1898] THE LEAF AND SPOROCARP OF PILULARIA ee 
around dorsally (4, figs. 37-33), and the sori, which origi- 
nate in cells having a lateral position (fig. 74), come later to 
have a position such that the soral canals open nearly terminally 
figs. 31-33), thus making the longitudinal axis of these canals 
nearly parallel to that of the stalk. 
On the dorsal side of the sporocarp in the meantime there is 
formed a small protuberance at the upper end of the stalk (/7/, 
Jigs. 31, 32), which is later found to be supported by a mass of 
thick-walled cells extending inward nearly to the vascular 
bundle (fig. 33). This protuberance, as was pointed out by 
Russow in P. minuta, is apparently homologous with the lower 
tooth of the capsule of Marsilia quadrifolia. Just above this 
tooth there is a rather narrow, but deep depression (4 9, jigs. 
31, 34), which according to Russow corresponds to that found 
between the upper and lower teeth in MZ. guadrifolia, but there is 
no marked increase in height of the epidermal cells above 
this that might represent the upper tooth of the Marsilia capsule 
(see Russow ’72, and Johnson ’98). ' 
It was this bending backward of the young capsule, per- 
haps, which led Meunier to think that the sori were primarily 
terminal in position, and this may account also for the view of 
Campbell that one of the valves is developed from the apical 
cell of the sporocarp, but it is not very difficult to follow out the 
details of development satisfactorily if the unchanging sagittal 
plane is used as a guide. 
During this change in the general form of the capsule the 
protodermal layer throughout, beginning on the dorsal side just 
above the basal pit, divides by periclines into epidermal and 
hypodermal layers, and then the latter divides again to form 
inner and outer hypodermis (e f,/ y, figs. 30, 31). These 
layers soon surround the whole capsule (Mettenius °46, Han- 
stein ’66) except for the stomata, most of which are near the 
base of the capsule. At these openings both hypodermal layers 
are wanting, the guard cells being as usual derived from the 
epidermis (s ¢, fig. 33). 
The development of these highly specialized tissue layers 
