62 
Fishery Bulletin 11 7(1-2) 
1000 
-£ 750 
500 
13 
■D 
> 
'~o 
_c 
-C 
o 
-t—' 
CO 
O 250 
I 
I 
I 
1 
_ 
I 
■ I ■.■ ■ 
Species 
| Alewife 
Blueback 
herring 
2007 2008 2009 2010 2011 2012 2013 2014 2015 
Figure 2 
Number of alewife (Alosa pseudoharengus ) and blueback herring (Alosa 
aestivalis ) captured per day in tributaries of the Potomac River with hoop 
nets from 2007 through 2015 and by electrofishing in 2007 and 2008 by 
the Potomac Environmental Research and Education Center of George 
Mason University. 
otolith readings but on only 792 of 828 
scale readings. Otolith readings were 
compared with scale readings to verify 
ages of scales by using age bias and 
precision analyses in RStudio 10 , vers. 
1.0.153 (RStudio, Inc., Boston, MA) and 
the R package FSA, vers. 0.7.3 (Ogle, 
2015). Plots of age bias were created 
by plotting ages agreed upon by otolith 
readers versus ages from scale read¬ 
ings to visually examine data for sys¬ 
tematic bias in aging scales (Campana 
et al., 1995). Ages within the age-bias 
plot were analyzed by using a f-test 
to determine if ages from scales agree 
with ages from otoliths (Campana et 
al., 1995). To statistically test for sym¬ 
metry, McNemar’s test, Evans-Hoenig 
test, and Bowker’s test were used to 
determine the differences around the 
main diagonal of the age-bias plot (Ev¬ 
ans and Hoenig, 1998). Average coef¬ 
ficient of variation (ACV) was used to 
find the variability in ages determined 
by using ages from otoliths versus ages 
from scales (Campana, 2001). 
Statistical analyses 
Relationships between SL and FL, SL and TL, and 
FL and TL were estimated on the basis of linear re¬ 
gression analyses. Length-at-age data were used to 
determine growth curves in 9 different growth mod¬ 
els: von Bertalanffy (von Bertalanffy, 1938), Gompertz 
(Gompertz, 1825), Laird-Gompertz (Laird, 1964; Zwe- 
ifel and Lasker, 1976), Richards (Richards, 1959), lin¬ 
ear (Haddon, 2011), logistic (Ricker, 1975), Ratkowsky 
(Ratkowsky, 1986), Francis (Francis, 1988), and Cer- 
rato (Cerrato, 1990). All analyses were conducted in 
Microsoft Excel (vers. 16.7; Microsoft Corp., Redmond, 
WA) by using Solver, an add-in tool available in Excel 
(Haddon, 2011). Growth models were run using SL as 
the measure for length to increase sample size in this 
study because some of the captured river herring had 
damaged caudal fins, making TL and FL unreliable or 
unattainable measures. However, past documentation 
of growth parameters for river herring were done with 
FL or TL. To directly compare this study’s results with 
those of past studies, the von Bertalanffy growth func¬ 
tion was run with FL and TL for each species; mean 
asymptotic lengths (L„) are reported in parentheses in 
the “Discussion” section when applicable for compari¬ 
son with results of other studies. 
The best-fit model was determined by using the AIC 
(Akaike, 1974; Hilborn and Mangel, 1997; Burnham 
and Anderson, 2002). The use of the AIC allows non- 
10 Mention of trade names or commercial companies is for 
identification purposes only and does not imply endorse¬ 
ment by the National Marine Fisheries Service, NMFS. 
nested models to be compared and over-parameteriza¬ 
tion of a model to be taken into context (Hilborn and 
Mangel, 1997). The AIC was then transformed to AIC 
weights to determine which model was furthest from 
the true AIC value (Burnham and Anderson, 2002). 
An analysis of residual sum of squares was con¬ 
ducted on the best-fit model to determine if there are 
sex-specific differences in growth parameters. Finally, 
a likelihood ratio was used to test which growth pa¬ 
rameters are responsible for any differences between 
sexes. The likelihood ratio is a chi-square distribution 
with degrees of freedom (df) that compares the sums of 
squares of the models for each combination of growth 
parameters by estimating each parameter individually 
through the growth model while holding some param¬ 
eters constant and calculating the sums of squares for 
each combination of parameters. 
Results 
Alewife (n = 1707) were captured in tributaries of the Po¬ 
tomac River for 9 consecutive years from 2007 through 
2015, and 598 of these fish were dissected for aging. 
Blueback herring (/? = 1159) were captured in 2007 and 
2008 and from 2011 through 2015, and 304 of them 
were dissected for aging. In 2015, the catch of both spe¬ 
cies was an order of magnitude higher than the catch 
of any other year during this study (Fig. 2). Methods 
and locations of sampling were consistent from 2009 
through 2015, except that Cameron Run was added 
from 2013 through 2015, resulting in the capture of 
1 alewife, 6 alewife, and 16 alewife in each of those 
3 years, respectively. Therefore, the increased catch of 
alewife in 2015 was not due to increased effort. 
