THE FOSSIL OSMTJNDACE^]. 777 



continuity with the outer. It is quite probable that the medullary endodermis was 

 already present before ever the stele opened at a branch-gap. In fact, Seward and Ford's 

 discovery of an occasional internal endodermis in Todea hymenophylloides (I.e., p. 249) 

 that never comes into contact with the external endodermis seems to give support to 

 this view. If, on the other hand, we assume that the two endodermes are really 

 homologous, it is true that initially the phloem must have found its way into the 

 medulla before the endodermis ; but it by no means follows that the two tissues should 

 subsequently keep pace with each other in their downward extension. The endodermis 

 may have outstripped the phloem. 



However it may eventually be settled, the whole question discussed above provides 

 an instructive example of the importance that should be attached to the proper 

 determination of the morphological status of the various tissues of the stem. The 

 question as to whether the internal ground-tissue of the Osmundacese is to be held as 

 stelar or not may at first sight appear to be a distinction of merely academic interest ; 

 but, nevertheless, the settlement of this point decides whether the order as a whole is 

 to be regarded as an ascending or a descending series. 



The Ancestry of the Osmundacese. 



The presence of several vertical series of pits on the broader walls of the tracheides 

 of both modern and fossil Osmundacese becomes a matter of considerable importance 

 when considering their ancestry, especially since in the primitive fossil species 

 Osmundites Dunlopi the tracheides suggest the presence of a reticulate or even porose 

 pitting. It is clear, in fact, that the presence of such markings on the tracheides of 

 any particular fossil can no longer be regarded as an objection to an Osmundaceous 

 affinity. According to the views already expressed above, the stele of such an ancestral 

 type would have a continuous ring of solid xylem, or, in a still more primitive form, 

 even a solid central xylem mass. In both cases the leaf-trace would depart in a 

 protostelic manner, without interrupting the continuity of the xylem, and the proto- 

 xylems of the leaf-traces would be more or less decurrent into the xylem of the stem as 

 a mesarch strand. It is further probable, in the more primitive forms, that there would 

 be a peripheral exarch protoxylem of small elements proper to the stem, and apart from 

 the mesarch protoxylem of the leaf-traces. In many of the species already known, the 

 tracheides of the xylem ring diminish markedly in size towards the periphery (Osmunda 

 regedis, Osmundites Duulojn and 0. skidegatensis) ; although it must be admitted that 

 in the living Osmundacese the differentiation of the metaxylem elements takes place 

 quite irregularly after the decurrent leaf-trace protoxylems are once fully developed. 



If a fossil stem possessing the more primitive of the above-described characters were 

 ever to be found, it would very probably be classed at first sight with the Botry- 

 opteridess. This is a conclusion with which we would at once agree, for we regard the 



