Elzey et al: Comparison of 4 aging structures for Alosa sapidissima 
51 
Table 1 
The precision of age estimates obtained from otoliths, 
scales, vertebrae, and opercula of American shad (Alosa 
sapidissima) was tested between readers, within re- 
sults for a single reader for a subsample of 100 fish, 
and between structures for each reader. All samples 
were collected in May and June during 2008-10 from 
the Merrimack River in Lawrence, Massachusetts. The 
percent agreement and coefficient of variation (CV) are 
presented for each comparison. 
Agreement 
CV 
(%) 
(%) 
Otolith 
76.2 
2.99 
Between 
Scale 
66.9 
4.28 
readers 
Vertebra 
63.4 
4.59 
Operculum 
49.8 
7.07 
Otolith 
74.0 
2.93 
Within 
Scale 
65.0 
5.10 
reader 2 
Vertebra 
56.0 
5.96 
Operculum 
49.0 
7.32 
Otolith vs. scale 
50.6 
6.87 
Reader 1 
Otolith vs. vertebra 
55.6 
5.95 
Otolith vs. operculum 
34.0 
9.80 
Otolith vs. scale 
53.2 
6.63 
Reader 2 
Otolith vs. vertebra 
53.2 
6.39 
Otolith vs. operculum 
34.2 
9.52 
Table 2 
American shad ( Alosa sapidissima ) collected in May and June during 
2008-10 from the Merrimack River in Lawrence, Massachusetts, were aged 
by 2 readers using otoliths, scales, vertebrae, and opercula. The percent 
agreement between readers, within results for a single reader for a sub- 
sample of 100 fish, and between structures for each reader is shown as 
exact agreement (±0), within 1 year (±1), 2 years (±2), and 3 years (±3). 
±0 
±1 
±2 
±3 
Otolith 
76.2 
97.6 
99.8 
100.0 
Between 
Scale 
66.9 
95.9 
99.4 
100.0 
readers 
Vertebra 
64.3 
95.9 
98.9 
100.0 
Operculum 
49.8 
93.5 
98.9 
99.8 
Otolith 
74.0 
100.0 
100.0 
100.0 
Within 
Scale 
65.0 
96.0 
99.0 
100.0 
reader 2 
Vertebra 
56.0 
96.0 
98.0 
100.0 
Operculum 
49.0 
93.0 
99.0 
100.0 
Otolith vs. scale 
50.6 
90.9 
98.3 
99.6 
Reader 1 
Otolith vs. vertebra 
55.6 
93.3 
99.4 
100.0 
Otolith vs. operculum 
34.0 
86.8 
97.8 
99.8 
Otolith vs. scale 
53.2 
89.4 
98.3 
99.8 
Reader 2 
Otolith vs. vertebra 
53.2 
92.0 
99.1 
99.8 
Otolith vs. operculum 
34.2 
87.0 
98.7 
99.4 
tebral ages and otolith ages was within 1 year (93.9% 
and 92%) and 2 years (99.4%, 99.1%), results that are 
better than the agreement between scale ages and oto- 
lith ages (1 year: 90.9% and 89.4%; 2 years: 89.3% and 
89.3%). More than 99% of all age disagreements were 
within 3 years (Table 2). 
Wilcoxon rank sum tests between readings from oto- 
liths and those from other structures showed that both 
readers significantly overestimated the ages of age-4 
and age-5 fish when they used scales and vertebrae. 
Both readers also underestimated the ages of fish age 
6 and older with the use of opercula, as well as ages 
of fish age 7 and older with the use of scales (Table 3; 
Fig. 3). Although significance was not tested for age-3 
or age-11 fish, because of a sample size of 1 for each 
age, the limited data for these 2 fish indicate that the 
trends outlined previously continue into these ages. 
Discussion 
On the basis of the presence of strong annular marks 
on otoliths as well as the lowest CV and the highest 
percent agreement between readers and within readers 
for readings from otoliths in this study, the otolith was 
deemed the best structure for determining the age of 
American shad. Furthermore, using fish marked with 
oxytetracycline as did Hendricks et al. (1991), Duffy et 
al. (2012) were able to validate the accuracy of aging 
American shad with the use of otoliths. Additionally, 
otoliths are moderately easy to re- 
move and require less preparation for 
aging than any of the other structures 
examined. 
The vertebra was the second-most 
preferred structure for aging Ameri- 
can shad. When compared with oto- 
lith ages, vertebral ages were less 
biased than scale or operculum ages. 
In a study similar to ours, Yilmaz and 
Polat (2002) compared the precision of 
age estimates from otoliths, vertebrae, 
opercula, subopercula, and scales from 
pontic shad and found that the most 
precise aging was performed with 
vertebrae. However, Visnjic-Jeftic et 
al. (2009) found that the precision of 
aging the pontic shad was dependent 
more on the experience of the reader 
than on the structure being aged. 
Readers in our study had minimal 
prior experience in aging vertebrae. 
Therefore, with practice for readers, 
vertebral ages for American shad 
could approach the level of precision 
seen with otolith ages. However, the 
work to remove and process vertebrae 
was more labor-intensive than the ef- 
fort required for otoliths, and there- 
