ANDROLETttAL FACTORS IN CENOTHERA

The genetic behavior of ~Enothera biennis and O. syrticola (0. muricata L.) as uniform and constant species is now generally explained by the assumption of the presence of two gametic lethal factors. The reciprocal hybrids of these forms are unlike each other and the double reciprocal hybrids return to the grandparental types. I t is assumed that there are two kinds of gametes, but that one of them is killed in the pollen and the other in the ovules. In fertilization the two kinds would meet, producing a complex constitution not unlike that of dioecious and heterostylous and of numerous selfsterile plants. In tracing a pedigree of O. Lamarckiana and its best studied relatives Boedijn proved that O. biennis is probably a derivative from that species, whereas O. syrticola belongs to another branch of the pedigree-tree. If this is true, the presence of analogous gametic lethals on such distant lines would indicate a more general occurrence in a latent condition. This supposition would be supported by the fact that of late a high degree of latent mutability has been found to exist in O. biennis. Among the mutants of O. Lamarckiana the apparition of gametic lethals is a common phenomenon and might be regarded as in some way connected with such a general latent condition. As a matter of fact, however, O. Lamarckiana itself is not known to contain gametic lethals. Its reciprocal hybrids with uniform species embrace the same types. On the contrary it has two zygotic lethals, which kill the germs within the young seeds, whenever the same lethal enters that germ from both parental sides. This produces the empty seeds, which in this species amount to one-half or more of the total number. Of these zygotic lethals one is linked to 109