84 
Fishery Bulletin 106(1 ) 
Marginal increment widths were only analyzed for the 
months of March through November because of the 
low sample sizes of dolphinfish scales from the winter 
months (December through February). To differenti- 
ate between changes in the marginal increment width 
attributed to potentially sampling different cohorts 
of age-1 dolphinfish, we calculated the monthly mean 
fork length of all age-1 dolphinfish whose scales were 
measured for a marginal increment width. 
A subsample (n = 50) of dolphinfish that was deter- 
mined to be >age-0 by using scales was further exam- 
ined for the presence of annual marks by using otoliths. 
We prepared transverse cross sections of sagittal oto- 
liths using methods described above. These sections 
were viewed under the light microscope (first at 100 x, 
then 400 x) to determine if annual marks could be de- 
tected in these structures. 
The von Bertalanffy growth curve was fitted to two 
dolphinfish age-length data sets: 1) daily ages from 
age-0 dolphinfish with a fork length less than 650 mm 
and annual marks on scales from >age-0 dolphinfish 
by using absolute ages, and 2) daily ages from age-0 
dolphinfish with a fork length less than 650 mm and 
relative scale ages. Relative scale ages were assigned 
by adding the number of days after the fixed birth 
date of 15 April (middle of estimated southeastern U.S. 
spawning season) when the dolphinfish was caught to 
the absolute annual age determined from scales. The 
15 April birth date was chosen according to the trends 
in gonadosomatic indices in Florida and North Carolina 
and back-calculated hatching dates (Beardsley, 1967; 
this study). 
The von Bertalanffy growth parameters were es- 
timated separately by nonlinear regression for male 
and female dolphinfish and were compared by using 
the likelihood-ratio test (Kimura, 1980; Cerrato, 1990; 
Haddon, 2001). To detect if any significant changes in 
growth had occurred since the last dolphinfish aging 
study in North Carolina, the mean size-at-age values 
from Rose and Hassler (1968) were plotted with the 
von Bertalanffy growth curve fit (relative age data 
set) and compared qualitatively. Additionally, von Ber- 
talanffy growth functions estimated from past stud- 
ies within different regions were plotted together for 
comparison. 
Reproduction 
Gonadosomatic indices and back-calculated hatching 
dates were used to determine timing of spawning, and 
maturity staging was used to determine length-at- 
maturity. When available, intact gonads were removed, 
weighed to the nearest 0.1 g, and assigned a maturity 
stage determined by gross examination of the gonads. 
Maturity stages for both male and female dolphinfish 
have been described (Beardsley, 1967; Oxenford, 1985). 
Female dolphinfish were considered mature or immature 
on the basis of the criteria developed by Beardsley (1967). 
Male dolphinfish were classified as mature on the basis 
of the presence or absence of milt in their gonads. 
A gonadosomatic index (GSI) was calculated as go- 
nad weight /(body weight - gonad weight ) separately 
for male and female dolphinfish pooled for 2002-04. 
Because dolphinfish body weight and length are cor- 
related with GSI values (Chatterji and Ansari, 1982) 
and dolphinfish size differed significantly by month 
(see below), ANCOVA was used to compare In ( gonad 
weight) by month with In ( fork length ) as the covariate 
for males and females separately. Log transformations 
were used to meet assumptions of ANCOVA. To deter- 
mine which months had significantly different gonad 
weights, the length-adjusted mean In ( gonad weight ) 
value was compared among months for both male and 
female dolphinfish by using ANCOVA univariate test of 
significance for planned comparisons. Significance levels 
were adjusted by the standard Bonferroni technique to 
account for multiple comparisons. 
Hatching dates were determined by subtracting age 
in days (determined from age-0 otoliths) from the catch 
date. Because the daily deposition of increments in 
dolphinfish sagittal otoliths begins on the hatching 
date (Uchiyama et ah, 1986; Massuti et al., 1999), and 
because ripe eggs hatch within 50-60 hours after fer- 
tilization (Palko et al., 1982), back-calculated hatching 
dates provide an estimate of spawning dates for surviv- 
ing offspring. 
The length at which 50% of the fish had become ma- 
ture was determined for both sexes by using a logistic 
model. The model was fitted by using nonlinear regres- 
sion analysis based on the following equation: 
% Maturity - 1/(1 + e ( ~^ * (L ~ L 50^), 
where Q = model parameter; 
L = fork length (mm); and 
L 50 - fork length (mm) at 50% maturity. 
Results 
Collections 
Dolphinfish were collected mostly from the recreational 
charter fishery (ti = 611, 76%), but also from the com- 
mercial fishery {n = 45, 6%), sportfishing tournaments 
(rc = 130, 16%), and from four fishery-independent trips 
(77 = 16, 3%). There was a seasonal trend in the total 
amount (number) of dolphinfish collected by month, 
with nearly half (n- 364, 45%) of all dolphinfish col- 
lected in the months of June, July, and August. Only 
17 fish (2%) were obtained in the months of November, 
December, and January (Table 1). The majority of the 
dolphinfish were sampled from catches in Morehead City, 
NC (ti = 676, 84%). 
The size range for the pooled sample of dolphinfish 
was 89 to 1451 mm FL (mean=736 mm FL, standard 
error (SE) = 9.3). Males (n- 257) ranged in length from 
310 to 1451 mm FL (mean length and weight of all 
males sampled was 855 mm FL [SE = 16.0] and 6.44 
kg [SE = 0.4]) and females (n = 422) ranged in length 
