152 
Lady Isabel Browne. 
since they are characteristic of the Upper Devonian and Lower 
Carboniferous rocks. But last year Mr. White described a form, 
coming from the Devonian of New York, in which the stem bore 
both the Lepidodendroid and the Sigillarian types of marking (28). 
The fact that such an ancient type should combine the characters 
of Sigillaria and Lepidodendron suggests that it, or allied forms, 
were among the direct ancestors of these two genera ; but this fact 
tells strongly against the tempting view that Lepidodendron and 
Sigillaria were evolved separately from a type resembling Bothro- 
dendron. Unfortunately Mr. White has not yet published a full 
account of this curious fossil, so that we do not know how closely 
it agrees with Mr. Kidston’s definition of Archaeosigillaria. 
Another striking peculiarity found within the order is the 
pitting of the tracheides in the stem of Renault’s Sigillariopsis 
Decaisnei (21). This species is the only one of its phylum in which 
such pitted tracheides are known, though they are common in other 
phyla of Pteridophytes. Their significance is thus best considered 
in discussing the inter-relationships of the phyla. 
The leaves of Lepidodendron and Sigillaria are ligulate and 
and linear-lanceolate; those of Archaeosigillaria appear to have 
been deltoid-lanceolate, and no ligule has yet been recognized in 
them, though its presence may yet be demonstrated. Renault 
was the first to describe a double leaf-trace in the leaves of his 
Sigillariopsis Decaisnei. This bifascicular character of the leaves, 
coupled with the occurrence of pitted tracheides in the stem, had 
thrown doubt on the Lepidodendraceous affinity of this species. The 
pitted tracheides of the stem have not so far been recognized in any 
other examples, but Dr. Scott described in 1904-two better preserved 
examples of Sigillariopsis, leaving little room to doubt the Lepido¬ 
dendraceous affinity of these leaves (16). Further, Mr. Kidston’s 
and Mr. Newell Arber’s recent investigations make it very probable 
that the leaf of Sigillaria normally contained two bundles, and that 
the numerous recorded instances of Sigillarian leaves with a single 
bundle are due to the fact that the bundle entering the leaf 
remained single for a varying portion of its course (10). Until Dr. 
Scott had described his examples of Sigillariopsis (or Sigillaria as 
we may now call them) from the Lower Coal Measures, the only 
example of a bifascicular Lepidodendraceous leaf was Renault’s 
Permian 5. Decaisnei ; it was therefore natural to regard this 
character as a later development having no relation to the phyto¬ 
geny of the more primitive members of the order. But now that 
