246 Holden,— Some Fossil Plants from Eastern Canada. 
horizontal after it leaves the pith is demonstrated by the fact that a given 
transverse section of a stem sometimes includes a trace almost from pith to 
cortex. This ‘ hump ’ of the foliar strand is a perfectly constant phenomenon, 
each scar being characterized by the resulting slit. 
The nature of the pith escaped the notice of Weiss, but, as recorded by 
Potonie, it is marked by anastomosing diaphragms. This feature has been 
referred to in connexion with Figs. 6 and 9 ; Fig. 14, a longitudinal section 
of the specimen represented in Fig. 10, also demonstrates their presence. 
The ligneous structure of this interesting fossil was ascertained both 
from sections cut from undoubted Tylodendra, and from pieces of petrified 
wood from the Sable River. It agrees exactly with that described by 
Dawson from Mr. Bain’s specimen as Tylodendron Bainp and with that 
described by Potonie as A raucarioxylon rhodeanum? Goepp. As repre- 
sented by Figs. 17 and 18, the tracheides are pitted throughout their entire 
length, usually by one (Fig. 18), rarely by two (Fig. 17) rows of closely 
approximated pits. Further, the mouth of each pit is circular. As shown 
in Fig. 19 the rays are abundant, low, and normally uniseriate, though 
rarely diseriate. 
There have been numerous theories as to the affinities of Tylodendron. 
Its closely compressed and alternating pits clearly affiliate it with Araucario- 
xylon, Krauss. The question then arises whether it be related to the 
Cordaitales or to the Araucarineae. There are several criteria on which it 
has been proposed to differentiate the woods of these groups. Felix 1 2 3 
attempted to do so on the number of pits on the radial wall of the tracheides : 
in Cordaioxylon there are generally three or four rows, while in Araucario- 
xylon there are normally but one or two. Gothan 4 enumerates several 
differences : in Cordaioxylon the mouth of the pit is elliptical, while in other 
Conifers, at least in the spring wood, it is usually circular. Further, the 
medullary rays of the Cordaitales are usually thirty to fifty cells high and 
two cells in width, as opposed to the Araucarineae, where they are rarely 
over ten cells in height or a single cell in width. . On all these criteria, 
Tylodendron is an Araucarian Conifer. Further, as pointed out by Potonid, 
the nodal swellings and instanding protoxylem strands causing the ridges 
and furrows of the pith casts are identical with similar structures in Araucaria 
and Agathis. Moreover, such characteristically Cordaitean features as the 
double leaf-trace, and the broad transitional region between primary and 
secondary xylem, are replaced by the Araucarian single trace and small 
amount of primary wood. Impressions present more evidence for merging 
1 Dawson : New Plants of Erian and Carboniferous, and Character and Affinities of Palaeozoic 
Gym., p. 13. 
2 Potonie, loc. cit, 
3 Felix : Ueb. d. verst. Holzer von Frankenberg, 1883. 
4 Gothan, W. ; Zur Anat. lebender u. foss. Gymn-Holzer, 1905, p. 12. 
