Saari et al.: Regional differences in the age and growth of Lutjanus campechanus in the Gulf of Mexico 
263 
34° N 
32°N 
30°N 
28°N 
26°N 
24°N 
22°N 
20°N 
18°N 
-100°W -98°W -96°W -94°W -92°W -90°W -88°W -86°W -84°W -82°W -80°W -78°W 
Figure 1 
Map of the Gulf of Mexico showing the 6 regions where red snapper (Lutjanus campechanus) were col- 
lected for age and growth analysis from recreational catches in 2009 and 2010. The 6 regions were South 
Texas, North Texas, Louisiana, Alabama, Northwest Florida, and Central Florida. 
d = the ordinal number (1-31) of the day of the 
month of capture. 
It was assumed that annulus formation began on 1 
January, and the periodicity of opaque zone formation 
was verified with edge analysis (Wilson and Nieland, 
2001; Fischer et al., 2004). To account for the uniform 
birthdate of 1 July, 182 days were subtracted from each 
age estimate. 
Analysis of variance (ANOVA) was used to compare 
mean TL, TW, and age among regions, and all statistical 
analyses were performed with SAS analytics software, 
vers. 9.1.3 (SAS Institute Inc., Cary, NC). TL, TW, and 
age data were first transformed with natural logs (In) 
to meet the assumptions of normality and homogene- 
ity of variance. Tukey’s honestly significant difference 
(HSD) test was used for post-hoc pairwise comparisons. 
Frequency distributions of size and age were compared 
by region with the Kolmogorov-Smirnov 2-sample (KS) 
test (Tate and Clelland, 1957). A chi-square (x 2 ) test 
was used to determine whether sex ratios differed from 
a 1:1 ratio, overall and by region. Relationships of TW 
to TL were described first by fitting a linear regression 
to the In-transformed variables and then by obtaining 
estimates (from the fitted regression coefficients) that 
corresponded to the parameters of the power function 
of TW - aTL h . Analysis of covariance (ANCOVA) was 
used to compare the linearized slopes and intercepts 
among the regions. 
Mean size at age (weighted by sample size) of the 
most common ages (3-7 years) was compared through 
the use of ANOVA with a Tukey’s HSD test for post-hoc 
comparisons. Growth was modeled for observed TL and 
TW at age with von Bertalanffy growth equations, and 
tested for coincident curves with likelihood ratio tests 
(Kimura, 1980; Haddon, 2001) and Bonferroni’s correc- 
tion for multiple pairwise comparisons. Growth models 
were fitted with nonlinear regression by least squares 
in the following form: 
where TLt 
TW t 
L M 
k 
t 
b 
TL t = LoJl— e-*«>), (2) 
TW t =Wj l-e- kit) ) b , (3) 
TL at age t\ 
TW at age t\ 
TL asymptote; 
TW asymptote; 
growth coefficient; 
age in years; and 
exponent derived from TW-TL regressions. 
