Butler et al.: Feeding ecology of Thunnus thynnus in North Carolina 
59 
calculated mean biomass; the assumed coefficient of 
variation was 30% (Overholtz, 2006). 
It is unknown what proportion of the age-6+ bluefin 
tuna biomass occurs off North Carolina during winter. 
Therefore, we used data from a preliminary Ecopath 
model of the South Atlantic Bight which indicated that 
biomass levels ranged between 5% and 25% of the total 
age-6+ western Atlantic bluefin tuna population (Butler, 
2007). A uniform distribution with this range of ?NC 
values was used for the Monte Carlo simulation. 
A pert distribution (Overholtz, 2006) was used to 
model the estimate of DR from the current study. Mini- 
mum and maximum estimates of DR were calculated 
as one standard deviation above and below the mean, 
respectively (Overholtz, 2006); the standard deviation 
of DR was calculated by the Delta method (Williams 
et al., 2002). 
The proportion by weight of Atlantic menhaden in 
bluefin tuna diet (W Mh ) was modeled by a uniform dis- 
tribution where the minimum and maximum values 
were the lowest and highest W Mh values observed annu- 
ally in this and Kade’s (2000) study. A uniform distribu- 
tion was chosen because it allows an equal probability 
of occurrence for the estimated proportion by weight 
value (i.e., W Mh ) between the minimum and maximum 
observed values. 
The maximum number of days (T NC ) that both Atlan- 
tic menhaden and bluefin tuna overlap in North Caro- 
lina is unknown. Adult Atlantic menhaden form large 
spawning congregations in the shelf waters off North 
Carolina from November through March (Checkley et 
al., 1999). Bluefin tuna are commercially harvested 
off North Carolina beginning in late November un- 
til the end of January. However, Boustany (2006) has 
shown that bluefin tuna fitted with pop-up satellite 
tags may extend their residency in these waters until 
late May. Thus, the maximum Tnc was assumed to 
be 120 days (i.e., late November through late March) 
and a minimum was chosen arbitrarily at 30 days. A 
uniform distribution covering this range was used for 
the simulations. 
To fully restore the western Atlantic bluefin tuna 
population, the International Commission for the Con- 
servation of Atlantic Tunas (ICCAT) has recommended 
a targeted biomass level equivalent to that in 1975 
(ICCAT, 2007). Thus, a new distribution of C Pop for a 
“restored” population was determined by using abun- 
dance-at-age data from 1975 for age-6+ bluefin tuna to 
calculate a new distribution of B bft . All other distribu- 
tions for the “restored” analysis were identical to the 
distributions used for the current population model. 
Results 
Diet analysis 
The stomach contents of 448 bluefin tuna were exam- 
ined. Of these, 124 (100 nonempty) were large medium 
and 324 (252 nonempty) were giant tuna. Samples were 
further categorized by year and within-winter time 
period (i.e., 1-14 December, 15-31 December, and 1-31 
January). The two December time periods were chosen 
because of good sample sizes throughout December and 
perceived changes in diets within that month. January 
was not split into two time periods because most Janu- 
ary fish were caught within the first two weeks of the 
month during each year of the study. 
Overall, stomachs of bluefin tuna contained four- 
teen families of teleosts, five species of portunid crabs, 
cephalopods (mainly Loligo pealeii), one species of elas- 
mobranch ( Mustelus canis), and unidentified algae. 
Atlantic menhaden ( Brevoortia tyrannus) was the most 
common prey item by %0 for both large medium (Table 
1) and giant bluefin tuna (Table 2). By weight, Atlantic 
menhaden similarly dominated the diets of both size 
classes (Tables 1 and 2). Although Atlantic needlefish 
(Strongylw'a marina) were not important to large medi- 
um bluefin tuna, they were the second most identifiable 
prey item of giants (7.14% O, 3.16% W ), mainly in one 
year (2005-06). Despite the occurrence of individual 
portunids and cephalopods in both size classes, they 
contributed little in terms of biomass. Other prey in- 
cluded several teleost species, elasmobranchs, bivalves, 
and algae that were rare items and that contributed 
little to the diet. 
Sample sizes were adequate to describe the within- 
winter diet of giant bluefin tuna, as well as between 
winters for both size classes and size class comparisons. 
All within-winter periods for giant bluefin tuna, with 
the exception of January 2006, reached an asymptote 
(Table 3). Both large medium and giant size classes 
reached asymptotes when data were pooled by winter 
(Table 3). However, the within-winter analyses of large 
medium bluefin tuna were likely biased because of the 
low sample sizes. Randomized cumulative prey curves 
did not reach an asymptote for any of the within-winter 
periods for large medium bluefin tuna (Table 3). Given 
the difficulty in collecting large numbers of stomachs 
over short periods, the lack of a defined asymptote is 
not uncommon in diet studies of other apex species 
(Bethea et al., 2004). 
Atlantic menhaden was the dominant prey of large 
medium bluefin tuna throughout the winter in both 
years (Fig. 1, A and B). Diets of large medium bluefin 
tuna were independent of within-winter time period 
(Table 4). Although Atlantic menhaden varied in im- 
portance, variability in other prey groups (e.g., por- 
tunids, teleosts) likely drove the within-winter effect 
of the giant size class (Fig. 2, A and B; Table 4). The 
diets of large medium bluefin tuna were dominated by 
Atlantic menhaden in both winters (Fig. 1C; Table 4). 
The diet of giant bluefin tuna did differ between years 
(Table 4) owing to the increased occurrence of At- 
lantic needlefish and cephalopods in 2005-06 (Fig. 
2C). Overall, the two size classes differed in diets 
(Table 4). This result was likely driven by the higher 
occurrence of non-Atlantic menhaden prey types in 
giants than in large medium bluefin tuna (Figs. 1C 
and 2C). 
