Quinn et al Origin and genetic structure of Oncorhynchus tshawytscha 



515 



New Zealand 

 rivers 



and not in the Battle Creek population 

 (HAGH*). 



Mitochondrial DNA: genetic 

 variation among populations 



There were no new base substitutions 

 found in the 171-bp section of the mtDNA 

 control region of the NZ chinook sequence 

 that we sequenced, compared with mtDNA 

 types reported for Sacramento River 

 chinook (Table 1 in Nielsen et al., 1994b). 

 However, because the base change re- 

 corded at bp 183 contained a cryptic de- 

 letion that was difficult to score consis- 

 tently in the NZ fish, we pooled mtDNA 

 types 1 and 4 (as previously reported), 

 which were originally differentiated by 

 a single deletion at bp 183. The same 

 cryptic nucleotide deletion was found in 

 haplotype 3 in the NZ chinook; therefore, 

 the base cytosine (C) and the deletion 

 found at this site were pooled for all popu- 

 lations, yielding five mtDNA types used 

 in this analysis (Table 6). 



The NZ salmon populations were domi- 

 nated by mtDNA haplotype 1, represent- 

 ing 82% in the pooled sample compared 

 with 69% in the Sacramento River popu- 

 lation (Table 7). Haplotype 2 was rare in 

 the NZ fish (only one sample was found, from the 

 Waimakariri River) and detected only in fall-run fish 

 from the Sacramento River. Haplotype 3 was detected 

 at low frequencies in NZ and in the Sacramento River 

 fall-run fish. This haplotype was found at higher fre- 

 quencies in the 1994 spring-run samples taken at 

 Butte Creek and not at all in the Deer and Mill 

 Creeks spring-run collection (1991-93) or the win- 

 ter-run California chinook. Fish carrying haplotype 

 5 in the Clutha and Waitaki NZ chinook shared a 

 close frequency distribution to that observed in the 

 Sacramento River fall-run fish. The frequency dis- 

 tribution of haplotype 5 in the Sacramento River fall- 

 run population was primarily due to an increase in 

 its relative abundance in one year class ( 1994). Hap- 

 lotype 6 was found only in the winter-run Sacramento 

 fish and not in any NZ population. 



The Fisher's exact tests, which compared the total 

 NZ chinook population with each of the three Sacra- 

 mento River runs of chinook were compared, indi- 

 cated that the NZ chinook differed significantly from 

 the winter- and spring-run fish (P<0.001) but not 

 from the fall-run samples (P=0.08). Haplotype fre- 

 quencies varied among the four NZ populations, but 

 the Clutha River population, with only two mtDNA 



Sacramento 

 River drainage 



Genetic distance 



Figure 4 



Neighbor-joining tree (Saitou and Nei, 1987) displaying relationships 

 among chinook salmon populations from California and New Zealand 

 based on a matrix of genetic distance measures (Cavalli-Sforza and 

 Edwards, 1967; Table 5). Dashed circles enclose the three New Zealand 

 and four Sacramento River populations. 



variants (1 and 5), was the only population to differ 

 (P<0.01) from any other population (Rakaia; Table 7). 



Discussion 



Contrasts among population groups 



The allozyme and mtDNA data sets provided comple- 

 mentary and independent comparisons for consider- 

 ing the origin and differentiation of the NZ chinook 

 salmon populations. At the broadest level, the Sac- 

 ramento and NZ populations share a common lin- 

 eage (Figs. 3 and 4). The collection from the Eel River, 

 though geographically adjacent to the Sacramento 

 drainage, is typical of a distinct coastal lineage, dis- 

 tinguished by high frequencies of GPI-2* variation 

 and monomorphism for GPIr* (Utter et al., 1989; 

 Bartley and Gall, 1990; Bartley et al., 1992). At a 

 finer level, some details of the separate Sacramento 

 and NZ subgroups were apparent from both data sets. 

 The most obvious distinction was the greater degree 

 of divergence among the Sacramento populations on 

 the basis of comparative genetic distances (Table 4; 

 Fig. 3), allelic heterogeneity and partitioning of gene 



