place: when we move it to another region, there is no knowing how it will alter. We do assume, however, that, if it alters, it alters bit by bit, gradually, continuously, and not by sudden jumps. Perhaps this assumption is unjustified. It belongs to the general outlook of relativity, which is that of continuity. No doubt it is owing to this outlook that relativity is unable to account for the discontinuities in physics, such as quanta, electrons and protons. Perhaps relativity will conquer these domains when it learns to dispense with the assumption of continuity. Finally we come to geography, in which I include history. The separation of history from geography rests upon the separation of time from space; when we amalgamate the two in space-time, we need one word to describe the combination of geography and history. For the sake of simplicity, I shall use the one word geography in this extended sense. Geography, in this sense, includes everything that, as a matter of crude fact, distinguishes one part of space-time from another. One part is occupied by the sun, one by the earth; the intermediate regions contain light waves, but no matter (apart from a very little here and there). There is a certain degree of theoretical connection between different geographical facts; to establish this is the purpose of physical laws. It is thought that a sufficient knowledge of the geographical facts of the solar system throughout any finite time, however short, would enable an ideally competent physicist to predict the future of the solar system so long as it remained remote from other stars. We are already in a position to calculate the large facts about the solar system backwards and forwards for vast periods of time. But in all such calculations we need a basis of crude fact. The facts are interconnected, but facts can only be inferred from other facts, not from general laws alone. Thus the facts of geography have a certain independent status in physics. No amount of physical laws will enable us to infer a physical fact unless we know other facts as data for our inference. And here when I speak of “facts” I am thinking of particular facts of geography, in the extended sense in which I am using the term.