ass of the people; in morals with avowed precepts,
not with the effective ideals; in institutions with official rules, not
with the real practice. On all these subjects the knowledge of
conventional forms must some day be supplemented by a parallel study of
the real habits.

It is much more difficult to embrace in a single formula a habit which
is composed of real actions, as is the case with economic phenomena,
private life, politics; for we have to find in the different actions
those common characteristics which constitute the habit; or, if this
work has already been done in the documents, and condensed into a
formula (the most common case), we must criticise this formula in order
to make sure that it really represents a homogeneous habit.

The same difficulty occurs in constructing the formula for a group; we
have to describe the characteristic common to all the members of the
group and to find a collective name which shall exactly designate it. In
documents there is no lack of names of groups; but, as they have their
origin in usage, many of them correspond but ill to the real groups; we
have to criticise these names to fix their precise meaning, sometimes to
correct their application.

This first operation should yield formulae expressive of the conventional
and real characteristics of all the habits of the different groups.

In order to fix the precise _extent_ of a habit we shall seek the most
distant points where it appears (this will give the area of
distribution), and the region where it is most common (the centre).
Sometimes the operation takes the form of a map (for example the map of
the _tumuli_ and the _dolmens_ of France). It will also be necessary to
indicate the groups of men who practised each habit, and the sub-groups
in which it was most pronounced.

The formula should also indicate the _duration_ of the habit. We shall
look for the extreme cases, the first and the last appearance of the
form, the doctrine, the usage, the institution, the group. But it will
not be enough to note the two isolated cases, the earliest and the most
recent; we must ascertain the period in which it was really active.

The formula of an evolution ought to indicate the successive variations
in the habit, giving in each case precise limits of extent and duration.
Then, by comparing all the variations, it will be possible to determine
the general course of the evolution. The general formula will indicate
when and where