FISHERY BULLETIN: VOL. 82, NO. 4 



drift) or, 4) some combination of the above. Un- 

 fortunately, the short duration of most elec- 

 trophoretic studies and/or inadequate sample 

 sizes for various age classes often prevent or limit 

 investigations of such age-dependent changes in 

 allele frequency. No year class heterogeneity in 

 transferrin allele frequencies was detected in 27 

 consecutive year classes of Atlantic cod by 

 Jamieson ( 1975). Similarly, no year class variation 

 was reported for mackerel EST by Smith et al. 

 (1981). On the other hand, changing patterns of 

 allele-frequency distribution with increasing age 

 (size) have been noted in several fish species, in- 

 cluding the blenny, Anoplarchus purpurescens 

 (Johnson 1971), the eelpout, Zoarces viviparous 

 (Christiansen et al. 1974), the mummichog, Fun- 

 dulus heteroclitus (Mitton and Koehn 1975), the 

 plaice, Pleuronectes platessa (Beardmore and 

 Ward 1977), and several New Zealand commercial 

 fishes (Smith 1979a; Gauldie and Johnston 1980). 

 In many of the latter cases, it has been suggested 

 that natural selection is the underlying cause of 

 the shift in allele frequencies with age. Our analy- 

 sis of the pink snapper data (Table 4) revealed that 

 both alcohol dehydrogenase and Ldh-C exhibited 

 significant differences in allele frequency between 

 young and old fish. However, we do not have any 

 direct information that these changes are due to 

 natural selection. 



There have been few electrophoretic studies of 

 commercial fishes which have addressed the ques- 

 tion of temporal stability of allele frequencies. In 

 the American eel, Anguilla rostrata, data for 4 

 successive years revealed that patterns of genetic 

 differentiation were generally unchanged 

 through time (Koehn and Williams 1978). On the 

 other hand, studies of pink salmon (Aspinwall 

 1974) have demonstrated substantial temporal 

 heterogeneity in allele frequencies due to the exis- 

 tence of distinct spawning stocks isolated in time. 

 Temporal differences in gene frequencies related 

 to different spawning stocks have also been 

 suggested by Kornfield et al. (1982) for Atlantic 

 herring. The present analysis of opakapaka allele 

 frequencies through time (Table 5) revealed no 

 temporal differences. 



There can be no doubt that the Hawaiian Is- 

 lands offer a sharply discontinuous distribution of 

 adult habitats for Pristipomoides fHamentosus 

 given that the species is almost completely re- 

 stricted to water depths of 50-200 m in Hawaii 

 (Ralston 1980) (200 m contour lines in figure 1 of 

 Shaklee and Samollow 1984). Furthermore, there 

 is no reason to believe that adult pink snapper — 



strictly demersal fish — migrate through the open, 

 oceanic waters between islands. It is therefore 

 somewhat surprising that no evidence of genetic 

 subdivision of opakapaka among localities within 

 the Archipelago was observed. This seems particu- 

 larly surprising given that there are several re- 

 ports of stock heterogeneity in other demersal 

 marine fishes such as Atlantic cod (Jamieson 1975; 

 Cross and Payne 1978; Jamieson and Turner 1978), 

 walleye pollock (Grant and Utter 1980), New Zea- 

 land snapper (Smith et al. 1978), and two species of 

 flatfishes (Fairbairn 1981a, b). The genetic 

 homogeneity observed for pink snapper in the pres- 

 ent study would seem a singular exception when 

 compared with the above results were it not for the 

 fact that numerous other demersal species exhibit 

 similar patterns of apparent genetic homogeneity, 

 e.g., plaice (Purdom et al. 1976; Ward and 

 Beardmore 1977), hake (Smith et al. 1979), ling 

 (Smith 1979b; Smith and Francis 1982), and New 

 Zealand hoki (Smith et al. 1981). However, it must 

 be remembered that the embryonic and larval 

 stages of P. filamentosus and most, if not all, of 

 these other species are pelagic and, therefore, 

 serve as a dispersal phase. Given that this pelagic 

 stage apparently lasts for 1-2 mo in pink snapper, 

 it would not seem unreasonable that larval disper- 

 sal among localities due to wind-driven currents 

 within the Hawaiian Islands would be of sufficient 

 magnitude to ensure adequate gene flow among 

 adult populations to prevent detectable genetic 

 differentiation (Lewontin 1974). 



If this interpretation is correct, it would mean 

 that although the total pink snapper harvest in 

 Hawaii is derived from numerous small 

 geographically separated fisheries, each as- 

 sociated with discontinuous patches of suitable 

 habitat (islands, reefs, and banks), the pink snap- 

 pers themselves are all members of a single, large 

 panmictic population distributed throughout the 

 Hawaiian Archipelago. Based on the above infor- 

 mation, it seems appropriate to manage the pink 

 snapper fishery in Hawaii as a single unit stock — 

 including both the main Hawaiian Islands and 

 NWHI. However, since the observed genetic 

 homogeneity does not unambiguously establish 

 the existence of a unit stock (it is merely consistent 

 with this hypothesis), this management policy 

 must remain open to reevaluation in the future 

 especially if data of another kind should suggest 

 stock heterogeneity. 



Genetic aspects of population structure have 

 now been studied in four marine species through- 

 out the Hawaiian Archipelago. In three of these 



710 



