448 
Fishery Bulletin 113(4) 
Table 2 
Analysis of molecular variance (AMOVA) based on 24 microsatellite loci genotypes of hogfish (Lachnolaimus maximus) 
(number of specimens=719) collected from the southeastern United States (from northwest Florida to the Carolinas) between 
November 2005 and August 2013. Numbers joined by hyphens indicate sampling areas (as geographically defined in Fig. 1) 
combined into clusters and analyzed on the basis of either the corrected or uncorrected genetic distances (Fst)> which were 
indicated by principal coordinate analysis. Only the proportions of variation computed among clusters (Fqt) for 3-cluster and 
2-cluster combinations are given here. Probability of finding a more extreme variance by chance alone (1000 permutations) 
was <0.001 for all computed FC T values. 
Observed partition 
AMOVA results 
Among 3 clusters 
Cluster 1 
Cluster 2 
Cluster 3 
Variance 
Total (%) 
Act 
Ast 
1-2-3-4-5-6 
7-8 
9 
0.14768 
2.07 
0.0208 
Corrected 
1-2-3-4-5 
6-7-8 
9 
0.14331 
2.02 
0.0202 
Uncorrected 
Between 2 clusters 
Cluster 1 
Cluster 2 
Cluster 3 
Variance 
Total (%) 
Act 
Ast 
1-2-3-4-5-6 
7-8 
- 
0.11203 
1.56 
0.0156 
Corrected 
1-2-3-4-5-6 
- 
9 
0.16073 
2.30 
0.0230 
Corrected 
- 
7-8 
9 
0.28595 
3.93 
0.0393 
Corrected 
1-2-3-4-5 
6-7-8 
- 
0.10629 
1.48 
0.0149 
Uncorrected 
1-2-3-4-5 
- 
9 
0.17683 
2.56 
0.0256 
Uncorrected 
6-7-8 
9 
0.22038 
3.01 
0.0301 
Uncorrected 
500 
400 
300 
200 
100 
o 
K 
Figure 3 
For K= 1 through K= 9, 10 simulations were conducted with 2 million 
Markov chain Monte Carlo replicates after a burn-in period of 1 
million runs. L(K) denotes the log probability of the data at a given 
modeled K value; AA is based on the rate of change in L(K) be- 
tween successively modeled K values. STRUCTURE HARVESTER 
was used to compute the ad hoc statistics L(K) and AA - so that we 
could determine the most plausible base value of K (number of ge- 
netic clusters) within the 9 sampling areas where specimens of hog- 
fish ( Lachnolaimus maximus ) were collected from November 2005 
through August 2013 from the southeastern United States (from 
northwest Florida to the Carolinas). 
Summary of genetic structure 
The PCA, AMOVA, and Bayesian popula- 
tion-assignment analyses elucidated a con- 
cordant pattern of genetic structure within 
the studied geographical range. Three geo- 
graphically based clusters were delineated 
as 1) the eastern Gulf of Mexico, 2) the Flor- 
ida Keys and southeastern Florida, and (3) 
the Carolinas (Fig. 1). There was no indica- 
tion of genetic isolation by distance in the 
eastern Gulf or over the geographic range 
of specimens collected. The area west of the 
Florida Everglades (sampling area 6) ap- 
peared to serve as a genetic break between 
clusters 1 and 2. The nature of the appar- 
ent genetic break between the Florida Keys 
and the Carolinas requires further study be- 
cause of a discontinuity between sampling 
areas 8 and 9. 
Discussion 
The results of this study indicate the ex- 
istence of at least 3 genetically distinct 
hogfish stocks in the southeastern United 
States: the eastern Gulf of Mexico, south 
Florida and the Florida Keys, and the Caro- 
linas. Specimens were collected over a rela- 
