Darden et al.: Population genetics of Rachycentron canadum 
27 
Table 1 
Summary of data for tagged and released cultured Cobia ( Rachycentron 
canadum) in South Carolina by release year, with number of broodstock 
used for production, size at release, mean total length (TL), and number 
of individuals released. Small juveniles were released during the summer 
of the production year, large juveniles were released during the fall of the 
production year, and yearlings were released the following spring. 
Year 
class 
Number of 
broodstock 
Size at 
release 
Mean TL 
(mm) 
Number 
released 
2004 
8 
Small juveniles 
97 
1128 
Large j uveniles 
328 
679 
Yearlings 
496 
93 
2005 
7 
Small juveniles 
56 
3200 
Large juveniles 
230 
516 
Yearlings 
545 
385 
2007 
16 
Small juveniles 
82 
53,264 
Large juveniles 
250 
409 
Yearlings 
541 
59 
2008 
11 
Large juveniles 
249 
2000 
Yearlings 
530 
54 
2009 
9 
Large juveniles 
235 
1392 
lequin 3.11 (Excoffier et al., 2005) with default param- 
eters. The frequencies of potential null alleles at each 
locus were estimated in Cervus, vers. 3.0 (Kalinowski 
et ah, 2007). Significance levels for all simultaneous 
analyses were adjusted with a sequential Bonferroni 
correction (Rice, 1989). 
To confirm the utility of the marker suite for genetic 
evaluation and parentage analysis (i.e., identification 
of genetic tags), loci were examined for genetic diver- 
sity and polymorphism, for the ability to distinguish 
between related individuals, and for adherence to the 
principles of Mendelian inheritance. Basic molecular 
diversity indices, including number of alleles per locus 
(N a ), and allelic size range were calculated for each lo- 
cus through the use of Arlequin software. Cervus was 
used to estimate the average parent-pair and identify 
nonexclusion probabilities for the loci suite — indices 
that measure the probability that a set of markers will 
match erroneous parents to offspring and the probabil- 
ity that a set of markers will not be able to distinguish 
between related individuals, respectively. 
To determine whether hatchery individuals contrib- 
uted to the southeastern Atlantic Cobia populations, 
parentage analyses were conducted that incorporated 
all field samples and hatchery broodstock. Simula- 
tions (n= 5) for the “sexes known” parentage analysis 
in Cervus consisted of 10,000 offspring and 8 candi- 
date parent pairs per year (100% sampled) with allele 
frequencies generated from all Cobia samples. Critical 
delta values were determined with 95% confidence for 
the relaxed criteria and with 99% confidence for the 
strict criteria. All parentage analyses were run with 
the modal simulation file. Although all project samples 
genotyped at 8 or more loci were sub- 
jected to parentage analyses, contribu- 
tion calculations were limited to only 
those sampled within the stocking 
area. The percentage of hatchery con- 
tribution is reported as [S( W+S) -1 ] 100, 
where S is the number of stocked in- 
dividuals and W is the number of wild 
individuals as designated by Cervus 
at the strict confidence level because 
no additional offspring were identified 
with the relaxed criteria. Contribution 
is reported in terms of both population 
(all samples) and year class (on the ba- 
sis of known-age fish). All sampled fish 
identified as being of hatchery origin 
were removed from population struc- 
ture analyses. Identity analyses (as 
implemented in Cervus) were also con- 
ducted to compare the 2008 and 2009 
samples in order to identify recaptures 
of both hatchery and wild individuals. 
For the Mendelian inheritance tests, 
25 offspring from two parental families 
of the 2007 hatchery production year 
were compared with the 2007 brood- 
stock by using Probmax, vers. 3.1 parentage analysis 
software (Danzmann, 1997) to verify the contributing 
parent pairs. The genotypes of the contributing parents 
were merged into a single file and imported into FAP 
3.6 (Taggart, 2007), to generate all the possible prog- 
eny genotypes associated with these parental crosses. A 
chi-square analysis (% 2 ) was performed to compare the 
observed genotypic frequencies from the progeny data 
set with the expected genotypic frequencies from FAP. 
Population genetic analyses 
For all population genetic analyses, samples were 
partitioned into those from the inshore aggregations 
(defined as being captured landward of the barrier 
island, either along the coast or in the estuary) and 
Table 2 
Distribution of Cobia (Rachycentron canadum) collected 
during 2008 and 2009 with a known collection location 
that were included in the analyses of population genetic 
structure. Dashes indicate locations where no sampling 
occurred. 
Location 
2008 
2009 
Virginia inshore 
35 
North Carolina offshore 
90 
56 
South Carolina inshore 
103 
109 
South Carolina offshore 
71 
55 
