420 EPIGONIANTHEA, 
Capsula in pedicello brevi vel seepius sat longo suffulta, raro spheerica, 
seepius oblonga vel subcylindrica, valida, 2-8-strata, ad basin usque 
4-valvis. Elateres dispiri, paucis ex parte vel tota longitudine 
monospiris raro adjectis. Spore parvule—minute. <Andrecia 
spicata, raro hypogyna, rarissime ad ramulum proprium relegata, 
plerumque in caulis ramive medio apiceve posita; bractez foliis 
consecutive, pro more minores, 1—-10-andre. 
In the subtribe Trigonanthee the leading character was to have the 
perianth flattened from the front, and its primary angles at the axis, or 
fold, of the three complicate flower-leaves; hence its trigonous form, with 
a flat front, an.angle on each side, and a third angle at back. But in the 
subtribe Epigonianthee the primary angles are at the marginal sutures 
ot the flower-leaves, and not at their axis ; so that where there are only 
two such leaves, and they are pressed together by their faces, a complanate 
perianth results, at right angles to the normal plane of the leaves, with 
one of its edges antical, the other postical, as in the large genus Plagio- 
chila. Where an underleaf exists and the flower-leaves are three, the 
perianth may be trigonous, with the third angle in front, as we see it in 
Lophocolea ; or it may still be complanate and ancipitous, if the underleaf 
be folded along the middle, so as to present only one edge at back, instead 
of two, of which we have an example in Levoscyphus. 
In all these genera the perianth is as wide at the mouth as in the 
middle, or even wider; but in some genera, of which Jungermania proper 
is one, it is constricted—or even before the protrusion of the capsule quite 
closed—at the apex ; never complanate, and rarely only trigonous; usually 
more or less inflated, and traversed by from 3 to 10 angles, or folds. In 
a few species, mainly of the subgenus Lophozia, the perianth is terete—or 
tereti-polygonal—so that it is impossible to say whether the compression 
is lateral or frontal, except in the very early stage, where it is mostly dis- 
cernible.* In Marsupella, however, the compression is decidedly frontal ; 
yet the genus is so closely related to Nardia (where the compression of 
the perianth is lateral) that, in a natural arrangement, it must perforce 
stand next to it, for Nature is continually overstepping our artificial 
limits. 
In Trigonanthew nearly every species has cladogenous and postical ? 
flowers; yet a few species of Cephalozia have terminal flowers, and some- 
times both terminal and postical on the same individual. In Epigoni- 
anthew, on the contrary, the rule is to have acrogenous @ flowers; yet in 
Chiloscyphus they are distinctly cladogenous, while, without flowers, the 
plants so much resemble Lophocolee—or sometimes Leiosceyphi—that it is 
often impossible to assign them to their proper genus. 
The leaves of Epigonianthee are succubous in every species. In Trigon- 
anther they are succubous or transverse in Cephalozia and its immediate 
allies, but incubous in all others, 
I have throughout endeavoured to collocate the Marsupiocarpous genera 
by the side of those which so nearly resemble them in every respect, except 
that their perianth is not a pendulous pouch, as to leave no doubt of their 
not-very-remote origin in a common ancestor. This is the natural method ; 
although it would doubtless be much easier, and is more convenient for 
* See also below the note to Jungermania cordifolia. 
