Sorting and blending of unit characters. 231 



repeats itself. The repeated form is modified by laeta characters (the 

 "triplet") and rarely a small percentage is modified by the velutina 

 characters ("quadruplet"). This is strong additional evidence that the 

 twin and triplet hybrids in the first generation of crosses between the 

 wild species, Oe. ptjcnocarpa and Oe. nutans, represent a type of segre- 

 gation different from that manifested in mutation crosses, and does not 

 appear capable of explanation on the same hypothesis. 



2nd. Theory of a dissimilar gametic constitution. Some 

 of the exceptional results obtained in cultures of the evening primroses, 

 it has been suggested, are due to a dissimilar constitution of the ga- 

 metes. In de Vries' interesting studies on the behavior of heterogamous 

 species he has developed a very ingenious method for the analysis of 

 the gametic constitution of these species which he terms "gamolysis" 

 (61, 1913). This analysis appears to show that the pollen constitution 

 of such species as Oe. biennis, muricata and most other heterogamous 

 species is not only quite different from that of the egg cell constitution 

 of the same species, but that in crosses between such species the pollen 

 constitution predominates. Since the reciprocal hybrids between biennis 

 and muricata strongly resemble the pollen parent, the heritable char- 

 acters of each species are almost wholly transmitted by its pollen. The 

 facies of biennis, or muricata made up of the patent characters, repre- 

 sents the pollen type, or constitution, of the species. The egg cell type, 

 or constitution, of such species can only be determined by crossing with 

 other species over which the unknown egg cell constitution is dominant. 

 This can only be determined by trial, but the analysis shows that the egg 

 cell type of biennis is very different from that of the pollen type. 



The patent characters, or constitution of biennis, or muricata, is 

 not exactly the same as the pollen type of the species. In self fertili- 

 zation the pollen type, or constitution, is somewhat modified by the egg 

 cell type. Since the phenomenon of partial sterility of the pollen and 

 ovules is widely distributed in the Oenotheras, de Vries (87, 1913) 

 suggests, as one of the possible explanations of patrocliny in hetero- 

 gamous species, a qualitative differentiation of the pollen and egg cells 

 during the meiotic divisions preceding their formation. According to 

 this hypothesis certain pollen grains would bear only the paternal here- 

 ditary characters, wliile others would bear only the maternal hereditary 

 characters, and a like differentiation would take place preceding the 

 formation of egg cells. Now if in biennis for example all the pollen 

 grains bearing the maternal hereditary characters should be sterile, and 



