222 Journal of Agricultural Research voi. xv. No. 4 



infecting it easily. Freeman {8) did similar work and came to the same 

 conclusions. No attempt has yet been made to repeat the work of these 

 investigators, but the writers have observed that some species of Bromus 

 are hosts for most biologic forms of P. graminis. Bromus tectorum, for 

 instance, can be infected by all the common forms of P. graminis in the 

 United States. It is possible that the conditions of experimentation on 

 which the idea of bridging is based were not rigid enough to exclude all 

 possibility of working with a mixture of biologic forms. On the other 

 hand, it is quite possible that the brown rust of bromes with which 

 Ward and Freeman worked was an unstable, easily changed form. 



Salmon (16) added considerable evidence to the concept of the efficacy 

 of bridging hosts in widening the host range of biologic forms by his 

 experiments with Erysiphe graminis DC. on various species of Bromus. 



Freeman and Johnson (9) applied the principle of bridging hosts to 

 P. graminis. They state {p. 20) that — 



The barley stem rust enjoys the widest range of any of the biologic forms of the 

 cereal rusts. On the other hand, a transfer of any of the other stem rusts to barley 

 widens the range of that rust. We have here, then, a decided reaction of host upon 

 parasite, enabling the latter to adapt itself to hosts not ordinarily congenial; for 

 instance, W >B >0. 



Johnson (u, p. 10) obtained similar results with timothy rust. He 

 states — 



A small number of experiments to test whether or not the timothy rust can be 

 transferred by means of bridging hosts to various cereals which are not successfully 

 infected directly from timothy were tried, and it was found that by using Avena sativa 

 as a bridging host the rust easily transferred to Hordeum vulgare (4 *imes in 10 trials); 

 and by using Festuca elaiior it transferred to Hordeum vulgare (twice in 10 trials) and 

 to Triticum vulgare (once in 10 trials); and by using Dactylis glomerata it transferred 

 to Triticum vulgare (once in 5 trials). By the use of the bridging hosts the rust im- 

 doubtedly could be made to transfer to many grasses on which it will not grow when 

 coming directly from timothy, but on which it might continue to grow after such a 

 transfer. That this takes place to some extent in natxu-e is very probable, and these 

 trials, together with recent experiments of a similar nature on the rusts of grains, 

 throw much light on the possible origin of many of the so-called "physiological 

 species" of rust. 



Pole Evans (7) stated that hybrid wheats could also act as bridging 

 hosts, enabling P. graminis to infect the susceptible parent more vigor- 

 ously and even to attack the highly resistant or almost immune parent. 

 Bififen (2), however, obtained no evidence of such remarkable changes. 



Arthur {i, p. 22J-228) cited evidence to show that barberry (Berberis 

 spp.) may also act as a bridging host, enabling "racial strains" of Puc- 

 cinia poculiformis (J acq.) Wettst. (=P. graminis Pers.) to increase 

 their range of infection capabilities. Bolley and Pritchard (j) and others 

 attributed to barberry a " reinvigorating function " for the rust, although 

 not necessarily a bridging function. 



