52 MR. J. С. WILLIS ON А NEW NATURAL 
(3) Simple moss-like leaf.—Complex compound leaf. There are in the 
Podostemacez simpler unbranched leaves, as for example in all the Indian 
forms, and in a few South American forms, but even so, these leaves, with 
their sheathing bases, and with the peculiar mode of (stem) branching that 
obtains are closely allied to, and must probably be considered derived from, 
the more complex leaves of the family. It is extraordinarily difficult to derive 
the one type of leaf from the other. Even if we suppose the ramuli or short 
shoots to be the precursors of the compound leaves of Podostemaceæ, the 
gap between the two families is a very large one, and there is no evidence 
that such eould be the case, nor is there any parallel case in the vegetable 
kingdom. 
(4) Exstipulate leaf.—Stipulate leaf. This is also a difficult change to 
make even between the most closely similar leaves of the two families, for 
the exstipulate Podostemaceæ of India are probably derived from the 
stipulate members of the same family, their leaves agreeing in their sheathing 
bases and in other ways, and the plants that possess them being highly 
modified, and having the peculiar branching. 
(5) Simple branching.—Branching in the “axil” of a lower stipule. 
This also is a point in which there is a very wide gap between the two 
families, with apparently no intermediate stages. 
(6) Secondary shoots simple, or slightly thalloid, with leaves on the 
upper side as well as on the margins.—Secondary shoots complex, with 
leaves only on the margins. Another point in which no intermediate stages 
occur, and in whieh it is almost impossible to derive one from the other. 
Practically it comes to this, that almost the only thing the two families 
have in common in the morphology of the vegetative organs is the origin of 
the shoots from creeping roots. 
There are also many minor details in which the Podostemacez differ from 
the Tristichaceæ, but the above seem amply sufficient to warrant the 
separation, even if the criterion of an order be that of Bentham and Hooker, 
rather than that of Engler and Prantl. 
The only form that shows any signs of closer relationship to the Podo- 
stemaceæ among the Tristichaceæ is Weddellina, but this is only in a very 
few points, such as more numerous stamens and a bicarpellary ovary. 
Finally, we may give the detailed characters of the two new families, as 
separated. 
TRISTICHACEÆ. Flower $, fully hypogynous, small, inconspicuous, 
regular, ог slightly irregular by absence of one or two stamens, mono- 
chlamydeous, without spathe. Perianth 3—5-merous, free or united, equalling 
or exceeding the ovary, imbricate, sepaloid, marcescent. Stamens as many 
аз perianth, or 4-5 times as many ( Weddellina), or reduced to 2 or 1, usually 
alternate with the perianth ; anthers bilocular, introrse ; pollen simple. 
