﻿284 BOTANICAL GAZETTE [april 



Fig. 2. — De- 

 tails of flower. 



leaves several, 3 or 4 inches long, involute, tapering to a sharp point, 

 upper surface rough, margin scabrous, otherwise smooth. Stem leaves 

 few, the upper larger and embracing the lower part of the panicle. 

 Inflorescence a panicle of spikes, a few inches to a foot long. Spikes 



r 



several, an inch or two long, distributed along the axis. Spikelets 



several, arranged in two rows along one side of a 

 flattened axis, i -flowered, appressed. Empty glumes 

 2, i-nerved, smooth, thin, and light colored, pointed 

 but not awned, lower about half as long as upper, 

 which is green on the margins. Flowering glume 

 about the length of the upper empty glume, obtuse 

 and awnless, 3-nerved, the lateral pair near the margin 

 and densely ciliate with long white hairs, also pubescent between 

 nerves. Palea nearly as long as flowering glume, 2 -keeled, the nerves 

 ciliate like the flowering glume. The base of the flower extends down 

 into a short, hairy callus. Stamens 3. Caryopsis oblong, smooth, 

 margined, and apiculate above. 



Distribution: Ennis and Beeville, Texas, /. G. Smithy 1897. 

 The specimens were found among the undetermined forms of Leptochloa 

 in the Herbarium of the U. S, Department of Agriculture, The plant from 



r 



Ennis may be considered the type. I have referred this species to Hackel's 

 genus Willkommia as it agrees with the description in all essential 

 particulars, and in generic characters with the specimen of Willkommia 

 annua which is in the Department Herbarium. In floral characters it differs 

 chiefly in the awnless flowering glume. The distribution of the genus is 

 peculiar, as the three previously described species occur in west South Africa. 



The genus Willkommia was described by Hackel, Verh. Bot. Ver. Bran- 

 denburg 30 : 145. 1888.. This name had already been applied to a group of 

 Composit^e (Sch. Bip. ex Nym. Consp. 357.1879), which is now generally 

 referred to Senecio. Hackel was aware of this, but, from the fact that the 

 earlier name was now a synonym, considered himself justified in taking up the 

 same name for his new genus of grasses. 



This genus is aUied to Schedonnardus Steud. and Craspedorachis Benth., 

 but is distinguished from the former by the broad, hyaline, unequal empty 

 glumes and the densely ciliate flowering glume and palea, and from the lat- 

 ter by the unequal empty glumes and the flowering glume as long as the 

 upper empty glumes. Craspedorhachis has equal empty glumes twice as 

 long as the flowering glume and palea. 



I am of the opinion that Willkommia should be referred to Craspedorha- 

 chis Benth. (Hook. Icon. PL HI. 4 : 58. pi. 1377. 1882), but as Hackel, who 

 has had the opportunity of examining all the African species, has kept the two 

 genera separate I have done the same. 



