460 
Fishery Bulletin 1 10(4) 
Szedlmayer, 2012). Red snapper were collected with 
hook and line and fish trap from each reef. Hook-and- 
line sampling was standardized to 30 min and 2 indi- 
viduals who fished. Fishing time was suspended when 
problems occurred (e.g., internally hooked fish) and 
continued once both individuals could resume fishing. 
For hook-and-line fishing, double 6/0 J hooks, 27.2- 
kg test monofilament line, 45.3-kg test monofilament 
leader were used, and whole Gulf menhaden (Brevoortia 
patronus ) as bait. After completion of hook-and-line 
sampling, additional fish were collected with a baited 
fish trap (1.2x1. 5x0. 6 m; Collins, 1990). In the fish trap, 
both Gulf menhaden and whole squid ( Loligo spp.) were 
used as bait. All fish traps were set for 15 min. After 
collections of red snapper reached ~50 individuals per 
reef, a diver released the number of fish caught above 
50 fish by opening the trap door and allowing random 
individuals to escape at the surface — with one exception 
(73 red snapper were kept on 5 May 2010 because of the 
possibility of area closures that might have resulted 
from the Deepwater Horizon oil spill). When the mini- 
mum target of 30 individuals per reef was not reached 
after the first fish trap set, the trap was fished at least 
one additional time. All red snapper collected from a 
reef were immediately packed on ice and returned to 
the laboratory for further processing. 
After fish collections were completed, 2 scuba div- 
ers completed visual counts and photographic (Nikon 
D200 2 , Nikon Corp., Tokyo) and video (Sony CCD- 
TR101, Hi8, Sony Corp., Tokyo) recordings to use in 
estimation of the remaining red snapper at each sample 
site. A clear plastic jar containing cut Gulf menhaden 
was used to attract surrounding red snapper into ag- 
gregations during visual surveys for increased accuracy 
of total counts. Divers completed at least 3 full-circle 
point-and-count surveys, where the divers counted all 
fish within visual circular range, and the highest count 
was used for estimates of total abundance. Poor vis- 
ibility at some sites limited these estimates. In addi- 
tion, when sharks were present, diver operations were 
suspended, and visual estimates were completed later 
within 30 days of the original fish collections. 
Laboratory analyses 
Red snapper size (standard length [SL], fork length [FL], 
and total length [TL] in millimeters) and total body 
weight (0.1 g) were measured in the laboratory within 
24 h of capture. For red snapper >250 mm TL, otoliths 
were removed with a Bosch fine-cut electric saw. For 
red snapper <250 mm TL, otoliths were removed with 
a small knife. Both left and right otoliths were removed 
from each fish, cleaned, and stored in dry plastic vials 
for later analysis. Opaque bands were counted on all oto- 
liths for age estimates. For fish <7 years old, bands were 
counted on whole otoliths that were immersed in water 
2 Mention of trade names or commercial companies is for 
identification purposes only and does not imply endorsement 
by the National Marine Fisheries Service, NOAA. 
and viewed under a dissecting scope with transmitted 
light. If ages were >7 years, thin otolith sections were 
prepared and bands were counted at 40x magnification 
with a compound microscope (Szedlmayer and Beyer, 
2011). Opaque bands of sectioned otoliths were counted 
along the dorsal edge of the sulcus acousticus. Bands on 
each otolith were counted independently 4 times. After 
4 readings, 2 readers examined the remaining otoliths 
for which counts still differed and attempted to reach a 
consensus on age. If an agreement on age could not be 
reached for an otolith, it was rejected. A reference col- 
lection of hatchery red snapper that had been released 
in the wild as age-0 and recaptured as age-lfish (n= 22) 
along with a group that had been reared in captivity to 
age-1 (n = 13) were used to validate counting methods 
of wild caught age-1 fish. Some of the otoliths of these 
known age-1 fish showed a “false” annulus (i.e., had 2 
opaque bands) but showed age-1 otolith shape patterns 
(Beyer and Szedlmayer, 2010). Therefore, some of the 
wild fish <200 mm SL caught in this study that had 2 
opaque bands were defined as age-1, on the basis of their 
age-1 shape patterns appearing similar to the shape pat- 
terns of hatchery-reared fish as well as hatchery-born 
but wild-reared fish. 
Video recordings and digital photographs of the stud- 
ied reefs were examined in the laboratory for compari- 
sons and validation of the divers’ visual counts. In the 
laboratory, photographs that showed the highest num- 
ber of red snapper for a particular reef were selected 
for computer-based counting. All red snapper in these 
photographs were identified and counted through image 
analysis with Image-Pro Plus software (vers. 4.5, Media 
Cybernetics, Rockville, MD). Two screens were used to 
count fish in video recordings. A single frame captured 
from the video was displayed on one screen while the 
video played on the second screen. Because image qual- 
ity decreases when a single frame of video is captured, 
we used the full video on the second screen to identify 
all the fish in the captured frame on the first screen. 
The captured video image then was marked and the fish 
in it were counted with Image-Pro software. 
Data analyses 
Catch per unit of effort (CPUE) for each reef was calcu- 
lated for both hook and line (CPUE=number caught by 2 
individuals/30 min) and trap (CPUE=number caught/15 
min). The precision of age estimates between readers 
was compared with linear regression and average per- 
cent error (Beamish and Fournier, 1981). Densities of 
red snapper were estimated by adding the total number 
of red snapper caught (from both hook and line and 
fish trap) to the number of fish counted in the visual 
survey. Pearson’s correlation coefficient was used to 
compare densities (number of fish per cubic meter of 
reef surveyed) among reefs of different ages, with reef 
age determined by the number of months that a reef had 
been deployed before the month of sampling. Analysis of 
variance (ANOVA) was used to compare the SL, weights, 
