DeVries and Grimes: Spatial and temporal variation in age and growth of Scomberomorus cavalla 
699 
Table 2 
Annual length frequency distributions of aged king mackerel, 1986-1992, by region and sex. See Figure 1 for overall 1977-78 and 
1986-92 size distributions. 
Number aged 
Size Females Males 
interval 
Region 
FL (cm) 
86 
87 
88 
89 
90 
91 
92 
86 
87 
88 
89 
90 
91 
92 
Atlantic 
20-39.9 



— 
1 
— 
— 
1 
— 
— 
3 
— 
— 
— 
Ocean 
40-59.9 
23 
72 
7 
14 
14 
16 
17 
36 
50 
12 
5 
27 
20 
13 
60-79.9 
94 
83 
86 
129 
135 
193 
248 
102 
86 
91 
122 
115 
117 
239 
80-99.9 
57 
83 
91 
205 
230 
289 
315 
35 
88 
87 
195 
186 
107 
217 
100-119.9 
21 
58 
53 
127 
150 
116 
196 
3 
9 
9 
16 
34 
31 
26 
120-139.9 
5 
42 
21 
30 
57 
50 
67 
— 
— 
— 
— 
— 
— 
1 
140-159.9 
- 
2 
— 
— 
7 
— 
3 
— 
— 
— 
— 
— 
— 
— 
Total 
200 
340 
258 
505 
594 
664 
846 
177 
233 
199 
341 
362 
275 
496 
Eastern 
20-39.9 
Gulf 
40-59.9 
30 
26 
9 
93 
20 
58 
49 
7 
14 
1 
45 
22 
83 
23 
60-79.9 
50 
135 
94 
79 
99 
242 
267 
20 
69 
37 
58 
60 
128 
170 
80-99.9 
34 
60 
58 
65 
55 
124 
206 
31 
46 
53 
43 
53 
129 
96 
100-119.9 
30 
73 
55 
39 
66 
153 
133 
8 
11 
5 
22 
11 
20 
17 
120-139.9 
11 
72 
35 
35 
33 
56 
58 
— 
— 
1 
— 
1 
— 
1 
140-159.9 
7 
17 
8 
4 
6 
8 
1 
— 
— 
— 
— 
— 
— 
— 
Total 
162 
383 
259 
315 
279 
641 
714 
66 
140 
97 
168 
147 
360 
307 
Western 
20-39.9 



6 

4 




7 

2 

Gulf 
40-59.9 
— 
3 
11 
12 
14 
8 
29 
— 
3 
8 
9 
11 
6 
35 
60-79.9 
14 
54 
76 
68 
107 
171 
83 
20 
47 
43 
62 
58 
130 
103 
80-99.9 
19 
88 
101 
124 
70 
108 
104 
27 
69 
65 
64 
53 
56 
57 
100-119.9 
9 
42 
64 
51 
32 
42 
33 
1 
11 
7 
18 
6 
8 
4 
120-139.9 
6 
15 
36 
32 
10 
7 
2 
— 
— 
— 
— 
— 
— 
— 
140-159.9 
— 
1 
— 
5 
1 
— 
— 
— 
— 
— 
— 
— 
— 
— 
Total 
48 
203 
288 
298 
234 
340 
251 
48 
130 
123 
160 
128 
202 
199 
during 1986-92, western Gulf females grew faster 
and larger than eastern Gulf females according to 
the growth curves (Fig. 3). Among males, growth also 
appeared to be lowest in the Atlantic, although the 
differences were slight (Fig. 3 ). 
There was interannual variation in growth within 
a region and sex during 1986-92; however, it prob- 
ably reflected sample differences as much as any 
actual differences (Tables 1 and 2); therefore we did 
not test these growth curves statistically. 
Plots of the von Bertalanffy curves (Fig. 6) sug- 
gested that growth was slightly less in 1977-78 than 
in 1986-92 in the Atlantic and eastern Gulf for both 
sexes, whereas western Gulf females grew faster in 
1977-78 than in 1986-92. The average differences 
(and standard errors) in predicted size at age between 
1986-92 and 1977-78 for all ages above age 7 were 
1) eastern Gulf females: +2.0 ±0.1 cm; 2) eastern 
Gulf males: +2.9 ±0.1 cm; 3) Atlantic females: +5.8 ± 
0.1 cm; 4) Atlantic males: +1.7 ± 0.1 cm; and 5) west- 
ern Gulf females: -9.6 ± 0.6 cm. 
Discussion 
Our findings were based on data collected as part of 
a long-term, stratified, nonrandom sampling program 
(following the suggestions of Ketchen [1950]) de- 
signed to provide age-length keys for annual stock 
assessments. Most fish sampled were caught by hook 
and line and gill nets, both of which are size-selec- 
tive gears. Goodyear ( 1995), using computer simula- 
tions, demonstrated that samples equally stratified 
by length and those from size-selective fisheries of- 
ten yield biased estimates of mean size-at-age; for 
this reason he recommended that only simple ran- 
dom sampling be used to generate models of fish 
growth. He found that most often mean lengths-at- 
age were overestimated by 5-15% for all but the 
youngest age classes, which were sometimes under- 
estimated. Goodyear explained that at the youngest 
ages, the smaller individuals of an age class are of- 
ten sampled disproportionately to their true abun- 
dance, whereas for older ages, the same happens for 
