XIII] THE TAPETUM 259 



appears outside the definitive sporogenous group as an ill-defined band 

 consisting of several layers of tabular cells. But in the smaller sporangia 

 of the Leptosporangiate Ferns the cells which go to form the more definite 

 tapetum are cut off from the sporogenous cell or cell-group. There is good 

 reason to believe that there has been a progressive change of origin of the 

 tapetum within certain circles of affinity. Indefinite and non-specialised 

 nutritive arrangements are characteristic of larger and probably primitive 

 sporangia: but more definite tapetal layers are found in the smaller and are 

 probably derivative. Thus in the large sporangia of Lycopodiuin the tapetum 

 originates outside the sporogenous group : in the smaller Selaginella it is cut 

 off from the superficial cells of the group. The same difference is seen in Ferns 

 on comparing the Marattiaceae and Ophioglossaceae with the Leptosporan- 

 giate Ferns. In the former indefinite layers of slab-like cells outside the sporo- 

 genous group become disorganised, and act as a tapetum. In the latter 

 the tapetum appears as two definite layers of tabular cells, which spring 

 from the " archesporium " itself The difference of origin seems to depend 

 on size. 



As the spore-mother-cells round off prior to the tetrad-division, the 

 tapetal cells fuse together to form ?i plasviodiwn, their nuclei retaining their 

 identity, and even multiplying by fragmentation. As the spore-mother-cells 

 separate this nucleated plasmodium intrudes between them, so that they 

 are suspended, usually isolated and rounded off, in a rich nutritive medium 

 (Fig. 255). In all the more primitive Ferns the material of the plasmodium 

 is absorbed into the developing spores : but in certain advanced types much 

 of it may remain as a deposit on the outer wall of the spore, and is then 

 called the '' perispoj'e." 



The tetrad-division itself has been described for Dryopteris (p. 22): it 

 is essentially the same for Ferns generally, excepting that the grouping of 

 the spores in the tetrad may follow either of two plans. In the first the 

 division of the spherical spore-mother-cell is by six walls, giving the spores the 

 tetrahedral form (Fig. 256, B): in the second the spore-mother-cell divides 

 by three walls to form spores of wedge-shape, with two flattened sides 

 (Fig. 256,^). Though these seem so different they may both occur within 

 nearly related circles. The Marattiaceae and the Schizaeaceae include both. 

 The spore-form is thus an uncertain guide to affinity. 



The spore-wall is often marked by characteristic sculpturing (Fig. 256). 

 But more important than this is the existence or absence of a deposit of the 

 perispore from the tapetum. Ferns may in fact be divided into two groups 

 according to the presence or absence of a perispore. None is seen in the 

 Eusporangiate Ferns, nor in the Osmundaceae, Schizaeaceae, Gleicheniaceae 

 Hymenophyllaceae, Cyatheaceae, Davallieae, or in Ceratopteris. In fact it is 

 absent from the relatively primitive Ferns. Of the more advanced Lepto- 



17—2 



