288 EVOLUTIONARY BEARINGS OF THE RESULTS [ch. 



earlier mutability in all directions." The results of intensive study of Ferns 

 detailed in these Volumes appear to make the demand for mutability in all 

 directions needless for them. Both externally and internally the changes in 

 certain well-marked instances show a definite trend, readily related to the 

 conditions, rather than a "mutability in all directions." 



No change in the Ferns points more clearly to a decision as to the origin 

 of an inherited character than the shifting of the sorus from, the margin of 

 the sporophyll, which comparison shows to have been its original position, 

 to a place on the lower surface which may be held as derivative, and physio- 

 logically advantageous. There is reason to believe that this change has 

 happened independently in many distinct phyla, sometimes early in the 

 evolutionary history (Marattiaceae, Gleicheniaceae): sometimes it appears 

 to be now in transition (Pteroid and Schizaeoid Ferns). That the change is 

 associated with a widening of the leaf-blade is suggested by a comparison 

 of the closely related genera, Osmunda and Todea (Vol. I, Fig. 210). The 

 Schizaeaceae are also very suggestive, for in them, though the sporangia all 

 originate from marginal cells, they acquire in the course of the individual 

 development a superficial position (Vol. I, Figs. 213, 214). But the most 

 illuminating results come from the Dicksonioid-Pteroid series, in which the 

 comparison yields many transitional steps. Starting with the ancient Jurassic 

 type of Thyrsopteris, the gradate sori are marginal upon the narrow leaf- 

 segments, and remain so till maturity (Vol. II, Figs. 527, 529). In Dicksonia 

 and Cibotiuni the receptacle and the first sporangia are seen to be still 

 marginal (Figs. 534, 535), but the bi-labiate and gradate sorus is deflected 

 downwards in the course of individual development on the widened blade 

 (Fig. 530). In Hypolepis the lower indusium is absent or only vestigial, and 

 the receptacle, with its mixed sporangia, is not marginal even in origin, but 

 superficial ; and it is spread over the lower surface of the leaf, thus giving a 

 structure indistinguishable from the superficial sorus oiPolypodinm (Chapters 

 XXIX, XXX, XXXVI, Figs. 584-587). These closely related Dicksonioid- 

 Dennstaedtioid Ferns thus illustrate steps in a "phyletic slide" of the 

 originally marginal sorus to the lower surface of the blade. The protection 

 thus derived gives the biological justification for the change, which has 

 become heritable. It is significant that along with the change of position 

 go also the progressive change to a mixed sorus, and abortion of the lower 

 indusium. 



A second series which shows a like progression is found in the Pteroid 

 Ferns (Chapter XXXVIIl). These are again Dicksonioid derivatives, but 

 they constitute a separate phylum distinguished from the first by prevalent 

 soral fusions to form linear coenosori. The series starts from Pteridinin and 

 Paesia : both of these genera have a marginal receptacle and a two-lipped 

 sorus, as in the Dicksonioids. As in them also the sorus is deflected down- 



