48 CONTRIBUTIONS FEOM THE NATIONAL HERBARIUM. 



fifth the length of the spikelet, truncate or obtuse; second glume slightly shorter 

 than the sterile lemma, both 7-nerved, palea of the sterile floret obsolete; fruit 2.1 

 mm. long, 1 mm. wide, elliptic, apiculate. 



This very distinct species is known from a single collection only, that distributed by 

 Pringle, mentioned above. 



17. Panicum dichotomiflorum Michx. 



Panicum miliaceum Walt. Fl. Carol. 72. 1788, not L. 1753. Since Walter does not 

 give Linnaeus as authority nor use his diagnosis, this is evidently intended as a new 

 species. No specimen of this is found in Walter's herbarium,** but the description 

 indicates P. dichotomiflorum, which, together with Walter's name, Elliott & refers 

 to P. geniculatum Muhl. 



Panicum dichotomiflorum Michx. Fl. Bor. Amer. 1 : 48. 1803. " Hab. in occi- 

 dentalibus montium Alleghanis . " The type is in the herbarium of Drake de Castillo. 

 It was sent by Richard, having been collected by Michaux "ad occidentum montium 

 Alleghanis. " The specimen of this in the Michaux Herbarium is labeled "in regione 

 Illinoensium." Both of these specimens are the common glabrous form of the United 

 States as represented by Chase in Kneucker, Gram. Exs. no. 546. 



Panicum geniculatum Muhl. Cat. PI. 9. 1813. Based on P. dichotomiflorum Michx. 

 The specimen in the Muhlenberg Herbarium is in folio 181, marked "Panicum 

 geniculatum (dichotomiflorum) M. 114." 



Panicum aquaticum Poir. in Lam. Encycl. Suppl. 4: 281. 1816. "Cette plante 

 croit a Porto-Ricco ; elle m'a ete communiquee par M. Ledru." The type, in the 

 Cosson Herbarium, is from Porto Rico. In the description the species is compared 

 to P. melicarium Michx., and the label of the type bears the abbreviated statement* 

 "aff. P. melicario Mich." In the Desvaux Herbarium there is a similar specimen 

 from the Antilles, labeled "P. aquaticum Desv. in Poir. Enc. Suppl." In both 

 specimens the spikelets are nearly 3 mm. long as in Wright 3861. In the original 

 publication there is no indication that Desvaux is the author of the species. This 

 name was erroneously referred by Hitchcock c to P. elephantipes. The fruit is not 

 acuminate as in that species 



Panicum multiflorum Poir. in Lam. Encycl. Suppl. 4: 282. 1816. "Cette plante 

 croit a la Caroline; elle m'a ete communiquee par M. Bosc." We take the specimen 

 labeled "bosc. Caroline, " in the Cosson Herbarium, to be the type. Another specimen 

 of the same collection is in the Desfontaines Herbarium. It is labeled, "Am. Sept. 

 Bosc," and also "Panicum brachiatum Bosc." These are the typical form. 



Panicum brachiatum Bosc; Spreng. Syst. Veg. 1 : 321. 1825, not Poir. 1816. The 

 locality given by Sprengel is "Ins. Bemud." As indicated above, the specimen of 

 P. multiflorum from Bosc is labeled P. brachiatum Bosc. There is also in the Delessert 

 Herbarium a specimen so labeled, collected in South Carolina by Bosc. We have 

 seen no specimen of Bosc's from the Bermudas, the published locality, and we 

 find no record that Bosc visited the Bermudas. The meager description applies 

 to P. dichotomiflorum. 



Panicum chloroticum Nees; Trin. Gram. Pan. 236. 1826. Trinius describes a variety 

 " cc (agreste N. ab Es.)" which is the equivalent of the species, and "/? (sylvestre 

 N. ab Es.)," both from Brazil, "V. 'utriusque spp. Brasil (N. ab Esenb. Langs- 

 dorff)." The latter differs in having a more open panicle, larger spikelets, and 

 narrower leaves. Nees d described the same species later, with three varieties, 



« For an account of Walter's grasses see Hitchcock, Rep. Mo. Bot. Gard. 16 : 31-56. 

 1905. 



&Bot. S. C. &Ga. 1: 117. 1816. . 

 cContr. Nat. Herb. 12: 218. 1909. 

 dAgrost. Bras. 164. 1829. 



