That the principle of selection is, in the end, the ability to modify behavior through the anticipation of possible consequences, appears from the fact that the category of stimulus and response is otherwise found to be unworkable. It is true that in the simpler forms of behavior stimulus and response may be correlated without practical difficulty. But when we deal with what has been called "delayed overt response," the matter becomes more complicated and the theoretical difficulty becomes more prominent. The behaviorist would not seriously undertake to record everything that happens between stimulus and response. He proceeds selectively, taking the relation of stimulus and response as his clue. He is properly interested in the movements which result from the application of the stimulus only in so far as they constitute response. Otherwise his study is not a study of behavior, but a study of movements. But when does a movement constitute a response? Do we label as stimulus the spoken word which results in overt action a week later, or the visual perception which sets a complicated and long-drawn-out problem, for no other reason than that it appears somewhere as an antecedent in the causal chain of events? If so, there is no obvious reason why the event which occurred just before or immediately after the _soi-disant_ stimulus should not be regarded as the true stimulus. Unless a satisfactory reason is forthcoming, it would seem better to substitute cause and effect for stimulus and response and to drop the term behavior from our vocabulary. Psychology then becomes a study of certain causal relationships, but is still without a principle for the selection of those causal events which are supposed to constitute its peculiar subject-matter. Even if we manage to become reconciled to this situation, however, our troubles are not yet at an end. There still remains the difficulty in certain cases of showing that the event which is selected as stimulus or cause bears any significant relationship to the event which figures in our scheme as the response. The stimulus is supposed to have a causal connection with the response, but how are we to know that this is the fact? How are we to know that the engineer who solves a problem for me at my request might not have done so anyway? No behaviorist can possibly show that the air waves set in motion by my vocalization were an indispensable stimulus. We doubtless believe that the spoken word was in fact the spark which lit the fuse and finally exploded the mine, but this belief involves a complication of causes which it is wholly beyond our power to control or to verify.