If this theory of memory be correct, the occurrence of any present bare sensation itself suffices to recall, in some sense, the whole past. But this is no use for practical purposes, just as the whole of the fact given in present perception is useless for practical purposes until it has been analysed into qualities. According to Bergson we treat the material supplied by memory in much the same way as that supplied by perception. The whole field of the past which the present calls up is much wider than what we actually remember clearly: what we actually remember is arrived at by ignoring all the past except such scraps as appear to form useful precedents for behaviour in the present situation in which we find ourselves. Perhaps this explains why sometimes, at the point of death, when useful behaviour is no longer possible, this selection breaks down and the whole of the past floods back into memory. The brain, according to Bergson, is the organ whose function it is to perform this necessary work of selection out of the whole field of virtual memory of practically useful fragments, and so long as the brain is in order, only these are allowed to come through into consciousness as clear memories. The passage just quoted goes on to speak of “the part played by the brain in memory.” “The brain does not serve to preserve the past but primarily to obscure it, and then to let just so much as is practically useful slip through.” But the setting of the whole past, though it is ignored for convenience, still makes itself felt in the peculiar qualitative flavour which belongs to every present fact by reason of its past. Even in the case of familiar facts this flavour is no mere repetition but is perpetually modified as the familiarity increases, and it is just in this progressively changing flavour that their familiarity consists.