REED: SPECIALIZATION OF PARASITIC FUNGI 385 
Salmon (123, 126, 127) has made extensive tests with the mildew 
on the brome grasses using conidia from a number of different species. 
Some of Salmon’s more important results are indicated in the following 
table: 
Plants Used as Hosts 
Source of Conidia 25 ae 25 <8 22 = § =:2 aN = = 
Seo RS SS | Se Re oe ke le x38 ss 
S8 S8 S& S$ | ss] 8 = S38 S38 Sk 
ANTS as aS AN Seiad FBGA ae as aS 
as ae} ae: | iS Qe | ia Q 2 5 ie Sis 
a re Sr ra ie (lege 
1. Bromus interruptus..... 3/3" |48/48|10/10| 0/42| 0/29 16/16 | 0/20) 0/25 | 11/26! 
2. Bromus hordeaceus var. | | | | 
DIOOKESCENS 2 2 a. ae | 33/33|34/34 38/47| 0/28, 27/27 | 0/13| 9/22" | 10/16 
3. Bromus commutatus... . | 0/7.| 0/9 |21/21| 0/7 | 4/8! | o/7 | 6/7 6/7 
4. Bromus arvensis....... | 0/9 8/8 | | 
5. Bromus tectorum....... | | | 8/8 | 7/8 
From his data, Salmon concluded that four, or possibly even five, 
specialized races exist within this genus. In looking over his data, 
however, there seems to be very little, if any, difference between some 
of them. The races on B. interruptus and B. hordeaceus differ only 
in their capacity for infecting B. commutatus, the mildew on B. horde- 
aceus infecting this host, while that on B. interruptus does not. The 
race on B. commutatus differs from that on B. hordeaceus in not in- 
fecting B. mollis and B. interruptus. The race on B. arvensis infects 
this same species but not B. mollis. Finally the race on B. tectorum 
differs from that on B. hordeaceus in being able to infect B. sterilis. 
It is evident that these races are not distinctly marked off from one 
another. 
But little work has been done using ascospores from various hosts. 
Marchal (98) mentions the following results: (1) ascospores from 
Hordeum vulgare infected H. vulgare, H. distichon, H. trifurcatum and 
H. Zeocriton, but not Avena sativa, Secale cereale nor Triticum vulgare; 
(2) ascospores from Secale cereale infected S. cereale, but not Hordeum 
vulgare nor Triticum vulgare; (3) ascospores from Triticum vulgare 
infected 7. vulgare but not Agropyron caninum, Avena sativa, Hordeum 
vulgare nor Secale cereale. Salmon (124, 132) found that ascospores 
from Hordeum vulgare infected H. vulgare, H. trifurcatum and H. 
Zeocriton, but not H. bulbosum, H. jubatum, H. maritimum, H. secali- 
num, Avena sativa, Secale cereale nor Triticum vulgare; ascospores from 
Bromus commutatus infected B. commutatus and B. hordeaceus, but not 
10 The denominator indicates the number of leaves inoculated, the numerator 
indicates the number infected. 
1! Subinfection. 
