Cc. A. BARBER. 9 
figure on Plate IV supports this view. There is nothing of this 
in the haustorium of Santalum, Although the upper limit of 
the nucleus is quite sharp there, the cells are not thus enlarged, 
being little if any larger than those of the ground tissue above 
them (Sant. IJ, Plates I—III). 
On either side of the nucleus the procambial strands are 
clearly visible ; and their cells are dividing vertically only, 
so as to become elongated in this direction (e), 
The lower part of the nucleus is divided into two parts, 
an outer (f), with small elongated nucleated cells, and a median 
strand with dense granular protoplasm, arranged as an epi- 
thelium on both sides of a narrow slit (7), This is presumably 
the gland in a late stage. 
The young sucker is formed of the extension downwards 
of the lower part of the nucleus, and it has penetrated half-way 
through the host’s cortex on its way to the vascular cylinder. 
Details of all these tissues may be followed on Plate IV, 
where enlarged drawings are made of the cells in different 
regions. 
7. The next stage, where the sucker has reached the 
woody cylinder of the host’s root but has itself not yet formed 
vessels, is frequently met with, and is very characteristic of 
these haustoria (Plate V, fig. 1), 
The cortical folds are assuming their permanent form. 
Growth in length has ceased, but a considerable thickening 
is noticeable on the inner side. Such growth in thickness, 
which is frequent in Cansjera, has not been noticed in either 
Olax or Santalum, 
The collapsed layers are now much thicker, and separate 
the outer, empty cells of the cortex from the central, more or 
less protoplasmic mass. As before, the collapsed layers do 
not pass to the ends of the cortical folds, and abut on the host’s 
bark half-way between the edge of the fold and the point where 
the sucker enters. The upper limit of the nucleus is particu- 
larly sharp, the cells of this region being, as usual, prominent. 
