day the action he had taken the day before, we still, without perhaps perceiving it, supposed him to be guided by memory in all the details of his action, such as his taking down his hat and going out into the street. We could not, indeed, deprive him of all memory without absolutely paralysing his action.

Nevertheless new ideas, new faiths, and new actions do in the course of time come about, the living expressions of which we may see in the new forms of life which from time to time have arisen and are still arising, and in the increase of our own knowledge and mechanical inventions. But it is only a very little new that is added at a time, and that little is generally due to the desire to attain an end which cannot be attained by any of the means for which there exists a perceived precedent in the memory. When this is the case, either the memory is further ransacked for any forgotten shreds of details, a combination of which may serve the desired purpose; or action is taken in the dark, which sometimes succeeds and becomes a fertile source of further combinations; or we are brought to a dead stop. All action is random in respect of any of the minute actions which compose it that are not done in consequence of memory, memory, real or supposed. "So that random, or action taken in the dark, or illusion, lies at the very root of progress.

I will now consider the objection that the phenomena of instinct and embryonic development ought not to be ascribed to memory, inasmuch as certain other phenomena of heredity, such as gout, cannot be ascribed

to it.

Those who object in this way forget that our actions fall into two main classes: those which we have often repeated before by means of a regular series of subordinate actions beginning and ending at a certain toler

ably well-defined point-as when Herr Joachim plays a sonata in public, or when we dress or undress ourselves; and actions the details of which are indeed guided by memory, but which in their general scope and purpose are new-as when we are being married or presented at court.

At each point in any action of the first of the two kinds above referred to there is a memory (conscious or unconscious according to the less or greater number of times the action has been repeated), not only of the steps in the present and previous performances which have led up to the particular point that may be selected, but also of the particular point itself; there is, therefore, at each point in a habitual performance a memory at once of like antecedents and of a like present.

If the memory, whether of the antecedent or the present, were absolutely perfect; that is to say, if the vibrations in the nervous system (or, if the reader likes it better, if the molecular change in the particular nerves affected-for molecular change is only a change in the character of the vibrations going on within the molecules-it is nothing else than this)-if the vibrations in the particular nerves affected by any occurrence continued on each fresh repetition of the occurrence in their full original strength and without having been interfered with by any other vibrations: and if, again, the new waves running into the faint old ones from exterior objects and restoring the lapsed molecular State of the nerves to a pristine condition were absolutely identical in character on each repetition of the occurrence with the waves that ran in upon the last occasion, then there would be no change in the action, and no modification or improvement could take place. For though indeed the latest performance would always have one memory more than the latest but one to guide

it, yet the memories being identical, it would not matter how many or how few they were.

On any repetition, however, the circumstances, external or internal, or both, never are absolutely identical: there is some slight variation in each individual case, and some part of this variation is remembered, with approbation or disapprobation as the case may be.

The fact, therefore, that on each repetition of the action there is one memory more than on the last but one, and that this memory is slightly different from its predecessor, is seen to be an inherent and, ex hypothesi, necessarily disturbing factor in all habitual action-and the life of an organism should, as has been sufficiently insisted on, be regarded as the habitual action of a single individual, namely, of the organism itself, and of its ancestors. This is the key to accumulation of improvement, whether in the arts which we assiduously practise during our single life, or in the structures and instincts of successive generations. The memory does not complete a true circle, but is, as it were, a spiral slightly divergent therefrom. It is no longer a perfectly circulating decimal. Where, on the other hand, there is no memory of a like present, where, in fact, the memory is not, so to speak, spiral, there is no accumulation of improvement. The effect of any variation is not transmitted, and is not thus pregnant of still further change.

As regards the second of the two classes of actions above referred to-those, namely, which are not recurrent or habitual, and at no point of which is there a memory of a past present like the one which is present now— there will have been no accumulation of strong and well-knit memory as regards the action as a whole, but action, if taken at all, will be taken upon disjointed fragments of individual actions (our own and those of

other people) pieced together with a result more or less satisfactory according to circumstances.

But it does not follow that the actions of two people who have had tolerably similar antecedents and are placed in tolerably similar circumstances should be more unlike each other in this second case than in the first. On the contrary, nothing is more common than to observe the same kind of people making the same kind of mistake when placed for the first time in the same kind of new circumstances. I did not say that there would be no sameness of action without memory of a like present. There may be sameness of action proceeding from a memory, conscious or unconscious, of like antecedents, and a presence only of like presents without recollection of the same.

The sameness of action of like persons placed under like circumstances for the first time, resembles the sameness of action of inorganic matter under the same combinations. Let us for the moment suppose what we call non-living substances to be capable of remembering their antecedents, and that the changes they undergo are the expressions of their recollections. Then I admit, of course, that there is not memory in any cream, we will say, that is about to be churned of the cream of the preceding week, but the common absence of such memory from each week's cream is an element of sameness between the two. And though no cream can remember having been churned before, yet all cream in all time has had nearly identical antecedents, and has therefore nearly the same memories, and nearly the same proclivities. Thus, in fact, the cream of one week is as truly the same as the cream of another week from the same cow, pasture, etc., as anything is ever the same with anything; for the having been subjected to like antecedents engenders the closest

similarity that we can conceive of, if the substances were like to start with. Same is as same does.

The manifest absence of any connecting memory (or memory of like presents) from certain of the phenomena of heredity, such as, for example, the diseases of old age, is now seen to be no valid reason for saying that such other and far more numerous and important phenomena as those of embryonic development are not phenomena of memory. Growth and the diseases of old age do indeed, at first sight, appear to stand on the same footing. The question, however, whether certain results are due to memory or no must be settled not by showing that two combinations, neither of which can remember the other (as between each other), may yet generate like results, and therefore, considering the memory theory disposed of for all other cases, but by the evidence we may be able to adduce in any particular case that the second agent has actually remembered the conduct of the first. Such evidence must show firstly that the second agent cannot be supposed able to do what it is plain he can do, except under the guidance of memory or experience, and secondly, that the second agent has had every opportunity of remembering. When the first of these tests fails, similarity of action on the part of two agents need not be connected with memory of a like present as well as of like antecedents; when both fail, similarity of action should be referred to memory of like antecedents only.

Returning to a parenthesis a few pages back, in which I said that consciousness of memory would be less or greater according to the greater or fewer number of times that the act had been repeated, it may be observed as a corollary to this, that the less consciousness of memory the greater the uniformity of action, and vice versa. For the less consciousness involves the memory's