Cc. A. BARBER. 15 
In older haustoria of Santalum the cells along the top of 
the nucleus become crushed and form a well-marked collapsed 
layer (Sant. IT, Plate XI, fig. 2), and this is sometimes the case 
in Cansjera. But in the latter case the transverse band of 
cells more frequently becomes separated irregularly and dis- 
integrated to a clear, yellowish brown mass (Plate IX, fig. 1, 5). 
More interesting is the presence in many sections of a small 
cap of brilliant vellow substance of the end of the host’s dis- 
rupted cortex, and it is in most cases quite easy to see that this 
cap 1s made up of crushed haustorial cells (Plates V, fig. 2, 2, 
and VII, fig. 4). In Olax and Santalum crushed masses of 
tissue are usual at this place, but they remain as such and are 
not fused into a yellow mass. 
With these cases in view, it 1s not surprising that the col- 
lapsed layers themselves are sometimes converted into yellow 
bands, and haustoria have been noted in which a clear transi- 
tion can be traced from the white walls of freshly added cells 
on the inner side of the yellow fused mass of older tissue out- 
side. On Plate VI (fig. 6) a haustorium is drawn in which 
the collapsed layer has undergone complete disorganisation. 
The cell walls of which it was composed have become a clear 
yellow band. This yellow band is continuous downwards 
along the inward projection to the yellow cap at the end of 
the host’s wing, and this last may be traced, through the clear, 
yellow substance which lines the inner side of the host’s wing, 
to the yellow mass which is always found opposite the end of 
the sucker lobes. We have thus an instructive continuity 
between the yellowing of decaying haustorial tissues and the 
film of yellow ‘‘secretion’’ which spreads over the host’s 
tissues wherever they are in contact with the sucker cells. 
This substance is again referred to in paragraph 18, where an 
attempt is made to trace its origin. 
11. Internal Meristems.—The vellowing of the tissues 
described in the last paragraph is not infrequently accom- 
panied by activity in the neighbouring cells which had, to all 
2 
