772 
Fishery Bulletin 96(4), I 998 
Table 3 
Mitochondrial (mt)DNA nucleon and intrapopulational nucleotide sequence diversities among samples of greater amberjack ( Seriola 
dumerili) from the Gulf of Mexico and Atlantic Ocean 
Locality 
Number of 
individuals 
Number of 
haplotypes 
Nucleon 
diversity 
Nucleotide 
sequence 
diversity (+SD) 7 
Port Aransas, TX 
58 
17 
0.886 
0.561 ± 0.443 
Freeport, TX 
44 
18 
0.899 
0.515 ± 0.387 
Port Fourchon, LA 
43 
16 
0.901 
0.579 ± 0.435 
Gulfport, MS 
7 
5 
0.905 
0.587 ± 0.415 
Pensacola, FL 
24 
12 
0.906 
0.601 ± 0.449 
Panama City, FL 
50 
17 
0.872 
0.525 ± 0.415 
St. Petersburg, FL 
42 
15 
0.879 
0.519 ± 0.395 
Sarasota, FL 
32 
12 
0.845 
0.483 ± 0.370 
Florida Keys 
55 
14 
0.893 
0.619 ± 0.419 
New Smyrna, FL 
53 
14 
0.861 
0.537 ± 0.427 
South Carolina 
36 
10 
0.846 
0.502 ± 0.342 
Total 
444 
49 
0.905 
0.548 ± 0.412 
1 In per cent. 
0.906 in the sample from Pensacola, Florida. 
Intrapopulational nucleotide sequence diversity 
within samples (Table 3) ranged (mean ±SD) from 
0.483 ±0.370 in the sample from Sarasota, Florida, 
to 0.619 ±0.419 in the sample from the Florida Keys. 
The latter values are all within one standard error 
(estimated from bootstrap analysis) of one another, 
indicating that levels of mtDNA variation are essen- 
tially identical throughout the geographic area sur- 
veyed. Levels of mtDNA variability, as measured by 
nucleon diversity and intrapopulational nucleotide 
sequence diversity, are commensurate with those 
estimated for several other marine fish of commer- 
cial or recreational value (Gold et ah, 1993). 
Results of bootstrap analyses and log-likelihood 
tests of spatial homogeneity in mtDNA haplotype 
frequencies are shown in Table 4. Tests were carried 
out 1) among all sample localities, 2) among samples 
from the Gulf, and 3) between defined groups where 
samples were pooled. In the last category, defined 
groups were based on region, i.e. Gulf and Atlantic. 
Gulf samples included the following: Port Aransas, 
TX; Freeport, TX; Port Fourchon, LA; Gulfport, MS; 
Pensacola, FL; Panama City, FL; and Sarasota, FL. 
Atlantic samples included New Smyrna, FL, and 
South Carolina. Two sets of pooled comparisons were 
carried out: in one, the sample from the Florida Keys 
was included with Gulf samples; in the other, the 
sample from the Florida Keys was included with At- 
lantic samples. We used this approach because the 
sample from the Florida Keys is located on the geo- 
graphic boundary between the two putative stocks 
of greater amberjack (Cummings and McClellan 3 ), 
Table 4 
Tests for spatial homogeneity in mtDNA haplotype frequen- 
cies among greater amberjack (Seriola dumerili) from the 
Gulf of Mexico and U.S. southeastern Atlantic. F ST is a 
measure of variance in haplotype frequencies. Number of 
samples is in parentheses. 
Number of 
haplotypes 
Homogeneity tests 7 
Test group 
r RB 
^ST 
All samples (11) 
49 
0.158 
>0.050 
0.005 
Gulf samples (8) 
Pooled comparisons 
Gulf + Florida Keys 
45 
0.250 
>0.050 
0.003 
vs. Atlantic (2) 
Atlantic + Florida 
49 
0.627 
>0.050 
0.004 
Keys vs. Gulf (2) 
49 
0.042 
-0.002 
0.009 2 
; P RB is probability from randomization (bootstrap) approach of 
Roff and Bentzen (1989); P G is probability from log-likelihood 
(G) test (Sokal and Rohlf, 1969). 
2 Value differs significantly (P=0.007) from 0.00; all other F ST val- 
ues are nonsignificant. 
and we could not place a priori the sample from the 
Florida Keys into either stock before testing the two- 
stock hypothesis. Significant heterogeneity was 
found only in the pooled comparison of samples from 
the Atlantic and Florida Keys versus samples from 
the Gulf; tests of homogeneity among all samples, 
among samples from the Gulf, and between pooled 
samples from the Atlantic versus those from the Gulf 
plus the Florida Keys were nonsignificant (Table 4). 
Estimates of F gT revealed the same pattern; the F gT 
