380 BROOKLYN BOTANIC GARDEN MEMOIRS 
Ochropsora Ariae (Fuck.) Syd., Puccinia graminis Pers., P. 
Caricis (Schum.) Rebent., P. Polygoni-amphibit Pers., Uro- 
myces caryophyllinus (Schrank) Winter, and Uromyces Pisi 
(Pers.) Winter. 
Certain other rusts might also be looked for in this group as 
Puccinia coronata (Corda) Kleb., P. loliit Niels. (P. coronifera Kleb.) 
and P. bromina Eriks. 
2. Heteroecious rusts in which the uredo and teleuto host (or hosts) 
is common to several races which occur on the aecidial hosts: 
Gymnosporangium tremelloides Hartig, Melampsora populina 
Lév., M. Tremulae Tul., Puccinia Bistortae (Str.) DC., P. 
mammillata Schroet., P. sessilis Schneid., P. silvatica Schroet., 
Uromyces Dactylidis Otth, U. Scirpi (Cast.) Burr. 
3. Heteroecious rusts in which the specialized races are recognized by 
their selection of both aecidial and uredo and teleuto hosts: 
Puccinia Centaureae-Caricis Tranz., P. coronata Corda, P. 
dispersa Eriks. and Henn., P. extensicola Plowr., P. glumarum 
(Schm.) Eriks. and Henn., P. Ribesii-Caricis Kleb., P. Stipina 
Tranz., Uromyces Poae Rebent. 
4. Autoecious rusts and those whose life history is incompletely 
known. The following belong in this group: Melampsora 
Euphorbiae (Schub.) Cast., M. Euphorbiae-dulcis Otth, Phrag- 
midium disciflorum (Tode.) James, Puccinia Absinth DC., 
P. bullata (Pers.) Winter., P. Carduorum Jacky, P. Centaureae 
Mart., P. Chaerophylli Purt., P. Epilobii-tetragoni (DC.) 
Winter., P. Heliantht Schw., P. Hieracit (Schum.) Mart., 
P. Leontodontis Jacky, P. Petroselini (DC.) Lindr., P. Pulsatillae 
Kalchbr., P. Ribis DC., Uromyces Fabae (Pers.) de Bary, U. 
proeminens (DC.) Lév. 
POWDERY MILDEWS—ERYSIPHACEAE 
The first work carried on to determine whether a host special- 
ization occurs among the species of the powdery mildews was that 
of Neger (107), the results of which were published in 1902. Since 
then a number of workers have contributed to the evidence for special- 
ized races in this well-defined group of parasites. In fact, at the pres- 
ent time, one or more species of five of the six genera of the Ery- 
siphaceae have been tested. In most cases, however, the data are 
very meager and it is not possible to draw any definite conclusions. 
In a few cases, notably for Erysiphe graminis and E. cichoracearum, 
the facts are better established. 
Erysiphe cichoracearum DC. The host relations of this species 
was first reported upon by Neger (107) who obtained the following 
