4 o8 McNicol.—On Cavity Parenchyma and Tyloses in Ferns . 
disintegrate, and its place may not be taken to any great extent by the 
xylem-sheath cells, so that frequently an actual cavity is seen. 
In the Schizaeaceae Boodle recorded the tissue in Aneimia Phyllitidis . 
I examined A fraxinifolia , which is regarded by Hooker as a subspecies of 
A. Phyllitidis ; in this form there were four groups of cavity parenchyma 
in the petiolar bundle ; the cells were lignified, and the nuclei very distinct. 
I found the tissue also in A. tomentosa but none in Lygodium . 
In Ophioglossum I found no cavity parenchyma, though the protoxylem 
cells become somewhat crushed by the soft cells adjoining them, and 
in Botrychium Lunaria and B. virginianum the same is the case. In 
Helminthostachys cavity parenchyma was recorded by Farmer as noted 
above. Such parenchyma occurs both in the fertile and sterile leaf-stalks, 
but is more evident in the sterile stalks. There are generally four or five 
enlarged cells to be seen in connexion with the protoxylem elements in any 
transverse section. Longitudinal sections show that the cells apparently 
crush the spiral vessels by lateral pressure ; I found no definite tylose-like 
growths, such as occur, for example, in Pteris , though the edges of some 
of the cells were slightly crenulated. Fig. n shows a semi-diagrammatic 
view of a cross section of Helmmthostachys and Fig. 12 is a longitudinal 
section of a single vascular strand. 
In the group of Marattiaceous Ferns Brebner (’ 02 ) has recorded and 
described cavity parenchyma in Danaea. He says: The cavity due to the break- 
ing down of the protoxylem is filled with parenchyma owing to the increase 
in size, accompanied by a greater or less amount of division, of the adjacent 
living cells ... It is probably not of any physiological importance, being 
simply a case of non-pathological hernia, so to speak * (p. 544). In Marattia 
laxa and Angiopteris evecta cavity parenchyma is well developed. In the 
former the tissue can be seen opposite most of the bundles, though not 
always opposite the very small ones. There may be one or more groups 
to each bundle, according to whether it is short or elongated. Generally 
the cells lie in a little ‘ bay ’ and occupy only the place formerly occupied 
by protoxylem, not extending into the phloem ; they do not, in spite of the 
large dimensions of the plants, attain the great size of the same cells in the 
Tree Ferns ; as a rule they are only a little wider than the phloem cells. 
In Angiopteris the tissue is more strongly developed than in Marattia and 
the cells may be three or four times the width of the ordinary xylem-sheath 
cells. The strand is about one cell wide. Fig. 13 shows a transverse 
section of a single bundle of Angiopteris evecta having one protoxylem 
group and a corresponding cavity-parenchyma strand cp : the nuclei 
are distinct. 
In Marsilia quadrifolia among the Water Ferns the tissue is very well 
developed ; it occurs most distinctly near the protoxylem strand at the 
apex of the triangular bundle, as shown in Figs. 14 and 15. At the other two 
