Relation between Host and Parasite 21 



tion of the chief herbaria. I have been enabled to observe the oc- 



currence of species of the ErysipJiaccac on 1002 different species of 

 host-plants. "*" To make the list, as far as possible, complete, I have' 

 incorporated '^(>J additional host-plants recorded in the literature 

 of the subject, and have also noted all those cases in which a dif- 

 ferent species of mildew to that personally observed on any host- 

 species has been recorded by authors. In both the latter cases 

 the plants are distinguished in the host-index by an asterisk, 

 and the, authority for each record is indicated in the list of host- 

 plants given under the respective species of fungus. The names 



1 



of the host -plants have been, so far as possible, brought into ac- 

 cordance with those of the Index Kewensis. 



The restriction of the parasite in its choice of host-plants varies 

 greatly in the species of the ErysipJiaceae. Many species, e. g,, 

 ErysipJie tortilis on Cormts sangiiinca^ Uncinula goiiculata on Morns 

 rubra, and PodospJiacra biiincinata on Hauiaindis Virginiana ]\a\'e 

 a more or less wide geographic distribution, and yet, so far as at 

 present known, are absolutely confined to their respective host- 

 plant. About as many instances occur in w^hich the hosts of the 

 fungus are limited to one genus ; e. g.^ Unciiuda aceris, its van 



Tulasnei, and U. circinata are confined to species of Acer ; U. 

 flexuosa to species of ^scnlus, and SpJiaerothcca lancstris to species 

 of Qucrcus. A few species are limited in their choice of host- 

 plants to certain families, e, g,, Uficinula salicis on Salicaccae^ 

 SpJiaerothcca pannosa on Rosaceae^ Erysiphe graminis on the 

 Gramhicae, Finally, many species occur on a great number of 



A word may be given here on tlie right manner of usinj the host-index, Rightly- 

 used, it affords, when the name of the host-plant is known, a useful clue toward iden- 

 tifying an unknown species, and in a certain sense serves as a key to all the species of 

 the Erysiphaceae^ wrongly used, it may easily be the means of perpetuating errors. To 

 illustrate, if there be but one species of fungus recorded on the host-plant in hand, it is 

 then necessary merely to turn to the description of the fungus and see if the specimens 

 to be named agree in the characters given. The practice of naming a fungus from a 

 host-index, without examination, cannot be too strongly condemned. If several spe- 

 cies have been recorded on the host, the description of each must be consulted until the 

 fungus is identified. It is, perhaps, not unnecessary to add that if the name of the 

 host -plant in hand is not to be found among those of the host-index, it is not to be as- 

 sumed that this gives any reason for supposing that the fungus is a new species, as the 

 host Hst is necessarily very incomplete ; in such cases the keys to the genera and spe- 

 cies must be consulted. 



