PAPERS OX BOTANY 177 



pycnidia of Rbabdospora possess thin-walled cells in> 

 stead of tliick-walled cells as in Septoria. Even this 

 separation seems artificial and insufficient for the two 

 genera. His definition of the genus Septoria, neverthe- 

 less, which appears to be more comprehensive and exact 

 than those quoted above, follows: ''Septoria umfasst 

 alle diejenigen Arten, deren Fruchtlager sich durch Aus- 

 bildung eiaer Decke in ehi pseudopyknidiales Gehause 

 umwandelt, das oben mehr oder wenig breit geoffnet ist. " 

 Phleospora, regarded as the only other very close rela- 

 tive of Septoria is stated by Diedecke to be synonomous 

 with Cylindrosporium. 



EELATIOX OF SPECIES TO HOST PLANTS 



Popular methods of naming Septoria species consist 

 of determinations of the generic characters and reliance 

 upon the host index in ascertaining whether the species 

 is new or old. The majority of the species of Septoria 

 have been reported from flowering plants and when we 

 consider that there are some 1,200 species of Septoria 

 and less than 280 families of flowering plants, it will be 

 seen that, if equally distributed, not less than four species 

 must have been described from every family of flower- 

 ing hosts in. existence. Actual facts, however, show that 

 from the more common families many more than tliis 

 number have been described while from the less common 

 families none at all are recorded. For instance: from 

 the family Compositae, there are listed 113 species ; from 

 the Gramineae, 86; Leguminosae, 50; Eosaceae, 53: Um- 

 beUif erae, 52 ; Labiateae, 33 ; CaryophyUaceae, 34 ; Cruci- 

 ferae, 22 and Solonaceae, 19.- Thus from nine families 

 of flowering plants there are listed 38 a- of the described 

 species of Septoria. 



It has been found from a comparative study of the 

 species that many reported from different genera of the 

 same family of hosts have the same spore length and 

 agree in aU other described morphological particulars. 

 From this it would seem as though the generic host limit 



- Species without spore measurements are not included in this sum- 

 mary. 



