III. When we speak of the same object as perceived on two occasions, we do not, as has been noticed, mean that what is actually in the sense at the two different times is similar. Nevertheless, we find here, too, the notion of similarity, for the two experiences are not considered merely in themselves, but as elements in a group or series, and as each representing the whole series. When, therefore, we have the two experiences, we regard ourselves as having in them two experiences of the one series; which means, to be more explicit, that we have in mind on the two occasions two complexes which are similar, and which, when thought of together, are related to each other as the memory image and its original are related. Here the likeness lies in what is represented, not in the representatives. IV. As in the second kind of sameness, so in the fourth, the reference to similarity is unmistakable. We call qualities or things the same when they are of the one kind, when they are observed to resemble each other. V. The relation of representative and thing represented evidently implies the notion of similarity. It is quite true that we often recognize as in this relation things that we do not think of as being similar at all, and yet a little reflection will show that one thing can stand for another only _in so far_ as it resembles it. The resemblance may lie in the qualities of the things in themselves considered, or it may lie in external relations of which the things are capable, or functions which they may serve. The very notion of a proxy is a something which, for the purpose in hand, may be regarded as capable of assuming the functions of another. In so far as it can do this it is like the other. Things wholly different (if things could be wholly different) could not represent each other.