Roelke and Cifuentes: Use of isotopes to assess groups of Scomberomorus cavalla 
545 
and Florida sites. Spines of individuals from 
Florida were typically 13 C-depleted in contrast 
with those from the Mexico, Mississippi, Louisi- 
ana, and Texas sites. 
After the nonsignificant variable (sex) was 
eliminated from the GLM (Eq. 3), the season of 
sample collection (P=0.Q001), fork length (P=0.Q31), 
and collection site (P=0.0001 ) influenced the varia- 
tion in nitrogen isotope data (F= 9.27; P=Q.00Q1 
for the overall model). Only collection site 
(P=0.Q28) significantly influenced the revised 
GLM (Eq. 3) for 5 13 C (F= 2.38; P=0.0281 for the 
overall model). A GLM was also constructed for 
the three regions in this study: Florida, Mexico, 
and northwestern Gulf of Mexico (Eq. 4). These 
regions were determined on the basis of previous 
king mackerel stock structure studies (Baughman, 
1941; Trent et ah, 1987; Johnson et ah, 1994; May 2 ) 
and isotopic patterns observed in previous stud- 
ies (Fry, 1983; Macko et al., 1984) as well as in 
this study. The GLM (Eq. 4) for 8 15 N (P=26.42; 
P=0.0001) showed that collection site (P=0.0001) 
and fork length (P=0.0023) were statistically sig- 
nificant regionally. Collection site (P=0.Q23) and 
fork length (P=0.047) also had a significant re- 
gional influence on 8 13 C (P=4.04; P=0.011) (Eq. 4). 
Discussion 
Although the dorsal fin spines were divided into 
multiple sections, and isotopic trends were ob- 
served along a spine (Fig. 3, A and B), life history 
could not be determined from isotopic data because 
it was not possible to assign accurately an age to 
a particular portion of the spine. The length-at- 
age relation varies regionally and shows large in- 
dividual variation (DeVries and Grimes 3 ). For ex- 
ample, male king mackerel from the eastern Gulf 
of Mexico, with a fork length of 105-110 cm, ranged 
from 4 to 22 years of age (DeVries and Grimes 3 ). 
Additionally, female king mackerel are larger than 
males at a given age (Beaumariage, 1973; Johnson 
et al., 1983; DeVries and Grimes 3 ) and sex was 
not known for the majority of the fish analyzed (Table 
2 ). 
Isotopic differences within individual dorsal spines 
were studied. One would generally expect the king 
mackerel with the greater fork length to have a larger 
range of isotopic values within its dorsal spine ow- 
ing to variation in trophic-level feeding with size; 
however, no clear trends were found (Table 2). For 
example one 113-cm-FL female king mackerel from 
GulfPort, MS, exhibited little variation among the 
segments analyzed. The § 15 N varied by only 1.1 and 
(1) 
C 
Q_ 
o 
CD 
O) 
CO 
CD 
> 
< 
-12 1 
A 
-14 
- 16 - 
% • 
-18 - 
® - , 
. * Wn 
-20 
• 
V 
-22 
1 
40 
60 
80 
100 
120 
140 
o> 
c 
CL 
SS 
o 
O ) 
(0 
1 7 
1 6 
1 5 
14 
1 3 
12 
1 1 
1 0 
9 
B 
• ® 
v* 
*4 ' 
• •• • ® ® 
• *® ® « ^ • 
• »• 
1 1 1 1 
0 60 80 100 120 
140 
Fork length (cm) 
Figure 2 
King mackerel fork length versus (A) stable carbon and (B) 
nitrogen isotopes. King mackerel fork length showed a posi- 
tive relation to nitrogen isotopic values, but not to carbon iso- 
topic values. 
the isotopic ratio of carbon varied by only 2.5 %o. Con- 
versely, a 76-cm-FL female from Celestun, Mexico 
differed by 4.3%c in nitrogen and 3.3%c in carbon 
among the spine segments analyzed. 
Additionally, an isotopic trend of an individual 
spine becoming heavier over time (from tip to base) 
would be expected because of an increase in trophic 
level feeding; however, this trend was not generally 
observed in the sites or regions for either carbon or 
nitrogen (Fig. 3). A greater enrichment in nitrogen, 
compared with carbon, would be expected for an in- 
crease in trophic level feeding (DeNiro and Epstein, 
1978; DeNiro and Epstein, 1981); however this ten- 
dency was not observed. In fact, carbon isotopic 
trends were contradictory to this assumption, and 
no clear nitrogen isotopic trends could be detected 
