198 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 



652; Pringle 3133. Guatemala: Cook & Griggs 691. Costa Rica: Pittier 2407, 11246; 

 BioMey 7469; Tonduz 4867. Brazil: Commis. Geogr. S. Paulo 2800. 



In the collection of Haenke at the herbarium of the German University at Prague 

 there are, under Thysanachne peruviana, two specimens. One is accompanied by the 

 label, "Peruan. mont. guanoc. Hanke." This specimen corresponds to Presl's 

 description of this species and agrees with a duplicate in the Bernhardi Herbarium 

 at the Missouri Botanical Garden figured by Scribner.a The other specimen is 

 A. martinicensis Trin. 



17. TRISCENIA Griseb. Mem. Amer. Acad. n. ser. 8: 534. 1862. 



1. Trisc^nia ovina Griseb. Mem. Amer. Acad. n. ser. 8: 534. 1862. 



Banks of creeks, May 28. Wright 756. 



The Grisebach specimen is from eastern Cuba in 1859, no. 756. 



Grisebach b cites this number also under Isachne leersioides. This appears to be an 

 error, as in Grisebach 's herbarium this number occurs only with Triscenia ovina. 

 Wright's 756 in the Gray Herbarium is from "banks of Pinal Creek in small dense 

 tufts, Monte Verde, Aug. 10, 1859." 



18. REIMAROCHLOA gen. nov. 



Spikelets lanceolate, acuminate, solitary, subsessile along one side of a flattened 

 narrow rachis (the back of the fertile lemma turned toward it), forming few to several 

 slender racemes, approximate at the summit of the culm, spreading or reflexed at 

 maturity; glumes obsolete except in the terminal spikelet in which one glume is 

 frequently present; sterile lemma exceeding the fruit; fertile lemma scarcely indu- 

 rated, faintly nerved, long-acuminate, inrolled at the base only, the palea of similar 

 texture, free nearly half its length. 



Perennials of the tropics and Bubtropics of the Western Hemisphere. 



The genus Eteimaria as established by Fliigge on the single differentiating character 

 "uniglumis," included three species, the first two of which, R. Candida and R. elegans, 

 differ from Paspalum only in having spikelets without glumes, a character which is 

 unreliable in this group of Paniceae. The third species, R. acuta, together with 

 those added to Reimaria by later authors, constitutes a distinct genus distinguished 

 by the characters in the diagnosis above. Considering thai Fliigge's three spe- 

 cies are not congeneric, bu1 that the lirst two on the one hand and the third on the 

 other must be separated, ii is necessary thai the name Reimaria go with the larger 

 group. c Reimaria then becomes a synonym of Paspalum, or if the glumeless species, 

 P. candidum H. B. K., P. pulchellum II. 15. K.. P. elongatum Griseb., etc., be con- 

 sidered generically distinct, the name would apply to this group. Fori?, acuta and its 

 allied species the above name is proposed with Reimaria acuta Fliigge as the type: 

 Reimarochloa acuta (Fliigge). Paspalum vaginatum Sw. and P. distichum L. (in 



which both glumes ar icasionally present | in habit and texture of the acute fruits 



show a closer affinity to this genus than do the glumeless species mentioned above. 



Spikelets aboul 2 mm. lout;; J. A', brasiliensis. 



Spikelets about 5 mm. long 2. U. oligostachya. 



1. Reimarochloa brasiliensis (Spreng.). 



Agrostis brasiliensis Spreng. Nov. Prov. Hal. 45. 1819. 



Reimaria brasiliensis Schlecht. Bot. Zeit. 10: 17. IS52. 



Panicum oxyanthum Stetid. Syn. PI. Glum. 1: 41. 1854. 



Wright 3437; Isle of Tines. Curtiss497. 



Grisebach's specimen of this number is from '-savannas. Eanabana, May 22." 

 Another specimen in his herbarium from "low wet ground around ponds. Hanabana" 



a Rep. Mo. Bot. Gard. 10: pi. 6. 189"). 

 M' :l i. PI. Cub. 234. 1866. 

 American Code, Canon 15 (Bull. Torr. Club 31: L75. 1904); Vienna Code, Art. 45. 



