Let us imagine that we succeed in separating these two effects. It will be conceived, theoretically, that a separation of this kind will lay bare the hidden resemblances, giving to each collaborator the part which belongs to it. The excitement, for instance, will be suppressed, and the cognition will be retained. Is it possible to make, or at least to imagine, such an analysis? Perhaps: for, of these two competing activities, one is variable, since it depends on the constantly changing nature of the objects which come into relation with us; the other, on the contrary, is a constant, since it expresses the contribution of our nerve substance, and, though this last is of very unstable composition, it necessarily varies much less than the series of excitants. We consequently see faintly that these two elements differ sufficiently in character for us to be able to suppose that they are separable by analysis. But how could this analysis be made? Evidently not by chemical or physical means: we have no need here of reagents, prisms, centrifugal apparatus, permeable membranes, or anything of that kind. It will suffice to suppose that it is the consciousness itself that is the dialyser. It acts by virtue of its own laws--that is to say, by changes in intensity. Supposing that sensibility increases for the variable elements of the undulation, and becomes insensible for the constant elements. The effect will be the same as a material dissociation by chemical analysis: there will be an elimination of certain elements and the retention of others. Now, all we know of the consciousness authorises us to entrust this role to it, for it is within the range of its habits. We know that change is the law of consciousness, that it is effaced when the excitements are uniform, and is renewed by their differences or their novelty. A continued or too often repeated excitement ceases in time to be perceived. It is to condense these facts into a formula that Bain speaks of the law of relativity of cognition, and, in spite of a few ambiguities on the part of Spencer and of Bain himself in the definition of this law,[50] the formula with the sense I have just indicated is worth preserving. Let us see what becomes of it, when my hypothesis is adopted. It explains how certain excitements proceeding from the objects--that is to say, forming part of the variable element--cease to be perceived when they are repeated and tend to become constant. _A fortiori_, it seems to me, should the same law explain how the constant element _par excellence_, the one which never varies from the first hour, is never perceived. There is, in the concert of the sounds of nature, an accompaniment so monotonous that it is no longer perceived; and the melody alone continues to be heard.