Oct. 28, 1918 Plasticity of Biologic Forms of Puccinia graminis 225 



in a preliminary way, that this change brought about by a host plant 



should be manifested, not only in the parasitic capabilities of the rust, 



but also in the morphology. Extensive work was therefore done on 



this phase of the problem, and the results will be presented in a separate 



paper. 



EXPERIMENTAL METHODS 



The methods used in inoculating and incubating the plants were 

 similar to those described by Stakman and Piemeisel {21, p. 4^1-4^2). 

 On account of the fact that conclusions would be difficult to draw unless 

 the experimental methods excluded to the greatest extent possible 

 accidental infection, the utmost precautions were taken to prevent 

 accidental infection. This is rather difficult when working under the 

 conditions necessary in such an investigation. Unfortunately the 

 pure-culture methods of bacteriology and mycology can not be success- 

 fully employed for rusts. However, it is possible to reduce the number 

 of accidental infections to a very small minimum. 



The seedling plants used in the experiments were grown under cages 

 made of two layers of fine-mesh muslin separated by a dead-air space 

 about an inch wide. Immediately after inoculation they were placed 

 under bell jars, and as soon as the incubation period was over they were 

 again placed under cages similar to those under which the seedling 

 plants were grown. 



Every precaution was taken to prevent infection from the outside 

 by air-borne spores and to destroy all infected material as soon as 

 possible. When accidental infection did take place, it could practically 

 always be determined with certainty, although in a few cases it was not 

 possible to do this. When, for instance, a leaf of wheat, as very rarely 

 indeed happened, developed a normal rust uredinium as a result of 

 inoculations with P. graminis secalis from rye, and when the spores in 

 the uredinium morphologically and parasitically were in every way 

 typical of P. graminis tritici spores, it seemed more reasonable to assume 

 that accidental infection had taken place than to assume that bridging 

 had occurred or that the rust had mutated. This conservatism in 

 interpreting results seemed to be especially necessary when such a 

 tremendous preponderance of evidence accumulated showing that 

 bridging and mutations did not occur. 



It was soon found also that the greatest precaution was necessary in 

 being absolutely certain that the biologic form in use in a given experi- 

 ment was absolutely pure before the results could be clearly interpreted. 

 In making inoculations from barley and rye and many grasses from the 

 field, both the tritici and secalis forms develop quite often. It is possible 

 to inoculate all of the cereals and to obtain results which seem to show 

 that only one biologic form is present, then to inoculate one of the 

 common hosts for both forms, and then to find that a very small amount 



