68 LEJEUNEA. 
inflorescence has usually been considered lateral, in the “Synopsis Hepati- 
carum” and elsewhere. In reality it is a sort of uniparous, or mono- 
tropous cyme ; for a very little observation suffices to convince us that 
the flowering branch terminates with the apical flower, and that its 
apparent continuation is a lateral branchlet, often taking nearly the 
same direction. In some cases the innovation (as it truly deserves to 
be called) diverges considerably, and then its true origin is obvious. 
Not unfrequently there are two innovations, one adnate to each lateral 
bract, and thus a dichotomous inflorescence is generated. A simple ? 
branch is rarely found in the same subgenus as an innovant one, but 
there are cases where the two types of structure must necessarily stand 
side by side; where two species agree so nearly in every other respect 
that they cannot possibly be placed in different subgenera. Such cases 
exist in the groups I have called Prionolejeunea, Trachylejeunea, and 
Leptolejeunea. 
Where innovations are present, the true bracts are mostly reduced to 
a single pair ; but when absent, the bracts may be considered plurijugous, 
the lower ones passing insensibly into normal leaves, if the 2 branch be 
long enough to bear any such. In all cases the bracts are leaves, modified 
in much the same way as in Frullania, except that they were never 
united into a cup-like involucre. They are indeed rarely exactly oppo- 
site, so that although one of the pair may be connate with the intervening 
bracteole, the other will remain free from it ; and it is only in Drepano- 
lejeunea and Leptolejeunea that I have seen both bracts commonly so united. 
The usually pyriform and abruptly rostellate perianth goes through 
many phases of form, for a full description of which I must refer to the 
descriptions of the subgenera, some of whose most important characters 
are drawn from them. Ina few species it is as perfectly flattened as that 
of Radula ; in a good many others it is much compressed, but with a 
more or less salient postical keel, or with two or more keels ; in others, 
we find it keeled also in front, and almost regularly pentagonal; in a 
few others pluricarinate, and in a very few, perfectly terete and ecari- 
nate. The keels may be either a simple fold, acute or obtuse, or they 
(especially the lateral ones) may have a salient wing, which is often 
toothed, or spinose, or ciliate; and in a good many species they are 
dilated at the apex into an auricle or an ascending horn. In some 
species the perianth goes on growing after fruiting, especially at the base 
(so as to become spuriously stipitate), and assumes a purplish tinge, in 
which the calyptra participates. 
The pistillidium is uniformly solitary—a character not known in any 
other genus of foliose hepaticee, and the slender style is persistent. 
The calyptra is above half the length of the perianth, and nearly as — 
wide, obovate or pyriform, thin in texture throughout, or incrassated only 
at the very shortly tubular base. In dehiscence it is sometimes bivalvular 
at the apex, and often trivalvular, with one valve narrower than the 
other two. 
The pedicel, when full grown, is very little exserted, being only twice 
or thrice as long as the perianth. It is cylindrical, and is built up of 
cells in regular horizontal tiers. A transverse section shows 4 large 
axial cells, each the quadrant of a circle, so that their adjacent walls take 
the form of a cross; and 12 (or 16) smaller peripheral cells, 3 (or 4) ex- 
terior to each axial cell. And now we have to note a curious correlation 
of structure, not easily to be explained. In the Holostipe, or entire- 
