FISHERY BULLETIN: VOL. 81. NO. 4 



TABLE 2. — Protein-coding loci surveyed inyellowfin sole. Multi- 

 ple loci are numbered beginning at the cathodic end of a gel. Let- 

 ter designations after locus abbreviations show homologies with 

 proteins in other teleosts. Tissue abbreviations: M = skeletal 

 muscle, H = heart, L = liver, E = vitreous fluid of eye, B = braia 

 Sp = spleen, K = kidney, St = stomach muscle, Gi = gill. Go 

 = gonad. 



regions (R), stocks (S), and populations (P) such 

 that 



1 Substrate 

 2 Substrate 

 3 Substrate 



Phenylalanyl-proline. 

 Leucyl-glycyl- glycine 

 Leucyl-alanine. 



frequencies in the six earlier samples were compared 

 with allelic frequencies of the samples taken from the 

 same general areas in 1979 using the contingency- 

 table analysis. 



The standard genetic distance, D, (Nei 1972) and its 

 standard error (Nei and Roychoudhury 1974) were 

 calculated for each pair of samples which were 

 examined for all 31 loci. D is an estimate of the num- 

 ber of codon differences in DNA between each pair of 

 samples. 



Total gene diversity (H T ) of allelic frequencies, 

 pooled over samples, was partitioned into its com- 

 ponents at three levels of population subdivision, 



H T = H,, + D PS + D SR + D 



HT 



where H,, is the average heterozygosity over samples, 

 D,, s is the diversity due to differences between pop- 

 ulations within stocks, D SR is the diversity due to dif- 

 ferences between the north and south Bering Sea 

 stocks, and D RT is the diversity due to differences be- 

 tween the western North Pacific Ocean, the Bering 

 Sea, and the eastern North Pacific Ocean. Genetic 

 differentiation relative toH T was estimated for each 

 subdivision. Thus, G SR = D SR /H r was the proportion 

 of gene diversity due to subdivision into stocks within 

 regions. The model of population subdivision for 

 both the gene- diversity and the contingency-table 

 analyses is presented in Table 3. 



TABLE 3. — Model of population subdivision inyellowfin 

 sole used for contingency-table and gene- diversity ana- 

 lyses. Location codes correspond to numbers in Table 1 

 and Figure 1. 



Total 

 Regions 

 Stocks 

 Populations 



1 2 3 4 8 9 10 14 15 16 



2 3 4 8 9 10 14 15 16 



1 2 3 4 8 9 10 14 15 16 



12 3 4 



9 10 14 15 16 



RESULTS 

 Electrophoretic Variation 



Nineteen protein systems were examined, and 31 

 zones of enzymatic activity appeared to represent 

 gene products of single locus, which could be reliably 

 scored for population data (Table 2). In the absence 

 of breeding data, the Mendelian nature of the elec- 

 trophoretic variants may be inferred from the band- 

 ing patterns. Four guidelines were useful in 

 formulating genetic models: 1) Banding patterns had 

 to be consistent with the subunit structures of 

 homologous proteins in related teleosts, 2) models 

 were formulated by considering gene expression in 

 other teleosts, 3) whenever the same locus was 

 expressed in two or more tissues, the banding pat- 

 terns of the variants had to be consistent among 

 tissues, and 4) the frequencies of the phenotypes had 

 to fit Castle-Hardy- Weinberg proportions in most of 

 the spawning-area samples. This last criterion has 

 been criticized by Fairbairn and Roff (1980) because 

 of the low power of statistical tests to distinguish 

 among alternative hypotheses (e.g., random distribu- 

 tion of phenotypes) with samples sizes normally used 

 in population studies. 



670 



