He investigated the effects of heterogeneous hybridization also in fishes, which are a much more favourable object. The egg of the marine fish _Fundulus heteroclitus_ can be fertilized with the sperm of almost any other teleost fish, as Moenkhaus[37] first observed. This author did not succeed in keeping the hybrids alive more than a day, but the writer has kept many heterogeneous hybrids alive for a month or longer,[38] and found the same two striking facts which he had already observed in the heterogeneous cross between sea urchin and starfish: first, practically no transmission of paternal characters, and second, a sickly condition of the embryo which begins early and which increases with further development. The heterogeneous fish hybrids between, _e. g._, _Fundulus heteroclitus_ ♀ and _Menidia_ ♂ have usually no circulation of blood, although the heart is formed and beats and blood-vessels and blood cells are formed; the eyes are often incomplete or abnormal though they may be normal at first; the growth of the embryo is mostly retarded. In exceptional cases circulation may be established and in these a normal embryo may result, but such an embryo is chiefly maternal. [37] Moenkhaus, W. J., _Am. Jour. Anat._, 1904, iii., 29. [38] Loeb, J., _Jour. Morphol._, 1912, xxiii., 1. This incompatibility of two gametes from different species does not show itself in the case of heterogeneous hybridization only, but also though less often in the case of crossing between two more closely related forms. The cross between the two related forms _S. purpuratus_ ♀ and _S. franciscanus_ ♂ is very sturdy and shows no abnormal mortality as far as the writer’s observations go. If, however, the reciprocal crossing is carried out, namely that of _S. franciscanus_ ♀ and _S. purpuratus_ ♂, the development is at first normal, but beginning with the time of mesenchyme formation the majority of larvæ become sickly and die; and again the question may be raised whether or not the beginning of sickliness coincides with the development of mesenchyme cells. If we assume that the sickliness and death are due to the formation of a poison, we must assume that the poison is formed by the protoplasm of the egg, since otherwise we could not understand why the reciprocal cross should be healthy. All of these data agree in this one point, that the fusion by grafting or fertilization of two distant species is impossible, although the mechanism of the incompatibility is not yet understood. It is quite possible that this mechanism is not the same in all the cases mentioned here, and that it may be different when two different species are mixed and when incompatibility exists between varieties, as is the case in the graft on mammals. II. _The Chemical Basis of Genus and Species and of Species Specificity_