AXILLARY CLEISTOGENES IN SOME AMERICAN GRASSES 255 
the spikelets borne on the terminal panicle) of the same plant. Often, 
if their source were unknown, they would not be placed in the same 
tribe. The characters that are common to all are simplified structure 
and enlarged grain. In Triplasis the prophyllum is enlarged and 
indurate, and enfolds the entire spikelet; in Danthonia it is thin in 
texture, is split to the base, and simulates a pair of narrow glumes. 
(Repeated dissections have been made to decide whether these organs 
are parts of a prophyllum or a pair of glumes. In the very few imma- 
ture spikelets found they are evidently prophylla, but immature 
spikelets are very difficult to find. The keels are ciliolate as are the 
prophylla, instead of glabrous as are the glumes of the chasmogenes.) 
The type of spikelet characteristic of the genus Danthonia is shown 
in the sketch of Danthonia spicata, our commonest species (fig. 2). 
The two long glumes exceeding the several crowded florets, the short 
rachilla joints, and the 2-toothed lemma bearing a flattened awn tightly 
twisted below, are the distinguishing characters of the genus. The 
cleistogenes are all without glumes. In Danthonia spicata there is 
but one floret, sometimes with a slender rachilla joint bearing a minute 
rudiment of a second floret. The lemma is not toothed; it is usually 
merely pointed, but a few have been found with the point slightly 
lengthened, flattened, and somewhat twisted. (Note the relative 
size and shape of the grains). In some of the other species, especially 
the western Danthonia intermedia, D. americana, and D. californica, 
there are commonly two, three, or four widely separated florets, with 
slender rachilla joints almost as long as the florets, forming a striking 
contrast to the crowded florets characteristic of the genus. The 
lemmas are entire and awnless, or awn-tipped as in D. spicata. The 
genus Danthonia comprises something over 100 known species, 
Africa being the home of more than half of them. All the material 
of the genus in the National Herbarium was examined. (The exami- 
nation, because it necessitates some injury to the specimens, was by 
no means thorough.) The one Mexican and two West Indian species 
show no sign of cleistogenes; their wiry junciform habit would not lead 
one to expect them. Of the South American species, four, Danthonia 
chilensis, D. cirrata, D. montevidensis , and D. picta, were found with 
cleistogenes. But only one species from the eastern hemisphere, 
Danthonia semiannularis of New Zealand, was found to produce them. 
From our own species it would appear that these cleistogenes are 
produced after the maturity of the panicled spikelets. In Danthonia 
