io6o Stevens.—On the Development of the 
of cells or initial layer of the tapetum, and these surround a single cell, 
which is proved to be the archesporium. 
The tapetal initials now multiply by anticlinal divisions, and finally each 
undergoes a periclinal division to form a two-layered tapetum (PL LXXXIV, 
Fig. i, d, and Figs. 2, 4, and PI. LXXXV, Fig. 8). This agrees essentially 
with the formation of the wall, tapetum, and archesporium of Lygodimn as 
described by Binford, excepting that he finds a three-layered tapetum in 
places. Turning to the Ophioglossaceae for comparison, Bower (’ 96 ) has 
found that in Helminthostachys the tapetum is formed by division of the wall- 
cells alone, and Campbell (’ 05 ) reports that in Botrychium the tapetum 
comes partly from the wall-cells and partly from the sporogenous cells. 
While the tapetum is thus increasing, the archesporium is undergoing 
divisions that result in a sporogenous tissue, or group of grandmother-cells 
of the spores (compare PL LXXXIV, Fig. 4, and PL LXXXV, Fig. 8). The 
sporangium is growing rapidly during these stages, the number of tapetal 
cells increasing step by step with the growth of the wall. 
About the time the number of grandmother-cells of the spores, or gono- 
tokonts, to use the term proposed by Lotsy (’ 04 ), is complete (approximately 
65) the outer tapetal cell layer begins to break down, and this is soon 
followed by a similar behaviour of the inner layer, leaving the solid mass of 
gonotokonts suspended in a tapetal cytoplasm (PL LXXXIV,Figs.4,5,and 6). 
Prantl found a different behaviour in the sporangia of Aneimia phyllitidis 
studied by him, in that the inner layer of the tapetum formed a plasmodium 
while the outer layer remained intact (see his Fig. 120); and Binford finds 
the outer layer persisting in Lygodimn (his Fig. 30). In some instances 
I find the outer tapetal layer breaking down even before the archesporial 
divisions are far advanced (Fig. 3). 
There seems to be a consensus of opinion that in all such cases the 
tapetal cells serve as nurse cells to the gonotokonts through their meiotic 
divisions, and to the spores through their developmental processes; and 
there are some considerations which suggest that this function can be carried 
on better when the entire tapetum is in the condition of a plasmodium, since 
the nuclei are then free to distribute themselves amongst the gonotokonts 
after these have separated into relatively small groups or into single free 
individuals, and also since the tapetal cytoplasm in the form of a plasmodium 
is free to circulate, and so facilitate the distribution of nutritive materials 
that in part are held in solution in the cell sap, and in part are suspended in 
the cytoplasm, or even enter into its composition. The conclusion of 
de Vries (’ 85 ) that, since diffusion has been found inadequately slow, the 
transportation of materials must be assisted by the rotation and circulation 
of the protoplasm, is given recent support in the researches of W. Bierberg 
(’ 08 ), who found cytoplasm rotation to more than treble the rate of trans¬ 
portation above that achieved by diffusion alone. 
