74 GYMNOGRAMMOID FERNS [ch. 



500 sporangia observed. The annulus may be interrupted, incomplete, or even 

 doubled in varying degree. The induration may be very limited, and it is 

 stated to be absent sometimes in Ceratopteris. Such signs of instability, not 

 in one genus only but in all, are held to mark a transitional state between 

 large sporangia with an oblique annulus, and smaller sporangia with a vertical 

 annulus. A parallel is found in Platyzoina (Vol. II, p. 209, Fig. 492), which 

 is also a transitional type. It is also seen in Acrostichmn aureinn (Schumann, 

 I.e. p. 220) at the region where fertility of the leaf-surface begins. Lastly, the 

 spore-enumerations give very varied results. It will be remembered that the 

 spore-output of Plagiogyria is only 48, though that of the Osmundaceae and 

 Schi^aeaceae is relatively high. It is then interesting to note that mjaniesonia, 

 which like them has no scales but hairs, the spore-counts show the exception 

 of exceeding 64, so rare among Leptosporangiate Ferns with mixed sori: 

 the numbers recorded range between 48 and 72. On the other hand, the 

 typical numbers in the other genera vary widely. In Llavea and Crypto- 

 gramme crispa they are 48 to 64, but in Ceratopteins Benedict found 16 to 

 32 for C. thalictroides, but only 16 in his C. deltoidcs. Thompson says the 

 number may be anything between 12 and 24. These facts appear again to 

 point to a transitional state of the sporangium. The spores in all of these 

 Ferns — including also Plagiogyria and Osmunda, and most of the Schizae- 

 aceae on the one hand, and the Gymnogrammoid Ferns on the other — are 

 spherical-tetrahedral, often large, with very bold markings of the outer wall 

 (Fig. 631). According to Hannig's record {Flora, Bd. 103, p. 338-9), there 

 is a prevailing absence of a perispore in these Ferns, and this is found to be 

 so also in the Osmundaceae and Schizaeaceae {l.c. p. 342). 



The general result of these comparisons appears to be that the four genera 

 under discussion are probably related to one another, and present many 

 primitive features that point towards such antique types as the Osmundaceae, 

 Plagiogyriaceae, and in a less degree the Schizaeaceae. Further, they suggest 

 that they occupy a transitional position: but inasmuch as the archaic features 

 that they show are not concentrated in any one of them, being in fact 

 distributed irregularly between them, it is difficult to place any one genus 

 as the most primitive of them all. If any such indication can be held to have 

 emerged, it would point to a special relation of Llavea to Plagiogyria or to 

 Osmunda. The general result is, however, to suggest the probability that a 

 Gymnogrammoid state, such as these Ferns actually show, has originated 

 directly from types already having superficial sporangia, such as Todea or 

 Plagiogyria. There is no evidence in any of them of obliteration or of trans- 

 formation "of either or both of the indusia of a two-lipped marginal sorus, 

 such as is seen in the Pteroids. This being so, these primitive Gymnogram- 

 moids must be held as standing phyletically on their own feet, distinct by 

 descent from the Pteroid Ferns. Even the linkage of the sori of OnycJiinm 



