One of the most extraordinary cases known to me in this connection is the unconscious selection on the part of native Indian cultivators, perfectly ignorant of the principles involved, of spring and autumn forms of Rice, Wheat, Castor Oil, Sugar Cane, Cotton, and other crops. "It has been estimated that Bengal alone possesses as many as 10,000 recognisable forms of rice." Now there is not the slightest ground for doubt that these have been unconsciously bred from the semi-aquatic native species during the many centuries of Indian agriculture, and nevertheless they have, among other peculiar races, some hill-breeds which they cultivate on dry soils and without direct inundation. That is to say, they possess tropical and temperate races differing far more than our spring and summer wheats. Something has been gained, then, if we can show that there is nothing absurd or hopeless in the search for disease-proof or resistant races, and I think this can be done. We must not forget that the ideal usually set before himself by a breeder of plants has hitherto been almost exclusively some standard of size, form, colouring, and so forth, of the flower, or of taste and texture of the fruit, tuber, etc., though experiments with _Cinchona_, with brewery yeasts, and other plants remind us that variations in other directions have been attended to also. Now it is obvious that in breeding sour limes and sweet oranges the cultivator is selecting, and intensifying by selection, very different metabolic processes in the cell: he can test the results of these, and so the selection proceeds. The question is, Could he select at the same time those variations in cell activity which express themselves in properties of the flower, fruit, foliage, etc., he desires, as well as such variations as aid the cells in repelling fungi, insects, or exigencies of the non-living environment? That more or less disease-proof varieties could be selected if that object alone were kept in view can hardly be doubted; plenty of examples exist already which show that the necessary variations to work upon exist in just those secretions of protoplasm, etc., which we have seen are concerned in repelling or attracting parasites. The Sweet Almond has lost the power of producing amygdalin and prussic acid in its cells; Cinchona plants vary immensely in the quantity of quinine formed, and in European hot-houses may even form none at all; some varieties of Maize have sugar and dextrine instead of starch in their endosperms, or coloured instead of clear sap in the aleurone layer, and recent researches prove that they can transmit these peculiarities to hybrid offspring; non-poisonous bacteria have frequently been got from poisonous species simply by cultivation under special conditions, and pigmented forms can be bred into non-pigmented races.