Connection between Host and Parasite 17 



as well as for several of those described by later American 



F 



authors. 



The further reduction of species that has been made in the 

 present work may be ascribed mainly to two causes ; (i) The ne- 

 cessity of adopting a wide view of a species when dealing with 

 material from all parts of the world ; and (2) The adoption of the 



m 



principle that classification must be based primarily on morpholog- 

 ical characters. 



As a necessary consequence of adherence to the latter view, 

 the connection of a-parasitic fungus with a certain host-plant can- 

 not be considered as affording a character of specific importance. 

 If it were to be proved that two forms of a fungus, morpholog- . 

 ically indistinguishable from one another, occurring on different 

 species of host plants, were incapable of infecting any but their 

 respective hosts, then these two forms might be classified as 

 ''biological " varieties or species, and in such a case the connec- 

 tion between the parasite and its host would become a character 

 of systematic importance. No such evidence, however, is at 

 present adduced in connection with the members of the Erj'si- 

 phaccae. For the present, therefore, to mention one example, 

 such a species as SpJiacrothcca epilobii, supposed to be confined 

 to species of Epilobiitm, but morphologically indistinguishable 

 from certain forms of the polymorphic 5. humuH, cannot be main- 

 tained. 



The question of the systematic value of the connection be- 

 tween host and parasite Is one of great importance. The connec- 

 tion is frequently treated as affording a character of primary spe- 

 cific importance, and it \\\\\, perhaps, not be out of place here to 

 utter a protest. against this practice. The method of classification 

 commonly followed in dealing with parasitic fungi is well illus- 

 trated in the treatment the genus Phoma has received. Over 

 1,000 species are enumerated in Saccardo's ''Sylloge," and the 

 majority^ of these are named after the host-plant on which they oc- 

 cur. That the connection of parasite and host is here made the 

 character of primary importance can be seen from the treatment of 

 the genus given in local floras. Allescher in enumerating 570 

 species of Phoma, in Rabenhorst's Krj^ptogamen Flora Deutsch- 

 lands, has adopted an arrangement thus described : '' Die Arten der 



