be in
anthropological research. It justifies those who assert that existing
savagery or existing survival will supply evidence of man at the very
earliest stages of existence. It is the root idea of Dr. Tylor's
method of research, and it is an essential feature in the science of
folklore.

Evidence of this nature, however, needs to be exhaustively collected,
and to be subjected to the most careful examination, as otherwise it
may be used for the merest _a priori_ argument of the most mischievous
and inconclusive description. It involves consideration of whole human
groups rather than of particular sections of each human group, of the
whole corpus of social, religious, and economical elements residing in
each human group rather than of the separated items. Each human group,
having its specialised and dormant elements, must be treated as an
organism and not as a bundle of separable items, each one of which the
student may use or let alone as he desires. That which is
anthropological evidence is the indivisible organism, and whenever,
for convenience of treatment and considerations of space, particular
elements only are used in evidence, they must be qualified, and the
use to which they are provisionally put for scientific purposes must
be checked, by the associated elements with which the particular
elements are connected.

The human groups thus called upon to surrender their contributions to
the history of man are of various formations, and consist of various
kinds of social units. There is no one term which can properly be
applied to all, and it will have been noted that I have carefully
avoided giving the human groups hitherto dealt with any particular
name, and only under protest have I admitted the terms used by the
authorities I have quoted. I think the term "tribe" is not applicable
to savage society, for it is used to denote peoples in all degrees of
social evolution, and merely stands for the group which is known by a
given name, or roams over a given district. But the use of this term
is not so productive of harm as the use of the term "family," because
of the universal application of this term to the smallest social unit
of the civilised world, and because of the fundamental difference of
structure of the units which roughly answer to the definition of
family in various parts of the world. It is no use in scientific
matters to use terms of inexact reference. As much as almost anything
else it has led to fals