70 ROYAL SOCIETY OF CANADA 



many branches of Lepidodendi-a figured by authors, and of this type 

 might, if better known, be found to be branches of Lepidophloios. Yet, 

 though this is possible, there is an equal possibility that they may really 

 be Lepldodendron. These facts, howevei-, lead to the discussion of what 

 is known from structure, form and fructification, of the relationships of 

 the Lepidodendrea; and SigillariîP. in referring briefly to which subject 

 I shall depend chiefly on Canadian examples in my own collections, as 

 the current descriptions and figures of fragmentary specimens by authors 

 abroad do not always furnish reliable data for comparison. 



IV. — EeLATIONS to LePIDODENDRON, UlODENDRON and SiGlLLARIA. 



We have already seen how easy it would be to refer fragments of 

 Lepidophloios to the genus Lepidodejidron, and in regard to internal 

 structure it is probable that branches of Lepidophloios are scarcely dis- 

 tinguishable from those of the simpler styles of Lep)idodendron, in which 

 an outer cylinder of radiating wood is either absent or developed only on 

 the larger stems. The difficulty is added to by the fact that some Lepi- 

 dodendra, as for example, L. ornatissiiymm and Veltheimiamim of Stern- 

 berg, bear sessile lateral cones on stems or large branches, and it seems 

 certain that some j^lants of this group, bearing sessile cones in two rows, 

 which have been referred to the genus Ulodendron of Lindley and Hut- 

 ton, are really portions of Lepidodendra of this tyjie. In my original 

 description of Lepjidophloios Aeadianus in 18G5, 1 was so far influenced by 

 these apparent connections as to include under this genus not only the 

 Lomatophloios of Corda, which is no doubt a synonym, but also the Uloden- 

 dron of Lindley and Hutton, of which two species or varieties are found 

 in Nova Scotia, and some j)lants with leaf-scars, similar to those of Lejndo- 

 phloios, but without the long pendant leaf-bases, and which are now usually 

 classed by paUx'obotanists with the Sigillaria'. With regard to the 

 Ulodendron, it may be stated that the Ulodendra and Ulodendroid Lepi- 

 dodendra cannot be distinguished by the two-ranked cone-scars, because 

 these occur also on Lepidophloios, but rather by the fact that the cones 

 were not stalked but sessile by a broad base,' and that the leaf-bases 

 and leaf-scars were of different form. I illustrate this by figures of 

 two species or varieties of Ulodendron^ corresponding to U. majum 

 and U. minus L. and H., which are found, though rarely, in Nova 

 Scotia. (Plates XI., XII.) In one of them (Plate XII.) the leaves 

 are present, and are more like those of Lepidophloios than that of Lepl- 

 dodendron, while the leaf-bases resemble in general form those of a Sigil- 

 laria of the Clathraria type. The plant is like S. discophora of Kœnig, 



1 It is quite likely that were the structures of the.se cones perfectly known, other 

 differences would appear. 



