514 



Fishery Bulletin 94(3), 1996 



sented), and a neighbor-joining tree (Fig. 4) also dis- 

 played similar clustering, that separated the NZ and 

 Sacramento River populations. 



Relationships from the subset of 10 polymorphic 

 loci (Fig. 3B) paralleled those indicated with 24 loci 

 (Fig. 3A). Each of the additional three samples from 

 Utter et al. (1989) were grouped within the cluster- 

 ing of Sacramento River collections reported by 

 Bartley et al. (1992). Notably, the spring-run collec- 

 tion from the Feather River Hatchery was indistinct 

 from fall-run collections from the Feather River and 

 from other Sacramento River hatcheries. Further- 

 more, the stability of allele frequencies at these loci 

 was evident for fall-run fish collected as juveniles in 

 1982 (Utter et al., 1989) and 1987 ( Bartley et al., 1992 ). 



Decreasing heterogeneity among the 24 loci was 

 apparent when chi-square tests within subgroups 

 were contrasted with those involving all collections 

 (Table 4). Within the total grouping, all but two loci 

 were heterogeneous (P<0.01). A lower level of sig- 



nificance occurred for 9 loci within the Sacramento 

 River and for all but one locus among the NZ collec- 

 tions. Standardized measures of these differences 

 indicated more than a doubling of heterogeneity for 

 the Sacramento subgroup relative to the NZ sub- 

 group, and a near doubling of the total heterogene- 

 ity relative to the Sacramento collections. 



Much of the total heterogeneity reflected the dis- 

 tinction of the South Fork Eel River collection from 

 the remaining group, notably owing to allele fre- 

 quency differences approaching or exceeding 0.4 at 

 GPI-B2* and GPIr*, and to only slightly lower dis- 

 tinctions at sMEP-1*, and MPI* (see Table 3). Allelic 

 differences between the Sacramento and NZ sub- 

 groups were more subtle. Nonoverlapping frequen- 

 cies at many loci (GPI-B2*, GPIr*, IDDH 2«, IDH- 

 2*, PGM-2*, PGK-2*, TPI-2.1* ) contributed to their 

 differentiation. In addition, some variants in the NZ 

 subgroup were either absent in all the Sacramento 

 subgroups \GR*) or occurred at a very low frequency 



