422 
Fishery Bulletin 108(4) 
tophorus, L. carponotatus, L. gibbus, L. fulviflamma) . 
Sufficient specimens were collected for comparisons 
among regions for only one species (L. carponotatus) 
and specimens for all other species were pooled across 
regions and years for analyses. 
Fish processing 
All fish were measured to the nearest millimeter fork 
length (FL) and weighed to the nearest gram before 
dissection. Otoliths were dissected from a subsample of 
specimens, cleaned, and stored dry in paper envelopes. 
Otoliths were embedded in epoxy resin and cut trans- 
versely through the primordium with a diamond-tipped 
blade on a low-speed saw to produce a thin section of 
300-400 pm. Sections were mounted on glass slides 
by using Crystalbond adhesive (Aremco Products Inc., 
New York). Otoliths were read under reflected light at 
40 x magnification and opaque growth increments were 
counted from the primordium to the proximal (in situ) 
margin along the ventral ridge of the sulcus acousticus. 
Sectioned otoliths were read by one or more individuals. 
Where more than one reader was used, otoliths were 
read at least once by each of two independent read- 
ers and the age was accepted if these first two counts 
agreed. A third count was made by one of the two read- 
ers if the first two counts did not agree and a match 
between this third count and either of the previous two 
was accepted as the age of the fish or, if the third count 
did not match either of the first two counts, the median 
count was assigned as a final age estimate. The remain- 
ing otoliths were read by a single, experienced reader 
who counted each otolith at least twice with a minimum 
of 24 hours between consecutive counts. If the first two 
counts did not agree, a third count was completed to 
assign a final age through count agreement or to assign 
a median age, as described above. 
Opaque increments were assumed to be annuli for the 
five species examined. Supportive evidence for annual 
periodicity in opaque increment deposition in otoliths 
has been demonstrated for the majority of tropical reef 
fish (Fowler, 1995; Choat and Robertson, 2002), lutja- 
nids (e.g., Cappo et al., 2000), and for most of the study 
species: L. vitta (Davis and West, 1992; Newman et al., 
2000a), L. carponotatus (Newman et al., 2000a; Kritzer, 
2002), L. adetii (Newman et al., 1996), L. fulviflamma 
(Grandcourt et al., 2006), and A. virescens (Pilling et 
al., 2000). Age estimates derived from the otoliths for 
L. gibbus and S. nematophorus have not been validated 
hitherto, but we assumed that increment deposition also 
occurs on an annual basis for these species, as has been 
demonstrated for the majority of other tropical lutjanids. 
A violation of this assumption is likely to result in bi- 
ases in the parameter estimates we present, such as for 
the von Bertalanffy parameter K and the instantaneous 
total mortality coefficient, Z. The magnitude of bias 
would likely inflate estimates for older-age species, and 
these estimates would vary among parameters, with a 
likely smaller influence on estimates of L x (mean as- 
ymptotic fork length) (Choat et al., 2009). Gonads were 
examined to define the sex and maturity of individuals, 
but low sample sizes prevented meaningful analysis of 
detailed histological staging or gonadosomatic index 
analyses. Gonads were dissected and either frozen or 
preserved in FAACC (formaldehyde 4%, acetic acid 5%, 
calcium chloride 1.3%) or 10% buffered formalin imme- 
diately after removal. Tissues were preserved in FAACC 
before May 1999 and 10% buffered formalin thereafter. 
The sex and maturity of L. carponotatus was estab- 
lished by macroscopic methods (Kritzer, 2004). The sex 
and maturity for a subset of individuals of other species 
were further assessed by histological analysis. Trans- 
verse histological sections were taken from the medial 
region of all gonads after the procedures outlined by 
Adams (2003). Female maturity was determined by 
the presence of vitellogenic oocytes, and the presence of 
brown bodies, atretic oocytes, vascularization, and the 
relative thickness of the gonad wall, all of which may 
indicate prior spawning (Sadovy and Shapiro, 1987; 
Ferreira, 1995; Adams, 2003). Male maturity was deter- 
mined by the presence of sperm in the sperm sinuses. 
Estimation of demographic parameters 
Length- and age-frequency distributions were constructed 
from pooled samples and compared among species quali- 
tatively. Differences in length and age frequencies by sex 
were assessed by using two-sample Kolmogorov- Smirnov 
tests where the largest of absolute discrepancies between 
relative cumulative frequency distributions (c? max ) was 
compared with an approximate two-tailed critical value 
(D\ Sokal and Rohlf, 1995). Growth was modeled with 
the von Bertalanffy growth function (VBGF) fitted by 
using a nonlinear least-squares regression of FL on age. 
The form of the VBGF was 
where L t = 
L„ = 
K = 
*0 = 
the length at age t\ 
the mean asymptotic fork length; 
the growth coefficient or rate at which 
is approached; and 
theoretical age at zero length. 
The VBGF model was fitted to L t on t data groups where 
t 0 was either constrained to zero (t 0 =0) or unconstrained 
(estimated). This was done to examine the effects of 
the absence of smaller individuals in the sample and 
(for t 0 = 0) to allow biologically sensible comparisons of 
growth among species if lengths at age of younger fish 
(not sampled) extrapolated by the model were biologically 
unrealistic. VBGFs for L. carponotatus for all regions 
were compared by likelihood ratio tests. 
Analysis of covariance (ANCOVA) was used to com- 
pare log-transformed length data from individuals 
across a common age range (5-12 years) among the 
five species with age as a covariate. A Tukey’s post hoc 
test (P=0.05) was used to test for differences among 
species when a significant difference among the groups 
was detected by the ANCOVA. 
