262 
PACIFIC SCIENCE, Vol. XIV, July I960 
Fig. 6. Portion of growth curve of albacore obtained 
by Wal ford’s method. Observed data are superimposed. 
low estimate of the upper asymptote (Fig. 5) 
and for the deviation of the observed data from 
the upper portions of both curves (Figs. 6, 7). 
Some studies have been made in the past on 
the age and growth of albacore, such as those by 
Uno (1936^; 1936b), Aikawa and Kato 
(1938), Partlo (1955), and Figueras (1955), 
all of which were based on the vertebral method 
of age estimation. Otsu and Uchida (1959 b) 
also attempted to determine the age of alba- 
core by this method and concluded in their 
study that the rings found on the centrum are 
probably not annuli as assumed by the several 
workers, but rather are growth marks laid down 
randomly with respect to time, and that, there- 
fore, aging could not be accomplished by this 
method. In addition, they thought that the verte- 
TABLE 3 
Growth Rates of Albacore Derived by Fitting 
Tag Recovery Data to Growth Curves by 
the Methods of Walford and Riffenburgh 
AGE* 
( years ) 
LENGTH (cm.) 
Walford 
Riffenburgh 
1 
— 
7.5 
2 
— 
17.3 
3 
26.3 
31.5 
4 
46.8 
46.5 
5 
62.7 
62.5 
6 
75.1 
75.0 
7 
84.8 
86.0 
8 
92.3 
94.5 
9 
98.2 
101.0 
10 
102.7 
106.0 
* The ages given are based on an arbitrarily selected origin 
(Fig. 7) and are therefore tentative. 
bral method generally yielded growth curves 
that were too linear to be typical and growth 
rates that were too slow when compared with 
the evidence from the then meager tag returns. 
The results of this study substantiate their crit- 
icisms. 
The growth curves in question are compared 
with the curve obtained in this study (Fig. 8). 
Since the ages assigned in this paper (Fig. 7, 
Table 3) were arbitrarily defined, and since 
there are obvious discrepancies in the ages as- 
signed by various workers, the mean lengths 
are plotted at yearly intervals without regard 
to the specific ages given. In other words, the 
curves in Figure 8 may be shifted freely in either 
direction along the abscissa. It is noted that the 
several growth rates do not differ markedly in 
magnitude. 
It is clearly seen, however, that the presently 
derived curve, at least a portion of which is 
believed to be a fair representation of the actual 
albacore growth curve, is curvilinear, while the 
others, with the possible exception of Uno’s 
( 193 6b) results, are almost perfectly linear. This 
difference in curvature supports the contention 
that the age of albacore could not be determined 
by the vertebral method. The linearity suggests 
that the rings found on the vertebrae are as- 
sociated with growth rather than with age, and 
that ring formation is a linear function of fish 
growth. 
Aside from growth studies by the vertebral 
by Riffenburgh’s short method. The curve obtained by 
Walford’s method (dashed) is superimposed. Observed 
data are superimposed on the Gompertz curve. 
