Locus of 

 subunits 



FISHERY BULLETIN: VOL. 87, NO. 2, 1989 



2,3 

 1,3 



Sample 

 number 



10 



Appendix Figure 2.—Gpi-l{H) variation in chinook salmon from extracts of skeletal muscle. Samples 

 2, 6, and 8 are H/H homozygotes. Genotypes of other samples are unknown for the Gpi-lfH) allele. 



Appendix Table 1,— Observed and (in parentheses) ex- 

 pected numbers of Chinook salmon progeny from parents 

 of known Gpi-1 phenotype assuming Ivlendelian Inheritance 

 of subunits having differential heterodlmer forming capa- 

 bilities. Phenotypic designations are based on Figure 3. 



delian variant affecting heterodimer formation 

 between Gpi-1 and Gpi-3 subunits. We therefore 

 conclude that individuals with such five-banded 

 phenotypes are homozygous for an allele at the Gpi-1 

 or Gpi-3 loci that affects dimer formation between 

 subunits of these loci. The present data give no in- 

 formation regarding which locus encodes the mutant 

 subunit. However, the polymorphism has been re- 

 corded and analyzed as a third allele at the Gpi-1 



locus [Gpi-l(H)] because of the low frequency of 

 mobility variants at this locus, none of which oc- 

 curred in populations where the allele affecting 

 heteromeric combinations was observed. The cor- 

 rect locus could probably be identified through in- 

 duced gynogenesis of eggs from females hetero- 

 zygous for mobility variants at Gpi-1 and Gpi-3, and 

 Gpi-l(H) heterozygotes. The gene-centromere dis- 

 tance for Gpi-l(H) would match that of the mobility 

 variant of the appropriate locus assuming that gene- 

 centromere distances differ for the loci encoding the 

 mobility variants (e.g., see Thorgaard et al. 1983). 

 Because of the difficulty in distinguishing the com- 

 mon and the Gpi-l(H) heterozygous phenotypes the 

 recorded allele frequencies are based on the square 

 root of the Gpi-l(H) homozygous (i.e., five-banded) 

 phenotypes under the assumption that the samples 

 where these phenotypes are observed are in Hardy- 

 Weinberg equilibrium. This assumption is supported 

 by the preponderance of genotypic frequencies in 

 Hardy- Weinberg equilibrium at other polymorphic 

 loci. This restriction results in an inevitable under- 

 estimation of the frequency of this allele when its 

 frequency is too low for homozygous expression at 

 reasonable sample sizes. 



264 



