Gold and Richardson: Population structure of Senola dumerili 
775 
12 3 4 
Correlograms based on frequencies of mtDNA 
haplotypes found in eight or more individuals among 
samples of greater ambeijack, Seriola dumerili , from 
the northern Gulf of Mexico. Abscissas: distance 
classes 1-4 (left to right); ordinates: mean auto- 
correlation coefficients (Moran’s I values) for each 
distance class (±SE). (A) Equal frequencies/distance 
class; (B) equal distances between distance classes. 
sample localities. Absence of structure in MP analy- 
sis is not inconsistent with results of homogeneity 
testing, F gT values, and AM OVA analysis, in that 1) 
presumed restrictions in gene flow between subpopu- 
lations could be relatively recent, i.e. there has been 
insufficient time for haplotypes (e.g. haplotype 6 ) that 
differ in frequency between subpopulations to become 
reciprocally monophyletic, or 2) there may be lim- 
ited gene flow between subpopulations. Finally, there 
was no indication of spatial autocorrelation (positive 
or negative) in common haplotypes among samples 
from the Gulf. This finding indicates the absence of 
an isolation-by-distance effect among greater amber- 
jack in the Gulf and is consistent with the hypoth- 
esis of continuous gene flow across the northern Gulf. 
Divergence in mtDNA between Gulf and Atlantic 
subpopulations has been documented for a variety 
of marine species. In some (e.g. American oysters, 
toadfish, black sea bass, and to a lesser extent, horse- 
shoe crabs), major phylogeographic discontinuities 
between Gulf and Atlantic subpopulations were 
found, leading to the hypothesis that the similar 
vicariant patterns may have stemmed from episodic 
changes in environmental conditions during Pleis- 
tocene glaciation (Avise, 1992). In addition, the pres- 
ence of phylogeographic structure was taken to indi- 
cate that current-day gene flow between subpopula- 
tions is very restricted, if it occurs at all. In species 
such as red drum (Gold et al., 1993), king mackerel 
(Gold et al., 1997), and greater amberjack (our study), 
mtDNA differences between Gulf and Atlantic sub- 
populations are documented but are limited to ei- 
ther a frequency difference in a single mtDNA hap- 
lotype, a small (but significant) difference in mtDNA 
haplotype distribution, or both. The relatively small 
genetic differences observed between subpopulations 
in these species, along with the absence of phylogeo- 
graphic structure, both between regions and among 
haplotypes, suggest either that limited gene flow 
occurs between subpopulations or that separation 
between subpopulations is fairly recent. In the case 
of greater amberjack, the former is consistent with 
mark-and-recapture data that suggest exchange 
rates of 1.5% between Gulf and Atlantic stocks 
(Cummings and McClellan 3 ). 
For greater amberjack, the boundary between Gulf 
and Atlantic subpopulations appears to be between 
the Florida Keys (included in the Atlantic subpopu- 
lation) and somewhere off the central-western 
Florida coast, possibly the Florida Middle Ground, a 
series of north-south reef structures located about 
150 km south of the north Florida coast and about 
160 km northwest of Tampa Bay (Hopkins et al. 6 ), 
and the primary source for amberjack fishermen 
located in Sarasota and St. Petersburg. Separation 
between Gulf and Atlantic subpopulations (stocks) 
of greater amberjack could stem from a number of 
causes that involve historical or recent interactions 
between dispersal capability and impediments to 
gene flow, or both. Among present-day alternatives 
are 1) offshore currents that are not conducive to 
unrestricted movement between regions; and 2) ab- 
sence of suitable habitat or difference in ecological 
(biogeographic) provinces between regions. A third 
(historical) possibility might be that subpopulations 
were separated (e.g. during Pleistocene glaciations) 
and have only recently (in geological time) begun 
exchanging genes. Rates of approach to genetic ho- 
mogeneity under this last hypothesis are, in part, 
time-dependent, and one could speculate that there 
has been insufficient time for accumulated genetic 
differences to disappear. 
6 Hopkins, T. 1981. Florida Middle Ground. In R. Rezak and 
T. J. Bright (eds.), Final report: northern Gulf of Mexico topo- 
graphic features study, p. 1-5. Tech. Rep. No. 81-2-T, Dep. 
Oceanography, Texas A&M University, College Station, TX. 
