Elhins and Wieland — On Devonian Wood. 69 



as narrow bands, one to three tracheids in thickness ; these 

 tracheids are distinguished by their radially shortened dia- 

 meters, fig. 2, photomicrograph 1. On either side of the ring 

 the tracheids quickly though not abruptly attain the normal 

 tracheid dimensions. The slight curvature of the rings affords 

 the only means of judging which may be the outer ones, and 

 indicates a trunk of large size. This type of ring is practically 

 identical with that of modern Araucarise. The significance of 

 this structure in Cordaites is not fully known, though in the 

 judgment of both the present writers it obviously indicates 

 some kind of seasonal or other climatic variation, such rings 

 not so far being known to have any relation to fructification. 

 Whatever the cause, it often failed to affect cells in the growth 

 ring area ; for bridges of full sized tracheids are frequently 

 observed crossing the growth rings. Scattered cells with 

 strikingly short radial diameters also occur between the rings. 



The tracheids in cross section are usually rectangular ; some, 

 however, are circular, evidently at the points of constriction, 

 which as one may plainly see in longitudinal sections both 

 radial and tangential occur quite regularly near the tracheid 

 ends. There is, however, so much variation in size that it 

 scarcely seems feasible to use tracheid dimensions as a specific 

 character, except as the upper and lower limits of size varia- 

 tion may corroborate the diagnosis of a species already deter- 

 mined from more fixed characters. This size variation does 

 not include rows of seemingly small tracheids extending 

 radially, which are in reality normal tracheids appearing in 

 this section near their termination, where the diameters, par- 

 ticularly the tangential diameters, are shortened (figs. 2 (£), 5 

 and 7). The thickness of the walls, also, can hardly be deter- 

 mined accurately since they either vary or may, as silicified, 

 appear somewhat irregular and ragged; although Penhallow 

 in his description of Dadoxylon species gives the thickness of 

 cell walls and cell diameters as specific characters. 



By far the most interesting and finely preserved structure is 

 the bordered pit. These pits are strikingly arranged in radial 

 bands not varying markedly either in width or distance apart 

 (fig. 3, photomicrograph 1). The number of vertical rows of 

 pits varies generally with the size of the tracheid from one to 

 four. The position of the growth ring can readily be dis- 

 cerned in a radial section by the occurrence of one or more 

 tracheids of comparatively short radial diameter, containing a 

 single vertical row of pits sometimes slightly smaller than the 

 others, figs. 3 and 4. But the single row usually conforms to 

 the radial band arrangement of pit grouping in each of these 

 lesser tracheids. The number of pits to a group may be as few 

 as three in the growth ring tracheid up to as many as forty on 



