430 Boodle and Hiley . — On the Vascular Structure of 
island, which, at its base, is separate from the inner sclerenchyma, endo- 
dermis, and phloem of the solenostele, and also corresponds in position and 
structure to the nodal islands of other species of Gleichenia . This seems to 
show that in G. pectinata solenostely has not arisen by decurrence from the 
leaf-trace, but independently of the nodal islands. 
i a. Decurrence of phloem and cortical tissues from the axil of a branch 
or dichotomy of the stele is not known in any species of protostelic ferns 
(without pith), and there would not be any probability that decurrence 
should occur at the junction of two solid cylindrical steles, transverse elon- 
gation, lateral constriction, and nipping off being the natural mode of 
subdivision of a solid stele. 
2. In some rhizomes of G. flabellata the tracheae were found to be less 
frequent in the central part of the xylem, the parenchymatous elements 
being complementarily more numerous. In this way, by the exclusion of 
tracheae from the central region, the protostele might become medullated. 
It would then be possible for ramular gaps to be formed by the departure 
of the branches. Through these gaps the cortical tissue, phloem, &c., might 
become decurrent in the pith, thus forming an internal rod of these tissues 
(island in transverse sections), which would be central, if derived from 
a dichotomy. The decurrent tissues might be extended until they met the 
similar tissues at the next lower ramular gap, and a uniform stelar structure 
would be established, which, supposing leaf-gaps to be subsequently formed, 
would then become solenostelic. It may be objected that this theory 
requires an excessive and unlikely amount of decurrence, and for this reason 
it seems more probable that the converse of decurrence, i. e. an acropetal 
extension of tissues (which may be named ‘ procurrence ’), may have been 
more largely concerned in the origin of the solenostely. At a ramular gap 
a slight upward as well as downward ‘ intrusion ’ of the tissues lying outside 
the xylem might occur (cf. the case of Osmunda cinnamonea , which may be 
an example of this kind ; Faull, ’01, Tansley, ’07, p. 263, Kidston and 
Gwynne-Vaughan, ’07, p. 775 ), and, when the meristem has begun to pro- 
duce a certain type of structure, the acropetal extension of the latter would 
not perhaps be very surprising. 
3. It may also be suggested that the solenostely has been developed 
by the rhizome, independently of the leaves and branches. The protostele 
having become medullated in the way described above, some elements of 
the medulla may have been differentiated as sieve-tubes forming a peripheral 
zone of phloem. Then it may be supposed that the inner medulla became 
sclerenchymatous, a change which necessitated the formation of an envelop- 
ing endodermal layer, when the formation of leaf-gaps would complete 
the solenostelic structure. 
On general grounds the rarity (except in cases apparently to be ex- 
plained by reduction) of a primarily disconnected tissue carries some weight 
against the likelihood of the last supposition (No. 3). 
