In addressing ourselves to the criticism of Neo-Lamarckism we shall neglect as far as possible all the different psychological principles concerned in it--which in any case would need rather a great amount of epistemological sifting--and shall keep to those hypothetic facts which are supposed to be such as may be actually observed in nature. All of you know that the so-called inheritance of acquired characters lies at the root of Lamarckism; and from this hypothesis our critical analysis is to start, disregarding a larger or smaller number of psychological principles that are brought into the field. The name of “acquired characters” may *a priori* be given to three different types of phenomena: firstly, variations including mutations; secondly, disease or injuries; and thirdly, the results of the actual process of adaptation of every kind. In the first of these groups, the true problem of the inheritance of “acquired” characters appears only with certain restrictions. All variations and mutations are indeed “acquired” by one generation so far as the earlier generation did not possess them, but mutations, at least, cannot be said to be acquired by the actual adult personality: they are innate in it from its very beginning, and therefore may better be called congenital.[147] Congenital properties of the mutation type are, in fact, known to be inherited: their inheritance does not present any problem of its own, but is included in the changes of the hereditary condition to which they are due altogether.[148] All properties of the variation type, on the other hand, having been studied statistically, are known to be inherited, to a certain small extent, as we have seen already whilst studying Darwinism, though they are possibly always liable to reversion. Modern science, as we know,[149] regards them as due to changes of nutrition, in the most general meaning of the word. Under such a view variations might indeed be said to belong to the acquired group of organic specifications; their inheritance, as will be seen later on, would hardly be quite a pure instance of what we are searching for. In no case can true variations claim to be of great importance in problems of transformism. [147] This would not be true, if the varieties of plants produced by Blaringhem, Klebs, and MacDougal by means of *external* agents were really “mutations” (comp. page 238, note 3). [148] Of course, the inheritance of mutations would imply a certain sort of “inheritance of acquired characters,” on the condition stated in the preceding note. But, probably, the germs of the next generation might be regarded here as being directly affected by the external agent, in a manner that will briefly be mentioned later on in the text. [149] Comp. page 238, note 2.