222 
Fishery Bulletin 117(3) 
approach to stock boundary identification (Begg and Wald- 
man, 1999) with results that are more robust than those 
from the use of either method alone. 
Materials and methods 
Tagging methods 
Tag-recapture data for cobia from 7 tagging programs 
were included in the analysis: the Virginia Game Fish 
Tagging Program, North Carolina Division of Marine Fish¬ 
eries, South Carolina Department of Natural Resources 
(SCDNR), Hilton Head Reef Foundation, NOAA South¬ 
east Fisheries Science Center Cooperative Tagging Cen¬ 
ter, Mote Marine Laboratory, and the Sport Fish Tag 
and Release Program of the Gulf Coast Research Labo¬ 
ratory, University of Southern Mississippi (Table 1). All 
7 programs employed similar methods, including the use of 
nylon or stainless-steel-tipped dart tags and reliance on a 
trained group of recreational anglers and charter boat cap¬ 
tains to tag and release cobia. Anglers were asked to submit 
a data card for each tagged fish that included the follow¬ 
ing information: tag number, date, release location, fish 
length (converted to millimeters in FL where necessary), 
and weight (converted to kilograms), as well as other per¬ 
tinent information that differed by tagging program. Upon 
recapture of a tagged fish, anglers reported similar data, 
including whether the fish was released or harvested. In 
this study, we determined general migratory trends over a 
large geographic area on the basis of recapture data. There 
were minor differences in tag types (stainless versus nylon 
anchors) and incentive structure between programs and 
within programs during different time periods that could 
affect tag retention and reporting rates and could, there¬ 
fore, influence estimates of mortality or other quantitative 
measures. However, these difference are unlikely to influ¬ 
ence the analysis of large-scale migratory patterns based 
on capture and recapture location. Therefore, recapture 
data were pooled into a comprehensive data set. Tag and 
recapture locations were assigned a GPS coordinate on the 
basis of location description, if latitude and longitude were 
not otherwise provided. In those instances, we assigned 
latitudes and longitudes by using the best available infor¬ 
mation. These coordinates were assigned solely for the pur¬ 
pose of graphical depiction of recapture locations and not 
for analyses; all analyses were based on zone assignment. 
Only cobia at large for greater than 30 d were included in 
the final analysis to minimize the location bias of fish that 
were tagged and immediately recaptured. Where programs 
reported total length (TL), rather than FL of fish tagged, 
the formula FL=13.52399+(0.878671xTL) (SEDAR, 2013b) 
was used to transform data. 
Data partitioning 
Tagging areas and recapture areas were partitioned into 10 
different zones for analysis of movement patterns (Fig. 1). 
The East Coast of the United States north of Florida was 
Table 1 
Number of cobia (Rachycentron canadum) tagged and recap¬ 
tured by each of 7 tagging programs throughout the Gulf of 
Mexico and Atlantic Ocean off the southeastern United States 
between 1988 and 2017. The tagging programs were those of 
the following organizations: Virginia Game Fish Tagging Pro¬ 
gram (VGTP), North Carolina Division of Marine Fisheries 
(NCDMF), South Carolina Department of Natural Resources 
(SCDNR), Hilton Head Reef Foundation (HHRF), NOAA 
Southeast Fisheries Science Center (SEFSC), Mote Marine 
Laboratory (MOTE), and Gulf Coast Research Laboratory, 
University of Southern Mississippi (GCRL). 
Data 
source 
Number 
tagged 
Number 
recaptured 
Recapture 
rate (%) 
Years 
active 
VGFTP 
3899 
433 
11.1 
1995-2017 
NCDMF 
73 
5 
6.8 
2017 
SCDNR 
1194 
216 
18.1 
1990-2014 
HHRF 
95 
14 
14.7 
2007-2012 
SEFSC 
1557 
159 
10.2 
1986-2014 
MOTE 
920 
100 
10.9 
1991-2001 
GCRL 
18,129 
1197 
6.6 
1988-2017 
Total 
25,867 
2124 
8.2 
1986-2017 
segmented into 5 zones: the areas within and extending 
offshore of states north of Virginia, Virginia, North Caro¬ 
lina, South Carolina, and Georgia. All tagging and recap¬ 
ture events that occurred within the Chesapeake Bay were 
assigned to the Virginia zone. Because preliminary genetic 
results (Darden 5 ) indicated a break along the east coast of 
Florida, the area was segmented into 3 zones: north of Brevard 
County, encompassing the area from the Florida-Georgia 
border south to the border of Volusia and Brevard Coun¬ 
ties; Brevard County, including Cape Canaveral; and south 
of Brevard County, encompassing the area from the border of 
Brevard and Indian River Counties to Biscayne Bay. County 
lines were chosen because of their congruency with recre¬ 
ational and commercial catch data used in the stock assess¬ 
ment process. To the south, the Florida Keys zone covered the 
area from Biscayne Bay around to Marco Island in southwest 
Florida. Although genetic structure and movements of fish 
between locations in the COM to the Atlantic Ocean were 
evaluated, specific movements of fish solely within the COM 
were beyond the scope of this paper, resulting in the entire 
GOM from Marco Island to the Texas-Mexico border being 
assigned a single zone. Dippold et al. (2017) provide a thor¬ 
ough analysis of cobia movements within the GOM. Analyses 
of movement between zones as well as recaptures within a 
zone were largely qualitative and focused on identifying sea¬ 
sonal trends and broad patterns in movement that indicate 
potential breaks in biological stocks. 
Genetics 
The SCDNR Genetic Tissue Collection currently houses 
more than 5050 archived genetic samples of cobia col¬ 
lected by numerous researchers and anglers around the 
