Fungus of Loliuni, 293 



unlikely that two races of l>oth Lolium temulentum an«l Lolium 

 perenne exist, one with a symbiont, the other fun£,nis free. An occa- 

 sional grain of either species may show the absence of hyphae, but 

 this would be accidental in character, so that instead of an ever- 

 increasing numljer of the latter type, they wouhl always tend to 

 .remain at a more or less stationary minimum. 



Summary. 



The foregoing investigation has led to the following, results : — 



(1) The occurrence of the fungus in the genus Lolium is wider 



and more constant than has hitherto been demonstrated. 



(2) Colour of the grain cannot be regarded as a diagnostic 



character in regard to tlie presence or absence of the 



fungus. 

 '(3) The fungus is intra-cellular or endophytic in nature. 

 '(4) Ihe distribution in the grain is not a result of any special 



method of infection, but is a result of the function of 



the fungus during the grain's development. 



(5) It is present in the embryo-sac at or innnediately after 



fertilisation. 



(6) Ihe fungus increases in quantity at the expense of the 



nucellus, and the cells of the carpel wall. This is only 

 a temporary phase. On the formation of endosperm 

 the fungus is absorbed as a source of food-supply to the 

 developing embryo. 



(7) The endosperm is- formed* by the division of its outer layer. 



This layer functions as a kind of cambium. I have 

 termed it the endospermic camhium. The cells which 

 are cut off. always to the inner side, increase in size, re- 

 main tliin-walled, and become packed witli starcli. Tliis 

 outer meristematic layer is constantly receiving and 

 absorbing hyphae, whicli, if present in any <|uanrity, 

 are finally crushed into a layer around the ])eri})hcry of 

 the endosperm. If the fungus does not kt-e]) ]>ace witli 

 the al)Sorbing power of the endosperm, no 1iy])lial layer 

 is formed in the ripe grain, but hyphae « an tlu-n be 

 found in the scutellum and embryo. 

 '(8) The endospermic cambium after it has ceased to divide 

 persists as the aleurone layer, which, in turn, reveivcK 

 a supply of nntrini'Mit from tlu- fun^jal system. 



ir>A 



