Craig et al.: Population biology and harvest of Acanthurus lineatus 
683 
In the second method, individual growth rates were 
calculated for a subset of the naturally marked fish 
described in the field mortality study (see next sec- 
tion). The 57 fish selected were those for whom a 
time series of 4-20 size estimates was available for 
each fish. Lengths were estimated visually under- 
water; a comparison of visual estimates with actual 
sizes of the same fish when caught by spear indi- 
cated that the average error was 8.1 ± 1.5% (mean 
and SE throughout text, n=19, 6-20 cm FL). 
The fish were initially selected from three general 
size classes according to their size at settlement onto 
the reef (2.5-5 cm) and approximate state of matu- 
rity based on dissection data (juveniles 6-14 cm, 
adults 15-23 cm). Sample sizes were 11 newly settled 
fish (monitored 1.7 ± 0.3 mo), 28 juveniles (5.2 ± 0.6 
mo), and 18 adults (14.1 ± 1.0 mo). These fish were 
grouped into eight size classes of 2.5-cm intervals on 
the basis of their initial sizes. By using the mean 
growth rate of each size class, we calculated the time 
needed to grow to the next size class. These growth 
increments were plotted sequentially, forming a 
single growth curve for the population. A Gulland- 
Holt (1959) plot of growth increments of individual 
fish produced estimates of and K. 
Mortality 
Total mortality (Z), which equals natural mortality 
(M) plus mortality caused by fishing (F), was esti- 
mated by monitoring the gradual loss of 145 marked 
fish for three years at the Afao site and by analyzing 
the length and age composition of fish taken in the 
artisanal fishery. 
Field mortality Earlier work had shown that A. 
lineatus was highly site-attached (Craig, 1996); thus 
we initially assumed that a fish had died if it failed 
to re-occupy its territory or nearby area. Individual 
adults (n=45) and juveniles (n=50) were recognized 
by distinctive line patterns behind the eye and on 
the cheek. Sexual dimorphism was not apparent, thus 
males and females were not distinguished in the field. 
Newly settled fish {n= 50) were identified by a com- 
bination of their specific location, size, color phase, 6 
and line pattern when discernible. Because newly 
settled fish were selected on the basis of identifiabil- 
ity rather than first appearance in the study area, 
the time elapsed since settlement was not known. 
6 Color phases of newly settled A. lineatus are not described in 
the literature. The “light” phase is that of adult coloration; the 
“dark” phase is light-to-dark gray (overlying a faint adult color 
pattern) and the caudal fin is orange (differentiated from dark 
newly settled Ctenochaetus striatus which have orange only on 
caudal fin tips). 
However, surveys were conducted frequently; there- 
fore most newly settled fish had probably arrived 
within the previous week or two. 
On average, about 35 fish were monitored at any 
one time; new individuals were added when others 
either outgrew their size class or were not relocated 
after three successive surveys. Small fish were in- 
spected at least twice each week, larger fish about 
once per week. All three size groups were intermixed 
on the outer reef flat. 
To calculate mortality, all fish within a size group 
were aligned to a common starting date. For each 
size class, mortality at any given time equalled 1 - 
(no. fish alive + no. fish outgrowing size class)/(ini- 
tial no. fish in that size class). This approach 1) un- 
derestimated mortality for newly settled fish if there 
had been high mortality during the first days of 
settlement before observations began, or 2) overesti- 
mated mortality if observed fish emigrated from the 
study area. Total mortality (Z) was calculated as the 
slope of the descending limb of the “catch curve” (a 
plot of the natural logarithm of fish remaining each 
year versus relative age). Annual mortality was es- 
timated as 1 - e~ z (Ricker, 1975). 
Total mortality Total mortality for harvested fish 
was estimated in several ways: 1) length-converted 
catch curves (Pauly, 1983); 2) the relation between Z 
and mean length of fish in the catch: 
Z = K(L m - L c )/(L c - L'), 
where L c = the average length of fish greater than 
length L'; and 
L' = the size at which fish are assumed to be 
fully recruited to the fishery (Beverton 
and Holt, 1957); 
and 3) Hoenig’s ( 1983) empirical relation between Z 
and a population’s longevity: 
ln(Z) = 1.46-1.01 ln(* m(W ), 
where t = maximum age. 
Natural mortality Natural mortality (M) was esti- 
mated with two empirical equations: 
log M = 0.007 - 0.279 log + 
0.654 log if + 0.463 log T (Pauly, 1980), (1) 
where T = the average monthly water temperature 
in the study area (28.6°C), and 
In (M/K) = 0.30 ln(Tj - 0.22 
(Longhurst and Pauly, 1987). 
(2) 
