654 
Fishery Bulletin 99(4) 
Accurate information on the age structure of the red 
snapper populations in the GOM is essential for monitor- 
ing year-class strength, for conducting stock assessments, 
and for documenting population recovery. Previous efforts 
at estimating red snapper age have employed a variety of 
aging methods. Bradley and Bryan (1975) cited the long 
spawning season and a constant recruitment into the pop- 
ulation as reasons for the difficulty in assigning red snap- 
per ages from length-frequency data. Moseley (1966) used 
scale annuli to age red snapper to age 4 yr and advanced 
spawning as the causal factor in check formation. Among 
240 red snapper taken off the west coast of Florida, Futch 
and Bruger (1976) estimated red snapper ages of 1 to 5 yr 
from 200 readable whole otoliths; however, they postu- 
lated ages up to 20 yr for larger individuals whose oto- 
liths were unreadable. Comparisons of ages derived from 
whole otoliths, scales, and vertebrae by Bortone and Hol- 
lingsworth (1980) revealed all three hard parts to be of 
equal utility in aging red snapper at age 1 and 2 yr. Wade 
(1981) also used scales to age red snapper to 9 yr. Nelson 
and Manooch (1982) reported red snapper ages from 1 to 
16 yr based on both scales and sectioned otoliths and dem- 
onstrated once yearly scale annulus formation in June and 
July from monthly mean marginal growth. A recent study 
of red snapper otoliths significantly extended the hypothe- 
sized longevity of red snapper in the GOM to 42 yr (Szedl- 
mayer and Shipp, 1994). Render (1995) provided a pre- 
liminary validation of yearly opaque annulus formation in 
sagittal otoliths and reported ages from 0 to 53 yr for red 
snapper in Louisiana waters. Examinations of otolith sec- 
tions from 537 red snapper captured in the northwestern 
Atlantic Ocean from Beaufort, North Carolina, south to 
the Florida Keys manifested a maximum longevity of 25 yr 
(Manooch and Potts, 1997). Among 907 red snapper from 
the GOM off Alabama, Patterson (1999) reported opaque 
annulus formation from January through June and maxi- 
mum ages of 30 yr for females and 31 yr for males. Despite 
these efforts, the longevity of red snapper remains contro- 
versial. Small sample sizes, a paucity of older specimens, 
and the failure to present legitimate validations of ages 
from hard parts (Beamish and McFarlane, 1983) have var- 
iously hampered the above studies. It has further been 
speculated that larger and presumably older red snapper 
form numerous false annuli within otoliths (Rothschild et 
al. 3 ) . And both the timing of deposition and the position of 
the putative first annulus remain in question. 
Otolith analyses have proven consistent in estimating 
ages of many fish species, including several from the tem- 
perate waters of the northern GOM (Johnson et al., 1983; 
Barger, 1985; Beckman et al., 1988; Beckman et al., 1990, 
1991; Crabtree et al., 1992; Murphy and Taylor, 1994; 
Franks et al., 1998; Thompson et al., 1998). Herein we pres- 
2 (continued) oratory, Southeast Fisheries Science Center, National 
Marine Fisheries Service, 75 Virginia Beacli Drive, Miami, FL 
33149-1099. [Not available from NTIS.] 
3 Rothschild, B. J., A. F. Sharov, and A. Y. Bobyrev. 1997. Red 
snapper stock assessment and management for the Gulf of 
Mexico. Report submitted to the National Marine Fisheries 
Service, Office of Science and Technology, 1315 East-West High- 
way, Silver Spring, MD 20910 USA. [Not available from NTIS.] 
ent our interpretations of the use of sagittal otoliths to es- 
timate ages of red snapper from the GOM off Louisiana. 
Specifically we address the timing of formation and posi- 
tion of the first annulus, validation of the once yearly accre- 
tion of opaque annuli, longevity, and reader reproducibility. 
We further describe the growth of red snapper with von 
Bertalanffy growth models for both males and females. 
Methods and materials 
Red snapper from recreational and commercial catches 
were sampled from 1989 to 1992 and from 1995 to 1998 by 
personnel of the Louisiana State University (LSU) Coastal 
Fisheries Institute and the Louisiana Department of Wild- 
life and Fisheries (LDWF). Although the vast majority of 
our sampling efforts were targeted at both wholesale facil- 
ities and charter boat docks located in Grand Isle and 
Port Fourchon, LA, the area of coverage in the northern 
GOM extended from off the Mississippi River Delta in the 
east to off Galveston, TX, in the west. Morphometric mea- 
surements (total length [TL] or fork length [FL] in mm, 
total weight [TW] or eviscerated body weight [BW], i.e. 
mass with liver, digestive tract, and reproductive organs 
removed, in g) were recorded, both sagittal otoliths were 
removed, and sex was determined, when possible, for each 
specimen. Eviscerated body weight was converted to TW, 
when necessary, with the equation TW-1.10KBW) - 26.32 
(linear regression, df=418; P<0.001; ^=0.996) and TL was 
estimated from FL with the equation TL=1.073 (FL) + 
3.56 (linear regression, df= 1015; P<0. 001; r 2 =0.999) . 
All undamaged, intact otoliths were weighed to the near- 
er 0.1 g. The left sagittal otolith from each individual was 
embedded in an epoxy resin and subsequently sectioned 
with a low-speed saw equipped with a watering blade as 
described in Beckman et al. (1988). In those instances 
where the left sagitta was damaged or unavailable, the 
right sagitta was substituted. Examinations of otolith sec- 
tions were made with a compound microscope and trans- 
mitted light at 40x to lOOx magnification. Counts of an- 
nuli (opaque zones) were accomplished by reading along 
the medial surface of the transverse section dorsal or ven- 
tral to the sulcus; annuli were often inconsistent in other 
regions of the otolith section. Annuli were counted by two 
readers without knowledge of date of capture or morpho- 
metric data. The appearance of the otolith margin was al- 
so coded as either opaque or translucent (Beckman et al., 
1988). Sections were recounted a second time by both read- 
ers when initial counts disagreed. Rather than excluding 
the small number of individuals for which a consensus 
could not be reached after a second reading (n=27), the as- 
signed annulus count for these was that of the more experi- 
enced reader (reader 1). Reader 2’s annulus count and edge 
condition were used in those circumstances where reader 
l’s were missing (n= 2). Annulus counting error between 
the two readers was evaluated after both the initial and 
second readings of the otolith sections. Reproducibility of 
the resultant age estimates was evaluated with the coeffi- 
cient of variation, the index of precision (Chang, 1982), and 
average percent error (Beamish and Fournier, 1981). 
