504 
Fishery Bulletin 99(3) 
FL = observed fork length of the fish in cm; 
R l = radius of the ring calculated as the aver- 
age value observed in ring i (Fig. 2B); and 
R = dorsal spine radius. 
Back-calculated fork lengths were used in Ford-Walford 
(Gulland, 1983) and nonlinear (Ratkowsky, 1983) meth- 
ods to fit the von Bertalanffy growth function (VBGF) 
and to obtain vital parameters by sex. Analysis of the 
residual sum of squares (ARSS) was employed to com- 
pare the VBGF between sexes (Ratkowsky, 1983; Chen 
et al., 1992). 
Weight was related to fork length by using the power 
function, and analysis of covariance (ANCOVA) (Steel, 
1980; Zar, 1999) was conducted to examine differences 
between sexes. 
Results 
Spines from 1149 specimens ranging in size from 45.6 
to 189.2 cm FL were examined (Table 1, Fig. 3). There 
was 90% agreement between the readers’ counts of growth 
rings and second readings improved this agreement to 
95.6%, which resulted in discarding 51 specimens from 
analysis. 
The relationship between FL (cm) and weight (kg) is 
shown in Figure 4. The ANCOVA indicated no significant 
difference between males and females (P>0.05); thus the 
FL-W relationship with sexes combined was expressed 
as 
W = 3 x 10- 5 FL 2 9278 (r 2 =0.97, n=856). 
The relationship of first dorsal spine lengths (L ws ) and FL 
was (Fig. 5) 
Table 1 
Sample sizes, ranges of fork lengths (FL, cm), and sampling 
months and areas of bigeye tuna from the western Pacific 
Ocean. A, B, C, D, and E denote areas in Figure 1. 
Sampling 
Month area 
Sample 
size 
Minimum 
FL 
Maximum 
FL 
Feb 1997 
A 
80 
70 
174.5 
Mar 1997 
A 
104 
64 
169.5 
Apr 1997 
B 
70 
101.3 
171 
May 1997 
B 
54 
83.8 
157.4 
Jun 1997 
B 
71 
75.5 
162.8 
Jul 1997 
B, D 
131 
72.2 
165.6 
Aug 1997 
E 
94 
78.5 
187.7 
Sep 1997 
E 
98 
45.6 
189.2 
Oct 1997 
A, C 
115 
86.5 
176.6 
Nov 1997 
A, C 
116 
89.6 
161.1 
Dec 1997 
C 
123 
104.6 
162.1 
Jan 1998 
A 
93 
88.7 
159.1 
Total 
1149 
45.6 
189.2 
FL = 6.9367 L ws + 6.6667 (/-0.94, n= 567). 
The trend of the monthly percentages of terminal trans- 
lucent edges (Fig. 6) suggested that the period from Febru- 
ary to September was the long period of inhibited growth 
(translucent edge). From October to November, growth ap- 
peared to resume (opaque edge) and later, from December 
to January, a new translucent edge appeared; indicating 
the formation of one growth ring per year. 
Given the significant linear relationship between the 
dorsal spine radius and fork length (FL=26.455R + 19.916, 
r=0.94, n = 1098), we used spine measurements to back- 
