Wang and Heino: Growth and maturation of Trichiurus japonicus in the subtropical Pacific Ocean 
1 77 
evaluated differences in growth and maturation be¬ 
tween sexes using similar randomization tests. 
Results 
Length distribution and length at age 
T. japonicus sampled in Kengfang (NE) were on aver¬ 
age larger (mean pre-anal length: 216.0 mm [SD 60.3]) 
than fish from Tsukuan (SW; mean pre-anal length: 
200.3 mm [SD 38.8]). Length distribution for Kengfang 
also was broader than that found for Tsukuan (coeffi¬ 
cient of variation (CV): 28% versus 19%; Fig. 2). 
We observed an ontogenetic shift in growth varia¬ 
tion between the 2 areas. For immature fish at ages 
of 0 and 1 year, mean lengths were lower in Keng¬ 
fang than in Tsukuan (Fig. 3A). Subsequently, for ma¬ 
ture fish of both sexes at ages <2 years, differences in 
lengths between the areas were not significant (Fig. 3, 
B and C). However, the patterns of lengths at ages be¬ 
tween the areas were partly reversed for fish at ages 
>3 years with greater lengths at age in Kengfang than 
in Tsukuan (Fig. 3, B and C). 
Early juvenile growth 
We estimated average daily ages for 17 and 19 age-0 
fish in Kengfang (NE) and Tsukuan (SW), respective¬ 
ly (Suppl. Table 4) (online only). Average daily ages and 
pre-anal lengths of fish were higher in Tsukuan than 
in Kengfang (Suppl. Table 4; for daily ages, 1 1 1 =3.56, 
df=21, P=0.002; for pre-anal lengths, 1 1 \ =2.63, df=29.9, 
P=0.01). However, growth of very young T. japonicus 
was similar between the areas on the basis of both es¬ 
timates of daily growth rates (see Suppl. Table 4 for 
the ratios of pre-anal-length to daily age, lfl=0.41, 
df=17.0, P=0.68; for the regression-based daily growth 
rate, P=1.43, P=0.26). 
Growth model 
The magnitudes of area- and-sex-specific estimates of 
and K varied greatly between the 2 different model 
fits (e.g., Equations 1-2; Table 1). However, for out¬ 
put from both models, consistent growth patterns be¬ 
tween the areas and sexes were observed (Table 1; Fig. 
4). Generally speaking, L„ was higher and K lower in 
Kengfang (NE) than in Tsukuan (SW). Females had 
higher L„ and lower K than those of males (Table 1; 
Fig. 4). On the basis of randomization tests, the be- 
tween-area differences in both and K were signifi¬ 
cant for both models: for of Equation 1, P<0.001 
for males and P=0.002 for females; for L„ of Equation 
2, P<0.001 for males and P=0.008 for females; for K 
of Equation 1, P<0.001 for males and P=0.003 for fe¬ 
males; and for K of Equation 2, P<0.001 for males and 
P=0.003 for females (Table 1; results of randomization 
tests of between-sex differences in L„ and K are shown 
in Suppl. Table 5) (online only). These differences are in 
300 
cn 
c 
0) 
To 
c 
cp 
CD 
CL 
200 - 
100 
0 1 
Males 
B 
b B. 
B 
,0? 1 
B 
n—i—i—i—i—i—i—i—i—i—i—r 
KTKTKTKTKTKT 
0 1 2 3 4 5 
12 3 4 
Age (years) 
Figure 3 
Length-at-age comparisons between 2 fishing areas off 
Taiwan near the ports of Kengfang (K, open boxes) and 
Tsukuan (T, gray boxes), based on sex- and maturation- 
state-specific data for (A) immature (B) male, and (C) 
female cutlassfish, Trichiurus japonicus, sampled in 
2013-2015. The secondary x-axis shows the age group 
of fish. The letters a and b denote significant differ¬ 
ences between the areas based on f-test results for 
immature fish at age 0 (If 1=2.01, df=35.9, P=0.05), 
immature fish at age 1 (If I =4.67, df=61.4, P<0.001), 
males at age 3 (I f I =3.08, df=61.0, P=0.003), females at 
age 3 (1 1 1 =2.29, df=91.9, P 0.02), and females at age 5 
(If 1=15.73, df=2.7, P=0.001). 
