6 
Fishery Bulletin 99(1 ) 
Table 4 
Results of MANOVA and AN OVA showing significant differences in the length-corrected internal otolith morphometric variables 
between year classes for eastern (EGB) and western Georges Bank (WGB) haddock. 
Region 
Age 
(yr) 
Year-class 
comparison 
(yr) 
MANOVA 
Significant 
variable 
ANOVA 
F 
df 
P 
F 
df 
P 
EGB 
1 
1994 vs. 1995 
1.23 
6,6 
0.4034 




2 
1993 vs. 1995 
7.82 
8, 18 
0.0002 
— 
— 
— 
— 
3 
1993 vs. 1994 
2.22 
9,9 
0.1255 
H2 
11.83 
1, 17 
0.0031 
4 
1992 vs. 1993 
1.90 
9, 15 
0.1309 
perimeter 
11.57 
1,23 
0.0025 
circularity 
9.14 
1,23 
0.0060 
WGB 
2 
1994 vs. 1995 
1.76 
8, 17 
0.1549 
— 
— 
— 
— 
3 
1993 vs. 1994 
3.55 
9,43 
0.0023 
HI 
15.59 
1,51 
0.0002 
Table 5 
Results of MANOVA and AN OVA showing significant differences in the length -corrected internal otolith morphometric variables 
between haddock from eastern and western Georges Bank. 
Year 
Age 
(yr) 
MANOVA 
Significant 
variable 
ANOVA 
F 
df 
P 
F 
df 
P 
1995 
1 
1.60 
6, 11 
0.2357 
rectangularity 
5.76 
1, 16 
0.0289 
1996 
3 
1.18 
9,8 
0.4109 
— 
— 
— 
— 
1997 
2 
2.10 
8,26 
0.0726 
circularity 
10.37 
1,33 
0.0029 
3 
4.30 
9, 44 
0.0005 
Al 
22.76 
1, 52 
0.0001 
A2 
7.84 
1, 52 
0.0072 
H2 
7.41 
1, 52 
0.0088 
4 
3.22 
9,43 
0.0045 
A2 
3.97 
1,51 
0.0518 
5 
5.53 
9, 10 
0.0067 
Al 
9.70 
1, 18 
0.0060 
H2 
9.90 
1, 18 
0.0056 
analyses were conducted for samples of individual ages 
and year classes in order to minimize the effects of these 
confounding variables. 
Significant differences in the internal otolith morpho- 
metries between haddock sampled on eastern and west- 
ern Georges Bank were found in three out of six compar- 
isons (Table 5). Eastern Georges Bank haddock tended 
to have smaller otoliths (i.e. length, area, perimeter, cir- 
cularity and rectangularity) than western Georges Bank 
haddock; this finding was derived mainly from the first 
growth increment (Al) which also tended to be smaller 
(Fig. 3). For the three significant comparisons, scatter 
plots of the most significant individual internal otolith 
morphometries for samples from each age group typically 
showed separation patterns between eastern and western 
Georges Bank haddock, albeit with some overlap (Figs. 
4-6). 
Principal component analysis provided further support 
that haddock from eastern and western Georges Bank 
separated into two groups (Fig. 7). Eastern and western 
Georges Bank haddock samples, 1 and 3 (1994 year class) 
years of age were mainly separated on the first principal 
component (PC I) (AN OVA, P<0.06), whereas samples that 
were 2, 3 ( 1993 year class), and 5 years of age were mainly 
separated on the second and third principal components 
(PC II and PC III) (ANOVA, P<0.05). Principal compo- 
nents I, II, and III accounted on average for 38.7 ±5.3% 
SD, 22.6 ±5.9%, and 15.6 ±0.7% of the total variation in 
the data. Differences in length, area, and perimeter were 
mainly responsible for the observed separations along the 
first principal component, whereas Al, HI and A2 were 
the main variables responsible for separation along the 
second and third principal components. 
Discriminant analysis also indicated that haddock com- 
prise two groups on Georges Bank (Fig. 8). Significant dif- 
ferences in the discriminant (CV I) scores between eastern 
and western Georges Bank haddock were found for all age 
groups (ANOVA, P<0.05), except age group 4. Total clas- 
sification success varied from 61% to 83% for the different 
age and year-class combinations (Table 6). 
