OF CEYLON AND INDIA. 
407 
peculiar method of formation of the whole thallus by a lateral 
outgroAvth apparently of the entire primary axis, so that there 
is a possibility, though probably a very unlikely one, that we 
should regard the thallus as ‘‘ combined ” root and shoot, and 
perhaps the endogenous shoots as representations of its 
“ root ” element. The whole question, however, must be 
left for detailed developmental investigation. Castelnavia 
princeps has a shoot thallus, apparently derived from the 
primary axis, but in this there seems to be a combination of 
stem and leaf, which is not the case, so far as we can see, in 
Lawia. In Lophogyne arculifera the root-thallus bears 
secondary shoot-thalli, of similar morphological construction 
to the thallus of Castelnavia princeps. 
The other seven Indian genera all possess some kind of 
‘‘ root ” thallus, and this is also the case with most of the 
American genera. Taking first the Tristicheæ, Ave find in 
Tristicha ramosissima and Weddellina squamulosa a thallus 
to Avhich, Avithout any serious stretching of the common 
meaning of the term, the name of root may Avell be applied. 
It is a thin thread-like organ, which is probably endogenously 
developed from the base of the primary axis just as in 
Podostemon subulatus, in which a similar organ occurs. It 
is endogenously branched, at least in Tristicha, and the 
structure of the vascular cylinder, though slightly dor si - 
ventral, as might be expected, is not markedly different from 
that of an ordinary root ; the tip is covered by a root-cap of 
the ordinary kind. The only marked feature in which these 
roots are peculiar is the regular development upon them, in 
acropetal succession, of the endogenous leafy shoots, and this, 
though carried here to a high degree of perfection, is by no 
means uncommon in ocher roots. These secondary shoots 
reach a high degree of development and complexity, but are 
not otherwise remarkable among water plants. In Tristicha 
hypnoides, accordiug to Cario (8), there is no root-cap, but 
otherwise it seems to resemble the other Tristicheæ. 
Passing on to the more modified forms in the order, we 
find the thread-like root-thallus in nearly all the other 
