ae PROCEEDINGS OF THE NATIONAL MUSEUM. 401 
and nodding, while from the apical spikelets slender branches issued, 
terminated by a few, (1 to4) small spikelets (Tab. xLvu11, Fig. 1). In some 
specimens the spikelets were transformed into leafy shoots, thus rep. 
resenting the well-known variety ‘vivipara,” as described for many 
species of Graminee (Tab. xLvitt, Fig. 12).* 
Now, concerning the first question, whether the two awned glumes 
represent two neutral flowers or not, let us examine figure 8 on the ac- 
companying plate. This spikelet, of which the proper empty glumes 
have been removed, shows altogether three awned glumes, but of which 
only the two basal ones are now in question, They are both situated 
on the same rhachis, but at different heights, and we see farther that 
the uppermost one, that on the right side in the figure, partly encloses 
another smaller and awnless glume, which is a normal palet. Judging 
from the position and the shape of these two glumes in connection with 
the fact that one of them has been found enclosing a palet and flower, 
may we then not suggest that they both are true flowering glumes? 
There is no doubt that they correspond to the two similar glumes of 
the normal spikelet, because their position is exactly the same, and 
there is no essential difference in regard to their appearance; they were, 
it is true, merely hairy along the midrib, but this character is of but 
small importance. Several other variations were observed even in the 
same inflorescence, a circumstance probably due to the unusual moist 
locality where the plants were found growing. In some other spikelets, 
(Figs. 2, 9 and 10) only one of these glumes was developed, but it was 
easily recognized. We have now another case (Fig. 4) in which we see 
the same glumes again, but widely separated from each other on the 
same rhachis. Their form is here very different from the normal one, 
- since they are distinctly acuminate and but shortly awned. We meet 
here a fact which shows that their form may not be constant, and also 
that they may resemble the proper empty glumes. This very abnormal 
case would have been a good support to DOll’s theory that these glumes 
should not be equal to the fifth glume in the normal spikelet, since they 
are not only very different from this in regard to their shape, but in 
this case, they even resemble the empty glumes. We venture, how- 
ever, to oppose this supposition of D6ll by referring to the spikelet in 
Fig. 8. There are here three awned glumes, the uppermost one being 
a true flowering glume which has here simulated the shape of the 
others, and which really corresponds to the same glume of the normal 
perfect flower. It is situated upon the same rhachis as all the other 
glumes, the empty and the flowering ones of the neutral flowers, and 
encloses a palet and a rudimentary pistil, but no stamens. This was 
observed in several other spikelets, and we see it illustrated again in 
Fig. 6, where the uppermost flowering glume is easily distinguished by 
its long awn, although the glume itself is much smaller than the others, 
*Compare EK. H. Hunger: Ueber einige vivipare Pflanzen und die Erscheinung der 
Apogamie. Inaug. dissertation, Bautzen, 1887, 
Proc. N. M, 92 26 
