Anderson et al.: Evolutionary associations between Cy noscion arenanus and C. nothus 
21 
loci. This divergence is coupled with primarily non- 
significant estimates of admixture between species, 
although one individual did exhibit significant evidence 
of admixture (Q>0). Although it is plausible that this 
individual is an advanced backcross, this result more 
likely represents an outlier. Evidence for the outlier 
supposition includes the admixture model with mtDNA 
haplotypes used to improve clustering; in this scenario, 
the silver seatrout contribution for this individual is no 
longer significant. Additionally, no overlap occurred in 
the mtDNA haplotypes or in the two diagnostic morpho- 
logical characters examined in this study. Thus, these 
data indicate that hybrid formation between sand and 
silver seatrout is either rare or nonexistent in the off- 
shore Galveston Bay area, and enough generations have 
elapsed since divergence from an ancestral population 
that contemporary mtDNA haplotypes do not indicate 
either incomplete lineage sorting or admixture during 
the time of, or after, speciation. Altough the absence 
of hybridization seems to be the case in the Galveston 
Bay area, the present study was limited geographically, 
and the possibility of hybridization elsewhere within 
the overlapping ranges of these species must be further 
explored. Although narrow hybrid zones are rare in 
marine organisms (Palumbi, 1994), heterogeneity in 
the rate of hybridization has been demonstrated. For 
instance, the rate of hybridization in European shads 
(. Alosa spp.) tends to be heterogeneous across the range 
of A. alosa and A. fallax (Alexandrino et al., 2006). 
Estimates of genetic variability were systematically 
higher in sand seatrout than in silver seatrout, indicat- 
ing possible demographic differences between popula- 
tions of these species. For instance, both allelic richness 
and gene diversity are higher at almost every locus in 
sand seatrout. This observation is likely of biological 
significance, rather than a genetic ascertainment bias, 
because these markers were initially isolated from a 
species outside the genus Cynoscion. The direction of 
ascertainment bias would therefore be expected to vary 
from locus to locus, rather than to have comparable 
effects across a majority of loci. Demographic param- 
eters such as population census size, rates of migra- 
tion among neighboring populations, and variance in 
reproductive success can all affect effective population 
size (N e ), and heterogeneity in these parameters would 
result in different levels of measured genetic variability 
between the species. Thus the implication, based on 
systematically lower levels of genetic variability, is that 
silver seatrout have a lower N e than sand seatrout. 
The finding of higher genetic variability in sand seat- 
rout is not likely the result of differences in migration 
rates because both species spawn primarily offshore 
in the GOM (Shlossman and Chittenden, 1981; DeVr- 
ies and Chittenden, 1982); the pelagic nature of eggs 
and larvae in both species allow for long-distance dis- 
persal and gene flow throughout the GOM. It is also 
unlikely that a difference in census sizes is the cause 
of elevated genetic variability in sand seatrout. The 
data of McDonald et al. (2009) indicates dramatically 
a higher abundance of C. nothus in the Galveston Bay 
offshore region, and the western GOM in general. The 
most likely explanation for the elevated genetic vari- 
ability of sand seatrout is an overall higher level of 
individual reproductive success than that recorded for 
individuals in the silver seatrout population. Two key 
biological differences support this assertion. First, sand 
seatrout have an approximately 25% higher fecundity 
(mean 100,900 eggs per spawn than that of silver seat- 
rout (mean 73,900 eggs per spawn) in the GOM (Sheri- 
dan et al., 1984). Second, individual sand seatrout are 
more likely to spawn on multiple occasions during their 
lifetime. Sand seatrout live longer (2-3 years, Shloss- 
man and Chittenden, 1981) than silver seatrout (1-1.5 
years, DeVries and Chittenden, 1982) and have two 
peak spawns per year in the western GOM, compared 
to a single relatively definitive peak in silver seatrout 
(McDonald et al., 2009). Both species mature before 
age one, and both species experience high mortality at 
early life stages. Hedgecock (1994) referred to the com- 
bination of high fecundity and high mortality at early 
life stages as a sweepstakes strategy, which is likely 
common in marine fishes. Such differences in fecundity 
and longevity between species likely result in a higher 
variance in reproductive success in silver seatrout, 
which concurrently results in lower effective population 
size and a decrease in genetic variability (Hedgecock, 
1994; Hedrick, 2005). In any event, one caveat to this 
result is the fact that while the sand seatrout sample 
was collected during the course of three trawls, all 
silver seatrout specimens were collected from a single 
trawl. Thus sampling error or sample ascertainment 
bias cannot be ruled out completely as an explanation 
for differences in diversity estimates. 
The morphological and ecological similarities among 
sand seatrout, silver seatrout, and gray weakfish have 
resulted in difficulty in distinguishing the taxonomic 
status of these species (Weinstein and Yerger, 1976). 
However, there is a fundamental difference between the 
distributions of the three species; whereas sand seat- 
rout and gray weakfish are functionally parapatric in 
their distribution, inhabiting primarily the GOM and 
Atlantic Ocean, respectively, silver seatrout are found 
in relatively large populations in both areas. The shal- 
low divergence previously reported between sand seat- 
rout and gray weakfish (Weinstein and Yerger, 1976) 
was likely caused by highly stochastic sea level changes 
throughout the Pleistocene Epoch, resulting in regional 
differences between populations in the GOM and At- 
lantic Ocean. A similar pattern is typical among other 
marine, freshwater, and terrestrial vertebrates world- 
wide that diverged during this time period (Avise and 
Walker, 1998; Avise et al., 1998; Hewitt, 2000), and 
has been particularly well documented in peninsular 
Florida (Avise, 1992). However, these episodic sea level 
changes were likely not long enough for reproductive 
isolation between sand seatrout and gray weakfish to 
develop. Therefore, contemporary hybridization between 
these species is common on the Atlantic coast of Florida 
(Cordes and Graves, 2003); in contrast, no such pat- 
terns have been indicated between populations of sand 
