276 Transactions British Mycological Society. 



Mv thanks are due to the firms of Messrs Sutton & Sons, 

 Austin & McAslan, Drummond, Little & Ballantyne and Carter 

 for kindly providing me with the seed free. 



(B) As has been conclusively proved by the work of several 

 investigators including Neger (2), Marchal (3), Salmon (4, 5, 6) 

 and Reed(ii), many species of Erysiphaceae show that special- 

 isation of parasitism which results in the evolution of "biologic 

 forms," forms morphologically identical but totally distinct in 

 their infection powers, being, in most cases that have been 

 investigated, confined to a single species or to a few closely 

 related species of host plants and seldom having powers of 

 infection outside the genus. 



The experiments described in this paper were undertaken wdth 

 a view to extending the knowledge of this specialisation as far 

 as the morphological species Erysiphe Polygoni was concerned, 

 more especially in the case of the conidial form of that species 

 found on cultivated Brassicae. 



It may be stated here that throughout the whole investigation 

 the perithecial stage was never found on any Brassicae, so that 

 the oidium used in inoculations was onl}- identified as Erysiphe 

 Polygoni by its host, no other species of Erysiphaceae having, 

 so far as I know, been reported on the genus Brassica. 



The method of inoculation was the usual one of drawing the 

 edge of a sterile scalpel through the patch of mildew from which 

 the conidia were to be taken and then, after placing a drop of 

 water on the leaf to be inoculated, drawing the edge of the 

 scalpel through the drop thus leaving the conidia floating in it. 

 The whole plant was then kept under a bell-jar for twenty-four 

 hours after which the bell- jar was removed completely or raised 

 enough for ventilation. 



There was always a chance of a certain amount of natural 

 infection taking place especially towards the end of September, 

 although all the plants used were kept in a separate greenhouse 

 as soon as inoculated, so only those results were accepted as 

 positive where infection occurred at the exact spot inoculated 

 and the remainder of the leaf was entirely free from mildew; 

 if the remainder of the leaf showed any mildew the experiment 

 was discarded, whether there was infection at the inoculated 

 spot or not. 



All the plants used were grown from seed in pots in a green- 

 house. In each case the first sign of infection (when this occurred) , 

 was noticeable in about six days, when a small powdery patch of 

 mycelium, visible to the naked eye, appeared. 



A few general observations will first be given. 



The late summer and early autumn of 1913 were characterised 



