506 
Fishery Bulletin 107(4) 
Table 3 
Individual assignments based on the frequency test. Values represent the number (%) of individuals from the source populations 
in each assigned population. BZ = Belize; DT = Dry Tortugas; HN = Honduras; JP = Jupiter; PR = Puerto Rico. 
Assigned population 
Source 
population 
BZ 
DT 
HN 
JP 
PR 
Total 
BZ 
11 (22.0) 
11 (22.0) 
13 (26.0) 
11 (22.0) 
4(8.0) 
50 (20.4) 
DT 
7(17.5) 
9(22.5) 
12(30.0) 
8(20.0) 
4(10.0) 
40(16.3) 
HN 
11 (20.8) 
10 (18.9) 
7(13.2) 
15(28.3) 
10(18.9) 
53(21.6) 
JP 
10(18.2) 
14(25.5) 
14(25.5) 
11 (20.0) 
6(10.9) 
55(22.4) 
PR 
9 (19.1) 
8 (17.0) 
13 (27.7) 
9(19.1) 
8(17.0) 
47(19.2) 
Total 
48(19.6) 
52 (21.2) 
59(24.1) 
54(22.0) 
32 (13.1) 
245 
individual fish, all converged on a clear mode at K= 1, 
but extensive mixing around this value showed nonzero 
probabilities of K- 2. When geographic uncertainty was 
increased to 1° on each axis, 3 out of 10 runs converged 
on K~ 2. Nevertheless, maps of posterior probabilities 
from runs with K fixed at 2 (a representative one shown 
in Fig. 2) showed that the regions encircling our five 
sampling sites fell into a single genetic cluster at P=l, 
produced no strong discontinuities in the genetic land- 
scape, and showed that no map regions from which 
fish were collected grouped with genetic cluster 2 (not 
shown). Hence the results from both Geneland and 
STRUCTURE analyses support the existence of a single 
interbreeding population across the sampled range of 
L. analis (Fig. 2). 
Discussion 
Based on the results of this study, we cannot reject the 
null hypothesis that the sampled mutton snapper popu- 
lations constitute a single panmictic unit. Population 
genetic substructure is absent from the five sample loca- 
tions, ranging across approximately 2000 km. Further, 
because we could not differentiate genetically among 
the potential sources, it was not possible to ascertain 
the relative contributions of potential sources to 
“downstream” populations. 
Given these results there are two possible sce- 
narios for the population-level genetic structure. 
The first is that genetic structure does in fact ex- 
ist among mutton snapper populations but could 
not be resolved in this study. If so, future efforts 
to estimate the relative contributions of mutton 
snapper spawning aggregations to downstream 
populations must be explicitly designed to detect 
such weak structure (e.g., Johansson et ah, 2008). 
This research could require increasing the number 
of microsatellite loci; in particular, models show 
that loci exhibiting moderate allelic diversity (i.e., 
6 to 10 alleles) are most efficient for population- 
assignment techniques in weakly differentiated 
populations (Bernatchez and Duchesne, 2000). 
Thus, future microsatellite applications to mut- 
ton snapper could target these loci in place of 
the highly polymorphic loci used in our study. 
In addition, sampling of populations between the 
relatively distant sites we sampled, as well as 
outlying populations in the eastern Caribbean, 
would allow improved evaluation of the genetic 
landscape of this species. 
The geographic scale covered in our study does 
seem sufficient to recover genetic structure, if it 
is present. Yellowtail snapper (Ocyurus chrysurus) 
in Belize are genetically differentiated across a 
