Perkinson et at: Evaluation of the stock structure of Rachycentron canadum in the southeastern United States 
229 
Figure 4 
Genetically determined population ancestry plots for cobia (Rachycentron canadum) collected in the 
Gulf of Mexico (GOM) and Atlantic Ocean off the southeastern United States in 2006-2017, produced 
by using the program STRUCTURE. Each vertical bar represents an individual in the plot with 
shades or colors indicating percent ancestry to each genetic group. Genetic groupings or collections, 
as presented in Table 2, are geographically oriented from Texas (TX) on the left to Virginia inshore 
(VA) on the right. (A) Results for the complete data set when the number of distinct populations (K) 
was 2, with the South Carolina inshore (SCI) collection identified as distinct from the remaining data. 
(B) Results for the data set that excludes the SCI samples, indicating distinct populations {K= 4) for 
the GOM and Florida groupings (TX-FEL2) and the North Carolina inshore (NCI) and VA groupings. 
Analyses of samples from South Carolina and North Carolina offshore collections (SCOs and NCOs) 
indicate a homogenous population, and the Cape Canaveral, Florida, through Savannah, Georgia, 
collections represent a transition zone as indicated by the dashed oval. (C) Results for the data set 
that excludes SCI, GOM, and east coast of Florida and Georgia groupings (K= 2), with the NCI and 
VA collections grouped together and with a distinct population (K= 1) indicated for an Atlantic Ocean 
offshore group consisting of the remaining SCOs and NCOs. 
the Atlantic Ocean offshore (SCOs, NCOs) and Savan¬ 
nah (GA) samples, between Savannah (GA) and Jack- 
sonville-Brunswick (FLGA) samples, and between Ft. 
Pierce (FLE2) and Cape Canaveral (FLE3) samples were 
also significant but not as strong (f? ST =0.0067-0.0069, 
P=0.004-0.006), explaining 0.67-0.69% of the variation. 
The last AM OVA scenario (break between Jupiter Beach 
and Hobe Sound) did not partition a significant amount 
of variation among groupings (P=0.45Q). Therefore, the 
AMOVA also support the occurrence of a transition zone 
from Cape Canaveral through Savannah. 
Guided by these analyses, final sample groupings 
included GOM, South Carolina inshore, North Carolina- 
Virginia inshore, and Atlantic Ocean offshore populations, 
as well as the Cape Canaveral and Jacksonville-Savannah 
groupings (Table 2). Due to the lower sample sizes from 
some of the GOM and North Carolina inshore collection 
locations, deviations from HWE were evaluated to verify 
no substructure was being masked within these regions. 
No loci were out of HWE within the GOM population 
and only a single locus was out of HWE (P<0.001) in the 
combined North Carolina inshore and Virginia data set, 
supporting the groupings. Pairwise comparisons among 
these groupings confirmed significant differences between 
all groupings (P<0.00001-0.04), except comparisons with 
the Cape Canaveral and Jacksonville-Savannah group¬ 
ings with GOM and Atlantic Ocean offshore populations 
(P=0.07-0.96). Significant genetic differentiation ranged 
from an i? ST of 0.020 between South Carolina inshore and 
GOM populations to an P ST of 0.006 between the South 
Carolina inshore and Atlantic Ocean offshore populations 
(Table 7). The levels of genetic differentiation detected 
translated into effective number of migrants ranging from 
0.2 to 10.0 individuals/year between these populations. 
Therefore, the results indicate that the cobia stock bound¬ 
ary is occurring somewhere within a range from Cape 
Canaveral to northern Georgia, a location that is consis¬ 
tent with that of the current management stock boundary 
along the coast of the southeastern United States. 
Discussion 
A few major patterns became apparent when tagging data 
were evaluated together. Cobia in the area north of Florida 
were seasonally available as a pulse fishery in close proximity 
