Schueller et al.: Using a projector and microscope to read scales from Brevoortia tyrannus and B. patronus 25 
Table 1 
Precision and agreement statistics for ages estimated by examining scales on an Eberbach pro- 
jector and a stereo microscope for Atlantic menhaden (Brevoortia tyrannus) and Gulf menhaden 
(B. patronus) sampled from landings along the Atlantic and Gulf coasts of the United States. Scales 
were from Atlantic menhaden collected in 2013 and 2017, Gulf menhaden collected in 2005 and 
2017, and reference collections for each species. The statistical tests used to determine precision and 
agreement of age estimates included average percent error (APE), average coefficient of variation 
(ACV), and percent agreement. Total number of fish for which scales were used to determine ages 
(N), sample size used for analysis after exclusion of scales (7), and percentage of scale samples that 
were excluded are presented for each data set. 
Samples excluded (%) 
Data set Projector 
Atlantic menhaden 
All 
2013 
2017 
Reference 
Gulf menhaden 
All 
2005 
2017 
Reference 
Table 2 
Precision and agreement statistics for ages estimated by 
examining scale subsamples on an Eberbach projector and 
a stereo microscope for Atlantic menhaden (Brevoortia 
tyrannus) and Gulf menhaden (B. patronus) collected from 
landings in 2013 and 2017 along the Atlantic coast of the 
United States and in 2005 and 2017 along the Gulf coast, 
respectively. The statistical tests used to determine pre- 
cision and agreement of age estimates included average 
percent error (APE), average coefficient of variation (ACV), 
and percent agreement. Reading 1 is the first reading of 
the sample set on both the projector and microscope, and 
reading 2 is the second reading on both instruments. Val- 
ues in the “Projector” and “Microscope” rows are from com- 
parisons of the 2 readings of the scale subsamples on the 
same device. n=sample size used for analysis. 
Percent 
Data set ACV agreement 
Atlantic menhaden 
Reading 1 
Reading 2 
Projector 
Microscope 
Gulf menhaden 
Reading 1 
Reading 2 
Projector 
Microscope 
Percent 
Microscope agreement 
8% 88.9% 
8% 86.3% 
9% 91.2% 
4% 95.8% 
10% 89.7% 
9% 86.2% 
11% 94.0% 
0% 100.0% 
were 4.4% for samples taken in 2005 and 2.5% for samples 
collected in 2017. The ACV for age estimates for all of the 
samples of Atlantic menhaden was 3.8%, and the ACV for 
age estimates for all of the samples of Gulf menhaden was 
5.0% (Table 1). The ACVs for age estimates for samples in 
the reference collections were as low as 1.2% for Atlantic 
menhaden and 0.0% for Gulf menhaden. The ACVs for age 
data on Atlantic menhaden were 5.0% for samples taken 
in 2013 and 2.6% for the samples collected in 2017. The 
ACVs for age data on Gulf menhaden were 6.2% for sam- 
ples taken in 2005 and 3.6% for samples collected in 2017. 
Age estimates for the more recent samples, those taken in 
2017, for each species had better APE and ACV than the 
age estimates for the older samples from 2013 and 2005 
(Table 1). 
Average percent errors for the subsamples were below 
5% for all comparisons of age data except for 2 compar- 
isons of estimates for Gulf menhaden (Table 2). Specif- 
ically, when the first readings of a scale sample were 
compared between devices, the APE was 9.7%, and when 
the first and second readings with the Eberbach pro- 
jector were compared, the APE was 8.7%. The ACVs of 
age estimates for the subsamples followed the same pat- 
terns as the APEs, given that the ACV is a fixed propor- 
tion larger than APE when comparisons are between 2 
measurements. 
Percent agreement was 88.9% for readings of scales 
from Atlantic menhaden overall and 89.7% for readings of 
scales from Gulf menhaden overall (Table 1). Percent 
agreement varied by year of sampling with the readings of 
