138 Rhodora [AUGUsT 
"bright" or “dull.” It seems practically impossible to differentiate 
textures and the best that can be said of the coloring is that E. pilosa 
and E. peregrina seem to have a greater tendency to be somewhat 
“purplish” tinged than E. Purshii. It seems very doubtful if there 
is any real value in these points — certainly none to distinguish 
readily E. peregrina. 
A distinction, apparently not previously noted, is to be found in 
the behavior of the paleas on the maturity of the grain. It is well 
known that in Eragrostis the paleas are often persistent after the fall 
of their lemmas but this point does not seem to have been used very 
extensively (or possibly found constant) as a diagnostic character. 
In the course of field study embracing mostly E. Purshii and E. pere- 
grina it became apparent that in the former the paleas are very per- 
sistent and tightly appressed to the rachilla, even after the panicle 
has become completely dead and broken up. A very characteristic 
appearance is produced, somewhat suggesting a shriveled or desiccated 
spikelet. On the other hand it was found that in E. peregrina all the 
scales, including the paleas, immediately fall away with the ripened 
grain from the rachilla.. The denuded, close, short zigzag of the 
rachilla proves to be a character of considerable value. It is not to 
be supposed, of course, that every palea always falls away but the 
tendency is so very strong that if a mature panicle be pulled through 
the fingers any paleas still attached will invariably break away at 
once, showing their natural disarticulation. This process applied 
to E. Purshii rarely if ever disturbs a single palea. It has not been 
possible to examine satisfactorily the behavior of the paleas in Æ. 
pilosa, as field experience with this species has been too meager to be 
conclusive, but it would appear that the paleas are more or less de- 
ciduous. In E. peregrina and E. Purshii this point is definite and 
distinctive but in E. pilosa herbarium material suggests it to be a 
variable character. 
Of differences almost microscopic, but apparently distinctive in a- 
way, the size of the grain may be noted. "That of E. peregrina is the 
smallest, measuring about 0.5-0.6 mm. in length. That of E. pilosa 
and of E. Purshii averages about 0.7-0.8 mm. These measurements 
are too minute and too close to be of much practical service — particu- 
larly when some variation must be allowed for spikelets of different 
ages. 
It might seem from a critical examination of the foregoing comments 
