Coulson and Poad: Biological characteristics of Psettodes erume/ from the Indian Ocean 171 
of Okamura and Semba (2009) was applied to determine 
the periodicity of occurrence of otoliths of Indian hali- 
but with marginal increments falling within the lower 
30th percentile of values for each category of zone counts 
(Coulson et al., 2016, 2017). For this analysis, binomial 
models linked with von Mises circular distributions were 
fitted by assuming 1) no cycle, 2) an annual cycle, or 3) 
a biannual (twice yearly) cycle. When compared by using 
the Akaike information criterion (AIC), the model with 
the lowest value was deemed to best represent the data 
(Burnham and Anderson, 2002). 
The number of opaque zones in all sectioned otoliths 
were counted twice by the primary reader, with the 
counts from the second read considered the most accu- 
rate. The level of precision between the primary reader’s 
second counts and those of a second experienced reader 
(J. Poad) for a subsample of 280 otoliths was assessed by 
calculating the coefficient of variation (CV) (Chang, 1982; 
Campana, 2001): 
CV, = 100% x 
where CV; = the age precision estimate for the jth fish; 
X;; = the ith age determination for the jth fish; 
X; = the mean age estimate for the jth fish; and 
R = the number of times each fish is aged. 
Growth and sex ratio 
Each Indian halibut was assigned an age, on the basis 
of the number of opaque zones in sectioned otoliths, the 
date of capture of the fish, the “average” birth date (i.e., 
the approximate midpoint of the spawning period; see the 
“Reproductive biology” subsection in the “Results” section) 
of 1 December, and the time of year when the outermost 
opaque zone typically becomes delineated in the otoliths. 
Growth was described by fitting von Bertalanffy growth 
curves to the lengths at age of the females and males, 
with estimates of von Bertalanffy growth function (VBGF) 
parameters, and their 95% confidence limits, obtained by 
using a nonlinear function in R, vers. 3.4.1 (R Core Team, 
2017). The VBGF used was as follows: 
L = L,(1-e(-k(t - t))), (2) 
where L = the total length (in millimeters) at age ¢ (in 
years), 
L,,=the estimated asymptotic total length (in 
millimeters); 
k = the growth coefficient (year +); and 
ty) = the hypothetical age (in years) at which fish 
have zero length. 
A likelihood-ratio chi-square test was used to compare the 
growth curves from the VBGF for female and male Indian 
halibut. The test statistic was twice the difference between 
the log-likelihoods obtained by fitting growth curves to 
the lengths at age for each sex, separately, and by fitting 
a common growth curve to all lengths at age regardless 
of sex (Cerrato, 1990). The hypothesis that the growth of 
the 2 groups could be represented by a single growth curve 
was rejected at the level of significance (a) of 0.05 if the 
previously mentioned test statistic exceeded x2(q), where 
q is the difference between the numbers of parameters in 
the 2 approaches, for example, a difference of 3 numbers 
(Cerrato, 1990). The log-likelihood (A) for each curve, ignor- 
ing constants, was calculated as follows: 
d = (—n/2) In(ss/n), (3) 
where n = sample size; and 
ss = the sum of the squared residuals between the 
observed and expected lengths at age. 
Chi-square tests were used to determine whether the 
ratio of females to males in each of the main TL and age 
classes, as well as the sex ratio for all TLs and ages collec- 
tively, differed significantly from parity. 
Spawning time and maturation 
When identifiable as ovaries or testes, the mass of the 
gonads of each Indian halibut was weighed to the near- 
est 0.01 g. On the basis of its macroscopic characteristics, 
each ovary was allocated to one of the following 4 groups 
of maturity stages adapted from the criteria used by 
Laevastu (1965): virgin and immature or resting (stages 1 
and 2); developing and maturing (stages 3 and 4); pre- 
spawning and spawning (stages 5 and 6); and spent and 
recovering (stages 7 and 8). Ovaries at stages 3—7 in each 
year were considered likely to become mature (stages 3-5) 
or to have matured (stages 6—7) during that year; there- 
fore, for convenience, ovaries at these stages are referred 
to as mature. The prevalence of females and males with 
gonads at each developmental stage in each month was 
determined. 
Because all female Indian halibut >300 mm TL collected 
during the spawning period possessed mature gonads, the 
mean monthly gonadosomatic indices (GSIs) for females 
were determined by using all individuals greater than this 
length. For males, the mean monthly GSIs were deter- 
mined by using individuals with lengths greater than or 
equal to the TLs at which 50% of the males attained matu- 
rity (L;9) (see the “Lengths and ages at maturity” subsec- 
tion in the “Results” section). For both sexes, the mean 
monthly GSI was determined by using this equation: 
GSI = (GM /TM)/TM x 100, (4) 
where GM = wet gonad mass, and 
TM = wet total body mass. 
To confirm the macroscopic gonoad staging, histological 
sections were prepared from the gonads of a subsample of 
the Indian halibut collected in each month; the subsample 
of fish had a wide range of lengths, and their gonads were 
at all maturity stages found in fish in that month. Depend- 
ing on the size of the gonads, they were placed in Bouin’s 
fixative for 24—48 h, prior to being dehydrated in a series 
of increasing concentrations of ethanol. The mid-region 
