THE FOSSIL OSMUNDACE^E. 



663 



nature, and the additional evidence provided for the complete continuity of these 

 tracheae up to the very centre of the stele. In the face of these facts it seems to us 

 almost impossible to avoid the conclusion that the pith of the Osmundacese has been 

 derived directly from an originally solid xylem mass. This xylem, we believe, 

 primitively consisted of normal elongated tracheae throughout, and we regard Zalesskya 

 and Thamnopteris as representing one of the stages passed through by the central 

 tracheae on their way to become short-celled medullary parenchyma. In large and 

 solid masses of xylem the efficiency of the central elements as water-conductors seems 

 in general to decrease the farther they are from the periphery ; possibly owing to their 

 increased distance from any of the living cells of the stele, so that the total elimination 

 of this function in the most central elements would not be attended by any serious 

 disadvantage to the water current. Again, from their central position these elements 

 can render but little assistance towards the support of an erect stem, such as we believe 

 these stems to have been (this may also be a reason why the primitive and 

 mechanically inefficient porose and reticulate types of perforation are retained by the 

 central tracheae for a longer time than by the peripheral). It seems clear, therefore, 

 that an economy would be effected if their walls were no longer thickened at all, and 

 especially if they could be converted into some other kind of tissue with different and 

 more useful functions. The exact manner in which the change was brought about has 

 yet to be made clear. In Thamnopteris, however, there are distinct indications that 

 the central tracheae were originally arranged in vertical files, the end members of which 

 are more or less pointed. This would suggest that the still meristematic central 

 elements of the desmogen strand underwent septation during, or shortly after, their 

 elongation. 



In their further alteration into a parenchymatous pith the short central tracheae of 

 Zalesskya and Thamnopteris would have to completely lose their thickening, and in 

 consequence their pitting. This may take place in two ways : either the tracheae may 

 all become transformed simultaneously, resulting directly in a homogeneous pith, or 

 else some of the tracheae may lose their tracheal characteristics before the rest, which 

 would result first of all in the formation of a "mixed pith" of short tracheal elements 

 scattered amongst parenchyma. If the latter were the course pursued, Osmundacese 

 must have existed with a central stelar tissue very similar to that which occupies the 

 axis of the stele of a Zygopterid stem. Indeed, if the transformation of the central 

 xylem were to extend so far outwards as to reach the immersed protoxylems, a stelar 

 structure almost exactly similar to that of Zygopteris corrugata would be attained.* 

 If the steles of the Osmundacese and the Zygopteridese have evolved along parallel 

 lines from the solid xylemed stele of the common ancestor we have assumed for the 

 two orders, the central tissue of the Zygopterid stele must be regarded as a mixed pith 

 rather than as a mass of confluent diffuse protoxylems. Further, if the Zygopterid 



* A comparison that has also been made by Tansley in his lectures on the evolution of the Filicinean vascular 

 system, New Phytologist, vol. i., p. 260, 1907. 



