This sounds so simple and so convincing that we are tempted to try it. But first let us make sure that we have learned our lesson properly. In the course of long ages mankind has slowly accumulated enough experience to warrant the confident belief that Nature is uniform. Now, primitive man was of course a savage, and knew nothing of the Uniformity of Nature; he therefore could not have had sufficient evidence for believing anything in his experience. But it is on the accumulation of such experiences--every one of which we must reject because they were not based on the Uniformity of Nature--that our belief in the Uniformity of Nature is supposed to rest. In other words, it is based on them and they were based on nothing. This result of acting strictly up to the principle of not suffering anything to pass without sufficient evidence seems somewhat discouraging, until the man of science comes to our rescue and reminds us that just as we, without knowing it, have acted all our lives on the tacit assumption that Nature is uniform, so did primitive man; and that consequently there really was sufficient evidence and scientific proof for the savage's experiences, though of course he could not have framed it in words; and so, the bases of the Uniformity of Nature are really quite sound. But even now we are not altogether out of our difficulties, for granted that the savage, like ourselves, tacitly assumed Nature to be uniform, was there sufficient evidence for the assumption? and if so, what was the evidence? It could not be the Uniformity of Nature, because that is just the question; and, if it was anything else, it was not sufficient evidence. It really seems rather difficult to get sufficient evidence for the axiom, viz. the Uniformity of Nature, on which the whole of science is built. And yet we must have sufficient evidence for it, or else we shall have to conclude that Science has no more logical foundation than Religion. But once more the man of science comes to our assistance and explains that in the beginning, before the Uniformity of Nature is proved, it is only probable that what has once happened will happen again in similar circumstances, and at first perhaps not very probable; but when wider and wider experience still shows that what has once happened does actually happen again under the same circumstances, the Uniformity of Nature becomes more and more probable, until at last, if not actually proved, it is still the most probable hypothesis that we possess or can possess: "our highest and surest generalisations remain on the level of justifiable expectations; that is, very high probabilities."[26]