FINAL CONCLUSIONS 443

examinee's measurements with a standard series based on
simple averaging and specifying the type.

4. Simple practical methods are suggested for recording
either by a numerical coefficient or by a psychographic
chart or c profile ' the essential temperamental character-
istics and c type ' of individual persons.

5. By means of the methods described a large number
of persons of both sexes and of different ages have been
assessed for the bipolar factor which appears to underlie
the difference between the so-called introverted and extra-
verted types. For each group examined the distribution
proves to be not bimodal but continuous, the mixed or
relatively well-balanced type being commonest of all.
For the entire group the distribution is slightly asym-
metrical, showing a small excess of extraverted cases.
Within the component groups the amount of extraversion
differs according to age, sex, and social or intellectual
status, being greater among younger children and less
in older or more intellectual males. In view of the age,
history, and physical peculiarities of many of the introverts,
it is suggested that introversion rather than extraversion is
usually the positive characteristic, and may not infrequently
be a late pathological development rather than a direct
manifestation of inborn temperament. In any case, the
'factors ' undoubtedly represent highly mixed and complex
groups of causes.

6. For statistical purposes it is * proposed to measure
approximation to type, not by the saturation coefficient as
it stands, but by its inverse hyperbolic tangent. With this
transformation the frequency curve approximates closely
to the normal. Thus the more typical introverts and
extraverts appear in the main to be merely extreme cases
taken from the opposite tail-ends of a normal or nearly
normal distribution.