THE FILICALES 283 



naturally be expected, much more generally in Paleozoic types 

 belonging to the Filicales. 



Before proceeding further with the question of the organization 

 of the tubular or siphonostelic type of central cylinder, it will be 

 well to discuss the important problem of the origin of the pith 

 or medulla in the Filicales. The situation in this group is much 

 more favorable than it is in the Lycopsida, living and extinct, 

 because the facts at our disposal are much fuller and as a conse- 

 quence put the whole subject on a more satisfactory footing. 

 There are two views in regard to the origin of the medulla in 

 stems of the type here designated siphonostelic. One hypothesis 

 considers it as originating from the central region of the xylem 

 of the stele by the transformation of tracheary elements into 

 parenchyma. It has been shown in earlier chapters that there 

 can be no question that parenchymatous elements may be derived 

 from tracheids in the primary wood, and the evidence in this 

 direction clearly points to a possible origin of the medullary tissues 

 from transformed tracheids. Indeed, in the case of the lepidoden- 

 drids the facts which favor the tracheary origin of the medulla 

 are not without importance. Therefore, if we are to regard the 

 medulla as derived from tracheary tissue on the basis of the exist- 

 ence of transitional stages between tracheids and parenchyma, 

 it will be only fair on the other side to assume that where elements 

 clearly of a cortical nature occur in the region of the pith they 

 afford evidence for the other hypothesis of the origin of medullary 

 parenchyma, namely, its derivation from the included cortex. 

 It is, of course, illogical to admit evidence for a derivation of 

 medullary structures from tracheids and at the same time to 

 reject equally cogent data as to the origin of the pith from cortical 

 tissues. Even in the lepidodendrids the evidence for the trache- 

 ary derivation of the medullary structures is not complete, 

 for transitions from tracheids to parenchymatous cells are pre- 

 sented only by protostelic stems and are entirely absent in 

 tubular cylinders. Indeed, in cases where secondary growth 

 is not characteristic of the siphonostele in this group, the medullary 

 region of the stele is frequently occupied by sclerenchymatous 

 elements similar to those which occur in the cortex. The hypothesis 



