HITCHCOCK—GRASSES OF CUBA. 197 
mann 629, 630; Langlassé 608; C. T. Smith 1892; Bottert 730, 731, 734; Palmer 1264, 
1920. Guatemala: Heyde & Lur 3907; Maron & Hay 3526; Rock in 1887; World’s 
Fair Commission in 1893. Salvador: Renson 207. Costa Rica: Tonduz 9211. 
2. Arundinella martinicensis Trin. Gram. Pan. 62. 1826. 
Arundinella pallida Nees, Agrost. Bras. 465. 1829. 
Thysanachne scoparia Presl, Rel. Haenk. 253. 1830. 
Wright 3478. 
The Grisebach specimen is from eastern Cuba in 1860 and numbered 113=3478.” 
Wright’s 3478 in the Gray Herbarium is from San Juan de Buenavista, Nov. 21. 
These were compared with the type of Nees at Munich and that of Trinius at St. 
Petersburg (from Martinique, Sieber 262). This species has an elongated dense panicle, 
more or less folded blades, the bent awn twisted below. It ranges from Cuba and 
Mexico to Brazil, and is represented in the National Herbarium by the following: 
Porto Rico: Heller 934, 4355, 6256; Sintenis 361, 5797; Barrett 101. Santo Domingo: 
Wright, Parry & Brummel 626. Jamaica: Eggers 3514. Mexico: Palmer 434, Costa 
Rica: Pittier 11005; Tonduz 3672. Brazil: Glaziow 17433; Regnell 1414 (IIT). 
Thysanachne scoparia Presl was published in the Symbolae Botanicae, the title 
page date of which is 1882 and which, hence, is later than the Reliquiae Haenkeanac. 
But the part containing the above species must have been published earlier, for in 
the latter work@ is cited ‘7. seoparia. Presl de thysanachne, 1829. cum icone.” 
There appears to be no such work by Presl except the portion of the Symbolae (pages 
1] and 12 and:pl. 6) where Thysanachne and 7’. scoparia are described as if they were 
there originally published. The species is based on Sieber 264 from Martinique. 
Presl’s type from Mexico was examined at Prague. . 
Pilger? refers the Porto Rico species to A. hispida (Willd.) Kuntze (Andropogon 
hispidus Willd.), to which he also refers 1. brasiliensis Raddi. I have not seen the 
type of either of the last two species, but in Trinius’s herbarium is the type of Gold- 
bachia mikani Trin., which is included by Pilger in the list of synonyris of 1. hispida, 
This I consider distinct from A. pallida, as did Nees, and it is what I take to be 
A. brasiliensis Raddi. This is also the A. brasiliensis of Hackelin Martius’s Flora Bra- 
siliensis, as indicated by plate 38 and by specimens so named received from Professor 
Hackel. The spikelets are smaller and the awn shorter, sometimes scarcely exserted. 
This species is represented in the National Herbarium by the following: Mexico: 
Liebmann 622, 635. Colombia: Pittier 1527. British Guiana: Mount Roraima Exped. 
954. Brazil: Glaziow 17921, 20567 a; Henschen in 1868, Dusén 3875. Uruguay: 
Arechavaleta. 
3. Arundinella peruviana (Presl) Steud. Syn. Pl. Glum. 1: Il, 1854. 
Thysanachne peruviana Presl, Rel. Haenk. 253. 1830. 
Arundinella cubensis Griseb. Mem. Amer. Acad. n. ser. 8: 533. 1862.4 
Arundinella crinita Trin. Linnaea 10: 299. 1836. 
Wright 1552. 
Sheaths and blades pilose, the latter narrow and more or less folded, usually less 
than 0.5 em. wide; panicle rather loose, not elongated as in A. martinicensis; awn 
slender and bowed back like a shepherd’s crook, but not twisted. The type of 
A. cubensis from eastern Cuba, no. 1552 in 1859, is in the Grisebach Herbarium. 
Another specimen also from eastern Cuba, 1860, is numbered “1 1o= 1552.”7 
The types above cited, namely, those of Presl at Prague, Grisebach at Géttingen, 
and Trinius at St. Petersburg, agree with each other and are well characterized by 
the shape of the awn. Additional specimens in the National Herbarium are as fol- 
lows: Mexico: Liebmann 621, 625, 632, 634; Bourgeau 1660, 2223; Palmer 12, 526, 
@ Rel. Haenk. 253. ¢ Agrost. Bras. 465, 1829. 
b In Urb. Symb. Antill. 4: 80. 1905. @ Pl. Wright. 2. 
