518 



Fishery Bulletin 94(3), 1996 



1994; Quinn and Foote, 1994) and may have reduced 

 the effective population size further. 



Ancestral population of New Zealand 

 Chinook salmon 



Although the genetic and historical data clearly point 

 to a seeding of NZ chinook from a Sacramento River 

 population, identification of the specific ancestral 

 population remains difficult. In addition to possible 

 genetic changes in NZ chinook salmon over the years, 

 the California populations have also undergone com- 

 plex changes in abundance, habitat, and manage- 

 ment since the turn of the century. Moreover, the 

 seasonal runs are not monophyletic races but have 

 apparently evolved independently in suitable habi- 

 tats; hence major ancestral groups are generally 

 based more on geography than on the timing of a 

 run (Utter et al., 1993). 



Aside from confirming a Sacramento River origin, 

 the allozyme data provided no clues about the an- 

 cestral population. Both the absence of distinct alle- 

 les or of distinguishing frequencies of common alle- 

 les among contemporary temporal segments within 

 the Sacramento River and the high genetic drift as- 

 sociated with the presumed bottlenecking in NZ pre- 

 cluded resolution from these nuclear variants. The 

 mtDNA D-loop data, which theoretically provide 

 higher resolving capabilities owing to higher muta- 

 tion rates and greater population divergence of the 

 haploid genome (Brown et al., 1982; Hoelzel et al., 

 1991), gave some clues but no definitive answers 

 about the ancestral population (Table 7). The rate of 

 molecular base changes at any one locus can vary 

 according to the number and size of colonies, migra- 

 tion rates, sex ratios, type of parental transmission, 

 as well as the expected rate of genetic drift (Birky et 

 al., 1989; Martin and Palumbi, 1993). The haplotype 

 profile of the Sacramento River fall-run fish re- 

 sembled that of the combined NZ fish, sharing four 

 haplotypes ( 1 and 4, 2, 3, 5). In contrast, the winter- 

 run fish were the most distinct, lacking haplotypes 

 2, 3, and 5, and were represented by a unique haplo- 

 type 6 in two individuals. The spring-run sample 

 contained haplotype 3 at a significantly higher fre- 

 quency than that found for any NZ populations but 

 shared a similar frequency profile for haplotype 5 

 with the combined NZ sample. 



The genetic signatures of the spawning runs of 

 chinook salmon found in the Sacramento River have 

 probably changed since the transfers to NZ at the 

 turn of the century owing to population bottlenecks 

 (particularly the winter-run), hatchery manipulation, 

 impacts from fisheries, changes in habitat, and 

 changes in the thermal regime leading to overlap in 



temporal spawning populations. Prior to develop- 

 ment of the Sacramento River system, Battle Creek 

 had fall and spring runs of chinook salmon, and the 

 current populations in the hatchery may have 

 introgressed to some extent or been altered by inter- 

 change with chinook from other populations, despite 

 efforts to maintain the runs separately Cope and 

 Slater (1957) noted that changes in the river's tem- 

 perature regime delayed the arrival of spring-run fish 

 at Battle Creek and that "This behavior, together 

 with the fact that these two runs were forced to spawn 

 on the same riffles, with the blocking of the river at 

 Keswick Dam, is presumed to have brought about 

 some mixing of the two stocks . . . when their spawn- 

 ing periods overlap in September. ... At the hatch- 

 ery . . . the dividing point between the spawning sea- 

 sons of the two runs has been more or less arbitrarily 

 set at September 25." Contemporary chinook salmon 

 runs are defined by the timing of adult migration up 

 the Sacramento River (Healey, 1991; Fisher 8 ): fall- 

 run (September-October), winter-run (November- 

 February), and spring-run (March-May). These runs 

 also differ in their spawning periods: fall-run (Octo- 

 ber-December), winter-run (May^June), and spring- 

 run (August-September). Currently, there is also a 

 putative "late fall-run," spawning in winter; 

 unsampled in this study, these fish differed margin- 

 ally from the fall-run in haplotype frequencies 

 (Nielsen et al., 1994b). According to temporal crite- 

 ria, the type of chinook embryos sent to NZ at the 

 turn of the century were probably fall-run fish be- 

 cause Thomson (1922) reported that the shipments 

 from California arrived ". . . early in January." Both 

 the spawning season of the winter-run populations 

 (late spring to early summer) and their mtDNA hap- 

 lotype distribution indicate that they did not found 

 the NZ chinook runs. 



Some of the allozyme data raise additional ques- 

 tions about the ancestral population. Allelic variants 

 at two loci occurred in two of the three NZ collec- 

 tions that were either rare (HAGH*65) or did not 

 occur in any Sacramento River fish {GR*110). The 

 occurrence of both of these variants in two popula- 

 tions (Rakaia and Waimakariri) and at frequencies 

 as high as 0.135 (HAGH*65) minimizes the possibil- 

 ity of mutation after colonization. More likely, the 

 variants were present in the founder population. 

 If so, these alleles have drifted to a very low frequency 

 or out of the sampled California populations or they 

 were present in unusually high proportions in the 

 founder population in NZ or in the two rivers in 

 particular. 



K Fisher, F. 1995. Calif. Dep. Fish and Game, Red Bluff. 

 CA. Personal commun. 



