422 
Fishery Bulletin 109(4) 
Table 1 (continued) 
Microsatellite 
St. Croix 
St. Thomas 
Puerto Rico-east 
Puerto Rico-west 
Florida Keys 
Ral 
n 
93 
97 
96 
94 
118 
#A 
5 
4 
5 
6 
6 
Ar 
4.84 
3.97 
4.67 
5.83 
5.25 
h e 
0.592 
0.560 
0.558 
0.606 
0.567 
P HW 
0.877 
0.327 
0.150 
0.308 
0.410 
Fis 
0.037 
0.043 
0.142 
-0.019 
0.059 
Ra2 
n 
93 
94 
96 
94 
118 
#A 
19 
17 
18 
19 
17 
Ar 
18.52 
16.51 
17.13 
18.63 
16.04 
H e 
0.780 
0.767 
0.753 
0.778 
0.781 
P HW 
0.345 
0.873 
0.466 
0.549 
0.436 
Fis 
0.008 
-0.013 
0.073 
-0.012 
-0.020 
Ra6 
n 
93 
97 
92 
94 
118 
#A 
2 
4 
4 
3 
3 
Ar 
2.00 
3.65 
3.87 
2.85 
2.90 
H e 
0.053 
0.109 
0.134 
0.111 
0.105 
P HW 
1.000 
0.106 
1.000 
1.000 
1.000 
F [S 
-0.022 
0.050 
-0.052 
-0.051 
-0.046 
mtDNA 
n 
27 
26 
29 
25 
27 
#H obs 
11 
11 
11 
11 
10 
#H exp 
9.6 (7-12) 
9.9 (7-13) 
9.5 (7-12) 
9.6 (7-12) 
9.7 (7-13) 
Hr 
10.97 
10.99 
10.92 
11.00 
9.97 
Hoobs 
0.832 
0.812 
0.865 
0.873 
0.721 
HDexp 
0.835 
0.818 
0.873 
0.818 
0.835 
(95% Cl) 
(0.732-0.915) 
(0.709-0.900) 
(0.788-0.934) 
(0.693-0.897) 
(0.731-0.916) 
n D 
0.003 
0.004 
0.004 
0.004 
0.003 
Significant deviations from Hardy Weinberg equilib- 
rium were found before Bonferroni correction in nine 
of 80 tests (Table 1). Only one microsatellite (Lanll) 
at one locality (SC) deviated significantly (P=0.0019) 
from expectation after correction. Analysis with Micro- 
checker indicated an excess of homozygotes, indicating 
possible null alleles, at Lanll (in SC) and OchA (in FK), 
a finding reflected in the F IS values of 0.11 and 0.16 
(Table 1), respectively, for these microsatellites at those 
localities. Sizes of observed alleles were compatible with 
the stepwise mutation model (SMM) for all microsatel- 
lites, except sizes of Lca20, Lsy 13, and Prs248. Very 
rare alleles at Lca20 (allele 231) and Prs248 (allele 227) 
that differed by one base from their “regular” dinucleo- 
tide repeat were excluded from the analysis (Migrate) 
where a SMM was assumed. Two such alleles (alleles 
118 and 120) were found at Lsyl3; consequently, Lsyl3 
also was excluded from analysis with Migrate. 
A total of 25 mtDNA haplotypes were observed among 
the 134 individuals sequenced. The number and dis- 
tribution of mtDNA haplotypes across localities are 
given in Table 2; summary statistics for mtDNA may be 
found in Table 1. Haplotype richness ranged from 9.97 
(FK) to 11.00 (PR-west), and haplotype diversity ranged 
from 0.72 (FK) to 0.87 (PR-west). Results of bootstrap 
resampling analysis indicated that observed haplotype 
number and diversity at each locality did not deviate 
significantly from expectations in random subsamples 
of the overall data set (Table 1). Estimates of haplotype 
richness and nucleotide diversity were essentially iden- 
tical at all localities (Table 1). 
Exact tests of homogeneity of both microsatellite al- 
lele and genotype distributions among localities were 
nonsignificant (P=0.225, alleles; P=0.288, genotypes), 
and the among-localities component of molecular vari- 
ance (all microsatellites combined), estimated by AMO- 
VA, did not differ significantly from zero (<P g7 ,=— 0.0001, 
P=0.644). Nearly identical results were obtained for 
mtDNA; an exact test of homogeneity of haplotype dis- 
tribution was nonsignificant (P=0.590) and the among- 
locality component of molecular variance (from AMO- 
VA) did not differ significantly from zero (<£> ST =-0.010, 
