July 1906] Better Classification of the Uredinales 
153 
duced spores on living leaves, as in the case of Puccinia Poarum 
and P. rubigo-vera , or as stray spores no longer connected with 
a living host, as in P. Sorghi and P. graminis. In such cases the 
uredinia are usually followed each season by a greater or less 
development of telia, which serve little or no purpose in the 
propagation of the species, as the proper host plants for the 
aecia may be rare or absent. Such a condition explains the 
great prevalence of such species as Coleosporium Solidaginis 
and Melampsora Medusae when suitable coniferous hosts do not 
occur within hundreds of miles, and their aecia are rarely or 
never collected. Again in warmer regions a species is main¬ 
tained through its urediniospores alone, the other spore-forms 
rarely or never being seen. But these are not instances of ab¬ 
breviated life-cycle within the meaning here implied. They are 
a form of extended conidial propagation, the full life-cycle, when¬ 
ever circumstances permit it to be completed, showing all spore- 
forms. In other cases the completed life-cycle may show less 
than the full number of spore-forms, as in Puccinia umbellP 
ferarum, where the aecia are wanting, etc. 
While every genus heading a large section of the Uredinales, 
like Coleosporium, Melamspora,, Cronartium , Ravenelia , Phrag- 
midium, etc., is theoretically capable of division into four genera 
in accordance with the extent of the life-cycle, yet forms are 
not known in all cases to permit of such a division, and no uni¬ 
formity exists in regard to the proportional number of species 
falling into each of the newly delimited genera. Moreover, 
in many cases other characters demand recognition, and alto¬ 
gether it will be found that the admission of the life-cycle as a 
generic character does not result in a mathematical regularity 
of genera, throughout the order, as at first sight might be as¬ 
sumed. 
If we require that a genus should represent as fully as 
possible a group of organisms giving evidence of having been 
derived from the same ancestors, and therefore with species 
more closely related genetically to one another than to those of 
any other genus, it becomes necessary to explain a well known 
parallelism, brought to our attention by Fischer of Switzerland. 
He showed that in many cases the teliospores of a species having 
an extremely abbreviated life-cycle, e. g. Puccinia Leucanthemi, 
closely resemble in structure those of an autoecious species, e. g. 
P. Aecidii-Leucanthemi, in which the host of its aecia is the same 
or practically so as the host of the abbreviated species. Tranz- 
schel has successfully applied this rule of parallelism in predict¬ 
ing the host of the unrecognized aecia in certain heteroecious 
species. In such cases of parallelism there can be no doubt that 
the forms in question have truly descended from a common an¬ 
cestor, but dating a long way back, even to the early days when 
all the rusts had four spore-forms. Searching for an adequate 
