676 



Fishery Bulletin 98(4) 



buffer systems were used to resolve 16 enzymatic sys- 

 tems and muscular protein: continuous tris-citrate buffer. 

 pH 8.0 (TC8) following Selander et al. (1971), 150 V, 6 

 h; discontinuous tris-citrate, pH 8.3-sodium borate buffer, 

 pH 8.65 (POU) according to Poulik (1957), 250 V, 5 h; 

 and continuous tris-EDTA-borate, pH 8.0 (TEB) following 

 Shaw and Koen (1968), 200 V, 6 h. Tissue sources, elec- 

 trophoresis systems employed to resolve each enzyme or 

 muscular protein, and number of loci resolved in each 

 enzyme system are listed in Table 1. Staining procedures 

 were accomplished according to the methods of Schaal and 

 Anderson (1974), Shaw and Prasad ( 1970), Shaw and Koen 

 (1968), Abreu-Grobois (1983), and Rosa-Velez (1986). 



Allelomorphs were named A to F depending on their 

 anodal mobility, A being the fastest one. Zymogram inter- 

 pretation was achieved following recommendations by 

 Utter et. al (1987). In those cases where more than one 

 zone of activity was present in the gel, genetic hypotheses 

 were constructed to ensure correct interpretation. The 

 more complex pattern was displayed by the group of ester- 

 ases, which was interpreted by standardizing procedures. 

 These involved gel staining with each of the esters used in 

 the staining mixture (a-naphtyl acetate (black bands] and 

 b-napthyl acetate [red bands] ), but in separate assays, fol- 

 lowed by a joined assay of the same sample, as Laubier et 

 al. (1984) suggested. 



Swofford's (1989) BIOSYS-1 software was fed with raw 

 genotypic data from electropherograms to obtain allelic fre- 

 quencies, proportion of polymorphic loci at the 95*^ level, 

 observed and unbiased expected (Nei, 1978) heterozygos- 

 ity, chi-square goodness-of-fit test for testing conformity 

 to the Hardy-Weinberg equilibrium (H-W equilibrium) of 

 variable loci, and Nei's (1978) unbiased genetic similarity 

 and distance. 



The same set of raw data was used with GENEPOP (ver- 

 sus 1.2) (Raymond and Rousset, 1995), to compute an unbi- 

 ased estimate of the P-value of an F^,-based exact test for 

 the distribution of genotypes by means of the Markovian 

 chain method. PHYLIP's software (versus 3.5) was used to 

 perform bootstrap resampling of gene frequencies to obtain 

 a genetic distance UPGMA ( unweighted pair-group method 

 with arithmetic averaging) dendrogram. A Bonferroni cor- 

 rection was applied where multiple tests were carried out. 



Results 



A total of 32 loci were resolved from 16 enzyme systems 

 and muscular protein. Twenty loci (Acph-1. Aph-2. Gdh. 

 Got-1. Got-2, G3pd, Gpd, Idh. Ldh, Mdh-1, Mdh-2. Me, Pt-1, 

 Pt-2, Pt-4, Pt-5. Pt-6. Sdh-1. Sdh-2, and Xdh) were mono- 

 morphic across all populations sampled. Protein polymor- 



