Utter et al.: Genetic isolation of Oncorhynchus tshawytscha of Snake and Klamath Rivers 



771 



Utter et al. (1989) speculated that the anomalously 

 high degree of genetic similarity between Klamath and 

 Snake River populations was due to coincidentally high 

 frequencies of the same common alleles (possibly a 

 reflection of restricted gene flow among populations 

 and reduced population sizes over an extended time in- 

 terval) rather than to a recent common ancestral origin. 

 Of the 25 polymorphic loci examined, only 18 were 

 variable in either the Snake or Klamath River groups, 

 and populations from these two areas had the lowest 

 average heterozygosities (0.027-0.045; Utter et al. 

 1989, App. A) of any populations included in the study. 

 Utter et al. (1989) predicted that additional genetic 

 surveys would ultimately reveal divergent frequencies 

 of alleles in the two areas. If such differences were not 

 found in more extensive studies, alternate explanations 

 for this apparent similarity would be required. 



This paper retests and rejects the null hypothesis of 

 no genetic difference between these two groups based 

 on two recently published studies, which sample several 

 new populations and an additional 15 polymorphic loci. 

 Comparison of results for the new studies shows the 

 formerly-indistinguishable chinook salmon populations 

 of the Klamath and Snake River to be quite distinct, 

 with a mean genetic distance between populations of 

 each river (0.014) that is double that of the maximum 

 within-group genetic distance. In addition to resolving 

 the anomalous apparent similarity between these 

 chinook salmon populations of these geographically 

 separated areas, the new results illustrate the care that 

 must be used in drawing inferences from negative data. 



Materials and methods 



Our analyses used the data from Hartley et al. (1992) 

 for Klamath River populations and Waples et al. (1991) 

 for Snake River populations; comparisons also were 

 made with earlier data from Utter et al. (1989). Sam- 

 pling locations included 10 areas from the Klamath 

 River and 11 from the Snake River drainages (Table 

 1, Fig. 1). Samples of juvenile fish from hatcheries and 

 naturally-spawning populations were collected between 

 1986 and 1989 for the Klamath River, and 1989 and 

 1990 for the Snake River. Starch gel electrophoresis 

 for all three studies followed procedures described by 

 Aebersold et al. (1987). The data used in these analyses 

 were part of a larger baseline dataset used by manage- 

 ment agencies to help determine natal origins of 

 chinook salmon harvested in mixed-stock fisheries 

 (Shaklee and Phelps 1990). 



Genetic nomenclature and abbreviations followed a 

 system suggested by Shaklee et al. (1989). Data were 

 collected from 21 enzyme systems and 30 presumptive 

 gene loci that were polymorphic in at least one of the 



populations (Tables 2,3). The observed polymorphisms 

 were attributed to 26 disomic loci and 2 isolocus pairs 

 (sAAT-1,2* and sMDH-Bl,2* ; see Allendorf and Thor- 

 gaard 1984). A single, average allele frequency was 

 computed for each isolocus pair for purposes of com- 

 paring populations. 



Genetic data were analyzed using the BIOSYS pro- 

 gram of Swofford and Selander (1981). Analyses in- 

 cluded calculation of unbiased pairwise genetic dis- 

 tances between populations (Nei 1978), unweighted 

 pair group method (UPGM) projection of a matrix of 

 these distances (Sneath and Sokal 1973), average 

 heterozygosities, and the number of alleles per locus. 



Results and discussion 



Our analyses focused on a comparison of genetic char- 

 acteristics between chinook salmon from the Klamath 

 and Snake Rivers. Discussion of population structure 

 within these two areas appears elsewhere, as do more 



