Syc and Szedlmayer: A comparison of size and age of Lutjanus campechanus with the age of artificial reefs in the Gulf of Mexico 
461 
and ages of red snapper among the different ages of 
reefs. If significant differences were detected, a Tukey 
test was used to show specific differences. 
Growth rates were examined with an analysis of 
covariance (ANCOVA) that compared the mean length 
at age of red snapper <10 years old between old (2006) 
and new (2009 and 2010) reefs. This analysis was used 
to determine if old reefs were providing additional re- 
sources, a difference that would be reflected in faster 
growth rates of red snapper on old reefs than on new 
reefs. For additional comparisons, Pearson’s correlation 
coefficients were calculated for reef age with red snap- 
per SL, weight, and age. Also examined with Pearson’s 
correlation coefficient was the potential influence of 
nearby known reefs on the age and abundance of red 
snapper on the reefs that we surveyed in 2010. Nearby 
reefs were 0.17-1.7 km away (mean=0.72 km, n = 37), or 
less than the distance between the reef sites sampled in 
our study. In an effort to remove possible depth effects, 
the ages of red snapper collected at the same depth (30 
m) were compared between the 2006 and 2010 reefs 
with a t-test. Differences were considered significant 
at P<0.05, and all data were analyzed with Statistical 
Analysis System software (SAS, vers. 9.1, SAS Insti- 
tute, Inc., Cary, NC). 
Results 
Red snapper were sampled from April to November 2010 
from 37 artificial reefs (2006 reefs=18; 2009 reefs = 10; 
2010 reefs = 9). Visual surveys were completed by divers 
at later dates at 2 sites because of the presence of sharks 
on the original sampling date and were not completed 
on 7 reefs (3 of the 2006 reefs and 4 of the 2010 reefs) 
because of poor visibility (<1 m). 
A total of 1028 red snapper were collected, 439 by 
hook and line and 589 by trap. Mean ±standard de- 
viation (SD) CPUE for hook and line was significantly 
greater on the 2006 reefs (20.4 ±8.5 30 min -1 ) than on 
the 2009 (6.3 ±8.1 30 min -1 ) and 2010 reefs (2.6 ±4.6 30 
min -1 ; ANOVA: F 2 34 = 20.38, P<0.0001). No significant 
differences in CPUE were detected among reef years 
for trap collections (2006 = 10.6 ±10.9, 2009=16.6 ±19.9, 
and 2010 = 14.3 ±12.7; ANOVA: F 2 34 =0.6, P=0.55). The 
SL and weight of red snapper caught by hook and line 
(429.4 ±79.8 mm, 2531 ±1409 g) were significantly 
greater than those measures of fish caught by trap 
(232.6 ±77.6 mm, 538 ±726 g; SL t-test, t 1018 =39.56, 
weight t 101g = 29.41, P<0.0001). Red snapper ages also 
were significantly different between these 2 sampling 
methods (hook and line = 4.1 ±1.3 years, trap = 1.9 ±1.1 
years; /-test, t 1024 = 29.68, P<0.0001). 
The visual survey methods significantly affected 
counts of red snapper. Visual counts by divers (mean 
±SD=78.3 ±54.8) were significantly higher than counts 
from image-analysis methods (photograph counts=30.7 
±20.2, video counts=16.5 ±10.3; ANOVA, F 2 42 =13.37, 
P<0.0001). Because of these differences, total densities 
of red snapper were estimated by adding the number 
Table 1 
Average percent error for both sets of independent read- 
ings of otoliths from red snapper (Lutjanus campechanus ) 
caught in 2010 during our study on artificial reefs in the 
northern Gulf of Mexico. Included are the percentages 
of agreement for each difference between readings (first 
and second reading, coefficient of regression [r 2 ] = 0.83, 
P<0.0001; third and fourth reading, /- 2 = 0.96, P<0.0001). 
First 
and second 
readings 
Third 
and fourth 
Readings 
Average percent error 
7.85 
1.41 
Standard deviation 
0.12 
0.05 
0 
62.16% 
92.32% 
±1 
35.89% 
7.39% 
±2 
1.95% 
0.29% 
>3 
0% 
0% 
of captured fish (hook-and-line and trap samples) to 
divers’ visual counts. 
Age-1 red snapper composed the dominate age class 
on the 2010 reefs and recruited to these reefs in the 
early summer. Mean ±SD numbers of red snapper per 
cubic meter of reef structure increased as reef age 
increased (Pearson’s correlation coefficient [ /-] = 0.48, 
P= 0.008) and were significantly greater on 2006 reefs 
(22 ±13) than on 2009 reefs (12 ±6) and 2010 reefs (8 
±7; ANOVA, P 2 27 =4.25, P<0.025). 
All caught red snapper (n = 1028: 2006 reefs = 587, 
2009 reefs = 280, 2010 reefs=161) were used in the final 
age comparisons. Initial agreement between the first 
and second independent readings was 62.2% (639/1028). 
A third and fourth reading increased the accepted oto- 
liths to 92.3% (949/1028). Average percent error was 
calculated for both sets of independent readings (Table 
1). An age consensus was reached on all remaining 
otoliths (n- 79) through simultaneous examination by 
the 2 readers. The reference collection of age-1 hatchery 
(n=35, laboratory and wild reared) red snapper showed 
25.7% with 2 opaque bands, indicating that counting 
opaque bands for age-1 fish may not be reliable. Among 
fish that were <200 mm SL and showed 2 opaque bands 
(n=72), all were identified as age-1 based on shape, 
thickness, and location of the opaque bands (Beyer and 
Szedlmayer, 2010; Szedlmayer and Beyer, 2011). 
Mean ±SD red snapper SL, weight, and age were 
significantly different among 2006 reefs (373.3 ±107.8 
mm SL, 1883 ±1388 g, 3.5 ±1.2 years), 2009 reefs (250.2 
±114.7 mm SL, 852 ±1464 g, 2.0 ±1.7 years) and 2010 
reefs (222.3 ±78.0 mm SL, 480 ±711 g, 1.7 ±1.0 years; 
ANOVA, P 2 1025 =194.2, P<0.0001; Table 2; Figs. 2 and 
3). Reef age was positively correlated with red snap- 
per age (Pearson’s r=0.61, P<0.0001), standard length 
(r=0.71, P<0.0001), and weight (r = 0.47, P=0.0035). 
Comparisons of linear growth rates for fish <10 years 
old showed no significant differences between old (2006) 
