Evidence of this kind does not, of course, prove that literally _nothing_ is ever lost but it goes far towards upsetting the ordinary view that it is the rule for past experience to be annihilated and the exception for fragments here and there to be preserved in memory. The evidence which has so far been collected and which is rapidly accumulating at least seems to justify us in reversing this rule and saying rather that to be preserved is the rule for experience and to be lost would be the exception, if indeed any experience ever really is lost at all. This way of regarding the field of memory is further supported by such evidence as has been collected with regard to the influence of past experience in dreams, phobias and various forms of insanity, but in these cases, of course, it is only isolated past experiences here and there whose activity can be observed, and so, while helping to upset the most natural assumption that whatever cannot be recalled by ordinary efforts of memory may be assumed to have been destroyed, they do not lend very much support to the wider view put forward by Bergson, that no experience, however trivial, is ever destroyed but that all of it is included in the field out of which memory makes its practical selection. Taking all the evidence with regard to the preservation of past experience which is at present available, then, it is safe to say that, while it cannot, in the nature of things, absolutely prove Bergson’s theory of knowledge, it in no way conflicts with it and even supports it, positively in the sense that the theory does fit the facts well enough to explain them (though it goes further than the actual facts and makes assumptions which can neither be proved nor disproved by an appeal to them) and negatively in the sense that what we now know about memory actually conflicts with the “natural” view that past experience which we are unable to recall has been destroyed, which is commonly appealed to to show the absurdity of the rival theory put forward by Bergson.