30 STUDIES IN ROOT-PARASITISM, 
poor development will also probably account for the fact that 
the secondary growths, which form so marked a feature in the 
attacks of Olax scandens, are almost entirely absent in the 
roots entered by Cansjyera. Secondary growth of any kind 
(excepting of course the formation of corky separating layers) 
is very rare in the wings of the cortex, and the formation of 
secondary wood there is practically unknown. 
The formation of cork in the angle opposite the edge of 
the sucker lobe is frequently met with, and this is usually 
sufficient to bar further progress. The origin of this cork is 
well seen in one section through the root of Plumbago rosea. 
The phellogen appears to arise from a row of cells derived 
from the cambium. Instead of becoming lignified in the usual 
way, these cells divide by radial walls and thus form tangen- 
tial rows of cork. In certain cases noted, this corky protection 
of the cambium is neglected or insufficient, the action of the 
sucker cells extending far beyond the place occupied by them. 
Thus, in a couple of haustoria fixed on Alangium Lamarckii, 
the destruction of the cambium passes nearly all round the 
root, although the lobing of the sucker is, as usual, small. 
Secondary thickening of the woody cylinder of the host, 
after attack, occurs in a few cases, but it is rare. Is this 
because of the smallness of the haustorium or its compara- 
tively short duration ? Or is it due to the fact that these 
haustoria have been collected in a dry region where growth of 
any kind is slow ? 
Some of the roots attacked have abnormal structure. 
Considering the differences noted in the structure of haustoria 
attacking monocotyledon and dicotyledon roots (Sant. I, 14), 
it is interesting to mark their behaviour in these cases. In 
Arua lanata the wood occurs in concentric, more or less con- 
tinuous bands, and these are mainly composed of thick-walled 
elements. Where these bands extend right across the area of 
contact, the sucker treats them as ordinary woody cylinder 
and spreads along the cambium. If there is any gap, however, 
the sucker passes through it and attacks the band within, the 
