Biology Reference
In-Depth Information
FIGURE 8.5
Allometry of Tamias alpinus (A)
based on 15 landmarks, result exaggerated fivefold;
(B) based on 15 landmarks plus 85 semilandmarks,
result exaggerated fivefold.
shape. Depicting that effect (
Figure 8.5B
), magnified fivefold, shows that the impact of size
is particularly pronounced on the curvature of the angular process, the depth of the molar
alveolus and the width of the condyloid process.
COMPARING TWO MEANS
A classic problem in statistics is to determine whether or not the mean value of a mea-
surement (or a set of one-dimensional measurements) in one group is different from that
of another group. These groups are distinguished because they differ in some discrete, cat-
egorical variable, such as a treatment that we applied to them, or some property such as
sex. In such cases, we cannot subdivide the classifying variable; there is no treatment
between the ones that we apply, nor is there a value for sex between male and female. In
some cases, the classifying variable is not necessarily discrete and categorical, but it is trea-
ted as if it were. For example, populations from different localities might be treated as cat-
egorically distinct even though there are geographical coordinates between any two sites.
Similarly, populations sampled at two different times might be treated as categorically dis-
tinct even though there is a date in between any two others.
The comparison between two groups that differ categorically can be expressed as a
mathematical model of the form:
Y
i
5
A
ð
i
Þ
1
B
1
ε
i
(8.15)
