Sipe and Chittenden: A comparison of calcified structures for aging Parcilichthys dentatus 
633 
Table I 
Average preparation times (min), reading times (min) ± 
standard error (SE ), and confidence scores ( ±SE ) for summer 
flounder calcified structures. 
Preparation 
Reading 
Confidence 
Structure 
time 
time 
score 
Opercular bones 
4.63 
2.43 ±0.20 
2.31 ±0.16 
Scales 
10.50 
1.20 ±0.13 
3.21 ±0.15 
Sectioned otoliths 
Right 
5.86 
0.27 ±0.04 
4.91 ±0.04 
Left 
13.93 
0.57 ±0.09 
4.75 ±0.05 
Whole otoliths 
Right 
0.00 
0.45 ±0.06 
4.10 ±0.11 
Left 
0.00 
0.41 ±0.04 
3.84 ±0.10 
left whole otoliths; however this could be attributed to the 
reader learning to use reflected lighting more effectively 
during the second reading, because 7 of the 9 consensus 
readings for right otoliths agreed with the second reading. 
Within-reader agreement was lower with the use of scales 
(80%), but precision varied with age. Agreement in repeated 
scale readings was actually high for ages 0 to 4 (92%, n= 52), 
but it decreased to only 59% for fish over age 4 (/?= 29). Preci- 
sion was lowest by far in opercular bones (46%), where there 
were no patterns in agreement by age. Because opercular 
bones showed the lowest precision and the poorest mark 
clarity, we did not include them in further evaluations. 
Agreement in age determinations between readers also 
varied greatly among calcified structures. Precision be- 
tween readers was highest by far (96%) for right sectioned 
otoliths (Table 3). Agreement was somewhat lower (86% to 
88%) for left sectioned otoliths and whole otoliths. Agree- 
ment was lowest by far for scales (58%), reflecting the over- 
all poor clarity of marks and the resulting subjectiveness 
in scale age readings compared with otolith age readings. 
Comparison of right and left otoliths 
Differences in right and left radial lengths were observed 
for both whole and sectioned otoliths. The right radial 
length was significantly shorter than the left in whole 
otoliths (paired 7=17.59, df=73, P<0.0001; Fig. 1A). How- 
ever, for sectioned otoliths, the right radial length was sig- 
nificantly longer than the left (paired t =-11.72, df=43, 
P<0.0001; Fig. IB) because the right otolith is thicker at 
the focus, where the transverse cross section was taken. 
Right and left whole otoliths generally gave the same 
age readings. Reader one had high age agreement between 
right and left whole otolith readings (96%), and the null 
hypothesis that the slope of the line equals one was not 
rejected (P=0.077, Fig. 3A). 
Although right and left whole otoliths generally indicat- 
ed the same age, they differed in mark clarity. When the 
posterior field (secondary counting path) was used to verify 
Table 2 
Regression statistics for relationships between structure 
size and summer flounder total length (TL). Structure 
abbreviations are defined in the “Methods” section of the 
text, n = sample size. All regressions were significant at 
P< 0.001. 
Structure Equation n 100 r 2 
Opercular 
bones 
OpRL = 
-2.280 + 0.0772 TL 
66 
98 
Scales 
ScRL = 
-0.348 + 0.0126 TL 
81 
93 
Sectioned otoliths 
Right 
SORL = 
-0.015 + 0.0027 TL 
66 
85 
Left 
SORL = 
0.015 + 0.0018 TL 
47 
72 
Whole otoliths 
Right 
WORL = 
0.642 + 0.0089 TL 
76 
91 
Left 
WORL = 
0.601 + 0.0111 TL 
77 
93 
Right 
WOTL = 
1.280 + 0.0164 TL 
76 
94 
Left 
WOTL = 
1.530 + 0.0156 TL 
77 
91 
Table 3 
Average percent agreement, within and between readers, 
for presumed annual mark counts on summer flounder cal- 
cified structures. 
Structure 
Within reader 
Between reader 
Opercular bones 
46 
— 
Scales 
80 
58 
Sectioned otoliths 
Right 
97 
96 
Left 
95 
88 
Whole otoliths 
Right 
89 
86 
Left 
97 
87 
or determine the number of presumed annual marks, the 
right otolith was generally much easier to read than the left 
because of the greater distance between the focus and the 
posterior margin on the right otolith (Fig. 1A). This greater 
distance made the marks further apart and more easily dis- 
tinguishable on the right than on the left otolith. The dif- 
ference in mark clarity was greatest for older fish and was 
also reflected in significantly higher confidence scores for 
the right whole otolith than for the left (Table 1). 
Right and left sectioned otoliths also generally gave the 
same age readings. Reader one had high age agreement 
between right and left sectioned otolith readings (94%), 
and the null hypothesis that the slope of the line equals 
one was not rejected (P=0.393, Fig. 3B). 
