312 Professor Marshall Ward, On the Question of 



on further growth of the species of Bromes investigated, and to 

 examine certain anatomical points. These results may be sum- 

 marised as follows : 



More than 2000 infections have been made on 22 species and 

 varieties of Bromus, principally seedlings, carefully selected and 

 cleaned aud grown in pots under conditions in most cases so 

 satisfactorily controlled that the rust-pustules only appeared (with 

 very rare exceptions) on the spots where the sowings of spores on 

 the leaves were made. These 22 species and varieties belonged to 

 four out of the five sub-genera of the genus, viz. Festucoides, Steno- 

 bromus, Serrafalcus and Ceratochloa :■: ,1 have not yet been able 

 to experiment with B. arduennensis representing the sub-genus 

 Libertia of Hackel. 



In order to illustrate the nature of these infection experiments, 

 let us take the following series from among the latest of those 

 made during the past summer. 



Series Numbered 735, 736 and 738. (Table I.) 



The pots of seedlings were all under like conditions and in- 

 fected similarly, by means of uredospores gathered separately and 

 kept separate, and known to be vigorous on germination. 



The annexed Table I. summarises the facts. 



On comparing these results we find no room for doubt as to 

 their meaning. The spores employed were in all cases germinating 

 well : the infections were made at the same time, and the con- 

 ditions were in every respect — including the age and treatment 

 of the seedlings before infection — alike. 



Nevertheless, we find as I had frequently found before 1 , that 

 spores from B. sterilis, while successfully infecting B. sterilis and 

 its ally B. madritensis, failed on all the other plants of the series. 

 On the other hand, the spores from B. mollis infallibly picked out 

 their own species and its allies B. velutinus, B. commutatus, B. ar- 

 vensis, B. secalinus, B. interrwptus and B. racemosus, but failed on 

 B. sterilis and its allies B. madritensis and B. maximus. 



Series Numbered 741 and 744. (Table I.) 



In the experiments where spores from B. secalinus Avere em- 

 ployed for inoculation we meet with exactly similar facts, only 

 here the fungus attacks B. secalinus by preference and B. arvensis 

 next in order. B. sterilis, B. maximus, B. madritensis, on the 

 other hand, as also B. velutinus, B. interruptus and B. racemosus, 

 were either entirely immune — to all appearance — or were only 

 attacked very slightly. 



1 Details of the conditions and results of other experiments are in course of 

 publication. 



