LIKE a chemical compound, scientific knowledge is purified by recrystallization. When first published, each new grain of fact or theory shines from a mud of irrelevant or erroneous details. In subsequent discussion the grains are redissolved, and filtered. Finally, in books and treatises, the solution is allowed to precipitate into a single crystal where each atom seems inevitably to be in its proper place. The writing of books is thus as much part of the scientific process as watching oscilloscopes or solving differential equations.
Individual scientific facts are the leaves and twigs of a great tree. They must be connected downwards, into smaller and larger branches, into the limbs, and then into the trunk itself. To visualize the tree, we must see the connexions. At each major fork, we need to comprehend in sufficient detail all that is borne above it. But a unified picture can only be made by one person comprehending the whole scene. Until a few years ago, the whole of the science of the solid state could be condensed into a single volume. Now it has grown into a library shelf, beyond the capacity of one man to write alone. The recent tendency has been the other way, to the collection of numerous, short, review articles, in which the whole picture is as clear as in a jumbled jig-saw puzzle of which each piece is painted by a different artist. There is need for treatises covering, in reasonable detail, up to the level of active research, the major branches into which the subject has divided.