280 Salmoii,On Specialization of Parasitism in the Erysiphaceae. 



Turning again to Table 9, it is plaiii tliat on account of the 

 differences liere shown the four, or flve, Oidiu must, adopting current 

 terminology, be called ,biologie forms,' since althoiigh agreeing in 

 all morphological characters, they present certain biological peculia- 

 rities. The term ,biologic form' is preferable to that of ,biologic 

 species,' used by some authors, since the latter implies a fixity of 

 characters which may possibly not exist in these cases. One of the 

 most fundamental problems that arises in the investigation of these 

 jbiologic forms' may be presented in the form of the following 

 question : Does an Oidium on passing from one species of host-plant 

 to another acquii^e at once the same powers of infection as those 

 of the Oidium that has grown for a long time on the latter host- 

 plant? Consideration in detail of four hypothetical cases will help 

 to show more exactly the nature of this problem, and at the same 

 time show the difficulties that arise whtn we seek for a strict 

 definition of the term ,biologic form.' 



1. Let US suppose that an Oidium, Ol, occurs on a certain 

 host-plant a, and that this Oidium proves by experiment to be cap- 

 able of infecting another host-plant b, but is incapable of infecting 

 the plant d. Another Oidium^ 02, is found on the host-plant f/, 



Oias^ j^Ozd 



fig. 1. 



and is likewise found capable of infecting h, but not a. These 

 relations may be expressed by the following diagram: — (Fig. 1). 

 Since Ol a is incapable of directly infecting d, and 02 ö^ of 

 directly infecting a, these two Oidia would constitute ,biologic forms,' 

 — according to the definition of the term at present in use. Now, 

 both Ol and 02 infect h. AVe sliall have therefore on b either 

 (Iw), two ,biologic forms' with different powers of infection accor- 

 düig as they originated from Ol on a or 02 ou b (i. e. Ol A might 

 be able to infect a and not d\ similarly with 2 6); or, [Iß) we 

 may find on h an Oidium showing fresli characteristics in its powers 

 of infection. For instance, the Oidium on b, whether derived from 

 the spores of the Oidium on a {=Olb) or on d (=026) may 

 be rendered capable through its residence on the common host 6, of 

 infecting both a and d. At first sight it would then seem necessary 

 to recognize a tliird ,biologic form,' 03 on b, capable of directly 

 infecting a and d, and produced by either Ol or 02 passing on 

 to b. It might be ui'ged that in such a case it would be best to 

 consider all the Oidia represented in the diagram as belonging to 

 one jbiologic form,' Oabd, since although the Oidium on a cannot 

 directly infect d, it can do so by passing through b, and similarly 

 with the Oidium on d. The obvious objection to such a view is 

 that in speaking of Oabd as one ,biologic form' we are including 

 in the term individuals which show different physiological peculiar- 



