STRUCTURE AND DEVELOPMENT OF DENDROCEROS. 475 
suppressed, so that it remains permanently but one cell thick. 
The arrangement of the sporogenous and sterile cells, too, is 
much less definite than in the other genera. Occasionally ( Pl. 22. 
fig. 17) there is a fairly regular alternation of sporogenous and 
elater-cells, determined sometimes by oblique divisions of the 
primary archesporial cells into two of unequal size, the larger one 
becoming the sporogenous cell, the other, one member of an 
eater. Quite as often, however, some of the primary cells 
remain undivided and, enlarging rapidly, form at once the mother- 
cells of the spores. The others divide once longitudinally, and 
become elater-cells. 
The columella in the forms examined by me corresponds very 
closely in the arrangement of the cells to that of Anthoceros, 
showing in cross-section (Pl. 22. fig. 18) usually sixteen cells 
arranged very regularly. As in Anthoceros, these cells later 
separate somewhat, showing intercellular spaces at the angles. 
Leitgeb’s figures of D. cichoraceus show a more massive columella 
and a relatively slight development of the amphithecium. . 
Each sporogenous cell gives rise to four tetrahedral spores iu 
the usual manner. These spores are much larger than in the 
other genera, and may remain undivided until the dehiscence of 
the sporogonium (D. Breutelii), but in the second species 
examined (probably D. erispus, Nees) the spores germinate within 
the capsule and become multicellular bodies of considerable size 
before they escape. Leitgeb records similar multicellular spores 
in several other species, c. g., D. crassinervis, Lehm. & Lindenb., 
D. cichoraceus, Lehm. & Lindenb., and states that in the latter 
species both unicellular and multicellular spores are found. He 
infers, and probably with reason, that the wultieellular spores are 
simply accelerated phases of germination suchas occur 1n Pellia. 
In the cases observed by the writer where multicellular spores 
occurred, all the older spores showed this condition, and all the 
*pores ofa tetrad developed perfectly, there being no sueh abortion 
of two of the spores of the tetrad as Leitgeb describes and figures. 
As already indicated, no constant relation between sporo- 
senous and elater-cells can be made out. ‘The elaters, “oe 
well known, are not single cells, but are composed of two OF 
three cells joined end to end. ‘There is no indication that the 
primary elater-cells undergo further transverse division, but the 
tully-developed clater is composed of several of these primary 
elater-celly which simply cohere, as is the case in Anthoceros. 
7 9 N 
LINN, JOURN,—BOTANY, VOL. XXXIII. -> 
