582 



CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 



wide, ovoid in outline, the main axis pilose, the numerous delicate but rather stiff 

 branches ascondinj^ at a uniform angle (in herbarium specimens the branches con- 

 spicuously parallel), the numerous but not crowded 

 spikelets recurved at right angles on capillary pedi- 

 cels; spikelets 1.2 mm. long, about 0.7 mm. wide, 

 strongly plano-convex, turgid; first glume nearly as 

 long as the spikelet, 3-nerved, narrow, covering the 

 middle internerves of the sterile lemma, thin in tex- 

 tmre and so closely appressed as to be usually invisi- 

 ble, sparsely hispid; second glume inflated, gibbous, 

 pointed, 5-nerved, papillose-hispid, at maturity sub- 

 indurate, the hairs stiffening and a^ much as 1 mm. 

 long; sterile lemma equaling the second glume and 

 inclosing a 2-keeled palea, 5-nerv'ed, glabrous, the two 

 middle internerv^es thin; fruit 1 mm. long, 0.6 mm. 

 wide, very turgidly plano-convex, at first white, at 

 maturity brown, sparsely sprinkled with minute glob- 

 ular hairs. 



The spikelets of this unique species at maturity look like tiny burs or, as Lamarck 

 says, like the "seeds of Caucus." The bristly second glume sometimes falls, leaving 

 the turgid fruit, together with the first glume and sterile lemma, attached to the 

 pedicel. 



DISTRIBUTION. 



Damp shady places, Trinidad to 

 Brazil. 



Trinidad: Arima, jff lic^coc^ 10310. 

 P i a r c Savanna, HitchcocJc 

 10363. St. Joseph, Hitchcock 

 10177. Port of Spain, Hitch- 

 cock 10320. Without locality, 

 Bot. Gard. Herb. 3194. 



Fig. 148 — P. Mrtum. From type 

 specimen. 



Fig. 149.— Distribution of P. Mrtum. 



> 



EXCLUDED SPECIES. 



The study of gener'^ allied to Panicum and the examination of a 

 large number of type specimens has resulted in the identification of 

 most of the species of North America included now or in the past by 

 some authors in Panicum. Besides the valid species and the names 

 accounted for in synonymy within the genus Panicum there are a 

 great many names that, according to our present conception of the 

 genera of Paniceae, are referable to other genera. While the hst is 

 not complete it is so nearly complete as to enable us to account for 

 almost aU the names of North American species t];iat have been re- 

 ferred to Panicum. The second name in the column merely indicates 

 the genus to which the species belongs. In no case is any name in 

 the following hst a transfer of a species or a new combination. In 

 many cases the name, if transferred to the genus indicated, would 

 be untenable. 





,7.6, 



