Saccharum.\ 
CL VII. GRAMINE.ZE (Stapf). 
97 
branchlets unequal, divided again, glabrous except at the nodes. 
Racemes sessile or the lower more or less peduncled, up to 8 lin. long, 
narrow to oblong ; joints and pedicels filiform, long-ciliate, with 
thickened tips, the latter shorter than the joints. Sessile spikelet 
lanceolate, acuminate, 2-3 lin. long, surrounded by hairs of the same 
length, springing from the callus. Glumes subequal, membranous, 
the lower acuminate, minutely 2-toothed, 2-keeled, with or without 
1 or 2 short intracarinal nerves, back smooth and glabrous, keels 
scabrid ; the upper lanceolate, acuminate, minutely mucronate, 
1 -nerved, margins hyaline, ciliolate, back glabrous or sometimes 
with a few hairs, keels scabrid. Valve of lower floret lanceolate, 
acute, 1J lin. long, hyaline, 1- or obscurely 3-nerved, ciliate, of 
upper floret ovate-lanceolate, 1-1J lin. long, hyaline, 3-nerved, 
glabrous or ciliolate, with a bristle-like awn about as long as the 
spikelet. Valvule oblong-lanceolate, glabrous, hyaline, up to 1 lin. 
long. Lodicules glabrous. Anthers up to 1J lin. long. Stigmas 
purplish, up to 1 lin. long. Pedicelled spikelet like the sessile, 
but both glumes 3-nerved and hairy. — Sibth. & Sm. FI. Graeca, 
t. 52 ; Reichb. Ic. FI. Germ. fig. 1505. Erianthus Ravennce, 
Beauv. Agrost. 162 ; Roem. & Schult. Syst. ii. 323 ; Kunth, Enum. 
i. 479 ; Hack, in DC. Monogr. Phan. vi. 139 ; Durand & Schinz, 
Consp. FI. Afr. v. 695 ; Hook. f. FI. Brit. Ind. vii. 121 ; Chiov. in 
Ann. Istit. Bot. Roma, vi. 164, and vii. 58 ; Stapf in Kew Bulletin, 
1907, 208 ; Nees, Gen. FI. Germ. t. 90. Andropogon Ravennce , 
Linn. Sp. PI. ed. ii. 1481 ; Host, Gram. Aust. iii. t. 1. Ripidium 
Ravennce , Trin. Fund. 169. 
Nile Land. British Somaliland : between Hargaisa and Berbera, Robecchi 
Bricchetti ! Plains of Seek, north-east of Hargaisa, Riva, 49 ! between .Derra 
Godeli and Bernera, Riva, 19 ! Issituggan and generally along the margins of 
river beds, Drake- Brockman, 4 ! 35 ! 
Mediterranean countries to Turkestan and N.W. India. 
The Somaliland specimens quoted here were referred by Dr. Chiovenda and 
myself to Hackel’s subvariety purpurascens (Hack. l.c. 140: Erianthus pur- 
purascens, Anders, in Of vers K. Vet. Acad. Stockh. 1855, 161) which is supposed 
to differ from the original in having smaller spikelets with purple or violet 
glumes, greyish hairs, a less acuminate lower glume and usually ciliate upper 
valve. Although the European specimens have a fairly uniform facies, those 
of the Asiatic material vary so much in the characters mentioned that a differ- 
entiation of subvarieties may for the present be neglected. I would only state 
that the Somali specimens seen by me match completely with some of the 
Indian examples. 
26. EULALIA, Kunth, Rev. Gram. i. 160. 
Spikelets all alike or nearly so, 2-nate, one sessile, the other pedi- 
celled, on the articulate fragile rhachis of 2-nate, digitate or fascicled 
spike-like racemes, the pedicelled falling from their pedicels, the 
sessile deciduous together with the contiguous joint of the rhachis 
and the pedicel. Florets typically 2 ; lower reduced to an empty 
FL. TROP. AFR. VOL. IX. H 
A 
'He 
A. 
