178 



NATURAL HISTORY BULLETIN. 



(see fig. 5), and the host determined as Pan i aim virgatum . 

 A collection made from the type 

 locality on May 21, of the same year, 

 was distributed in Ellis & Everhart's 

 N. Am. Fungi, as No. 2888. This is 

 teleutosporic material, as would be 

 natural in a winter collection, but a 

 little search reveals the presence of 

 uredospores and paraphyses, and these, 

 together with the teleutospores, agree 

 exactly with those of the type collec- 

 Fig. 4- tion of Pncci7iia clavispora Ellis and 



I'REDO ALABAMENSIS Diet. ,-a , i c . , , -, 



From type collected at Au- Everhart, and of the two types already 

 bum, Aia., oct. i 4 , 1891. mentioned. The leaves of Panicum 

 virgatum and of Chrysopogon avcnaccus can not be certainly 

 told apart by any characters derived from the color, 

 texture or surface markings of the blades, but on the other 



hand the ligules are wholly 

 distinct, those of the Paniaim 

 being prominent and fringed 

 with long, white, silky hairs, 

 and those of the Chrysopogon 

 being still larger and entire. 

 Diagnoses drawn from both 

 rust and host show that these 

 two Kansas collections sup- 



pttccinia vnoATAjf. ,%E. From type posed to be on Paniaim are 

 collected in Rooks Co., Kans., jan. 23, 1892. identical with Pucciuia clavi- 



spora known to be on Chrysopogon, and that the host of 

 the type of the earlier name was not a Panicum but was 

 Chrysopogon avcnaccus. It appears that with the possible 

 exception of material found in the herbarium of the 

 Kansas Agricultural College no collection of the supposed 

 P. virgata on Panicum has been made since 1892, and it 

 was then only found "on one solitary tuft of grass about 

 two feet across" (see Bartholomew's "Kansas Uredinese" 

 in Trans. Kans. Acad. Sci. 16:183). 



