142 W or s dell . — The Structure and Origin of the Cycadaceae. 
dendron , in which the sinuous 1 character can always be more or less clearly 
traced, Fig. 9), and the character of the entire general structure, I regard 
the single large monostele of Heterangium and, I would add, of Megalo- 
xylon , as equivalent to a single stele of Medullosa. The very sinuous 
contour of many Medullosean steles is due to two causes: firstly, to the 
perpetual anastomoses occurring between neighbouring steles, and secondly, 
to the constant abstriction of concentric leaf-trace and medullary bundles ; 
but these two causes will not account for the sinuosity of the central 
cylinder of Lyginodendron. There 
may be traced a gradual reduction 
in the number of steles composing 
the cylinder: (1) the very numerous 
ones of M. Solmsii ; (2). the three 
or four much larger steles of M. 
anglica or M. Leuckarti ; (3) the 
single stele of Heterangmm and 
Megaloxylon which, owing to 
space-relations, is naturally and 
inevitably of much greater di- 
mensions than, while preserving 
an identity of structure with, the 
steles of other members of the 
group. Hence, on my view, each 
individual bundle and not, as Scott 
holds, the entire cylinder, oiLygino - 
dendron, is the homologue of the 
large stele of Heterangium , &c. 2 
At the bases of branches of Lyginodendron , as described by both 
Williamson and Lomax, the primary centripetal xylem-groups tend to 
become continuous (although complete continuity never obtains), while the 
Fig. 9. Lyginodendron Oldhamium : transverse 
section of secondary wood of stele of stem, with 
primary xylem-groups included, showing bulging of 
the ring opposite the primary bundles, due to other 
causes than the natural curvature of the cylinder 
(diagrammatic). 
1 Cf. Williamson’s Figs. 1, 2, 4 (Plate XXII), and 6, 7 (Plate XXIII), of Lyginodendroti with 
his Fig. 30 (Plate XXVIII) of Heterangium (Organ. Fossil Plants of Coal-Meas., Part iv, 1873). 
Also Williamson and Scott’s Fig. 1 (Plate XXI) and their Photo. 1 (Plate XVIII) of Lyginodendron 
in their joint paper of 1896 with Williamson’s Fig. 1 (Plate XXI) in his memoir of 1887. 
Williamson’s Fig. 2 (Plate XXII) of the same memoir, showing part of the transverse section of the 
stem of Heterangmm tiliaeoides, shows a somewhat sinuous contour of the ring, but it will be noted 
that even here the secondary xylem is never arched around each primary xylem-group as it always 
is in the case of Lyginodendron. Whenever a section of the entire cylinder is seen this striking fact 
becomes incontrovertible. 
2 Any argument for the close affinity of Lyginodendron and Heterangium drawn from the fact 
that they both bore a Sphenopteroid type of foliage can have little weight in view of the fact that 
several forms of Sphenopteiis-ioYx&ge belonged to the true Ferns bearing such fructifications as those 
named Renaultia , Discopteris , Dactylotheca , Oligocarpia. 
Again, the stem of some species of Davallia is polystelic, that of other species is solenostelic, 
yet the same type of foliage obtains throughout the genus. It is not, therefore, very strange to find 
a considerable latitude in the character of the stelar anatomy of Pteridosp^rms bearing Sphenopteris- 
foliage. 
