398 BROOKLYN BOTANIC GARDEN MEMOIRS 
would expect. Agrostis alba, however, seems to harbor a distinct race 
of rust and so the result recorded is not in line. 
Taking the results of aecidiospore inoculations as a whole, there 
seems to be no good reason for assuming that aecidiospores from the 
barberry, produced by teleutospores from a known grass, have any 
greater range of hosts than uredospores from the same grass. It 
appears that the racial strains of the black stem rust are not so sharply 
fixed in their host restrictions in either the uredo or aecidial stage. 
Further, the nature of the specialization is different in Europe from 
what it isin America. It is not surprising, then, that these races are 
able to grow on other hosts. There is, however, no clear indication 
that the barberry acts in any way as a bridging host and that it 
enables the races on different grasses to increase their range. 
The possibility of the aecidial host serving as a means for extending 
the host range of specialized races is quite apparent in Puccinia 
coronata Corda. Miihlethaler (102) records Rhamnus Imeretina as an 
aecidial host for specialized races of three of the main subgroups of 
the crown rust: Puccinia coronata (Corda )Kleb., P. coronifera Kleb. 
and P. alpinae coronata Mihlethaler. Rhamnus Purshiana is likewise 
an aecidial host for the races on two subgroups—P. coronata (Corda) 
Kleb. and P. alpinae coronata Miihlethaler. There is, however, no 
evidence at hand to indicate that, as a matter of fact, these species do, 
in any way, act as bridging hosts. 
There are many other cases where a particular species of plant is 
a host for two or more specialized races of a parasite and it might be 
possible for these to enable the different races to extend their host 
range. A few cases of this sort may be mentioned. According to 
Jaczewski (68) Agropyron repens and A. caninum are hosts for the 
races Tritici and Secalis of Puccinia graminis and this might serve to 
enable one race to pass over on to the hosts of the other. There is, 
however, no experimental proof in support of the suggestion. Stak- 
man and Piemeisel (149) record a number of hosts as common to several 
or all of the six races they studied. However, no bridging occurs, 
each race being distinct. According to Miihlethaler (102), Festuca 
elatior is a host for races Festucae and Loli of Puccinia coronata. 
Phalaris arundinacea is the only uredo and teleuto host for the special- 
ized races of Puccinia sessilis with their aecidial stages on Liliaceae, 
Orchidaceae, Amaryllidaceae and Araceae. Several species of Ribes 
are common aecidial hosts for the specialized races of Puccinia Ribesu- 
Caricis. Similar conditions are found among a large number of other 
forms—Uromyces Dactylidis, U. Poae, U. Fabae, U. Scirpi, Coleo- 
sporium Campanulae, Melampsora Larici-epitea, M. populina, M. 
Tremulae, etc. 
