A REMARKABLE FOSSIL TREE TRUNK 329 



the characters of the cuticle and axial impression of the rootlet 

 agree equally well. In its present condition of preservation on the 

 slab the flattened base of the trunk is subtruncate. It is therefore 

 impossible to determine with certainty whether the tree was pro- 

 vided with the four radiate principal forking roots characteristic 

 of the later Lepidodendreae and Sigillariae. However, since the 

 need for support for so large a tree seems to demand the aid of 

 some type of root we are justified in assuming that the trunk was 

 originally supplied with a more efficient root buttress than the small 

 rootlets now remaining; and since the latter appear to correspond 

 in essential characters to the rootlets of the Carbonic Stigmaria, 

 and since this type is known to be present in the Lower Carbonic, 

 it becomes highly probable that the Naples tree was sustained by 

 roots essentially Stigmarian in character. A notable feature of the 

 Devonic trunk is the large size of the rootlets and areoles as com- 

 pared with the size of the trunk. 



Enlarged base. As shown in plate I, the base of the trunk is 

 very much swollen in a way suggesting the boles of the royal palm. 

 Throughout the extent of this enlargement the cortex shows evi- 

 dent signs of expansion with the advanced growth of the tree by 

 which the epidermis was probably ruptured and the original rows 

 of leaf cushions were widely separated and displaced. This feature 

 is well shown in plate 4, in the lower part of which, beginning just 

 above the area shown in plate 3, figure 1, the cortical impression is 

 seamed by irregular longitudinal and subparallel narrow ridges 

 probably due to the presence of hypodermal strands. In the upper 

 part of plate 4 the impression of the lower side of the trunk is 

 marked in low rounded, ill-defined, and indistinctly rhomboidal 

 prominences, between which lie the now widely separated and 

 oblique rows of leaf cushions (P). The rupture of the distended 

 cortical tissue between the at first oblique and then mostly vertical 

 leaf cushion rows is more sharply defined a few centimeters higher 1 

 where, as illustrated in the lower part of plate 5, we have a cortical 

 condition similar to that often presented in the dilated basal portions 

 of the trunks of some Carbonic Sigillariae. 2 Whether the basal 

 dilation of the trunk in the Naples tree was concomitant with a 

 development of secondary (exogenous) wood, can not be deter- 

 mined in the absence of specimens so petrified as to show the 

 internal anatomy. I am, however, inclined to believe that it may 

 have attended a secondary growth. 



1 See plate 2. 



2 Similar seamed and welted phases of the distended and ruptured bark were described, 

 though not properly, by Lesquereux as Ulodendron. 



