We seem forced to the conclusion--and this is the result to which all this discussion is intended to approximate--that variations, heritable variations at least, arise spontaneously. That is, there are organic differences which have no causes, a conclusion against which all our habits of reasoning rebel. Yet it may be possible to argue that the problem of the causes of variations is really a pseudo-problem after all, and that there is no logical reason why we should be compelled to postulate such causes. When we think of organic variability, do we not think, surreptitiously it may be, of something that varies, that is, something that ought to be immutable but which is compelled to deviate? But what is given to our observation is simply the variations among organisms. Let us think of the crude minting machines of Tudor times which produced coins which were not very similar in weight and design. From that time onward minting machines have continually been improved, each successive engine turning out coins more and more alike in every respect, so that we now possess machines which stamp out sovereigns as nearly as possible identical with each other. Yet they are not quite alike, and this is because the action of the engine, in all its operations, is not invariably the same. In imagination, however, we make a minting machine which does work perfectly, and turns out coins absolutely alike, but this ideal engine is only the conceptual limit to a series of machines each of which is more nearly perfect than was the last one. It is unlikely that matter possesses the rigidity and homogeneity which would enable us to obtain this perfect identity of result; nevertheless this identity has a very obvious utility, and we strive after it, so that the result of our activity is the conception of a perfect mechanism, and of products which are identical. We assume that the reasons why our early and cruder machines were imperfect are also the reasons why our later and more perfect ones do not produce the results that we desire.