18 STUDIES IN ROOT-PARASITISM. 
offered as a full explanation of the phenomenon, and it is almost 
certain that other factors enter into the case. Such sharp lines 
of separation do not usually occur in homogeneous tissues rising 
from the same meristem, however much their contents may be 
wanted by the adjoining tissues (see for instance the action of the 
scutellum in emptying the endosperm of seeds), and there is 
evidently a profound alteration in the character of the cells in 
the clear zone at this time. Their after history shows that they 
are of no further use to the haustorium and take no further part 
in its activity. The remarkable clearness of the cells, taken 
together with their evident turgidity, lastly, suggests the 
possibility that the clear zone may act as a_ sort of 
transfusion tissue. It has already been noted that the lower 
part of the gland is widely removed from the vascular strands 
coming from the mother root (Plate VI, fig. 1). The formation 
of such masses of secretion as are seen in the final stage of the 
gland will require a considerable amount of water. Such asupply 
cannot presumably be rapidly passed from above—because of 
the absence of vascular continuity, the great development at the 
same time taking place in the upper part of the nucleus and the 
density of the glandular “cap.” It is possible then that the 
clear zone represents a water reservoir on which the lower swell- 
ing part of the gland may draw when it reaches its period of 
maximum activity. 
It should, however, be remarked that sharp lines of 
separation between starch-laden and clear cells are occasionally 
found in other parts of the haustorium of Olax scandens, although 
with less precision. In an older stage, for instance, as shown on 
Plate VIII, where the entry of the sucker into the host’s root 
may be presumed to draw rapidly upon the surrounding tissues, 
the inner cells of the cortex become emptied of their contents 
and the line of separation between the clear and starch-laden 
cells is very distinct (Plate VIII, figs. 1 and 2 at 2). 
11. For entrance into the host’s root the haustorium of 
Olux scandens appears, excepting in very small haustoria, to be 
