3 



far recovered in September, 1910, as to set some 

 seed; in 1911 all were still infected. One of 

 them had some branches with hermaphrodite 

 flowers containing large ovary, short styles and 

 stamens with little or no pollen, but without 

 Ustilago spores, while the rest of the plant had 

 flowers with anthers full of spores, and the ovary 

 and styles more reduced. 



These observations seem to prove that infection 

 with Ustilago can turn the female flower into the 

 apparent hermaphrodite, but that no production 

 of female organs takes place in a male flower 

 when it becomes infected. 



L. DONCASTEE 



CAMBRIDGE, ENGLAND 



The occurrence of uninfected hermaphrodite 

 flowers on one of Doncaster's original infected 

 plants may possibly indicate that this plant 

 was not a female previous to its infection, but 

 a hermaphrodite. If it were possible to secure 

 pollen from a ustilaginized female, certain 

 genetic problems of very great interest might 

 be solved. It is of great theoretic importance 

 to know whether infection by Ustilago affects 

 the genotypic nature of the host. If the effect 

 is purely somatic, as seems to me the more 

 probable, the offspring of a self -fertilized her- 

 maphroditic female, or of a normal female fer- 

 tilized by sperms from a hermaphroditic 

 female, should consist only of females (if un- 

 infected), and not of females and hermaphro- 

 dites, as I have shown to be the case when a 

 female is fertilized by a hermaphroditic male. 

 If infection by Ustilago produces a genotypic 

 modification, it would be interesting to know 

 whether such induced hermaphrodites are 

 homozygous like the females by whose modifi- 

 cation they are produced; they should in that 

 case yield only hermaphrodite offspring. Her- 

 maphroditic males produce both female and 

 hermaphroditic male offspring, because the 

 males are sex-heterozygotes. 



As I have been fortunate enough thus far 



