NOTE ON THE GENUS AMPHIDONAX. 39 
bifaria, Roxb., and two previously undescribed Grasses were added to 
a 
South-eastern China, though, as it has been found in Japan, its 
presence here might have been confidently predicted ; and in fact it is 
to be met with, though not very commonly, in the country around 
violet and white. Most careful examinations o , made 
independently by the J. C. Nevin and myself, prove beyond a 
that not only is it in every respect a true species of Arundo, but 
at it isin reality very closely allied indee onax, Linn, ; for 
the florets are all undoubtedly hermaphrodite, though from the some- 
what paler colour of some of the lower pistils it is possible they do 
not all ripen fr e glumes are equal, almost as long as the florets, 
the lower one 5-7-nerved at and throughout, the 
upper 3-nerved, and the lower glumella is dis b pon 
very 
an arista between the two teeth considerably longer than in A. Donay. 
But Nees also describes both 4A. Heynei, N. ab E., and A. tenella, Ww 
& A., as with homogamous florets, and Dr. Thwaites ae oe herma- 
phrodite ones to his 4. obtusifora. A. bengalensis, N. ab E., has, 
I believe, an entire lower glumella, and in this respect resembles 
Phragmites ; ; and when it is borne in mind that in P. Roxburghit, Kth., 
the lower floret is not always male, whilst in Arundo Reynauiana 
Kth. (which is the Asiatic form of A. madagascariensis, Kth.) K unth 
himself testifies to having found the lower floret neuter,}t it will be 
seen how very unsatisfactory, slight, and interchangeable are the differ- 
uishi € nee. 
ences dist g the various so-called To me, 
indeed, it seems that the logical outcome of the foregoing observations 
is that Amphid nu no raison itn’ and that we must 
and limits of genera are apt to be far more matters of tradition 
habit than we are fully aware, as, for example, in the admission even 
* Enum. pl. ar heater hg + Gram. Afr. Austr., 357. 
t Plant. vase. Gen. g Fi. Ind. bat. iii., 360. S 
q BL Ind. i 347. @ Ann. Mus. bot. Lugd.-Bat.ii., 278, 
+* ibes it as “lanceolata, trinervia, nervis in acumen 
