590 
Fishery Bulletin 99(4) 
Figure 2 
An example of a whole otolith that was unreadable (A) according to 
our criteria, but readable sectioned (B). Age was determined to be 
9 years. 
viewed whole. Some whole otoliths from central California 
were soaked in anise oil during initial handling to facilitate 
viewing annuli. The anise oil, however, left a residue that 
later caused problems in reading the otoliths. Therefore, 
these otoliths had to be sectioned before they could be read. 
For each otolith (either whole or sectioned), the two read- 
ers counted the annuli twice at different times, for a total 
of four independent readings. The birth date for a halibut 
was assigned to 1 January (Williams and Bedford, 1974). 
The mean and standard deviation of the four age deter- 
minations for each whole otolith were calculated. Because 
of our preliminary analysis, if the standard deviation was 
less than 1.5 years, the mean, rounded to the nearest in- 
teger, was the estimated age for that halibut. A standard 
deviation of 1.5 years was arbitrarily chosen as the upper 
limit for acceptance of an age determination. This devia- 
tion represented a difference in ages of 3 years or more 
among readings. If the standard deviation was 1.5 years 
or larger the whole otolith was judged unreadable. These 
otoliths were cut laterally through the nucleus with a di- 
amond blade on a Buehler low-speed Isomet saw. Three 
sections, approximately 0.38 mm thick, were cut from 
each otolith and mounted on a glass slide with Eukitt 
clear mounting medium. These sections were viewed with 
a compound microscope (25x to lOOx) with transmitted 
light. The best of three sections on a slide was chosen for 
age determination. Figure 2 is an example of a whole oto- 
lith that was unreadable, but readable when sectioned. 
The mean and standard deviation of the four age deter- 
minations for each sectioned otolith were calculated. In ac- 
cordance with our preliminary analysis, if the standard 
deviation was less than 1.5 years, the rounded mean was 
considered the estimated age. If the standard deviation 
was 1.5 years or larger, the otolith was not used in the 
analysis because it was felt to be unreadable. The index 
of average percent error was calculated to compare with- 
in-reader precision (Beamish and Fournier, 1981) for both 
whole and sectioned otoliths. 
Several analyses were made on total lengths and ages 
by sex and region. Comparisons of both length and age 
distributions between sexes for each region were made by 
using the two-sample Kolmogorov-Smirnov test (Holland- 
er and Wolfe, 1973). Length-at-age data for females were 
compared with those for males sampled from the same re- 
gion by using the nonparametric Mann-Whitney test. The 
Mann-Whitney test was also used to compare the length- 
at-age data between regions for each sex separately. 
The von Bertalanffy growth function was fitted to the 
length-at-age data for individual fish (Ricker, 1975); pa- 
rameters were estimated by nonlinear least squares by us- 
