Andropogon Gryllus, Linn. Sp. Plant. p. 1480; Gren. et Godr. Fl. de 
Fr. iii. 468; Woods, Tour. Fl. p. 395. 
(C.) Spec. Cuar.—Spikelets in pairs along a hairy axis, forming a 
pair of spikes, with long hairs below the point of junction. Sessile 
spikelet 2-flowered, lower outer glume many-nerved, coriaceous, hairy on 
back, wpper outer glume membranous, hairy above ; flowering glume of 
aborted flower (C 4) membranous, oblong-lanceolate ; lower flowering 
glume of hermaphrodite flower membranous, bifid, with long awn ; wpper 
flowering glume (C 5) minute, ovate, membranous, exaristate, about as 
long as the lodicules. Culm bearing many flower-branches, each sheathed 
in a leaf. 
Andropogon hirtum, Linn. Sp. Plant. p. 1482; Gren. et Godr. Fl. de 
Fr. iii. 469; Woods, Tour. Fl. p. 395. 
(D.) Spec. Cuar.—Differs from Andropogon hirtum, Linn., in its 
longer awns and spikes, and in having the stem glabrous below each 
pair of spikes. 
Andropogon pubescens, Vis. Pl. Rar. Dalm. p. 3; Gren. et Godr. Fl. 
de Fr. ii. 469; Woods, Tour. Fl. p. 395. 
(£.) Spec. Coar.—Spikelets in pairs, forming a digitate terminal pa- 
nicle of 4-8 spikes. Sessile spikelet 2-flowered, lower owter glume many- 
nerved, coriaceous, exaristate, wpper outer glume membranous, exaristate ; 
flowering glume of aborted flower (EK 3) membranous, exaristate ; lower 
flowering glume of hermaphrodite flower linear, tapering into a very 
long, rigid awn, upper flowering glume minute, only as long as the lodi- 
cules (E 5). Grain terete on back, embryo prominent on face. Culm 
simple or branched at the base. 
Andropogon Ischemum, Linn. Sp. Plant. p. 1480; Gren. et Godr. Fl. 
de Fr. iii. 465; Woods, Tour. Fl. p. 395. 
Haszitats.—(B.) From Kew Gardens, September 7, 1867. (A, C, 
D, E.) From Mentone, where I gathered them in the winter and spring 
of 1867. 
Remarxs.—There has been much discussion and some confusion 
about the structure of the hermaphrodite spikelets in Andropogon, some 
authors affirming and others denying the presence of a second flower or 
part of a second flower within each pair of outer glumes. There are dis- 
tinguished authorities on either side, and among those who failed to see 
any trace of a second flower was M. Parlatore himself. After a careful 
study, I find that the rudiment of a second flower is always present 
(in the species figured), but that the minute upper flowering glume is 
often wanting in some of the spikelets, even though fully developed in 
others on the same spike, and this reduces the number of parts in the 
deficient spikelets. According to the view taken here, we must regard 
the sessile spikelets of Andropogon (in the European species) as com- 
posed of two flowers, one of these being reduced to a single mem- 
branous flowering glume lying within and next to the lower outer 
glume. In some of the extra-European species this second flower is 
