124 BOTANICAL GAZETTE [FEBRUARY 
habit and the inflorescence are exceedingly well characterized. 
There is one point especially which has escaped the attention of 
later botanists, namely, that the spikelets are dimorphic in respect 
to the shape of the glumes. It is true that Torrey did not express 
this as plainly as he might have done, but his diagnosis certainly 
shows that he did observe the different structure of the spikelets, as 
shown by the following statement: “In the uppermost spikelet, and 
often in the middle one, these nerves are bearded with long white 
hairs towards the base; but the flowers of the lowest spikelet are 
usually quite naked.” That BENTHAM and Hooker considered this 
character of some importance is evident from their description of 
the flowering glumes as being “‘longe ciliatis v. in eodem fasciculo 
glabris.”’3 But none of the subsequent authors, not even the agros- 
tologists, appear to have noticed this point, and it so happens that 
the genus has been received as a plant of rather ordinary habit, dis- 
tinct from Monanthochloe by “Spikelets usually in threes, terminal 
in the axils of stiff spinescent leaves which project far beyond them.” 
Since Torrey called attention to this slight modification in struc- 
ture of the flowering glumes, it seemed natural to investigate the 
matter further, as it might lead to the discovery of additional charac 
ters sufficient to attribute a certain degree of dimorphism to the 
spikelets. A careful examination of a large number of specimens, 
studied in the field, revealed the fact that Munroa does exhibit such 
dimorphism constantly, and to a greater extent than hitherto observed. 
The deviation is not only due to the presence of long white hairs at 
the base of the flowering glume in the uppermost spikelet in contrast 
with the others, as first described, but the complete structure of these 
spikelets is quite prominently distinct. As a rule, each inflorescence 
(fig. 1) in Munroa consists mostly of three spikelets, one median and 
two lateral; and when only two are developed, one of them shows 
always the structure of a median and the other that of a lateral 
spikelet. Of the two spikelets here figured, fig. 2 represents the 
median and fig. 3 one of the two lateral. It is readily seen from 
the drawings that the structure of these spikelets is distinct. The 
empty glumes are very unequal (figs. 2 and 4) in the median, but 
3 Genera plantarum 3:1180. 1883. 
4 ScRIBNER and SouTHWoRTH, The true grasses. New York. 137. 1890- 
