172 NATURAL HISTORY BULLETIN. 



the UredinecB makes the accurate determination of the 

 host of the utmost importance : sometimes it is only 

 necessary to know the genus, but more often the species 

 must be unquestionably established in order to give full 

 value to a specimen. But it is not enough that the name 

 of the host should be written on the packet to make the 

 specimen of greatest worth: there should be within the 

 packet such fragments of the host, bits of inflorescence, 

 fruits, leaves, stipules, or other parts not too bulky, and 

 that may or may not be affected by the rust, as may 

 furnish characters for independent verification, or strong 

 presumptive evidence. Moreover, the limits of species and 

 genera among flowering plants are not unchangeably fixed, 

 and it sometimes becomes necessary to revise the nomen- 

 clature of the host as well as of the rust, when such 

 characteristic fragments become very important. The 

 more difficult the host the more imperative the need of 

 such helps. Among grasses and sedges, especially, we 

 have often had occasion to deplore their absence. The 

 ligules of grasses should in particular be included, and 

 bits of their inflorescence whenever possible. 



We have experienced much difficulty in securing good 

 material of the uredo stages for our distribution, and in 

 some cases even for purposes of study. The idea that 

 diagnostic characters are lacking in the uredo seems to be 

 so widely and firmly established, that collectors do not 

 search for this stage of a rust, and rarely collect it even 

 when conspicuous. The idea is erroneous, however, 

 although it seems to have support from the customary 

 omission, or subordination, of such characters in technical 

 descriptions. Sometimes the uredosori are much more 

 distinctive of the species than the teleutosori, and rarely 

 less so than are the aecidia. Specimens showing the best 

 development of the spermogonia, aecidia, uredosori, and 

 teleutosori should in most cases be independently collected 

 for each species. The finest specimens are those which 

 show one of these stages in optimum development, and 



