THE UREDINALES FOUND UPON THE ONAGRACEAE 557 
is known, occurs upon Gymnosperms. The telial stage further suggests 
a more primitive condition, resembHng the fern rusts, which Dietel 
considers to be the oldest rusts. The genera Epilobium and Cha- 
maenerion, the hosts of these species of Pucciniastrum, might therefore 
be considered to be older genera of the Onagraceae. Other facts, as 
indicated elsewhere in this article, point to the same conclusion. It is 
further to be noted that these two hosts harbor more species of rust 
than any other genera of the Onagraceae, including two heteroecious 
species, two autoecious long-cycled species, one with and one without 
uredinia, and three short-cycled species. 
The second group is that including the long-cycled species Uromyces 
plumharius and Puccinia Epilohii-tetragoni, and also the heteroecious 
species with its aecia upon the Onagraceae, Puccinia Veratri, the 
European species without uredinia, Puccinia Epilobii-Fleischeri, and 
the short-cycled species Puccinia Epilohii and Puccinia scandica. 
The morphological resemblances between the several species is so 
close that it seems quite logical to infer a relationship. 
Uromyces plumharius and Puccinia Epilobti-tetragoni differ but 
little in morphological characters in all the spore forms, save for the 
occurrence of one-celled and two-celled teliospores respectively. In 
other respects, however, the correlation does not hold as it does between 
certain other parallel species of Uromyces and Puccinia, for here 
a different set of hosts is attacked by the two species in question, and 
the geographical range of Uromyces plumharius, while largely including 
that of Puccinia Epilohii-tetragoni, extends far beyond it, the Puccinia 
species being, in North America, wholly western, the Uromyces ex- 
tending over the greater part of the United States. It is a curious 
fact that, so far as is known, the same species of host is not attacked 
by the two rusts. Indeed, unless they may meet upon the genera 
Oenothera, different host genera are attacked, that is, different genera 
as now subdivided. Twelve genera of the Onagraceae are given as 
hosts for Puccinia Epilohii-tetragoni; and seven genera as hosts for 
Uromyces plumharius; while related genera obviously occur in the two 
sets, yet all are different. While deductions must be vague, this fact 
would seem at least to indicate that a rather definite and distinct 
specialization has arisen within these two species of rust. It might 
be inferred further that this specialization has occurred in a somewhat 
different way in each species; in the Puccinia, over a larger number of 
host genera, but limited geographically, in North America, to the 
