776 



Fishery Bulletin 90(4|, 1992 



sive geographic separation of the two regions (Lewon- 

 tin and Hubby 1966). A subsequent study that found 

 previously-unknown genetic variants (Singh et al. 1976) 

 demonstrated clear genetic differences between popu- 

 lations of each region. 



The important message here is to beware of the 

 danger of drawing positive conclusions from negative 

 data. It should also be emphasized that problems of this 

 nature are not confined to genetic data; rather, the 

 limitations of nondiscriminatory information (i.e., the 

 power to reject the null hypothesis) should be con- 

 sidered in evaluating any kind of comparative data for 

 two or more samples. 



Similar allele frequencies among samples, then, sup- 

 port but do not confirm hypotheses that the samples 

 are drawn from a common breeding group. This well- 

 established principle requires restatement from time 

 to time (e.g.. Utter 1981, Waples 1991). Such aware- 

 ness serves to safeguard against a premature conclu- 

 sion of identity for groups that are distinct and thus 

 may be subject to different management criteria. 



In these instances it is important to recognize the 

 power of Mendelian data involving multiple polymor- 

 phic loci to detect differences between populations 

 when they do exist. For example, assuming that most 

 allozyme variation is neutral, it will take populations 

 that are divided into large units a considerable amount 

 of time before significant divergence will occur. Thus, 

 Atlantic herring Clupea harengus populations of the 

 eastern and western Atlantic Ocean that have likely 

 been isolated for thousands of years could not be distin- 

 guished because of similar allele frequencies at a num- 

 ber of polymorphic loci (Grant 1984). The observed 

 value for Wright's (1943) fixation index (F^i) of 0.0042 

 approximates an Fgt value of 0.003 expected for 

 neutral markers among populations of effective size of 

 1 million individuals separated over 3000 generations 

 (Nei and Chakravarti 1977). Such dynamics preclude 

 genetic distinction of these herring populations through 

 neutral genetic markers (and thus rejection of the null 

 hypothesis) even with very large samples of loci and 

 individuals. Under such circumstances, other criteria 

 (e.g., tagging data) are needed to determine whether 

 one or more populations is being sampled. 



Finally, we note the complementary nature of rela- 

 tionships among populations indicated by many pheno- 

 typic traits on one hand and by most molecular genetic 

 markers on the other hand. A strong selective compo- 

 nent appears to be involved in the maintenance of 

 phenotypic traits such as timings of spawning and 

 migration (e.g., Ricker 1972, Helle 1981); consequent- 

 ly, relationships inferred from such traits tend to 

 reflect relative similarities in adaptations among pop- 

 ulations. Conversely, the apparent absence of strong 

 selection at most electrophoretically detectable loci 



permits the estimation of relative degrees of gene flow 

 within and among regions (e.g., Chakraborty et al. 

 1978, Allendorf and Phelps 1981), and such estimations 

 provide useful insights about ancestral relationships. 

 In view of the complementary nature of these different 

 categories of genetic information, adequate sets of both 

 molecular markers (for clarifying ancestral relation- 

 ships) and phenotypic traits (for identifying adaptive 

 differences within lineages) should be included in 

 genetic surveys of a particular species whenever pos- 

 sible. Such adaptive differences have been noted within 

 a number of apparent ancestral groupings of chinook 

 salmon, including both spring- and fall-spawning migra- 

 tions within the Klamath River populations of the 

 species (Utter et al. 1989). 



Acknowledgments 



Research funded in part through contract DE-AI79- 

 89BP0091 with Bonneville Power Administration. 



Citations 



Aebersold, P., G. Winans, D. Teel, G. Milner, and F. Utter 



1987 Manual for starch gel electrophoresis: A method for the 

 detection of genetic variation. NOAA Tech. Rep. NMFS 61, 

 19 p. 

 Allendorf, F., and S. Phelps 



1981 Use of allelic frequencies to describe population struc- 

 ture. Can. J. Fish. Aquat. Sci. 38:1507-1514. 

 Allendorf, F., and G. Thorgaard 



1984 Tetraploidy and the evolution of salmonid fishes. In 

 Turner, B (ed.). Evolutionary genetics of fishes, p. 1-53. 

 Plenum Press, NY. 

 Allendorf, F., N. Ryman, and F. Utter 



1987 Genetics and fishery management: Past, present, and 

 future. 7« Ryman, N., and F. Utter (eds.). Population genetics 

 and fishery management, p. 1-20. Wash. Sea Grant Prog., 

 Univ. Wash. Press. Seattle. 

 Bartley, D., and G. Gall 



1990 Genetic structure and gene flow in chinook salmon pop- 

 ulations of California. Trans. Am. Fish. Soc. 119:55-71. 

 Bartley, D., B. Bentley. J. Brodziak, R. Gomulkiewicz. 

 M. Mangel, and G.A.E. Gall 



1992 Geographic variation in population genetic structure of 

 chinook salmon from California and Oregon. Fish. Bull.. U.S. 

 90:77-100 (authorship amended per errata. Fish. Bull. 90(3):iii). 

 Chakraborty, R.. P. Fuerst, and M. Nei 



1978 Statistical studies on protein polymorphism in natural 

 populations: 11. Gene differentiation between populations. 

 Genetics 88:367-390. 

 Gall. G.A.E. . B. Bentley, C. Panattoni, E. Childs, C. Qi. S. Fox, 

 M. Mangel, J. Brodziak, and R. Gomulkiewicz 



1989 Chinook mi.xed fishery project 1986-1989. Report of the 

 Department of Animal Science and the Department of Applied 

 Mathematics, Univ. Calif., Davis, under Interagency Agree- 

 ment C-1335 with Calif. Dep. Fish & Game, 192 p. 

 Grant, W. 



1984 Biochemical population genetics of Atlantic herr- 

 ing. Copeia 1984:357-364. 



