304 CYATHEACEAE [CH. 



This is in near accord with IVoodsia (Schlumberger, l.c. ¥\gs. 2, 3, 4), or with 

 Dryopteris filix-tnas ; such figures compare with the number of spores in the 

 sporangia, according to the table (Vol. I, p. 292). Records are not to hand 

 for C. dealbata, though these would be interesting in view of the low spore- 

 output in that species. In these details the Cyatheaceae appear to take an 

 intermediate place between the more primitive Ferns and the Leptospor- 

 angiates. The first segmentations of the embryo correspond with those 

 general in Leptosporangiate Ferns (Campbell, Mosses and Ferns, p. 391). 



Comparison 



The facts stated in the "Studies in the Phylogeny of the Filicales, II, III," 

 and elsewhere, have been summarised in the preceding chapters. They 

 formed the foundation of a distinction of the Gradatae, and other Ferns 

 holding a middle position between the Coenopterids and ordinary Lepto- 

 sporangiates, into two large series designated the "Marginales" and the 

 "Superficiales." If such a distinction is to be held valid it must be founded 

 upon consecutive sequences of related forms that are relatively primitive, 

 and the comparisons must refer not to one character or another, but to the 

 sum of such characters as can be observed. The comparison of later types 

 cannot be depended upon to give secure conclusions, owing to the frequency 

 of homoplasy. Since the validity of the two series thus characterised by 

 the position of the sorus has been doubted, the opportunity will now be 

 taken of restating the position, and of marshalling the facts that support 

 it. In particular the reasons for segregating the Cyatheaceae phyletically 

 from the Dicksoniaceae will be reconsidered, for it is round these families 

 that the question chiefly turns (von Goebel, OrganograpJiie, II Aufl., 2 Teil, 

 p. II 54). 



The comparative treatment of the Cyatheaceae naturally starts from the 

 Gleicheniaceae, a Family that dates back certainly to the Mesozoic Period, 

 and possibly it was represented in Palaeozoic time by the Fern Oligocarpia. 

 Here the naked sori were clearly superficial, and constructed on a plan closely 

 similar to that of the living species of Dicranoptcris, such as G. {D.)flabellata. 

 If the ancestry of the Ferns bearing those sori had ever borne their sporangia 

 at the margin (as they probably did in the first instance), the transition from 

 the margin to the surface must have taken place in Palaeozoic times. A 

 comparison of Oligocai^pia, G. flabellata, G. linearis ^.nd pectinata, LopJiosoria, 

 and AlsopJiila shows a progression from a radiate uniseriate, simple sorus, 

 through an unpractically crowded type, to a gradate state: but always the 

 position was superficial, and the sorus naked. What more probable biological 

 step should then be taken than that, in high-growing tree-ferns with exposed 

 leaves, a basal indusium should be formed as a new structure, to protect the 

 young sporangia at the base of the elongated receptacle? This would then 



