110 
At the earliest stages of their development which I examined, the 
haustoria form slight excrescences from the cortical portion of the 
Gerardia root, in such a way that the epidermis forms an unbroken 
covering of both root and haustorium. On young haustoria I noticed 
some papillose superficial cells which looked exactly like short root- 
hairs, but their principal function evidently is to take hold of some 
contiguous foster-root. The youngest haustoria do not contain any 
vascular tissue, but soon there appear in the axial portion single 
threads of reticulated vessels connecting the apex of the haustorium 
with the fibro-vascular tissue of its motlier-root. At the same time 
the haustorium increases in all its parts and finally appears as figured 
on the plate. 
As I called the haustorium an outgrowth of the root, and as the 
tissues of the latter are continuous with those of the former, I will, in 
the first place, give a brief description of the root of our Gerardia. 
In the centre of the root there is a fibro-vascular cylinder; its 
diameter is usually equal to about one-half the diameter of the entire 
root. This cylinder contains very wide reticulated vessels and thick- 
walled wood-cells; it is surrounded by a narrow zone of cambium 
, beyond which we find large parenchymatous cells with many inter- 
cellular spaces. Irregularly scattered in the parenchyma we see 
numerous sclerenchyma cells, singly or in groups from 2 to 6, mostly 
of prismatic shape; their walls are exceedingly thick and pierced by 
branching canals. The zone of parenchyma around the central 
wood cylinder is divided about midways by a layer of closely arranged 
cells. This layer seems to be a true endodermis; its cells are very 
narrowcompared to the surrounding parenchyma cells; the partitions 
separating them are strictly radial and frequently undulating, and 
the cell-walls resist the action of sulphuric and even of chromic acid. 
Inside of this ring the starch granules in the parenchyma cells are 
much smaller than in those outside of it, where they are quite large. 
At the circumference of the root there is a single layer of epiderm- 
oidal* cells the outside walls of which are thickened and even more 
resistant against the action of the reagents mentioned above than are 
the walls of the endodermis. The surface view of the root shows 
that these cells are arranged in oval, frequently elongated hexagonal 
plates or patches. The middle portion of such a group is formed of 
rows of two or three cells each; the cells are separated by radial (and 
vertical) partitions which are still more wavy than those found in the 
endodermis. 
At first I was inclined to consider this epidermoidal layer as the 
true epidermis, and to explain the absence of root-hairs by the para- 
sitic nature of the plant; but, on closer examination, I soon found 
that the supposed epidermis was covered with a continuous layer of 
very long, exceedingly thin-walled, empty cells, which, on the cross- 
section, show quite an irregular outline (Figs. 5 and 7, e). Owmgto 
the tenderness of their membranes it is not easy to obtain transverse 
sections that will show these cells completely encircling the root, and 
♦This term is used by v. Hohnel, Olivier and others to designate a layer o 
cells close to the epidermis, that resembles in most respects the endodermis. 
