Bower.— On Medullation in the Pteridophytci . 559 
(= L. selaginoides ) from the Coal Measures, where the centre of the stele is 
occupied by a ‘mixed pith’ composed of parenchyma and tracheides. A 
further step is seen in L. Harcourtii , also from the Coal Measures, where 
there is a bulky parenchymatous pith. It is true that Professor Jeffrey, in 
his Memoir on Eqnisetum} states in connexion with his view of the extra- 
stelar origin of pith, that where the pithed Lepidodendroid axis branches, 
the pith and the cortex are in continuity. But he does not appear to have 
gone further into detail on the question of the origin of the pith in the 
Lepidodendreae. It is to be remarked that such continuity, where it exists 
at the branchings, is only one condition necessary for the intrusion of the 
cortex into the stele. It does not demonstrate that it occurs, and it would 
be equally consistent with extrusion of the pith into the cortex, or with no 
such changes at all. In the absence of more cogent evidence than that 
which Professor Jeffrey has hitherto brought, the reiteration of the state¬ 
ment that the origin of the pith is extrastelar does not weigh against the 
comparison of the structure of forms in stratigraphical sequence such as 
those quoted. Palaeontological anatomists generally have accepted the 
stelar origin of the pith in the Lepidodendreae, which seems to me to be 
the only reasonable reading of the facts. 
In presence of this evidence of the origin of an intrastelar pith in the 
Lepidodendreae, the next natural step in testing the truth of the conclusion 
should have been to see whether in their nearest modern correlatives there 
is any evidence of a like process, viz. in Isoetes and Selaginella. I do not 
think that these comparisons have yet been made, though they would seem 
to be a necessary preliminary to any statement in terms of ‘ must in all 
cases’. In Isoetes the anatomical and physiological conditions are so com¬ 
plicated and peculiar that it would not be well to lay much stress upon the 
structure of its primary stelar xylem. It appears to have the character of 
a ‘ mixed pith * with tracheides and parenchyma intermingled. But a better 
basis for comparison is in Selaginella , and it is not to be sought in the 
more specialized dorsiventral types of the genus, but in those of upright 
habit, with radial shoots, which comparison shows to be the more primitive, 
and therefore more directly comparable with the Lepidodendreae. 2 Such 
a species is S. spinulosa. Harvey Gibson 3 has long ago described the 
structure of its stele, and shown how it is first centroxylic, and that as it 
continues upwards the protoxylem strands pass to the periphery of the 
solid xylem-core. But he did not follow it upwards into the erect unbranched 
strobilus, and it is there that the point of structure appears which has 
a bearing upon our present question. If sections be cut through the axis 
below the strobilus, the stele may be found with a solid xylem-core as already 
described by Harvey Gibson, the protoxylem groups being peripheral. 
1 Mem. Eost. Soc. Nat. Hist., 1899, PI. XXVII, Fig. 3. 
2 Land Flora, pp. 300, 356. 3 Ann. Bot., vol. viii, p. 171. 
P p 3 
