592 



Fishery Bulletin 94(3), 1996 



The analysis of heterogeneity of allelic frequencies 

 showed relatively small but significant differences 

 (Table 4), with the total G=8.79 (P<0.05). The 

 weighted average ofF- was -0.013, not different from 

 zero (P>0.05), and the weighted average of F st was 

 0.039, different from zero (P<0.001). Considering the 

 F , value, the estimated number of migrants per gen- 

 eration was 6.2. 



Discussion 



North Pacific swordfish exhibit low heterozygosity. 

 The heterozygosity in swordfish, 0.020, is low com- 

 pared with the averages of 0.055 (Smith and Fujio, 

 1982) and 0.064 (Ward et al., 1994) reported for ma- 

 rine fishes. For bony fishes in general, Nevo (1978) 

 and Winans ( 1980) reported 0.051 and 0.048, respec- 

 tively. The Pacific blue marlin, Makaira nigricans, 

 another billfish, showed a mean heterozygosity of 

 0.06(Shakleeetal., 1983). 



Both PROT-2* and PROT-3* were polymorphic in 

 the Hawaiian population; however, the latter was 

 responsible for the significant differences in allelic 

 frequencies. Although PROT-3* drove the final re- 

 sult in the analysis of allelic frequencies, ODH* and 

 PROT-2* also contributed in a significant way to the 

 amount of divergence as denoted by F .. 



The locus IDH* showed a high number of heterozy- 

 gous organisms in the Hawaiian samples during the 

 preliminary analysis, but that result was doubtful 

 considering the delicate nature of that enzyme 

 (Richardson et al., 1986). Future studies should use 

 tissues that are as fresh as possible. 



The observed heterogeneity in allelic frequencies 

 among the two samples was an unexpected result. 

 Recently, Grijalva-Chon et al. ( 1994) reported no dif- 

 ferences in genotypic frequencies in a restriction analy- 

 sis of mtDNA of North Pacific swordfish and could not 

 reject the null hypothesis of a single population. Chow 

 (1994) earned out PCR-RFLP analysis on the control 

 region of mtDNA and found no differences in the hap- 

 lotype frequencies between western (Japan) and east- 

 ern (Baja California) Pacific specimens of swordfish. 



A widespread opinion is that mtDNA analysis is a 

 more sensitive approach for defining population 

 structure than are allozymes (Ferris and Berg, 1987). 

 In fact, many studies using mtDNA analysis in sev- 

 eral species have confirmed preliminary results of 

 stock identification or have shown differences that 

 other methods have not detected (e.g. Avise et al., 

 1986; Kornfield, 1986; Hanzawa et al., 1987; 

 Kornfield and Bogdanowicz, 1987; Ward et al., 1989). 



Recently, incongruent results have been reported 

 with mtDNA and allozymes. Ferguson et al. (1991) 



did not find evidence of population structure in brook 

 char, Salvelinus fontinalis, using mtDNA analysis. 

 Allozymatic variation showed, however, a significant 

 divergence among sampled localities. Stott et al. 

 (1992) concluded that American plaice, Hippo- 

 glossoides platessoides, comprise a single population 

 in the Canadian Atlantic coast; mtDNA analysis gave 

 less resolution than did allozymes. Also, Ward et al. 

 (1994) reported that yellowfin tuna, Thunnus 

 albacares, from the Pacific Ocean form at least two 

 groups or populations according to allozyme analy- 

 sis, comprising five polymorphic loci and contrast- 

 ing with the homogeneity suggested by mtDNA re- 

 striction analysis. 



Results obtained here could be explained by a re- 

 cent population "bottleneck," which would have had 

 a bigger effect on mtDNA than on nuclear genes 

 (Ferris and Berg, 1987). However, it is very difficult 

 to fix the time of such a bottleneck from allozyme 

 data because, where a specific genetic distance value 

 is obtained, the divergence time interval is very wide 

 (HillisandMoritz, 1990). 



Another possible explanation lies in the popula- 

 tion sex ratio. Birky et al. (19891 pointed out that 

 extranuclear genes may show a more subdivided 

 population than do nuclear genes, but this feature 

 can be reversed if females are in excess. The sex ra- 

 tio in our samples indicate a skew toward females 

 (4.8:1 in samples from Mexico and 2.7:1 from Ha- 

 waii), and data from several fishing campaigns in 

 Baja California during 1992 and 1993 indicate con- 

 sistently high female:male ratios (-5.8:1, Castro- 

 Longoria 1 ). The observed sex ratio in Baja Califor- 

 nia waters, however, is not sufficiently skewed to 

 explain the observed difference. In accordance with 

 Birky et al. (1989), the number of females must al- 

 ways be more than seven per male for nuclear genes 

 to show more subdivision than extranuclear genes. 



Because the gene flow suggested by N e m was 

 higher than the required value to prevent differen- 

 tiation due to genetic drift, the third alternative ex- 

 planation could be some kind of selection. In our case, 

 the hypothesis is that selection supported a signifi- 

 cant heterogeneity in the face of genetic flow. The 

 proof of this hypothesis is beyond the scope of this 

 paper. In accordance with Lewontin (1991), it is not 

 possible to discriminate between selectionism and 

 neutralism with static-type data such as ours. 



For fishery management purposes, it is important 

 to establish clearly the presence or absence of ge- 

 netic differentiation of fishery resources among geo- 



Castro-Longoria, R. 1994. Universidad de Sonora, CICTUS, 



Rosales y Nirios Heroes s/n., Hermosillo, Sonora, Mexico. 

 Personal commun. 



