I am inclined to think that, in the present state of physiology, the introduction of the engram does not serve to simplify the account of mnemic phenomena. We can, I think, formulate the known laws of such phenomena in terms, wholly, of observable facts, by recognizing provisionally what we may call "mnemic causation." By this I mean that kind of causation of which I spoke at the beginning of this lecture, that kind, namely, in which the proximate cause consists not merely of a present event, but of this together with a past event. I do not wish to urge that this form of causation is ultimate, but that, in the present state of our knowledge, it affords a simplification, and enables us to state laws of behaviour in less hypothetical terms than we should otherwise have to employ. The clearest instance of what I mean is recollection of a past event. What we observe is that certain present stimuli lead us to recollect certain occurrences, but that at times when we are not recollecting them, there is nothing discoverable in our minds that could be called memory of them. Memories, as mental facts, arise from time to time, but do not, so far as we can see, exist in any shape while they are "latent." In fact, when we say that they are "latent," we mean merely that they will exist under certain circumstances. If, then, there is to be some standing difference between the person who can remember a certain fact and the person who cannot, that standing difference must be, not in anything mental, but in the brain. It is quite probable that there is such a difference in the brain, but its nature is unknown and it remains hypothetical. Everything that has, so far, been made matter of observation as regards this question can be put together in the statement: When a certain complex of sensations has occurred to a man, the recurrence of part of the complex tends to arouse the recollection of the whole. In like manner, we can collect all mnemic phenomena in living organisms under a single law, which contains what is hitherto verifiable in Semon's two laws. This single law is: IF A COMPLEX STIMULUS A HAS CAUSED A COMPLEX REACTION B IN AN ORGANISM, THE OCCURRENCE OF A PART OF A ON A FUTURE OCCASION TENDS TO CAUSE THE WHOLE REACTION B. This law would need to be supplemented by some account of the influence of frequency, and so on; but it seems to contain the essential characteristic of mnemic phenomena, without admixture of anything hypothetical.