changing, and must adapt our policy to the problems of the future.
In doing this, we shall find that different fields of research
require very different policies as regards co-operation and
subordination. It will be profitable to point out those special
differences, because we shall thereby gain a more luminous insight
into the problems which now confront the scientific investigator,
and better appreciate their variety, and the necessity of
different methods of dealing with them.

At one extreme, we have the field of normative science, work in
which is of necessity that of the individual mind alone. This
embraces pure mathematics and the methods of science in their
widest range. The common interests of science require that these
methods shall be worked out and formulated for the guidance of
investigators generally, and this work is necessarily that of the
individual brain.

At the other extreme, we have the great and growing body of
sciences of observation. Through the whole nineteenth century, to
say nothing of previous centuries, organizations, and even
individuals, have been engaged in recording the innumerable phases
of the course of nature, hoping to accumulate material that
posterity shall be able to utilize for its benefit. We have
observations astronomical, meteorological, magnetic, and social,
accumulating in constantly increasing volume, the mass of which is
so unmanageable with our present organizations that the question
might well arise whether almost the whole of it will not have to
be consigned to oblivion. Such a conclusion should not be
entertained until we have made a vigorous effort to find what pure
metal of value can be extracted from the mass of ore. To do this