452 Bower . — Studies in the Phytogeny of the Fiticales. 
bore on their surface the bases of arrested leaves ; but some grew upwards 
at once. Anatomically they ultimately repeated the structure of the main 
axis. But at the extreme base Stenzel describes the stele as a simple strand 
(‘ einfacher Faden ’), and he explains that only where the branch emerges 
from the cortex of the parent plant does the stele widen out, ‘ like a filter/ 
and then, growing on as a cylinder, forms the vascular tube of the branch. In 
this case the structure of the young runner appears to have been more rudi- 
mentary at first than that in Lophosoria or Metaxya ; it corresponds more 
nearly with that in Cibotium Barometz , to be described below ; but the con- 
tracted region of the stele is longer continued in the Alsophila than in that 
Fern. Doubtless there is considerable variation between species in this, 
which is a feature probably dependent in some degree upon the strength of 
development of the individual runner. 
These facts from Alsophila and Hemitelia suggested a like inquiry for 
Cyathea , and naturally the soboliferous species Cyathea mexicana , Schlecht 
and Cham, was used, material being available from the Glasgow Garden. 
The buds are very numerous ; their position is sometimes below the leaf- 
base, as in Lophosoria ; sometimes more than one may be found there. 
Transverse sections of the upper part of one of these shoots show the 
ordinary Cyatheaceous structure. Towards the base it closes into a soleno- 
stele which narrows basally, but in the material examined it was not found 
to contract to the Lindsay a condition. It is thus seen that all the genera 
of Cyatheae may form runners, as a rule on the abaxial side of the leaf-base ; 
and in all of them the basal region shows a simpler vascular arrangement 
than the mature Cyatheaceous structure. In all of them the solenostelic 
type is present at least for a greater or less distance from the extreme 
base. 
In Lophosoria and Metaxya we see two genera naturally related to the 
Cyatheae, with which they have always been classified. But they differ 
from them (i) in their habit, which in Metaxya is permanently trailing, 
in Lophosoria it is temporarily so in the runners ; (ii) in their pronounced 
solenostely ; (iii) in their undivided, or in Lophosoria only slightly divided, 
leaf-trace ; (iv) in their dermal appendages being hairs, not scales ; (v) in the 
simple character of the sorus ; (vi) in the details of their sporangia. All 
these characters indicate for them a relatively primitive position at the base 
of the Cyatheoid series, while they link it to the Gleicheniaceae. It was 
concluded in a previous memoir (‘Ann. of Bot.,’ xxvi, p. 269) from com- 
parisons of Lophosoria on the one hand with the Gleicheniaceae, and on the 
other with the Cyatheae, that the creeping habit was relatively primitive, 
and that the upright axis of the Cyatheaceous Tree-ferns was a secondary 
derivative from it. This conclusion is very greatly strengthened by the 
additional facts from Metaxya. For this Fern shows dorsiventrality more 
fully and permanently than Lophosoria , and with it the solenostelic struc- 
