What I wish to suggest may be put as follows. Everything that occurs elsewhere, owing to the existence of an electron, can be explored experimentally, at least in theory, unless it occurs in certain concealed ways. But what occurs within the electron (if anything occurs there) it is absolutely impossible to know: there is no conceivable apparatus by which we could obtain even a glimpse of it. An electron is known by its “effects.” But the word “effects” belongs to a view of causation which will not fit modern physics, and in particular will not fit relativity. All that we have a right to say is that certain groups of occurrences happen together, that is to say, in neighboring parts of space-time. A given observer will regard one member of the group as earlier than the other, but another observer may judge the time order differently. And even when the time order is the same for all observers, all that we really have is a connection between the two events, which works equally backwards and forwards. It is not true that the past determines the future in some sense other than that in which the future determines the past: the apparent difference is only due to our ignorance, because we know less about the future than about the past. This is a mere accident: there might be beings who would remember the future and have to infer the past. The feelings of such beings in these matters would be the exact opposite of our own, but no more fallacious. The moral of this is that, if an electron is only known by its “effects,” there is no reason to suppose that anything exists except the “effects.” In so far as these “effects” consist of light waves and other electromagnetic disturbances, we may say that what is called “empty space” consists of regions where these disturbances are propagated freely. Every such disturbance, we find, has a center, and when we get very near the center (though still at a finite distance from it) we find that the law of propagation of the disturbance ceases to be valid. This region within which the law does not hold is called “matter”; it will be an electron or proton according to circumstances. The region so defined is found to move relatively to other such regions, and its movements follow the known laws of dynamics. So far, this theory provides for electromagnetic phenomena and the motions of matter; and it does so without assuming that “matter” is anything but systems of electromagnetic phenomena. In order to carry out the theory fully, it would no doubt be necessary to introduce many complications. But it seems fairly clear that all the facts and laws of physics can be interpreted without assuming that “matter” is anything more than groups of events, each event being of the sort which we should naturally regard as “caused” by the matter in question. This does not involve any change in the symbols or formulæ of physics: it is merely a question of interpretation of the symbols.