How, then, it may be asked, can we ever be certain that we are faithfully recalling the actual events of the past? Given a fairly good, that is, a cultivated memory, it may be said that in the case of very recent events a man may feel certain that, when the conditions of careful attention at the time to what really happened were present, a distinct recollection is substantially correct. Also it is obvious that with respect to all repeated experiences our memories afford practically safe guides. When memory becomes the basis of some item of generalized knowledge, as, for example, of the truth that the pain of indigestion has followed a too copious indulgence in rich food, there is little room for an error of memory properly so called. On the other hand, when an event is not repeated in our experience, but forms a unique link in our personal history, the chances of error increase with the distance of the event; and here the best of us will do well to have resort to a process of verification or, if necessary, of correction. In order thus to verify the utterances of memory, we must look beyond our own internal mental states to some external facts. Thus, the recollections of our early life may often be tested by letters written by ourselves or our friends at the time, by diaries, and so on. When there is no unerring objective record to be found, we may have recourse to the less satisfactory method of comparing our recollections with those of others. By so doing we may reach a rough average recollection which shall at least be free from any individual error corresponding to that of personal equation in perception. But even thus we cannot be sure of eliminating all error, since there may be a cause of illusion acting on all our minds alike, as, for example, the extraordinary nature of the occurrence, which would pretty certainly lead to a common exaggeration of its magnitude, etc., and since, moreover, this process of comparing recollections affords an opportunity for that reading back a present preconception into the past to which reference has already been made. The result of our inquiry is less alarming than it looks at first sight. Knowledge is valuable for action, and error is chiefly hurtful in so far as it misdirects conduct. Now, in a general way, we do not need to act upon a recollection of single remote events; our conduct is sufficiently shaped by an accurate recollection of single recent events, together with those bundles of recollections of recurring events and sequences of events which constitute our knowledge of ourselves and our common knowledge of the world about us. Nature has done commendably well in endowing us with the means of cultivating our memories up to this point, and we ought not to blame her for not giving us powers which would only very rarely prove of any appreciable practical service to us. NOTE. MOMENTARY ILLUSIONS OF SELF-CONSCIOUSNESS.