Life History of Cobia 
5 
present. Represented food items were drained, identified, counted, and 
weighed to the nearest gram. 
Importance of each prey item to the cobia diet was based on 
an index of relative importance (iri; Pinkas et al. 1971). Percent frequency 
of occurrence for each item in non-empty stomachs (/), percent total 
number of prey items («), and percent total mass of prey items (w) 
were calculated. The original iri formula was modified to use the 
mass of a prey item instead of volume, ( iri = f(n+w)). The results 
were examined for areal differences in diet (Beaufort Inlet and vicinity, 
Ocracoke and Hatteras inlets and vicinity, and offshore oceanic waters). 
To determine changes in cobia food habits with growth, specimens 
with food items were partitioned into arbitrary size classes (<4.5, 4.5- 
9.0, and >9.0 kg), and prey items were grouped into four categories, 
that is, shrimps, crabs, teleost fishes, and elasmobranch fishes. Percent 
iri ’ s were calculated as a percent of total iri within each cobia size 
class. 
Acetate impressions of cobia scales were difficult to interpret, 
therefore, sagittal otoliths of cobia were used to estimate specimen 
age. Sagittae of cobia were removed, washed in distilled water, and 
stored dry in individually labeled envelopes. Sagittae were embedded 
in 14x6x3-mm epoxy molds. Casts were affixed to a microscope slide 
with a drop of cyanoacrylate glue, then clamped to the arm of a 
circular low-speed saw. A 0.5-mm transverse section was made through 
the sagittal focus using a diamond-edge circular blade. The resulting 
wafer was permanently mounted to a microscope slide with a fixative. 
Sagittal sections were viewed on a dissecting microscope (16x) 
with transmitted, polarized light. Cross-sectioned sagittae had an opaque 
central core, followed by alternating translucent and opaque zones 
(Fig. 2). Although marginal increment analyses were precluded because 
specimens were unavailable throughout the year, most sagittae had 
an opaque edge, or an opaque zone in close proximity to the sagittal 
edge. Moreover, research in the northern Gulf of Mexico (Franks et 
al. 1991, Thompson et al. 1991) confirmed the validity of the formation 
of one translucent and one opaque zone on cobia sagittae each year. 
Thus, I assumed that one translucent and one opaque zone was deposited 
each year, and that opaque zones could be used to estimate cobia 
ages. 
Opaque zones along the ventral medial axis were counted as 
apparent annuli; estimated fish ages were based on opaque zone counts. 
I used the SAS NLIN procedure with the Marquardt option (SAS Institute, 
Inc. 1987) to estimate von Bertalanffy growth parameters based on 
individual fork lengths. Lengths referred to in the text are fork lengths. 
