) to provide data for spring-run fish from the Feather 

 River Hatchery that were not examined by Bartley 

 et al. (1992) and 2) for comparison among different 

 year classes of fall-run fish. Genetic nomenclature 

 followed Shaklee et al. (1990). 



Genetic data for NZ and California populations 

 were analyzed with the BIOSYS-1 program 6 to cal- 

 culate pairwise measures of genetic distances 

 (Cavalli-Sforza and Edwards, 1967; Nei, 1972), gene 

 diversity (F ST , Wright, 1969), chi-square comparisons 

 of heterogeneity, and to construct phenograms 

 through the unweighted pair-group method (Sneath 

 and Sokal, 1973 ). A neighbor-joining tree ( Saitou and 

 Nei, 1987) was constructed from a matrix of arc dis- 

 tances (from Cavalli-Sforza and Edwards 11967]) 

 through the NTSYS-pc program (Rohlf, 1994). Con- 

 formance of genotypic proportions of NZ collections 

 to those expected under Hardy-Weinberg (binomial) 

 equilibrium was tested by Levene's (1949) formula 

 for small sample size; the variant alleles were pooled 

 in tests involving loci with more than two alleles. 



mtDNA analysis 



Total genomic DNA was extracted from a small sec- 

 tion of caudal fin tissue from 172 juvenile NZ salmon 

 with Chelex-100 (Walsh et al., 1991 ) according to the 

 protocol in Nielsen et al. ( 1994a). A 2-pL aliquot of 

 the chelex-treated supernatant containing salmon 

 DNA was used as a template for amplification with 

 the polymerase chain reaction (PCR) and conserved 

 primers. Our PCR protocol used primers (S-phe and 

 P2; sequences in Nielsen et al., 1994a) that amplify 

 a highly variable segment of the mtDNA control re- 

 gion in salmonids. Amplification, purification, and 

 sequencing of salmon mtDNA were done according 

 to the protocol in Nielsen et al. ( 1994a). Sequencing 

 reactions were separated in linear 9% polyacryla- 

 mide-7 M urea gels and were autoradiographed for 

 24 to 72 h at room temperature. Mitochondrial DNA 

 sequences were scored with base-pair (bp) differences 

 found within the amplified control region sequence 

 (Nielsen et al., 1994a). 



Data on NZ chinook salmon mtDNA sequence fre- 

 quencies were compared with data from Sacramento 

 River chinook salmon with known spawning seasons 

 (Nielsen et al., 1994b) as follows: winter-run: 72 fish 

 taken from 1991 to 1993 broodstock program at the 

 Coleman Hatchery; fall-run: 359 fish taken from 1992 

 to 1994 at hatcheries on the American, Merced, and 



Feather rivers; and spring-run: 32 wild fish collected 

 in Deer and Mill creeks ( 1991-93) and 27 adults from 

 the Butte Creek spring-run population (1994). 



To test for differences in mtDNA frequency among 

 the NZ chinook salmon populations and between the 

 NZ and Sacramento River runs we used an unbiased 

 estimate of Fisher's exact test for population differ- 

 entiation with a Markov chain analysis (GENEPOP 7 ). 

 The Markov model in GENEPOP was run with 4 

 seeds for the pseudo-random number generator, and 

 the dememorisation number was set to 1,000, the 

 number of batches to 50, and with the number of it- 

 erations/batch to 1,000. 



Results 



Protein-coding loci: genetic variation within 

 populations 



Tests for conformance to Hardy-Weinberg genotypic 

 proportions in the NZ populations indicated no de- 

 viation from proportions expected in a random mat- 

 ing population. Two tests out of 42 deviated signifi- 

 cantly from expected proportions at the 5% level of 

 significance; deficiencies of heterozygotes were ob- 

 served for sAH* in the Waimakariri and for PEPA* 

 in the Waitaki populations. This proportion of de- 

 viations (0.048) would be expected by chance at this 

 level of significance. Similar findings were reported 

 by Bartley et al. (1992) for wild California popula- 

 tions, although somewhat higher rates of deviation 

 occurred among hatchery samples. No deviations 

 with Sacramento River samples were reported in 

 Utter et al. ( 1989). Levels of genetic variation were 

 generally higher in the Sacramento River collections 

 than in the NZ or South Fork Eel River samples 

 (Table 2). The highest mean number of alleles per 

 locus, percentage of loci polymorphic, and mean het- 

 erozygosity were seen in the Sacramento River col- 

 lections, and no overlap in percentage of polymor- 

 phic loci occurred between these and any of the other 

 collections. 



Protein-coding loci: genetic variation among 

 populations 



Allele frequencies at 24 polymorphic protein-coding 

 loci ( Table 3 ) varied considerably among populations. 

 The gene diversity among populations (F ST ) averaged 



6 Swofford, D. L, and R. B. Selander. 1989. BIOSYS-1: a com- 

 puter for the analysis of allelic variation in population genetics 

 and biochemical systematics, release 1.7. Illinois Natural His- 

 tory Survey, 43 p. 



7 Raymond, M, and F.Rousset. 1994. GENEPOP, version 1.0, 

 January 1994. Available through M. Raymond, Laboratoire 

 de Genetique et Environnement, URA CNRS 327, Place E. 

 Bataillon, 34095 Montpellier cedex 05, France. E-mail: 

 Raymond@univ-montp'2.fr. 



