1905] HOLM—MUNROA SQUARROSA 125 
almost equal in the lateral spikelets (fg. 3); the flowering glume is 
pubescent in both, but in the median there are three distinct nerves 
with a tuft of long hairs at the base (figs. 2, 5, and 6), while in the 
lateral the nerves are not visible, and no long hairs are developed 
(figs. 3 and 7); moreover, the apex of the flowering glume is long- 
awned and distinctly bidentate in the median (jigs. 5 and 6), but merely 
emarginate and short-awned in the lateral spikelets (jig. 7). Some 
distinction in regard to texture is also observable, the flowering glume 
1G. 1.—Munroa squarrosa.—The inflorescence, magnified. Fic. 2.—A terminal 
spikelet. Frc. 3-—A lateral spikelet. Frc. 4.—The empty glumes of a terminal 
sR Fic. 5.—Flowering glume of same, dorsal view. Fic. 6.—Same, side view. 
i ee Fiaaigaiaiag glume of lateral spikelet, side view. Fic. 8.—The prophyllon, 
sit 
and the palet being membranaceous in the median, but quite thick 
and coriaceous in the lateral spikelets. It may seem justifiable, 
therefore, to consider the spikelets of Munroa as being dimorphic, 
even if we have not been able to observe any deviation in regard to 
the distribution of the sexes within them. 
While Torrey’s description of the spikelets contains a good deal 
more than is recorded by subsequent writers, except BENTHAM and 
Hooker, there are some other points which may be added to the 
diagnosis of this singular grass. The mode of growth is very well 
