sterility which prevents two healthy gametes on uniting from producing a zygote with normal powers of growth and reproduction. For there are forms of sterility which are purely mechanical. The pollen of _Mirabilis jalapa_ cannot fertilise _M._ {152} _longiflora_, because the pollen tubes of the former are not long enough to penetrate down to the ovules of the latter. Hybrids can nevertheless be obtained from the reciprocal cross. Nor should we expect offspring from a St. Bernard and a toy terrier without recourse to artificial fertilisation. Or sterility may be due to pathological causes which prevent the gametes from meeting one another in a healthy state. But in most cases it is probable that the sterility is due to some other cause. It is not inconceivable that definite differences in chemical composition render the protoplasm of one species toxic to the gametes of the other, and if this is so it is not impossible that we may some day be able to express these differences in terms of Mendelian factors. The very nature of the case makes it one of extreme difficulty for experimental investigation. At any rate, we realise more clearly than before that the problem of species is not one that can be resolved by the study of morphology or of systematics. It is a problem in physiology. * * * * * {153} CHAPTER XIV ECONOMICAL Since heredity lies at the basis of the breeder's work, it is evident that any contribution to a more exact knowledge of this subject must prove of service to him, and there is no doubt that he will be able to profit by Mendelian knowledge in the conduct of his operations. Indeed, as we shall see later, these ideas have already led to striking results in the raising of new and more profitable varieties. In the first place, heredity is a question of individuals. Identity of appearance is no sure guide to reproductive qualities. Two individuals similarly bred and indistinguishable in outward form may nevertheless behave entirely differently when bred from. Take, for instance, the family of sweet peas shown on Plate IV. The F_2 generation here consists of seven distinct types, three sorts of purples, three sorts of reds, and whites. Let us suppose that our object is to obtain a true breeding strain of the pale purple picotee form. Now from the proportions in which they come we know that the dilute colour is due to the absence of the factor which intensifies the colour. Consequently the picotee cannot throw the {154} two deeper shades of red or purple. But it may be heterozygous for the purpling factor when it will throw the dilute red (tinged white), or it may be heterozygous for either or both of the two colour factors (cf. p. 44), in which case it will throw whites. Of the picotees which come in such a family, therefore, some will give picotees, tinged whites, and whites, others will give picotees and tinged whites only, others will give picotees