292 WILLIS : MORPHOLOGY OF THE PODOSTEMACBÆ 
and his pupil Wächter (41). I have also examined Prof. 
Goebel’s material, which he has kindly presented to me, and 
verified most of their observations. The fioral shoots here 
arise from the root thallus independently of the vegetative 
ones. The latter are branched in two ranks in one plane 
like Tristicha hypnoides, and bear large numbers of curious 
scale leaves, in many of whose axils are ramuli, while the 
shoots also branch into shoots of unlimited growth. The 
pairs of shoots seen in Tristicha do not appear here. The 
leaves of the ramuli, instead of having nothing in their 
axils, have what Goebel calls Kiemenbüschel, small shoots 
bearing a few slender leaf -like organs. Similar tufts occur 
on the surface of the thalli or leaves of some S. American 
Podostemaceæ ; they appear to be assimilatory organs, but 
whether they are homologous in all cases remains to be 
discovered. 
The Indian sub-genus Dalzellia is distinguished chiefly by 
the three stamens of the flower, the connate leaves at the 
base of the flower-stalk, and the absence of tristichy in the 
ramuli. The last point is led up to by the form lately dis- 
covered by Miss Lister in the Nile, T. alternifolia, var. 
pulchella, Wmg., in which the lower part of the ramulas has 
broad tristichous leaves, the upper narrow leaves irregularly 
arranged. 
In many respects the Indian species seems one of the 
most primitive among the Podostemaceæ, though its anemo- 
philous flowers may perhaps indicate it as reduced in some 
respects ; we know too little, however, about the phytogeny 
of the order to say whether it is descended from anemo- 
philous ancestry or not. At any rate, T. ramosissima lives 
in very much less violent water than most of the Podoste- 
maceæ, and has, much more than most of them, the 
appearance we are accustomed to look for in water plants ; 
and in view of the great difficulties in tracing the 
phylogeny of this family, it deserves most careful study in 
detail. 
