558 FILICALES 
cell is thus never actually immersed in the tissue of the receptacle as 
is the case in G/. flabellata. In this feature again, G/. dichotoma approaches 
the ordinary Leptosporangiate type. The central sporangia, above noted 
as occurring in this species, arise, as far as can be seen, simultaneously 
with the rest, and actually occupy the central area of the sorus from the 
first (Figs. 311 4, 2); this area is usually vacant in other species, and is 
sometimes vacant also in G/. adichotoma (Fig. 311 g). Since the sporangia 
originate in this central position, their presence there cannot be accounted 
for by displacement due to pressure; it is to be ascribed rather to extra 
development, or interpolation of one or more accessory sporangia, which 
arise in a position usually unoccupied in the genus. 
The divisions in the sporangial head to form the lateral cells of the 
wall, the cap-cell, the tapetum and definitive archesporial cell, follow with 
slight deviations the type general for Leptosporangiate Ferns (Figs. 311 d-g) : 
the archesporial cell is usually of tetrahedral form, but from ‘the first it 
is of relatively small size, while the tapetum, which soon divides periclinally 
into two layers, grows rapidly. The outer wall remains a single layer of 
cells, but the cells divide freely by anticlinal walls so that in the mature 
state the sporangial wall consists of very numerous cells (Figs. 311 d, /, 4). 
The annulus soon becomes differentiated, and it appears that part of the 
annulus owes its origin to the cap-cell, but the larger part to the lateral 
segments. The outer layer of the tapetum with occasional supernumerary 
cells near the attachment of the stalk remains small, and forms a narrow 
inner investment of the wall; it is permanent for a considerable time, 
and traces of it may be found even in the mature sporangium. ‘The inner 
tapetal cells enlarge greatly, and often become polynucleate; their pro- 
toplasm becomes aggregated, with the nuclei in close proximity to the 
sporogenous mass (Figs. 311 7, 2), while the cell-walls become absorbed. 
The definitive archesporial cell in G. jabellata undergoes successive 
divisions (Figs. 311 4, 7), but the divisions are continued beyond the limited 
number usual for Leptosporangiate Ferns; the result is a very considerable 
cell-mass, so that a single vertical section through a sporogenous group of 
an average sporangium may traverse as many as 46 spore-mother-cells 
(Fig. 311 7); 45 was found to be the mean of countings in sections 
through eight different sporangia. A section through a sporangium parallel 
to the surface of the leaf may traverse even a larger number, as in the 
sporangium of Fig. 311 &, where 66 are shown in section. This difference 
may be in part due to the section traversing the curved sporogenous 
mass obliquely, but this explanation will not account completely for the 
variation in number. Any one section will only traverse about one-eighth 
of the whole number of sporogenous cells, thus there will be about 
45 x 8=360 spore-mother-cells in a single sporangium, and the potential 
output of spores may be estimated at about 360x4=1440. Comparing 
this with the case of ordinary Leptosporangiate Ferns, it is plain that the 
potential productiveness of an average sporangium of G/. flabellata is far 
