Abstract.— An anomalous inabil- 

 ity to distinguish certain geograph- 

 ically-separated Chinook salmon On- 

 corhynchus tshaurytscha populations 

 of the Snake River and the Klamath 

 River from a survey of 18 polymor- 

 phic loci led to a prediction that 

 distinction would ultimately be found 

 through sampling of additional poly- 

 morphic loci. Recently published 

 studies involving pertinent groups 

 within each of these rivers included 

 data from an additional 15 polymor- 

 phic loci, and therefore allow a re- 

 examination of the relationships be- 

 tween these groups. Comparison of 

 results for the new studies shows the 

 formerly indistinguishable groups 

 from two areas to be as distinct from 

 one another as from other major 

 groupings of the species with a mean 

 genetic distance between popula- 

 tions of each river (0.014) that is 

 double that of the maximum within- 

 group genetic distance. Two newly- 

 resolved gene loci (niMDH-2* and 

 sMEP-1*) are particularly good at 

 distinguishing populations from the 

 two rivers. In addition to resolving 

 the anomalous similarity between 

 populations inhabiting geograph- 

 ically separated areas, the new re- 

 sults illustrate the care that must be 

 used in drawing inferences from 

 negative data. 



Genetic isolation of previously 

 indistinguishable Chinook salmon 

 populations of the Snake and Klamath 

 Rivers: Limitations of negative data 



Fred M. Utter 

 Robin S. Waples 

 David J. Teel 



Coastal Zone and Estuarine Studies Division 



Northwest Fisheries Science Center, National Marine Fisheries Service, NOAA 



2725 Montlake Boulevard East, Seattle, Washington 981 12 



Manuscript accepted 5 July 1992. 

 Fishery Bulletin, U.S. 90:770-777(1992). 



A variety of characteristics can be 

 useful in distinguishing particular 

 groups of organisms from other 

 related groups. In humans, for in- 

 stance, major ancestral groups can be 

 identified by heritable morphological 

 traits, as well as by characteristic fre- 

 quencies of alleles detected by molec- 

 ular or immunological procedures. 

 Conversely, although two groups 

 lacking any distinguishing character- 

 istics may, in fact, be closely related, 

 the possibility of undetected differ- 

 ences often prevents a conclusive 

 determination of the degree of re- 

 latedness. For example, two cryptic 

 species of bonefishes in Hawaii were 

 considered members of a common 

 gene pool until biochemical genetic 

 analysis revealed that the two forms 

 diverged perhaps 20 million years 

 ago (Shaklee and Tamaru 1981). 

 Other examples of genetic distinc- 

 tions between and within species of 

 fishes previously considered to be 

 homogeneous are listed in Allendorf 

 et al. (1987). 



The motivation behind our present 

 study was a puzzling instance of ap- 

 parent genetic similarity between 

 two geographically separated groups 

 of Chinook salmon Oncorhynchus 

 tfihawytscha. Indigenous chinook 

 salmon from the Klamath River and 

 spring- and summer-run chinook 

 salmon from the Snake River are 

 well differentiated from nearby 



populations at several protein-coding 

 gene loci (Utter et al. 1989, Hartley 

 and Gall 1990, Waples et al. 1991, 

 Hartley et al. 1992). However, a com- 

 parison of the two river groups by 

 Utter et al. (1989) failed to distin- 

 guish them despite their substantial 

 geographic separation. The mouths 

 of the Snake and Klamath Rivers are 

 separated by a distance of almost 600 

 river-ocean miles, and a number of 

 ancestrally distinct groups of popula- 

 tions (Utter et al. 1989) are found in 

 intervening areas. 



This apparent genetic similarity 

 was even more puzzling because of 

 substantial life-history differences 

 between chinook salmon from the 

 two rivers. The populations that were 

 not well differentiated in the Utter 

 et al. (1989) study included four 

 spring-run and two summer-run 

 populations from the Snake River 

 and two fall- and one spring-run 

 population from the Klamath River. 

 Utter et al. (1989) also sampled 

 fall-run fish from the Snake River, 

 but this population is genetically 

 quite different both from Snake 

 River spring- and summer-run fish 

 and chinook salmon from the Klam- 

 ath River. Whereas the fall-run fish 

 migrate to sea as subyearlings, the 

 other populations produce juve- 

 niles that spend an additional winter 

 in freshwater and outmigrate as 

 yearlings. 



770 



