10 
Fishery Bulletin 99(1) 
2.0 -| 
1.5- 
1 . 0 - 
o 
o 
o 
o 
o 
o 
■ 
0.5 -I — 
-4.2 
— i 1 — 
-3.7 -3.2 
Log, H 2 
-2.7 
o 
- 2 . 0 - 1 — 
-4.2 
-3.7 
-3.2 
Log, H 2 
-2.7 
- 2.0 1 
Q) 
O) 
o 
- J -3.0- o 
OHO ■ 
-3.6 -I — 
-4.2 
o 
-3.7 
-3.2 
-2.7 
2.2 
-4.2 
-3.7 
-3.2 
-2.7 
Log, H2 
Log, H 2 
Figure 6 
Scatter plots of internal otolith morphometric variables for 5-year-old haddock ( 1992 year class) show- 
ing grouping patterns of eastern Georges Bank (squares) and western Georges Bank (circles). 
Classification success similar to our present study was 
found between eastern and western Georges Bank had- 
dock in a comparable study where whole-otolith shape 
analysis was used (Begg and Brown, 2000). However, the 
use of internal otolith morphometries entailed a few more 
caveats than that of whole-otolith shape analysis; there- 
fore internal otolith morphometries was questioned as a 
preferred tool for stock identification. Although our ap- 
proach was semi-automated, with an aim to increasing 
objectivity and decreasing processing time, the use of in- 
ternal otolith morphometries still required some inter- 
pretation by the person taking the measurements. This 
was particularly true for growth zones, where interpreta- 
tion partly compromised our goals of increasing objectiv- 
ity and speed. Difficulties in interpreting measurements 
when growth zones are poorly defined can also be a source 
of uncertainty (Hopkins 4 ). These sources of uncertainty 
should be minimized, provided the person taking the mea- 
surements is also an experienced reader of otolith growth 
increments. 
Internal otolith morphometries are useful for stock iden- 
tification, but significant overlap among variables may 
preclude their use for stock discrimination, that is to 
say for classifying unknown fish according to origin of 
spawning component. Although mean values may differ 
between stocks, measurements from individual fish may 
not allow them to be classified to a particular stock be- 
cause of individual within-stock growth differences (Paw- 
son and Jennings, 1996). For example, slow-growing fish 
from a fast-growing stock may be incorrectly classified 
with a slow-growing stock (Campana and Casselman, 
1993). However, provided growth differences exist and 
analyses are conducted to account for samples of mixed 
ages and year classes to minimize the effects of these con- 
founding variables, internal otolith morphometries can be 
a useful phenotype-based stock identification tool. 
Results from this analysis on internal otolith morpho- 
metric differences have added to the evidence indicating 
separation between the eastern and western Georges Bank 
haddock spawning components. Although these differences 
do not provide a genetic basis for separation between the 
two spawning components, they do reflect the phenotypic 
characteristics of each spawning component, indicative of 
stock separation during life history. Studies such as this 
one, are needed to provide evidence of stock structure if his- 
torically established management units are to be changed 
in response to changing exploitation and resource patterns. 
Evidence of stock separation within the Georges Bank had- 
dock resource may be useful in forming stock definitions 
that can be used by both the U.S. and Canada in defining 
