I] SPECIALIZATION 25 



Another remarkably resistant species is Bromus mollis, yet the so-called 

 B. "■ kordeaceus" the seed of which was sent to Cambridge from Petrograd, and 

 which is morphologically indistinguishable from B. mollis, is nevertheless 

 susceptible to infection by the biological species which B. mollis is able to 

 resist. In other words the morphological species B. mollis includes two 

 groups or races possessingdistinct physiological (or constitutional) characters 

 and respectively immune and susceptible to infection. 



In the case of certain parasitic fungi, and especially oi Puccinia glumarum,, 

 the yellow rust on wheat, Biffen has shown that resistance to the attacks of 

 the parasite is a recessive character in the Mendelian sense. When a variety 

 susceptible to rust was crossed by another practically immune' from it, the 

 offspring was at least as much infected as the original rusty type. In the 

 next generation segregration took place in the ordinary way, three-quarters 

 of the plants being rusty and the remaining quarter standing green and un- 

 injured among them. In relation to this investigation Marryat showed that 

 the rust hyphae are checked after entering the stomata of the resistant plants 

 either by the death of the host tissue locally, accompanied by the starvation 

 and death of the parasite, or, after a more protracted struggle, by the gradual 

 degeneration of the invading hyphae. If, as has been suggested above, 

 resistance depends on the presence of an antitoxin, the dominance of sus- 

 ceptibility in this case must be taken to indicate that the development of 

 the antitoxin is inhibited by the presence of some additional factor in the 

 dominant forms. Cases may be expected where, in the absence of an 

 inhibitor, resistance is dominant and depends on the presence of the anti- 

 toxin or its progenitor. Such a case is suggested by the work of Biffen on the 

 inheritance of immunity to ergot, but here two factors appear to be involved. 



The susceptible races have in some instances an importance beyond that 

 implied in their own liability to infection ; it has been suggested that they 

 may serve as a bridge by which the fungus can pass from its original host 

 to a species resistant to direct attack. 



This was indicated by the work of Marshall Ward on Puccinia dispersa, 

 and by Salmon using Bromus " kordeaceus" as a bridging host. Inoculation 

 experiments showed that the form oi Erysiphe Graminis on Bromus racemosus 

 was incapable of developing on uninjured B. commutatus but that it never 



failed to produce full infection on B. "kordeaceus." The mycelium on its new 

 host gave rise in due course to conidia and some of these, when transferred 

 to B. commutatus, succeeded in establishing themselves and produced full 

 infection in eight days. 



