D. H. Scott. 
28 
of the wood, except where they bend out to contribute to the leaf- 
traces. This displacement of the primary wood inwards probably 
indicates some change of function. Pitys , however, is an example 
of the retention of the old form of wood (reduced and somewhat 
altered, it is true) in trees which probably had already become 
Gymnosperms. 
Poroxylon (Upper Carboniferous and Permian) is pretty 
completely known, so far as the vegetative organs are concerned, 
and is interesting from its evident affinity to the Cordaiteae. The 
stem-structure is about at the average grade of the genus 
Calamopitys. The leaf-trace xylem-strands are almost wholly 
centripetal in the upper part of their course, but the centripetal 
wood dies out about eight internodes below their entrance into the 
stele. The leaves, though simpler in form, were structurally much 
like those of existing Cycads. 
Dadoxylon Spenceri (probably Upper Carboniferous) has a dense 
wood of very Coniferous appearance; the leaf-traces pass out in 
pairs, as in the recent Maidenhair-Tree (Ginkgo). At the border 
of the pentagonal pith, near the angles, we find a few very small 
primary xylem-strands, which, where they approach their exit as 
leaf-traces, have a distinct mesarch structure, though elsewhere 
centripetal xylem cannot be detected. 
it is interesting to note that in Ptychoxylon (Permian) a genus 
somewhat remote from the series we are considering, but with 
undoubted Cycadean affinities, the leaf-traces only retain their 
centripetal wood in passing trhough the cortex of the stem, and have 
completely lost it before they enter the stele. 1 
Lastly, we come to Cordaites itself, which is anatomically 
on the same level as recent Cycads; centripetal wood has wholly 
disappeared from the stem, while it still forms the main constituent 
of the xylem in the bundles of the leaf. This statement applies to 
those specimens, from the Coal-measures and Permian, which have 
been referred with certainty to this family. As has been pointed 
out above, it is extremely probable that some of the stems which 
retained more or less of the old wood, also belonged here. 
’This stem bus a very complicated secondary structure, with 
bands of inverted vascular tissues around the pith, but, as 
Renault has shown, the centripetal wood which occurs in 
this position is secondary, and has nothing to do with the 
Cryptogamic or primary centripetal wood. See Flore 
fossile d’Autun et d’Kpinac. Pt. 2, p. 410 
