XXVI] A GENERAL REVIEW OF THE SIMPLICES 231 



the Eusporangiate and the Leptosporangiate Ferns, and their early fossil 

 history justifies this view. Not only are their soral and sporangial characters 

 in accord with it, but their anatomy with its tentative advances towards 

 medullation, solenostely, and in the largest types the final disintegration of 

 the stele, the primitive trace widening into the horse-shoe, and the hairs as 

 dermal appendages all support the position thus assigned. The gametophyte 

 and sexual organs point rather to a Leptosporangiate afiinity than to a 

 Eusporangiate, though still with archaic features. 



All the Families named in the foregoing paragraphs are regarded as 

 specially primitive. None of them can be related as direct progenitors to 

 any other living Family: so that it is in a general rather than a special 

 sense that these Ferns may be held as prefiguring those later developments 

 embodied in the Leptosporangiate Ferns. The Osmundaceae may, however, 

 be regarded as synthetic forms in one feature which has stamped itself deeply 

 on the later evolutionary progress. Of its two living genera Osmnnda bears 

 its sporangia normally in marginal tassels on the narrow fertile pinnae; 

 while Todea bears them on the lower surface of its expanded pinnae. But 

 where the fertile pinnae of Osmunda expand also, as they do sometimes 

 abnormally, to a broad form, the sporangia are in like manner found on the 

 lower surface. The two genera in their normal development prefigure on 

 the one hand the marginal type of sorus, on the other the superficial type; 

 and they supply a link between the two, while they suggest that an expanding 

 leaf-surface encourages a superficial insertion. It is, however, a fact based 

 upon wide observation that these two distinct soral positions are strictly 

 maintained in many large groups of Ferns. For instance, in the Gleicheni- 

 aceae, Matoniaceae, Cyatheaceae (excl. Dicksoniaceae) the son* are all super- 

 ficial, and remain so: in the Schizaeaceae, Hymenophyllaceae, Dicksoniaceae 

 they are all marginal at least in their origin, and (excepting in the Schizae- 

 aceae) they remain so up to maturity. Thus the distinction visualised 

 within the Osmundaceae becomes one of the most stable characters for 

 comparison of the early Leptosporangiates, and accordingly that ancient 

 Family may be held to be synthetic in respect of it. It applies at once to 

 the almost equally ancient Families of the Schizaeaceae and the Gleicheni- 

 aceae, the former being marginal, the latter superficial in origin of sori. 

 Further, the relatively late and derivative Leptosporangiate Ferns con- 

 stitute two main sequences related more or less directly to them. These 

 sequences have been designated respectively the Margina/es, retaining with 

 greater or less persistence the marginal position as seen in the Schizaeaceae; 

 this was probably the original position of the sporangia for all F'erns: and 

 the Superficiales, which have their sori borne on the lower surface of the 

 leaf, as in the Gleicheniaceae: this was probably in the first instance 

 a derivative state consequent upon broadening of the foliar surface. 



