Hitherto we have derived the variations of the determinants of the germ-plasm, upon which we based the process of germinal selection, from _chance local_ fluctuations in nutrition, such as must occur in an individual id, independently of the nutrition of the other ids of the same germ-plasm. But there are doubtless also influences which set up similar nutritive changes in _all ids_, and by which, therefore, all homologous determinants, in as far as they are sensitive to the nutritive change in question, are affected in the same manner. To this category belong changes in the external conditions of life, and particularly climatic changes. It is, then, germinal selection alone which brings about the presence of a majority of ids with determinants varying in the same direction, and personal selection has no part in the transformation of the species. Many years ago I instituted experiments with a small butterfly, _Pararga egeria_, and these showed that a heightened temperature so influenced the pupæ of this form that the butterflies emerged with a different and deeper yellow ground-colour, similar to that of the long-known southern variety _Meione_. More thoroughly decisive, however, were the experiments on _Polyommatus phlæas_, the small 'fire-butterfly,' which were carried on in the eighties by Merrifield in England and by myself almost at the same time. I shall discuss these later in more detail, and will only say here that this butterfly, whose range extends from Lapland to Sicily, occurs in two forms, the southern distinguished by a 'dusting' of deep black from the northern, in which the wing-surfaces are of a pure red-gold. The experiments showed that the southern form can be artificially produced by warmth, and the interpretation must be that the direct influence of higher temperature affects the quality of the nutritive fluids in the germ-plasm, and thereby at the same time the determinants of one or more kinds of wing-scales are caused to vary in all the ids in the same direction, in such a fashion that they give rise to black scales instead of the former red-gold ones. It is thus certain that there are external influences which cause particular determinants to vary in a particular manner. I call this form of germinal variation 'induced' germinal selection, and contrast it with 'spontaneous' selection, which is caused, not by extra-germinal influences, but by the chances of the intra-germinal nutritive conditions, and which will, therefore, not readily occur at the same time in all the ids of a germ-plasm, and so will not give rise to variation of the same kind in the homologous determinants of all the ids.