500 



Fishery Bulletin 100(3) 



comparisons (eleven) in each distance class. A total of 50 

 alleles (five at Sca-14, sixteen at Sca-23, four at Sca-37, 

 six at Sca-44, five at Sca-49, three at Sca-61, and eleven 

 at Sca-65) was tested, resulting in 250 Moran's I values. 

 Only 10 significant (P<0.05) Moran's I values were gener- 

 ated: one at Sca-14 (positive in the third distance class); 

 seven at Sca-23 (two positive in the second distance class, 

 four positive in the third distance class, and one negative 

 in the fifth distance class); two at Sca-37 (one negative 

 in the fourth distance class and one positive in the fifth 

 distance class), and one at Sca-65 (negative in the fifth 

 distance class). No significant Moran's 1 values were found 

 at Sca-44, Sca-49, and Sca-61. Only one of the "significant" 

 Moran's I values (a positive value for Sca-23*22 in the 

 third distance class) remained significant after Bonferroni 

 correction. The general absence of spatial autocorrelation 

 indicates that gene flow in king mackerel is consistent 

 with expectations of an island model (sensu Wright, 1943) 

 of population structure, meaning that there is roughly 

 an equal probability of gene flow between any of the 20 

 sample localities. 



Assignment tests generally were concordant with other 

 analyses of king mackerel microsatellites in that fish from 

 the Atlantic appeared to be weakly divergent genetically 

 from fish in the Gulf Each of the 20 samples included high 

 proportions of both Atlantic and Gulf fish, and there ap- 

 peared to be no strong geographic pattern in proportion 

 of fish assigned to either Atlantic or Gulf groups (Table 

 7). On average, samples from the Atlantic contained more 

 "Atlantic" fish (54%), whereas samples from the Gulf con- 

 tained more "Gulf" fish (51.2%). The three samples from 

 the Florida Keys, on average, contained more "Atlantic" 

 fish (53.3%), but this finding is misleading because the 

 proportion of "Atlantic" fish in the three samples from 

 the Florida Keys ranged from 37.2% to 64.0% (Table 7). 

 In addition, the estimated proportions of Atlantic versus 

 Gulf fish in samples from the Florida Keys were not con- 

 sistent with what might be predicted based on the time 

 of sampling and the spatial-temporal boundaries used 

 currently in king mackerel stock assessment. The KEY' 

 and KEY" samples were obtained in March, a time when 

 both would be considered Gulf stock, yet close to the tem- 

 poral boundary (1 April) when they would be considered 

 Atlantic stock. The estimated proportion of Gulf fish in 

 these two samples was 62.8% (KEYM and 41.4% (KEY'-'). 

 Alternatively, the KEY'^ sample was obtained in January 

 when king mackerel in the Florida Keys are considered 

 Gulf stock. The estimated proportion of Gulf fish in the 

 KEY3 sample was 36.09i . 



Synopsis and conclusions 



Genetic data obtained in our study are compatible with 

 the hypothesis that two, weakly differentiated "genetic" 

 subpopulations of king mackerel exist in waters off 

 Florida and that considerable, perhaps extensive, mixing 

 occurs between them. King mackerel sampled from the 

 Florida Keys cannot be assigned unequivocally to either 

 "genetic" stock; all collections tested appeared to be mix- 



tures, with approximately equal proportions of fish from 

 the two "genetic" stocks. These results are not consistent 

 with the current spatial-temporal boundaries employed 

 in stock assessment and allocation of the king mackerel 

 resource. Results are consistent with the hypothesis that 

 considerable gene flow occurs among all of the localities 

 sampled, and that differences in gene flow likely do not 

 arise as a function of geographic distance. Similar find- 

 ings were obtained by Gold et al. (1997) in their study 

 of variation in king mackerel mitochondrial DNA. The 

 genetic differences between king mackerel in the Atlantic 

 versus those in the Gulf most likely stem from reduced 

 gene flow (migration) between the Atlantic and Gulf in 

 relation to gene flow (migration) along the Atlantic and 

 Gulf coasts of peninsular Florida. This is consistent with 

 the notion based on studies of other marine fishes ( Avise 

 et al., 1992; Gold and Richardson, 1998) that the southern 

 Florida peninsula serves (or has served) as a biogeo- 

 graphic boundary. 



