Acanthurus triostegus sandvicensis — Randall 
cream in color, smooth, slightly compressed lat- 
erally, and closely applied to one another in the 
mid-line. The testes are white, strongly com- 
pressed, irregularly lobular, and in contact with 
each other only ventrally. 
It was apparent from sampling the catch of 
trap fishermen throughout the year that approxi- 
mately twice as many male manini as females 
were taken. In the month of May, 1953, 291 
trap -caught adult fish were sexed; 68 per cent 
of these were males. The traps in which the fish 
are caught are unbaited. A typical trap consists 
of a rectangular framework ( about 2X5X6 
ft.) of steel rod covered with chicken wire. A 
cone of chicken wire with a narrow slitlike 
opening extends inward from one end to about 
the center. Some trap fishermen prefer to leave 
one or two butterfly fish or other fishes of little 
economic value in the trap in the belief that 
other fishes will be more prone to enter. If male 
manini are more gregarious or less cautious or 
more migratory than females, they would be 
caught in traps with greater frequency than 
females; therefore sex ratios ascertained from 
trapped fish should be viewed with caution. 
The sex was determined for 221 adult manini 
from islands in the tropical Pacific in the col- 
lections of the U. S. National Museum, the Ber- 
nice P. Bishop Museum, and the University of 
Hawaii. Most of these specimens were taken 
with rotenone, seine, and throw net. These col- 
lective techniques are less selective than traps 
and the sample would seem to be a better indi- 
cation of true sex ratio. Of the 221 fish, 134 
(60.7 per cent) are females. A chi-square value 
of 10 results from testing whether 134 females 
could be obtained from 221 fish randomly sam- 
pled from a population in which the true sex 
ratio is 50-50. This high chi-square has a prob- 
ability greater than 0.01 that a sample more 
extreme than the above could be drawn by 
chance from a half male, half female popula- 
tion; thus the hypothesis of a 50-50 sex ratio 
is rejected. 
It is nevertheless possible that the apparent 
predominance of female manini is spurious. 
Most of the 221 specimens were taken in inshore 
areas. The traps from which 198 males of 291 
manini were caught were set in from 30 to 90 
ft. of water. Perhaps females are more abundant 
231 
in shallower water and the males predominate 
over deeper parts of the reef. More collections 
are needed to clarify this problem. 
The museum collections of manini offer some 
slight evidence for differential schooling by 
sexes. One sample of 7 manini from the outer 
reef at Eniwetok Atoll, Marshall Islands, is en- 
tirely male; another of 12 fish from the lagoon 
of Kwajalein Atoll in the Marshalls is wholly 
female. 
Spawning Cycle 
The manini in the Hawaiian Islands has a 
distinct spawning season. This has been ascer- 
tained by the examination of the gonads of adult 
fish and collections of young throughout the 
year. In more equatorial areas, on the contrary, 
the spawning of this surgeonfish appears to be 
year-round. There is also evidence both in Ha- 
waii and the Gilbert Islands that the pattern of 
the incoming young correlates with the lunar 
cycle. This is attributed to a lunar effect on 
spawning by adults. 
From October 31, 1952, to October 27, 1953, 
a total of 137 female and 156 male manini 100 
mm. or more in standard length were purchased 
in 41 samples from a fisherman for the purpose 
of examination of the gonads. The fish were 
caught in traps set in depths of 30-90 ft. from 
Kewalo Basin to Koko Head, Oahu. The length 
of the gonad on each side of the body cavity of 
each specimen was measured and the average 
length recorded. This average measurement was 
divided by the standard length and the quotient 
multiplied by 100. The range and mean of the 
resulting percentages are presented in Table 5. 
Although the monthly samples are small, it 
is nevertheless apparent that the gonads of both 
sexes undergo a significant change in relative 
length during the year. They are largest in late 
winter and spring and smallest in late summer 
and early fall. The change which takes place in 
the ovary in late July and August is more com- 
plex than would be indicated from a mere short- 
ening. The ovary (presumably following the last 
spawning) is red, flattened, and has a large lu- 
men. Residual eggs are gradually absorbed ( eggs 
being absorbed appear as amorphous masses of 
variable size ) , and the entire organ regresses to 
small size and a translucent gray color. There is 
