Anderson et al. : Evolutionary associations between Cynoscion arenanus and C. nothus 
19 
Table 3 
Genetic allele diversity, gene diversity, and the inbreeding coefficient (F IS ) for nine microsatellite loci in sand seatrout ( Cynoscion 
arenarius ) and silver seatrout (C. nothus ) collected from offshore Galveston Bay, Texas, in July 2007. No value of F is was signifi- 
cantly different from 0. Genetic divergence between species (measured as 0) was significant at each locus. Summary statistics for 
overall variability are included on the bottom rows (mean above, SE below). 
Locus 
C. 
arenarius 
C. nothus 
d 
Allele diversity 
Gene diversity 
F,s 
Allele diversity 
Gene diversity 
F, s 
SOC050 
11 
0.77 
-0.095 
9 
0.54 
-0.045 
0.059 
SOC243 
5 
0.30 
0.038 
8 
0.52 
0.081 
0.116 
SOC415 
40 
0.97 
0.021 
15 
0.90 
0.107 
0.066 
SOC410 
6 
0.60 
0.062 
4 
0.13 
-0.048 
0.464 
SOC412 
17 
0.85 
-0.056 
11 
0.80 
-0.106 
0.084 
SOC416 
30 
0.94 
0.045 
15 
0.89 
-0.031 
0.058 
SOC419 
16 
0.77 
0.050 
13 
0.86 
-0.070 
0.100 
SOC428 
3 
0.11 
-0.046 
2 
0.02 
0.000 
0.025 
SOC432 
20 
0.91 
-0.024 
12 
0.81 
0.014 
0.075 
Overall 
16.44 
0.69 
0.000 
9.89 
0.61 
-0.010 
0.117 
(±SE) 
(±8.01) 
(±0.20) 
(±0.037) 
(±3.00) 
(±0.22) 
(±0.045) 
(±0.087) 
Lane no. 
100 bp 
Figure 4 
Sample gel image from mitochondrial DNA (mtDNA) restriction 
fragment length polymorphism (RFLP) analysis. Lanes 1-6 are 
restriction digests of silver seatrout (Cynoscion nothus ) mtDNA 
with the enzyme Rsal, whereas lanes 8-12 are restriction digests 
of sand seatrout ( C . arenarius ) mt DNA. Lane 7 is a sizing ladder, 
where the 1000-base pair (bp) fragment is marked. The lightest 
ladder fragment is 100 bp. 
The performance of each of the three Bayes- 
ian assignment models was evaluated by ex- 
amining the average membership proportion 
of each species in either of two genetic clusters 
after assignment. Under the assumption that 
the best assignment model should result in the 
highest proportion of membership (POM) of in- 
dividuals to their correct species cluster (that 
is, assuming no admixture between the spe- 
cies), each model performed equally well for as- 
signment of sand seatrout (POM= 0.978), where- 
as the six-locus panel of markers (POM=0.989) 
performed better than the nine-locus panel 
(POM= 0.985) for silver seatrout. The highest 
average POM of 0.994 and 0.996 for sand and 
silver seatrout, respectively, was attained by 
assigning group membership a priori based 
on mtDNA haplotype, and by refining group 
membership scores with the six highest-rated 
microsatellite loci. Using the six-locus model 
without mtDNA information, we found signifi- 
cant coancestry between the species for a single 
individual identified as belonging to the sand 
seatrout cluster, and an estimated proportion 
of silver seatrout ancestry of 0.303 (Table 5). 
In six successive iterations, the mean Q for 
this individual was 0.297 (range: 0.291-0.305). 
This value of Q was higher than any obtained 
in simulated populations (P<0.05). However, 
when mtDNA haplotype data were used to 
improve clustering, probability of admixture was not 
significant in this individual (Q = 0.119). Furthermore, 
morphological evidence indicated no intermediacy at 
the diagnostic traits (anal-fin to eye-diameter ratio=1.8, 
anal-fin rays=12); both of these diagnostics indicated 
this individual was a sand seatrout. 
Discussion 
Morphological and genetic identification of white trout 
Two of four commonly used morphological diagnostics 
are useful in distinguishing conclusively between sand 
