52 
Fishery Bulletin 1 13(1) 
Table 3 
American shad (Alosa sapidissima), collected in May 
and June during 2008-10 from the Merrimack River in 
Lawrence, Massachusetts, were aged by 2 readers. Each 
reader used otoliths, scales, vertebrae, and opercula to 
determine the age of each fish. The average age assigned 
to each structure for each otolith age is shown. An aster- 
isk (*) denotes significant differences (P< 0.05) from the 
otolith ages determined with Wilcoxon rank sum tests. 
Significance was not tested for ages 3 and 11 because of 
insufficient sample size ( n ). 
Otolith age 
Vertebra 
Scale 
Operculum 
n 
Reader 1 
3 
3.00 
4.00 
3.00 
i 
4 
*4.48 
*4.95 
4.10 
21 
5 
*5.29 
*5.53 
4.98 
135 
6 
5.99 
6.09 
*5.73 
172 
7 
6.88 
*6.63 
*6.30 
80 
8 
7.84 
*7.00 
*7.00 
32 
9 
*8.40 
*7.07 
*7.20 
15 
10 
9.50 
*7.83 
*7.83 
6 
Reader 2 
3 
3.00 
5.00 
3.00 
1 
4 
*4.68 
*4.74 
4.26 
19 
5 
*5.25 
*5.42 
4.93 
130 
6 
6.03 
6.09 
*5.65 
159 
7 
7.00 
*6.72 
*6.22 
89 
8 
7.71 
*7.29 
*6.85 
41 
9 
8.86 
*7.36 
*7.71 
14 
10 
9.63 
*8.13 
*7.63 
8 
11 
10.00 
8.00 
7.00 
1 
fore the vertebra is a less ideal structure for produc- 
tion aging. 
In this study, age estimates obtained from scales of 
American shad were not as precise as those produced 
from otoliths and were more biased toward underag- 
ing older (>7 years) fish than were age estimates ob- 
tained from vertebrae. For these reasons, the scale is 
not viewed as a preferred aging structure. The use 
of scales for age determination had the advantage of 
being the only nonlethal method tested in this study. 
However, we found that many scales were regener- 
ated and, therefore, not suitable for age determina- 
tion. Although scales can provide data regarding re- 
peat spawning behavior (Cating, 1953), information 
that is used in the management of the American shad 
(ASMFC 2 ), the spawning marks left by reabsorption of 
the scale margin during a freshwater spawning run 
(Cating, 1953) can hinder the use of scales for accurate 
age determination. If enough of the scale is reabsorbed, 
annuli laid down in previous years can be very difficult 
to interpret. 
The results of our study agree with the findings of 
McBride et al. (2005) that indicate that scales produce 
biased ages, where readers tend to over-age young (<5 
o 
2 4 6 8 10 12 
"O 
a> 
2 4 6 8 10 12 
Otolith age (years) 
Figure 3 
Ages from otoliths of American shad ( Alosa sapidissi- 
ma) collected in May and June during 2008-2010 from 
the Merrimack River in Lawrence, Massachusetts, com- 
pared with the average ages assigned through the use 
of vertebrae, scales, and opercula by (A) reader 1 and 
(B) reader 2. The diagonal line represents agreement 
with otoliths. Filled symbols represent ages that are 
significantly different (P<0.05) from ages from otoliths; 
empty symbols are not. Significance was tested with 
Wilcoxon rank sum tests for ages 4-10. Significance 
was not tested for ages 3 or 11 because of a sample size 
of 1 for each age. 
years) fish and under-age older fish. Furthermore, Up- 
ton et al. (2012) showed a 50% error in scale ages of 
American shad of known age. This trend of otoliths 
providing better precision and a larger range of ages 
than those provided by scales has been shown in sever- 
al other species as well (Barnes and Power, 1984; Welch 
et ah, 1993; Secor et al., 1995; Sipe and Chittenden, 
2001; Zymonas and McMahon, 2009). 
Opercula resulted in the least reliable readings 
in this study of aging structures for American shad. 
Opercula were difficult to read, required more process- 
ing than scales or otoliths, and provided estimates of 
the lowest precision compared with results from the 
other structures examined. Contrary to the findings of 
