Elkins and Wieland — On Devonian Wood. 73 



this is inferred from the appearance of a single oblique slit in 

 each pit of the raj cell. In examining the pits on the walls of 

 the raj cells it is occasionallj possible to detect the border of 

 the pit in the adjoining tracheid wall ; or depending on the 

 plane of the section the border is indistinct and onlj a slit or 

 broader aperture is visible. However the relation of the slit — 

 semi-bordered pits to raj cell and tracheid can alwajs be clearlj 

 discerned, the slit appearing first when seen from the interior 

 of the raj cell, the semi-border when the view is from the 

 tracheid side. The shallowness of these pits and narrowness 

 of slit combined with greater distance apart renders them 

 virtually invisible in the transverse and tangential section, 

 even when highlj magnified ; though some evidence of them 

 maj be observed as illustrated in figure 8. 



I, of course, have some hesitation in describing the raj pit 

 features, as so thorough and careful a student of gjmnosperm 

 structure as Thomson (loc. cit.) sajs : 



" The medullar j raj cells are characteristicallj thin walled in 

 the Araucarians, and are comparable to those in the Cordaitean 

 forms. Thej are also unpitted in the Araucarineae as in the latter. 

 Often these cells have been wronglj spoken of as having pits on 

 the lateral walls where thej come into contact with the tracheids, 

 some anatomists even going so far as to distinguish the pitting 

 as of the bordered tjpe. Gothan has noted this especiallj in the 

 case of Penhallow when discussing the latter's statement that 

 there are raj tracheids in Cordaites Clarkii. He sajs that all 

 Penhallow's raj cells are tracheids on account of their pitting. 

 With this as a basis and the fact that the raj tracheids have not 

 been figured he goes so far as to question this unique example 

 of their occurrence in the Cordaitean forms. I have carefullj 

 examined the tjpe set of sections and have found no evidence to 

 contradict Gothan's conclusion." 



However this maj be, and the point is indeed of far-reach- 

 ing import in determining the relationships amongst ancient 

 gjmnosperms and the modern lines of descent, I can onlj 

 figure the structures as shown in fig. 8, which appears correct. 

 That the raj parenchjma cells varj greatlj in length and 

 development is of course true, as fig. 9 well shows. 



The contents of the raj cells occur as dark opaque masses 

 (tyloses) almost entirety filling the cell or scattered like little 

 droplets. These masses have been described as resinous or 

 starch j matter, and their appearance does stronglj suggest a 

 resinous vacuolation. Penhallow states that the raj cells of 

 C. Newoerryi are of one kind onlj, resinous and starch bear- 

 ing.* E. C. Jeffrej, in describing an Araucarian genus, sajs 

 that the rajs are not clear in transverse section ; certain dark 

 spots occur, which are not resiniferous parenchjma, but car- 

 * North American Species of Dadoxylon, p. 64. 



