20 
Fishery Bulletin 107(1) 
and silver seatrout. As previously described, the ratio 
of anal-fin base to eye diameter has a nonoverlapping 
distribution (DeVries and Chittenden, 1982), and this is 
due in part to a difference in eye size of each species, for 
which the mean diameter is lower in sand seatrout than 
in silver seatrout. Anal-fin soft-ray counts also are diag- 
nostic in these samples, contrary to previous marine fish 
keys, which indicate overlap of these characters (Robins 
et al., 1986; Chao, 2002; McEachran and Fechhelm, 
2005). The ineffectiveness of the remaining diagnostics 
(pectoral- to pelvic-fin length ratio, lateral-line scale 
counts) may be due to differences at various age classes; 
i.e., the previously described differences may be more 
apparent in older individuals. Also, it should be noted 
that several damaged pectoral fins (resulting in shorter 
fins) in the silver seatrout sample were encountered 
because of the nature of capturing fish with trawling 
nets. Therefore there was a smaller sample size for the 
Table 4 
Locus scores and relative scores from WHICHLOCI anal- 
yses, ranked in order of population assignment power. 
The locus score is the power of each locus as a ratio of cor- 
rectly assigned to incorrectly assigned individuals in data 
simulations. The relative score of each locus is obtained 
by summing all locus scores and then dividing an indi- 
vidual locus score into the total score. 
Locus 
Score 
Relative score (%) 
SOC415 
0.97 
20.8 
SOC416 
0.70 
15.1 
SOC419 
0.63 
13.4 
SOC410 
0.61 
13.1 
SOC432 
0.58 
12.5 
SOC412 
0.49 
10.6 
SOC50 
0.32 
6.8 
SOC243 
0.31 
6.6 
SOC428 
0.05 
1.1 
pectoral- to pelvic-fin ratio measurement. Nevertheless, 
the morphological data in conjunction with completely 
diagnostic mtDNA haplotypes, should prove useful for 
future species identification issues between sand and 
silver seatrout. 
In contrast to the diagnostic differences in morpholog- 
ical characters and in mtDNA haplotypes, no diagnostic 
microsatellite loci were identified in the present study. 
This is likely due to several factors. First, microsatel- 
lites have a more elevated rate of mutation than other 
loci (Ellegren, 2000) and tend to evolve in stepwise 
fashion (Ellegren, 2000; Xu et al., 2000), such that 
they may be subject to increased rates of homoplasy 
(Estoup et al., 2002). The result is that convergent 
electromorphs can obscure the actual rate of fixation 
for homologous microsatellite alleles. Second, the fixa- 
tion of different alleles between the two species may 
be confounded by recombination, resulting in longer 
evolutionary timeframes for lineage assortment to occur 
than what would be expected for mtDNA, which is typi- 
cally nonrecombinant and clonally inherited. Finally, 
the enormous population sizes characteristic of marine 
fish populations may partially mitigate the effects of 
genetic drift (Allendorf and Phelps, 1981), which is 
likely the main mechanism for divergence at neutral 
loci. Collectively these processes have resulted in simi- 
larity in allele ranges for all but a single microsatellite 
locus (SOC415), despite significant differences in 6 at 
each individual locus. Thus although no single micro- 
satellite is diagnostic, when used in concert they are 
adequate for reliable identification to species. Further, 
because of their high overall between-species divergence 
and within-species variability, these microsatellites are 
likely adequate for diagnosis of hybrids, in particular 
at the F x (first generation hybrid) level. 
Genetic variability and divergence 
within and between species 
Significant genetic divergence between these species 
(0=0.117) is indicated by examination of microsatellite 
Table 5 
Admixture class (proportion of genetic contribution from the opposing species, in rows) membership in a randomized popula- 
tion sample (n = 1000), compared to experimental populations (n = 60) of silver seatrout (Cynoscion nothus ) and sand seatrout (C. 
arenarius) collected from offshore Galveston Bay, TX, in July 2007. The admixture classes are defined arbitrarily by Q-score 
increments of 0.05. The value “P” represents the proportion of individuals in each population that fall into a particular admixture 
class (range of Q), and significance in experimental populations (indicated by a superscript “a”) was assessed quantitatively by 
comparing membership in each class to expected values as assessed by examination of the randomized population. 
Q-score 
Randomized population 
P 
C. nothus 
P 
C. arenarius 
P 
0-0.050 
980 
0.98 
59 
0.98 
56 
0.93 
0.060-0.100 
13 
0.01 
0 
0.00 
2 
0.03 
0.011-0.150 
5 
0.01 
1 
0.02 
1 
0.02 
0.016-0.200 
1 
0.00 
0 
0.00 
0 
0.00 
0.210-0.250 
1 
0.00 
0 
0.00 
0 
0.00 
>0.250 
0 
0.00 
0 
0.00 
1° 
0.02 
