vin] LJREDINALES 



21 I 



Christman and Olive have inferred that the ancestral type of the heteroecious 

 species was a form with teleutospores only, a lepto- or a micro- form occurring 

 on the host of the present gametophyte. Other investigators (Tranzschel 

 '04, Grove '13) who regard the aecidium as an essential constituent of the 

 primitive rust, have suggested that hetcroecism may have arisen in relation 

 to host plants with a short vegetative period. In such a case there would 

 hardly be time for the production of the full complement of spores and the 

 fungus might either sin >rtcn its elaborate life-histi iry.giving the micro- < >r simi- 

 lar forms ( Uromyces Scillarum on wild hyacinth, Pucciniafuscaon anemone), 

 or some of the spores might become adapted to life on a new host. This 

 might be the case in particular with the aecidiospores, the development of 

 which, owing perhaps to the recent fertilization stage, is especially vigorous. 

 Aecidiospores must fall on hundreds of leaves besides those of the host, 

 and the germ-tubes in their case enter through the stomata. If then an 

 aecidiospore germinated and penetrated a satisfactory host it is suggested 

 that a mycelium might develop and further adaptations might fix the 

 heteroecious habit. Again it is readily understandable that the Gramineae 

 and other hosts with similarly refractor) - cuticles are easily infected by the 

 germ tubes from the aecidio- and uredospores which pass through the 

 stomata but not by those of the basidiospores which, in a large majority of 

 cases, penetrate the walls of the epidermal cells. This fact may be significant 

 in relation to the return of the parasite to its gametophytic host each spring. 



There is reason to believe that some species have an autoecious and a 

 heteroecious variety and the study of such forms is likely to prove of great 

 assistance. 



Specialization of Parasitism. The parasitism of the rusts shows very 

 marked specialization so that biological species have arisen, which, though 

 they may be morphologically indistinguishable, differ from one another in 

 their power to infect different hosts. Injury to the host may break down 

 its resistance to attack and may render it liable to infection by a species to 

 which it is normally immune. 



1'nder favourable conditions rust appears suddenly, and spreads with 

 great rapidity. Eriksson believed such epidemics to depend on the presence 

 in the seeds or buds of the host plant of the protoplasm of the rust, 

 indistinguishably mingled with that of the host, a mixture to which the term 

 mycoplasm was applied. He considered that the protoplasm of the fungus 

 remained unaltered till the leaves were formed ; under appropriate conditions 

 it then separated itself rapidly from that of the host and developed into the 

 ordinary spore-bearing mycelium. The investigations of Marshall Ward 

 and others have not substantiated this hypothesis. 



Nuclear Division. It would appear, from the work of various observers, 

 that nuclear division in the Uredinales has undergone a process of simpli- 



14—2 



