1905] HOLM—MUNROA SQUARROSA 135 
palets exhibit a structure so distinct that when detached the spikelets 
may be easily assorted into terminal or lateral by their tissues alone. 
In the terminal spikelet the flowering glume and the palet are thin 
and membranaceous, as in the majority of Gramineae. The flower- 
ing glume has four prominent ribs on the dorsal face supported by 
heavy groups of stereome, and small scattered groups of this same 
tissue are observable also on the ventral face of the glume, but only 
between the two innermost mestome bundles. Otherwise the glume 
consists only of a single layer throughout, represented by the dorsal 
epidermis. The palet of this same spikelet—the terminal—is two- 
nerved, but very slightly bicarinate, and the structure is even more 
delicate than that of the flowering glume. 
In the flowering glume of the lateral spikelets the structure is 
very different, a large mass of very thick-walled stereome covering 
nearly the whole ventral face of the glume, and a colorless parenchyma 
occupying the larger inner portion. The four mestome bundles are 
imbedded in this tissue, thus forming no prominent ribs, as is the 
case in the terminal spikelet. A similar broad stereomatic tissue is 
observable in the palet on the ventral face, besides a colorless tissue 
nearer the dorsal. The two nerves are located in this tissue, which 
extends into two sharp keels with numerous epidermal projections. 
So great a diversity in structure observable in the glumes of 
spikelets belonging to the same inflorescence is evidently very rare 
within the order. It seems the more peculiar since the spikelets are 
otherwise identical, the number of glumes being the same, and the 
flowers perfect in both the terminal and lateral spikelets. The very 
firm structure, described above, which is especially marked by the 
unusual development of stereome, is also known in other genera. 
We might mention, for instance, the coriaceous or cartilaginous 
empty glumes in MaypEar and ZoystEAE; the cartilaginous flower- 
ing glume and palet in PANICEAE and OryzEAE; finally the promi- 
nent structural deviation noticeable in the unisexual spikelets of 
Buchloe and Amphicarpum. But none of these cases are really 
comparable with that of Munroa, where the only distinction between 
the spikelets depends upon position, terminal or lateral, and not upon 
the distribution of sexes, and still exhibiting such marked modification 
in structure. 
