REED: SPECIALIZATION OF PARASITIC FUNGI 399 



The occurrence of bridging hosts has been suggested in other 

 groups of parasites as well as in the rusts. Salmon (126, 127), in 

 connection with his work on the powdery mildew of the bromes, has 

 suggested the possibility that Bromus hordeaceus may act as a bridging 

 host for the mildew on Bromus racemosus and B. commutatus. He 

 found that the mildew on B. racemosus failed to infect B. commutatus 

 (0/12),^'' while it readily infected B. hordeaceus (34/34). Furthermore 

 conidia from B. commutatus failed to infect B. racemosus (0/36), while 

 infecting B. hordeaceus (40/49). Salmon supposes that B. hordeaceus 

 may act as a bridge for the mildew on B. racemosus to pass over to 

 B. commutatus and also the reverse. In one case, Salmon infected 

 B. hordeaceus with conidia from B. racemosus. The conidia produced 

 on the former were then used to infect B. commutatus. Salmon, how- 

 ever, did not test the infecting capacity of the conidia thus produced 

 on B. commutatus. 



Steiner (150), in his work with the mildew {Sphaerotheca Humuli 

 (DC.) Burrill) of Alchemilla, reports the occurrence of bridging hosts. 

 He states that conidia from Alchemilla connivens and A. pubescens 

 readily infected A. pastoralis and A. impexa but gave negative results 

 when sown on A. micans. On the other hand, A. micans is readily 

 infected with conidia from A. pastoralis and A. impexa. Accordingly 

 Steiner supposes that A. pastoralis and A. impexa may carry the mil- 

 dew over to A. micans from A. connivens and A. pubescens. Some 

 evidence is also given to indicate that A. impexa is a bridging host 

 between A. nitida and A. fallax. Steiner's conclusions, however, are 

 based on only a few tests. 



In the Erysiphaceae the question has been raised whether the 

 ascospores and conidia from a particular plant possess the same in- 

 fecting capacities. Marchal (98), Salmon (124, 132), and Voglino 

 (166) have, in a few cases, used ascospores for inoculation tests and 

 they report that the capacity of ascospores for infection is identical 

 with that of conidia from the same host. 



It is, however, somewhat surprising that the evidence that various 

 races of parasites may increase their range by means of bridging hosts 

 is so very meager, if such really occurs. In those cases where the 

 suggestion of bridging has been most emphasized one is not impressed 

 with the data supplied. In fact, in all such cases the races of the 

 parasite are not sharply limited in their host range. They may infect 

 some hosts more readily and more vigorously than others, but the 

 virulence of the parasite does not seem to be increased or decreased 

 by developing on congenial or uncongenial hosts. 



^^ The denominator of the fraction indicates the number of leaves inoculated 

 and the numerator the number infected. 



