388 
branches. Occasionally female spikelets also terminate the other- 
wise male branches of the lower whorls. As to the direction of the 
branches, the herbarium specimens, if cut from flowering or fruiting 
plants in their early stages when the joint cushions at the base of 
the branches are still active, cannot be relied upon. Such panicles 
are in fact as much contracted as similar panicles of Z. aquatica. 
But Turezaninow and Hance state from observation in the field 
differentiation of the male and female pedicels is much more 
ciliolate and oS eae on the back. The same difference is 
e following pair of glumes which in an equally 
sparingly hairy rim; in Z, aquatica on the other hand it is perfectly 
the bristle-like awn varying from one-fifth to the full-length of its 
glume ; rarely it is reduced to a mere cusp. In Z. aquatica it is 
usually absent, the glume ending with an acute point; but in some 
> Sree it is found to be drawn out into a fine, bristle-like mucro. 
Not much stress should therefore be laid on this feature taken by 
itself, The lodicules, stamens, and pistils of both species fail to 
“ ord any character that may be relied on for discrimination. 
ermaphrodite spikelets occasionally occur in both species. 
OA deere ig only grains which I have seen of Z. latifolia are 
: ose of Balansa S specimen (Pl. de Tonkin, no. 4727). They are 
inear oblong and 5 by 1mm. If this should prove to be their 
yee condition, they would indeed be easily distinguishable from 
_ the slender linear, long (9-15 mm.) grains of Z, aquatica. 
cg ae eg latifolia seems to have its home in Manchuria 
ee to Korea and eastwards to Japan. It is also 
3 ‘trom the neighbourhood of Pekin, but whether it occurs 
