Jeffers et al.: Habitat and bycatch effects on population parameters of Synodus foetens in the Gulf of Mexico 
421 
opportunistically in various locations to enhance 
sample sizes and the temporal coverage for oto- 
lith marginal condition analysis and aging (Fig 
1). Within study habitats, inshore lizardfish ap- 
peared in 91% of all trawl samples. Fish density 
was significantly different among sampling quar- 
ters and habitats (ANOVA, P<0.001 for both), 
but not between trawled and untrawled habitats 
(ANOVA, P=0.2754; Fig. 2). However, interpreta- 
tion of the main effects on density is complicated 
because of significant first-order interactions be- 
tween sampling quarter and trawl effects (ANO- 
VA, P<0.001) and habitat and trawling effects 
(ANOVA, P<0.001); there was no difference in sig- 
nificance for main effects or interactions between 
tests computed with or without the high shell 
habitat data. Inshore lizardfish were more abun- 
dant in sand habitat (mean density ±standard 
error [SE]=22.5 [±4.0] fish/ha) than in the other 
three habitats (mean density <13 fish/ha), but the 
level of difference was driven by two sand habitat 
samples taken outside the AR zone in summer 
2004 that yielded the highest estimated densities 
(105 and 65 fish/ha, respectively). Among study 
habitats not exposed to trawling, inshore lizard- 
fish densities were similarly high for sand and 
high shell habitats (Fig. 2A). 
Inshore lizardfish caught during quarterly sam- 
pling ranged in TL from 49-404 mm (Fig. 3A), 
whereas opportunistically sampled fish ranged in 
TL from 76 to 472 mm (Fig. 3B). Fish length was 
significantly different among sampling quarters 
(ANOVA, P<0.001), habitats (ANOVA, P<0.001), 
and levels of the trawling effect (ANOVA, P= 0.006; 
Fig. 3). However, interactions between sampling 
quarter and habitat (ANOVA, P<0.001) and habitat and 
the trawl effect (ANOVA, P<0.001) complicated inter- 
pretation of the main effects; there was no difference 
in significance for the main effects, or their interactions 
on fish size between tests computed with and without 
the high shell habitat data. Overall, mean (±SE) fish 
size was smaller in habitats exposed to trawling (230.7 
[±3.2] mm TL) than in habitats sampled within the AR 
zone (242.9 [±2.2] mm TL). However, high shell habitat 
outside the AR zone sampled in spring 2004 had the 
largest mean (±SE) size (320.0 [±44.4] mm TL) among 
all factor level combinations, although only three fish 
were captured in that habitat in spring 2004. Fish size 
was smallest (mean TL ±SE=204.7 [±4.6| mm) during 
summer 2004 among all habitats. That trend was most 
pronounced in sand habitat both inside and outside the 
AR zone where high densities (Fig. 2) of mostly small 
fish (Fig. 4) were encountered. 
Of the 970 otoliths prepared for age estimation (749 
from quarterly sampling and 221 opportunistically sam- 
pled), age could be determined for 967. Marginal condi- 
tion analysis demonstrated that otolith opaque zones 
generally began forming in November and continued to 
do so until February (Fig. 5). All samples were at least 
one year of age. The oldest fish sampled was a 424-mm- 
o 
Sand 
A 
Low-relief shell 
□ 
High-relief shell 
V 
Reef 
60 - 
40 - 
20 - 
€ 0 - 
•J 
£ 
ii $ 
n 
l f 
a 
o 
o° 
W 
>. 
Spring 04 
Summer 04 
Fall 04 
Winter 04 
Spring 05 
(/) 
S ioo- 
Q 
B 
80 ■ 
60 - 
« 
> 
40 ■ 
5 
20 - 
0 ■ 
*o v 
2 
o 
2 
2 
£ 
n 
A * 
Spring 04 
Summer 04 
Fall 04 
Winter 04 
Spring 05 
Quarter 
Figure 2 
Mean (±standard error) density (fish/ha) of inshore lizardfish 
( Synodus foetens) sampled from spring 2004 through spring 
2005 in study habitats in the northern Gulf of Mexico for (A) 
areas inside the artificial reef zone that were not subjected 
to trawling and (B) areas outside the artificial reef zone that 
were subjected to trawling. 
TL 9-year-old female that was sampled in June 2006 
onboard during the SEAMAP trawl survey. Of the fish 
sampled within study areas, the oldest was 8 years and 
nearly half of all fish were 3-year-olds. Reader agree- 
ment was judged to be good with an APE of 5.94%. 
Von Bertalanffy growth functions were not signif- 
icantly different between sexes (likelihood ratio % 2 - 
test, P- 0.998); thus size-at-age data were modeled 
jointly between sexes. The resultant growth equation 
was L t = 290.8(1 - e -0 - 486 (f-o.204)> (nonlinear regression, 
P<0.001; r 2 (coefficient of determination? If so, lower- 
case)^. 22) (Fig. 6). Estimates of M based on an ob- 
served t max of 9 yr were 0.51/yr, 0.49/yr, and 0.53/yr 
from the methods of Royce (1972), Hoenig (1983), and 
Pauly (1980), respectively. Before conducting catch 
curve analysis, the age distribution of the aged fish 
(n- 749) sampled in our study habitats was expanded to 
the samples collected that were not aged (>z = 490). This 
was accomplished by computing habitat- and sampling 
quarter-specific age distributions and by assigning age 
by means of a random number table for fish collected 
that were not aged. Then, catch curve analysis was 
computed for ages 3 yr and older because those were the 
fully recruited ages (Fig. 7). Total mortality of inshore 
lizardfish sampled within the AR zone was Z = 0.93/y 
