372 BROOKLYN BOTANIC GARDEN MEMOIRS 



Puccinia Pulsatillae Kalchbr. Bubak (22) distinguishes the fol- 

 lowing races of this rust, basing the separation on the distribution and 

 character of the sorus, not on cultural experiments: 



1 . Concortica on Pulsatilla alpina and P. sulphurea. 



2. Atragenicola on Atragene alpina. 



3. Genuina on Anemone silvestris and Pulsatilla patens. 



4. Pulsatillarum on Pulsatilla pratensis and P. vulgaris. 



Puccinia Ribis DC. Eriksson (39) distinguishes a specialized race 

 Rubri, for he found that teleutospores from Ribes rubrum would infect 

 this species but not R. nigrum nor R. Grossularia. 



Uromyces alchimillae (Pers.) Lev. Fischer (53) has carried out a 

 few experiments with this rust which tend to show the existence of a 

 host specialization. The results reported, however, do not indicate 

 any close correspondence between the plants infected with a particular 

 collection of uredospores and the systematic grouping of the host 

 plants within the genus Alchimilla. Fischer found that uredospores 

 from hosts belonging to the section Vulgares infected plants belonging 

 to sections Pubescentes and Splendentes. 



Uromyces caryophyllinus (Schrank) Winter. Fischer (49-51) 

 gives experimental evidence for the occurrence of specialized races in 

 the carnation rust. The rust is heteroecious, the aecidial stage oc- 

 curring on Euphorbia Gerardiana and the uredo and teleuto stages on 

 various Caryophyllaceae. Fischer finds that aecidiospores from 

 Euphorbia Gerardiana in one locality are able to infect only Tunica 

 prolifera, while aecidiospores collected on the same host in another 

 region are able to infect only Saponaria ocymoides. In still other 

 localities, however, a race of rust is found that is able to infect both 

 Tunica and Saponaria. 



Uromyces Dactylidis Otth and U. Poae Rabh. These rusts, 

 probably indistinguishable by any well-defined structural character- 

 istics, are further alike in having their aecidial stage on species of 

 Ranunculus. The former develops its uredo and teleuto stages on 

 Dactylis glomerata, while the latter has the corresponding stages on 

 several species of Poa, Agrostis alba also being listed as a host. A large 

 number of workers have contributed to our knowledge concerning the 

 heteroecism of these rusts. 



Krieg (93, 94) and Klebahn (89, 91) have worked with Uromyces 

 Dactylidis from the standpoint of specialization and the following 

 races are indicated : 



1. Aecidial stage on Ranunculus bulbosiis and R. repens. 



2. Aecidial stage on Ranunculus lanuginosus, to a slight extent on 



R. bulbosus {Uromyces lanuginosi-dactylidis Kleb.). 



3. Aecidial stage on Ranunculus aconitif alius, R. alpestris, R. glacialis, 



and R. platanifolius {Uromyces platanifolii-dactylidis Krieg.). 



