DESCRIPTIONS OF NEW AND RARE INDIAN PLANTS. 77 
section of Anemone and the species belonging to the old genus Pulsatilla to justify DeCandolle's 
group-name of Pulsatilloides, contrary to Janczewski’s dictum: “ Il est difficile de compendre pourquoi 
De Candolle a donné ce noma une section embrassant deux plantes africaines qué ne présentent 
aucune affinité avec nos Pulsatilles.” Тһе hairs on the ovaries, when present, are mostly rigid, 
as they are in the species which go to form the section Pulsatilloides; but in a variety of 
what I call subspecies ovalifolium from Kumaon the hairs are rather soft, whilst in undoubted 
forms of A. obtusiloba from Kansu the ovaries are in some plants densely hairy, in others perfectly 
glabrous without the plants differing from each other in any other detail. This abolishes, moreover, 
one of the distinguishing characters of the A. rupestris of the Flora of British India. As regards 
the latter it has to be mentioned that the carpels of the form referred to are usually more or less 
beset with stiff hairs, quite according to the manner of certain forms of A. obtusiloba from Chamba 
and Nepal. We have, therefore, to consider the value of the leaves in furnishing diagnostic characters 
to serve in the discrimination of the forms allied to А. obtusiloba. We begin with А. írullifolia 
H. f et T. There is now in the Calcutta Herbarium a large collection of specimens belonging to 
this form, gathered in Sikkim, Chumbi, and Setchwan. They all agree with each other as regards 
habit, structure of flower, indumentum, and the colour of the sepals; the leaves of most of them are 
like those shown in (A) of plate 106; the extreme forms (figs. 18 and 19) are placed side by side, but 
they are connected by forms with leaves like those represented in figures 14—18, The leaf shown in 
fig. 13 does not differ in form so very much from leaves of A. obtusiloba of the shape shown in fig. 
4. Franchet endeavours to resuscitate Royles A. discolor. He says in his Plante Delavayanae: “Je 
erois qu'il faut conserver le nom d’ A. discolor, Royle, à une plante trés-voisine de P А. obtusiloba, 
mais qui s'en distingue facilement par la forme nettement pentagonale, à l'angles assez aigus, de ses 
feuilles. L’ А. discolor est aussi moins velu.” The transition as to form of leaves. between those 
plants which, according to Franchet, would severally constitute the two species A. obtusiioba, Don, 
and А. discolor, Royle, is sufficiently illustrated by figures 1—6 of plate 106 to definitely dispose of 
A. discolor as species or even variety. There are a number of aberrant forms which, to a certain 
extent, may be distinguished from the typical A. obtusiloba by the leaves being more or less distinctly 
ovate in outline, not pentagonal nor semicircular or subreniform ; they gradually shade into the type; 
but forms like that depicted in (B) of plate 106 look very distinct. I unite them into a subspecies 
which I call А. ovalifolia. Although differing in their outline, such deeply divided leaves as those 
of some forms of A. ovalifolia appear to render the reduction of the A. rupestris of the Flora of 
Brit. Ind. imperative; nor does A. imbricata, Maxim., lie outside the group of forms which have 
diverged from the typical 4. obtusiloba. The A. rupestris of the Calcutta Herbarium—there is an 
original Wallichian type-sheet here—is different from the А. rupestris distributed under this name from 
Kew, and evidently the one described as such in the Flora of Brit. Ind. Wallich, as has happened 
pretty frequently with him, appears to have distributed quite different species under the same name. 
As, however, the form described by Hooker and Thomson in the Flora of Brit. Ind. has to be 
reduced to A. obiusiloba, I have changed its name into A. sazicola, leaving the name A. rupestris to 
the species to which doubtlessly Wallich originally meant to give it. The following is а classi- 
fication of the numerous forms which we thus see ourselves constrained to include in А. obtusiloba:— 
Subspecies I. TRULLIFOLIA, H. f. et T.; mostly villous; scapes erect ог more frequently ascending 
or prostrate, 5—20 cm. long; leaves subsessile or shortly or rather long-petioled (petiole some times as 
long as the blade), obovoid-subspatulate, more rarely rhomboid or suborbicular in outline, sometimes 
oblong-linear, cuneate or rarely obtuse or even cordate at the base, coarsely crenate-serrate, more or 
less distinctly 3—lobed; sepals elliptic, golden nus usually 2 to 3 times as long as the head of 
stamens; carpels densely hirsute. 
Var. а. LINEARIS; leaves oblong linear or oblanceolate, entire or scarcely serrate at the apex. 
Tachienlu region (Pratt n. 493 /). 
Var. В. sPATULATA; leaves obovate-subspatulate or rhombic, entire or very shortly trilobed, 
cuneate at the base, coarsely crenate-serrate. Sikkim (Н. f. and others/), near 
