FISHERY BULLETIN: VOL. 82, NO. 4 



lations in the NWHI were differentiated, and thus 

 independent, from populations in the main is- 

 lands? 



individually labeled glass vials which were capped 

 and stored at -75°C until the electrophoretic 

 analysis was completed. 



MATERIALS AND METHODS 



All specimens were obtained using commercial 

 handline gear and were either frozen or iced at sea. 

 Details of the collections are presented in Table 1. 

 One series of samples (from French Frigate Shoals 

 and Maro Reef) was filleted at sea, and the remain- 

 ing carcasses (containing the tissues of interest) 

 were preserved in an ice cold brine solution. This 

 means of sample handling had the unfortunate 

 effect of inactivating some of the enzymes (espe- 

 cially glucosephosphate isomerase and lactate de- 

 hydrogenase) so that these two enzymes could not 

 be reliably scored in these samples. Initial screen- 

 ing for polymorphic loci in the pink snapper was 

 conducted on extracts of white skeletal muscle, red 

 skeletal muscle, heart, eye, brain, and liver. Each 

 of these tissue samples was dissected from fresh or 

 frozen specimens and homogenized in an equal 

 volume of 0.1 M Tris-HCl pH 7.0 buffer (containing 

 1 x 10" 3 M EDTA and 5 x 10" 5 M NADP + ) using a 

 loose fitting, motorized pestle. Homogenates were 

 centrifuged for at least 20 min at a minimum of 

 20,000 x g (liver supernatants were routinely cen- 

 trifuged a second time to minimize lipid content). 

 The resulting supernatants were transferred to 



TABLE 1. — Collection details for total samples and individual 

 collections of Pristipomoides filamentosus used in the elec- 

 trophoretic analysis. 



'See figure 1 of Shaklee and Samollow (1984) for locality information. 

 2 Fork length, FL ( ±1 standard deviation) in mm. 

 *FL of 38 fish from collections a and b unknown. 

 *"FL of 79 fish from collection a unknown 



Electrophoresis 



The supernatants were analyzed by horizontal 

 starch gel electrophoresis (Selander et al. 1971). 

 Each enzyme system surveyed in the initial 

 screening for genetic variation was elec- 

 trophoresed on from two to eight different buffer 

 systems using extracts of several different tissues. 

 Following electrophoresis, isozyme patterns were 

 visualized using standard recipes (modified from 

 Shaw and Prasad 1970; Selander et al. 1971; 

 Siciliano and Shaw 1976). The umbelliferyl es- 

 terase (often called EST-D in the literature) was 

 visualized using 4-methylumbelliferyl acetate as 

 substrate. 



Gel Scoring and Data Analysis 



Patterns of enzyme variation which were consis- 

 tent with the subunit structure of the homologous 

 protein in other fishes (when known) and simple 

 genetic models were scored and recorded as 

 genotypes. Names of enzymes and Enzyme Com- 

 mission numbers follow the recommendations of 

 the Commission on Biochemical Nomenclature 

 (1973). For multilocus enzyme systems, loci were 

 given alphabetic designations to indicate homol- 

 ogy with known forms (e.g., Gpi-B and Ldh-C). 

 With the exception of one very rare allele (ob- 

 served once) for both ADH and UMB, each of the 

 polymorphic enzymes screened exhibited only 

 two detectable alleles. These two alleles are re- 

 ferred to hereafter by their relative electropho- 

 retic mobility from the origin as f (= fast) and s 

 (= slow). 



Tests of Hardy- Weinberg equilibrium and calcu- 

 lations of average heterozygosity (H) were ac- 

 complished as described in Shaklee and Samollow 

 (1984). A locus was considered polymorphic if the 

 frequency of the most common allele was =s 0.95. 



Two types of x 2 tests were used to test for 

 genetic differentiation and, therefore, stock 

 heterogeneity. First, for all polymorphic loci, con- 

 tingency tests of all possible pairwise combina- 

 tions of localities were conducted. Second, con- 

 tingency tests comparing pooled samples 

 representing the main Hawaiian Islands and the 

 NWHI were conducted for all loci. Wright's F ST 

 statistic (Wright 1965, 1978) was calculated using 

 the BIOSYS-1 computer program (Swofford and 

 Selander 1981). 



704 



