﻿Dodge: Effect of host on Gymnosporangium 539 



grounds of the Brooklyn Botanic Garden April 19, 1914. This 

 cedar bore a witch 's-broom of G. Ellisii and many leaves showed 

 son of G. fraternum. Dr. Olive has informed me that although 

 sori of G. Ellisii have appeared this spring, no G. fraternum has 

 been found. This, so far as it goes, suggests that G. fraternum 

 is annual. 



Against the first of these arguments we have Farlow's state- 

 ment that the number of cells in the spores of G. hiseptatum varies 

 according to the size of the branches upon which sori develop — 

 the smaller the branches the greater the number of two-celled 



As for the second point it should be remembered that only two 

 converse infections have been reported in connection with this 

 genus. Arthur (i) reports that it requires only one year to com- 

 plete the development of G. davipes on the common juniper. 

 Plowright (17) shows that G. davariaeforme needs two years. 

 These authors do not state the length of time previous to these 

 experiments the junipers had been under observation so as to 

 preclude the possibility of natural infections. To cite a case in 

 point: a seedling Chamaecyparis about four inches high bearing 

 a dozen sori of G. Ellisii was brought in from Newfield in May. 

 A month later it was difficult to locate with certainty the region 

 infected. Heald (7) was unable to infect the red cedar with the 

 common apple rust. It is plainly as yet entirely uncertain how 

 long a time elapses between inoculation and the formation of the 

 spindle-formed distortions by G. hiseptatum. 



The behavior of known biological species of other rusts is such 

 as might account for the difference in the susceptibility of Aronia 

 and Amelanchier to infection by G. fraternum and G. hiseptatum. 



Having determined the facts regarding the infection limits of 

 G. fraternum and G. hiseptatum in connection with Aronia and 

 Amelanchier, the most difficult part of the work still remains to 

 be accomplished if a relationship between the two species of rust 

 is to be established. We must know, first, whether " G. fraternum 

 on Amelanchier'' {Roestelia transformans? on Amelanchier) goes 

 back to the cedar as G. fraternum or as G. hiseptatum or as both; 

 second, does G. hiseptatum on Amelanchier {Roestelia Botryapites 

 on Amelanchier) go back to the cedar and reappear as G. hisepta- 



