Begg et al.: Use of otolith morphometries for identification of stocks of Melanogrammus oeglefinus 
5 
Table 2 
Internal otolith morphometric variables significantly correlated with fish length, and the corresponding regression coefficients ( b ) 
required to standardize the variables for fish length. 
Otolith morphometric 
variable 
Length x region 
Length 
b 
F 
df 
P 
F 
df 
P 
Al 
0.85 
15,200 
0.6195 
1.70 
1,215 
0.1936 

HI 
1.27 
13,183 
0.2333 
1.22 
1,196 
0.2713 
— 
A2 
0.92 
12,181 
0.5274 
2.36 
1,193 
0.1261 
— 
m 
1.19 
12,155 
0.2918 
11.62 
1,167 
0.0008 
0.01810 
Length 
1.16 
15,200 
0.3082 
103.65 
1, 215 
0.0001 
0.00922 
Width 
1.93 
15,200 
0.0222 
— 
— 
— 
— 
Area 
1.53 
15,200 
0.0971 
91.91 
1,215 
0.0001 
0.01514 
Perimeter 
1.48 
15,200 
0.1165 
88.05 
1, 215 
0.0001 
0.00837 
Circularity 
1.06 
15,200 
0.4007 
2.92 
1, 215 
0.0889 
— 
Rectangularity 
1.63 
15,200 
0.0695 
8.50 
1, 215 
0.0039 
-0.00092 
Table 3 
Results of MANOVA and ANOVA showing significant differences in the length-corrected internal otolith moi 
between age groups for eastern (EGB) and western Georges Bank (WGB) haddock. 
•phometric 
variables 
Region 
Year 
Age group 
comparison 
(years) 
MANOVA 
Significant 
variable 
ANOVA 
F 
df 
P 
F 
df 
P 
EGB 
1995 
1 vs. 2 
15.51 
6,10 
0.0002 
length 
19.16 
1,15 
0.0005 
area 
28.70 
1,15 
0.0001 
perimeter 
12.39 
1,15 
0.0031 
1996 
1,3 vs. 4 
3.98 
12,34 
0.0007 
length 
32.79 
2,21 
0.0001 
area 
127.78 
2,21 
0.0001 
perimeter 
23.13 
2,21 
0.0001 
circularity 
6.93 
2,21 
0.0049 
1997 
2,3,4 vs. 5 
4.05 
24,135 
0.0001 
Al 
6.87 
3,50 
0.0006 
H\ 
4.75 
3,50 
0.0054 
length 
4.90 
3,50 
0.0046 
area 
7.71 
3,50 
0.0002 
WGB 
1996 
2 vs. 3 
4.71 
8,9 
0.0163 
length 
11.99 
1,16 
0.0032 
perimeter 
9.19 
1,16 
0.0079 
1997 
2,3,4 vs. 5 
5.46 
24,297 
0.0001 
area 
10.81 
3,104 
0.0001 
circularity 
7.46 
3,104 
0.0001 
rectangularity 
4.96 
3,104 
0.0029 
variable. The same ANCOVAs also indicated that H 2 and 
otolith length, area, perimeter, and rectangularity were 
correlated with fish length (P<0.004), resulting in those 
variables being corrected for fish length by using their 
respective within-group regression coefficient (Table 2). 
Internal otolith morphometries were not significantly 
different between the sexes (MANOVA, P>0.05), result- 
ing in the sexes being pooled within each region, year 
and age strata to increase the statistical power used in 
subsequent analyses. In contrast, significant differences 
were found between the different age groups of haddock 
(Table 3). Not unexpectedly, haddock that were 1 and 2 
years of age tended to have significantly smaller otoliths 
than haddock that were 3, 4, and 5 years of age (HSD, 
P<0.05). Likewise, there were significant differences in- 
dicative of annual growth differences found in the in- 
ternal otolith morphometries of haddock sampled from 
different year classes (Table 4). Hence, the remaining 
