THE FOSSIL OSMUNDACE^E. 469 



important stele. The xylem mass is more or less oval in outline, without any prominent 

 angles, and is entirely composed of tracheae. There is, however, a very obvious distinc- 

 tion between the elements of the central region and those of the peripheral zone (fig. 29, 

 C. xy. and o. xy.). This difference is still better seen in longitudinal section (fig. 30), 

 which shows that the outer xylem consists of normal elongated pointed tracheae of fairly 

 large diameter, with porose pitting. Those of the central mass, on the contrary, are 

 short, blunt- ended, and vesicular. They vary considerably in size, but are narrower 

 than the elements of the peripheral zone, and they undoubtedly possess typical tracheal 

 markings (fig. 31). We have here, then, a Zygopterid stem with solid xylem fulfilling all 

 the requirements made by our supposition. 



The Zygopteridean Leaf-trace. 



The greatest stumbling-block in the way of this attempt to conceive of the Osmund- 

 acese and the Zygopteridese as converging towards a common ancestor, is the difficulty 

 found in deriving such remarkably specialised types of leaf-trace as those of the 

 Zygopteridese from the simple form held to be primitive in the Osmundacese. We 

 believe that a valuable suggestion to this end is provided by those leaf-traces of 

 Thamnopteris which in the neighbourhood of the stele possess two immersed proto- 

 xylems situated at points more or less apart from the centre of the strand (Part III., 

 PI. V. figs. 33 and 34). If these traces were to retain this structure throughout the free 

 petiole, and if their protoxylems were somewhat nearer the ends of the strand, the trace 

 would present an appearance very similar to that of the fossil petiole Clepsydropsis 

 antiqua (text fig. 7). Now, Dr Paul Bertrand, in his recently published monograph 

 on the Zygopteridean petiole,* has brought forward a great weight of evidence to prove 

 that Clepsydropsis must be included in the Zygopteridese, and, further, that it possesses 

 the most primitive type of Zygopteridean leaf-trace, from which all the other types may 

 be directly or indirectly derived. Dr Bertrand reaches this conclusion by a series of 

 arguments entirely different from those which we have employed, and we only utilise 

 his results in so far as they advocate the acceptance of Clepsydropsis as one of the 

 true Zygopteridese, although of a very simple type. Setting out, then, from our assumed 

 primitive Zygopterid leaf-trace with a more or less oval or oblong xylern strand and an 

 immersed protoxylem near each end (text fig. 6), as suggested by the above-mentioned 

 leaf-traces of Thamnopteris, it is only necessary to modify its form along two directions 

 in order to bring it into close conformity with the general ground-plan of the more 

 complex Zygopterid types : — (l) The two ends of the xylem strand (as seen in transverse 

 section) have to become more or less extended at right angles to its long axis. (2) Islands 

 of parenchyma must appear in relation to the protoxylems, which also become extended 

 in the same plane as the ends, each protoxylem dividing into two, which pass to the 

 anterior and posterior ends of their particular islands. 



* Etudes sur lufronde des Zygopteridees, Lille, 1909. 



