48 
Surfaces often glandular and the coste and costules above rusty-pubescent by short 
articulated Ctfenitis-hairs, while common dry whitish hairs or downs are always 
absent. By this character the species of the hirta- and ampla-group agree but the 
scales of the latter species are thinner, not blackish, hairlike and those of the co- 
stules beneath never bullate. In shape of lamina the two groups differs distinctly. 
D. ampla and its allies have a deltoid or deltoid-ovate lamina, i. e. the largest pinne 
are always the basal ones, which have their lower side much produced with the 
basal pinnula much the largest. — From the species of the next group (subincisae) 
D. ampla differs by the abscence of longer subulate hairs on the costules above and 
by its venation. The midrib of a posterior basal ultimate segment runs out from the 
costula at its very base or above it, not from the costa as in D. subincisa. It is in 
general not difficult to distinguish a specimen of D. ampla or its allies from one of 
the subincisa-group, not only by the differences mentioned, but also by the general 
aspect of the leaf (colour, thin texture, shape), and the two groups represents each 
a separate line of development and are certainly not closely related. Still there are, 
of course I dare say, intermediate forms. Such is especially D adenopteris which 
by its whole habit, texture and colour seems to be closely related to D. ampla. 
Nevertheless I refer it to the next group; certainly it has its nearest relatives there. 
All the species referred to the ampla-group are decompound. It is possible 
that certain bipinnatifid species placed in the group of D. submarginalis should be 
grouped together with D. ampla, f. inst. those species with a large tuft of basal 
scales (D. submarginalis and others), but none of these species show any tendency 
to a bipinnate-decompound state, which might more definitely lead to a righter 
understanding of their relation to D. ampla. 
No species of this group is known from South Brazil; all the known species 
are Andine-West-Indian. They are closely related and their right delimitation is very 
difficult. My treatment is scarcely quite satisfactory; it is probable that several 
pteridologists would prefer to unite into a single species D. ampla, D. excelsa and 
D. equestris, others have descibed a number of »species« which I unite with D. ampla. 
I am inclined to believe that my D. ampla includes several »geographical species«, 
but very. comprehensive collections of perfect specimens are needed for a delimita- 
. 
tion of such species. . 
| Key to the Species. 
I, saeget of the axils of pinnx bright-brown or reddish brown, slightly fimbriate 
r entire. 2 
cr OF 
ee Seales of the axils brown or nearly black, often glossy. 5 
