682 
Fishery Bulletin 95(4), 1997 
and only 3.5% live coral cover (Craig, 1996). Spawn- 
ing and nocturnal rest sites at Afao are shown in 
Figure 2 for those A. lineatus that maintained day- 
time feeding territories in the study area. 
Length, weight, sex, and maturity 
Length frequencies of fish in the artisanal catch were 
measured at 54 occasional intervals from 1987 to 
1995. During 1991-94, fish were purchased at local 
markets to determine fork length (to the nearest 
mm), weight (to 0.1 g), sex, and maturity. Pooled 
monthly samples for these years consisted of 18-33 
mature females, 18-49 mature males, and 9-116 
immature fish, for a total of 1,139 fish. Maturity was 
assessed by visual inspection of gonads and by 
gonadosomatic index (GSI = 10 2 x gonad weight/ 
whole body weight). To determine maturity-at-size, 
immature fish whose sex could not be determined 
(13% of total sample) were assigned in equal propor- 
tions to numbers of identified males and females 
because the sex ratio of identified males and females 
was equal. Limited samples of rotenone-treated fish 
were collected at several nearshore sites in August 
1990 to compare with sizes of fish taken in the 
artisanal fishery. 
Spawning 
Seasonal spawning patterns were determined from 
GSI trends and by conducting monthly visual sur- 
veys in the outer reef channel at the Afao site, 1993- 
94. Confirmation of spawning was determined by an 
upward rush of fish with the production of visible 
milt clouds. 
Settlement of young onto reef 
At each of the three study sites, newly settled fish in 
five 2 x 20 m permanent transects along the outer edge 
of the reef flat were censused monthly, approximately 
one week after the new moon. A repeated-measures 
multivariate analysis-of-variance (MANOVA with 
Pillai’s trace test statistic) was used to test for signifi- 
cant differences in settlement time among the three 
sites and to accommodate autocorrelation of counts 
among observed times (Tabachnick and Fidell, 1989). 
Condition factor fCFJ 
We used two measures of fish condition: 1) as 10 5 x 
body weight/length 3 , and 2) as the weight of the 
paired postabdominal fat bodies found in surgeon- 
fishes (Fishelson et al., 1985). To examine seasonal 
changes, monthly mean CF values and fat body 
weights were compared with trends in five environ- 
mental factors that might affect fish growth: 1) 
nearshore water temperature, 2) available feeding 
hours, 3) calm surf conditions, 4) rainfall, and 5) 
daylength. Available feeding hours were calculated 
as the number of feeding hours per month during 
the fish’s daily peak feeding period (1000-1800; 
Craig, 1996), minus losses of feeding time during the 
cooler season due to earlier sunsets and increased 
occurrence of spring low tides that prevented access 
to the fish’s intertidal feeding territories. An index 
of calm surf conditions was calculated as the inverse 
of wind speed, because seasonally strong winds in- 
crease turbulence in the surf zone, thereby decreas- 
ing feeding opportunities for A. lineatus (Craig, 1996). 
Rainfall was used as a possible indicator of the 
amount of nutrient input into coastal waters that 
might, in turn, enhance growth of the algae that the 
fish eat. Similarly, daylength might affect algal pro- 
duction. Monthly means of these five factors were cal- 
culated for the years 1991-94 and presented as per- 
centages of the maximum monthly value that occurred 
during this period, which were water temperature 
(29.6°C), available feeding period (225 h), wind speed 
(27 km/h), rainfall (72 cm), and daylength (13.0 h). 
Growth 
Growth data were fitted to the von Bertalanffy 
growth function (VBGF): 
l t = L„[ 1 - 
where l ( = length at age t; 
L ^ = asymptotic length; 
K = growth coefficient; and 
/ 0 = time when length would theoretically be 
zero. 
Two independent estimates of fish growth were made. 
In the first method, sagittal otoliths were used to 
estimate ages of 94 fish selected to span the widest 
size range possible (5.3-22.9 cm FL from pooled lo- 
cations) with the methods described by Choat and 
Axe (1996), who aged the same species (by including 
tetracycline verification) from the Great Barrier Reef. 
Estimates of and K were derived from a Ford- 
Walford regression of the age-length relation (Pauly, 
1983). Estimates of t 0 were made with Pauly’s (1979) 
empirical equation: 
log(-/ 0 ) = -0.392 - 0.275 log - 1.038 log K. 
Additionally, one of each otolith pair was weighed to 
10 -4 g for comparison with fish age. 
