133 



myrmecophilous Hydnophytum and the saprophytic Rafflesia, 

 with the additional advantage of a botanical laboratory, Ameri- 

 can homes, and American botanists at its base. 



{To be continued.) 



RELATIONSHIP BETWEEN ROESTELIA TRANS- 

 FORMANS AND R. BOTRYAPITES 



By B. O. Dodge 



Roestelia transformans is one of the few rusts of this type 

 that have remained unconnected with a Gymnosporangiiim 

 form. Several European works on the fungi contain statements 

 to the effect that Gymnosporangium Ellisii is conruected with 

 R. transformans but no one has reported making such a con- 

 nection, Fromme has recently shown that G. Ellisii and Aeci- 

 dium Myricatum are phases of the same rust. 



Farlow, in his work on the " Gymnosporangia of the United 

 States," describes a foliicolous form of G. biseptatum as having 

 "all or nearly all two-celled spores." Kern has made a new 

 species of this leaf form under the name G. fraternum. 



A number of infection experiments with G. biseptatum and 

 "G. fraternum'' have been carried out at Columbia University 

 and a more detailed report of the results obtained will soon be 

 published. It has been found that the leaf form will infect 

 both Aronia and Amelanchier. The roestelia developing on 

 Aronia are very characteristic of R. transformans I The basal 

 hypertrophies from which secondary horn-like galls arise, and 

 the long, strongly hygroscopic peridial cells, coarsely warted on 

 their inner face, have been generally regarded as unmistakable 

 characters of this species. 



The hypertrophies produced by the infection on the Amelan- 

 chier are quite unlike those on the Aronia. The galls from 

 which the roestelia arise do not ordinarily coalesce and are at first 

 merely flattened, wart-like growths which later elongate some- 

 what. The roestelia are very different from R. transformans , 

 resembling, in fact, R. Botryapites as commonly described and 

 distributed in exsiccati. 



