OP CEYLON AND INDIA. 
365 
Habitat . — The plant appears to grow in moderately rapid 
water, such as is affected by the Dicræas. I have found it 
mixed in my material with Lawia zeylanica, Podostemon 
Barberi, Dicræa stylosa kanarensis, and Tristicha ramosis- 
sima. 
Dry Season Appearance . — The general appearance of the 
plants at the end of the year is shown in the three smaller 
rocks in Plate XXV., where also may be seen the first 
indication of the very remarkable polymorphism, which is 
characteristic of this species. On the large lower stone is a 
large plant just exposed and flowering. This has a closely 
attached creeping branched thallus, with marginal flowers. 
A similar and larger plant, still in the vegetative condition, 
is shown on tbe large rock figured on the left. On the 
smaller stones above are thalli in fruit, evidently, though 
shrivelled, of the most various shapes, and much smaller 
than the creeping one below. Examination with a lens 
shows that some of them are flat upon the rock and branched, 
others cup-like, some like discs or cups on stalks, &c. The 
fruit is smooth, almost spherical, with one smaller valve 
splitting off obliquely from a larger persistent one. The 
seeds are shed upon the rocks in the usual way, and the 
early stages of the life are no doubt similar to those in the 
rest of the group. It is very much to be desired that this 
plant should be followed throughout its life- history. The 
material at my disposal was all collected at the end of the 
life, and contains no stages that show the primary axis or 
the development of the peculiar forms assumed by the 
thalli. 
Mature Structure . — The chief interest in this species 
centres in the thallus and its extraordinary polymorphism. 
In general it has the morphology of that of Dicræa. The 
growing point in all my specimens was more or less shrivelled, 
and I was not able to definitely make out its construction, 
but on the whole it seems to be like that of Dicræa stylosa, 
var. fucoides, broadly obtuse in shape ; whether there is a 
properly developed root-cap or not I was not able to decide* 
