404 Boodle . — Anatomy of the Schizaeaceae . 
of stelar structures are arrived at. One of these is that Van 
Tieghem’s three types of central cylinder are modifications 
of the siphonostelic type, i. e. a vascular tube with leaf-gaps. 
Basing one’s terminology on the position of the endodermis, 
this is evidently the case for the examples investigated, and 
is probably generally true for the polystelic type. But it 
must be recognized that the homology of different types 
of stelar structure is a difficult question, and is not at once 
disposed of by merely tracing the continuity of the various 
tissues. Perhaps in the present state of our knowledge it 
does not admit of any certain solution. 
One would be inclined to describe Fig. 31 as a horseshoe- 
shaped stele with a projection of ground-tissue in its con- 
cavity, and to regard the endodermis as limiting the ground- 
tissue towards the stele. This is no doubt a convenient 
description, but there is no proof that it is anything more 
than a physiological use of terms, which does not necessarily 
agree with the morphology of the parts. The central ‘ ground- 
tissue 5 is certainly continuous with, and similar to, the external 
ground-tissue, but one must regard it as possible that in this 
case the stele may really have a roundish outline, and that 
part of its tissue has been differentiated as parenchyma with 
the characters of ground-tissue, and limited by a piece of 
endodermis formed from stelar tissue so as to join the external 
endodermis, which is interrupted opposite the stelar paren- 
chyma. The endodermis is usually formed from the inner- 
most layer of the cortex 1 , but that the endodermis is not 
dependable as a morphological limit in certain other cases 
is sufficiently shown by the species of Equisetum 2 as pointed 
1 In several Ferns at any rate the peri cycle is formed by subdivision of the same 
layer, as described by Van Tieghem (’88, p. 404) for the stem of Hymenophyllum 
and the stolon of Nephrolepis, and the same is probably the case in Lygodium. 
Other cases are quoted by Strasburger (’91, p. 446), but these being polystelic 
involve other questions. 
2 The parenchyma-rays between the bundles would have a different morpho- 
logical value in nearly allied species (gamodesmic and dialydesmic), if the endo- 
dermis were regarded as always limiting stelar tissue. Jeffrey (’99, p. 157) regards 
Strasburger’s conclusions as invalidated by the structural facts cited by Van 
Tieghem (’90), but the case just mentioned appears to weigh heavily against the 
morphological value of the endodermis. 
