ae _ TRANSACTIONS OF THE [ Sess. LXXIII. 
itself. Sometimes all trace of this in the immediate 
vicinity is lost. But, as a rule, it persists with the radial 
walls slightly elongated where the cells abut on the point 
of attack. For some distance on either side of the gall 
the radial walls of the endoderm are unthickened. Further 
round the normal thickening reappears. Thus an attacked 
root in T.S. shows an endodermis consisting on one side of 
thin parenchyma, on the other are the usual horseshoe- 
shaped cells (Plate IV. fig. 2). 
The pericycle, under the irritation of Heterodera, divides 
tangentially (Plate IV. figs. 2, 3, p), and may form a layer 
three to four cells deep. 
The effect on the cortex too is distinct. In TS. there 
appears a broad proliferation of thin-walled cells stretching 
outwards from the endodermis and widening as it goes. 
This mass of cells must originate from the inner cortex 
while that is still capable of division. The sheath of 
sclerenchyma, which forms in Bromeliads a mid-cortical 
layer, never is found opposite a gall (Plate V. figs. 2-4). 
Heterodera in some way inhibits the deposition of thicken- 
ing matter on the cell wall. Two advantages in all 
probability directly follow. There is reduced pressure, 
while additional moisture may come in from the outside. 
Further, as rupture takes place latterly in this proliferated 
region the young Heterodere find prepared a passage to the 
soil where the early stage of the life-cycle is passed. 
One curious feature of this mass of thin-walled cells is 
that here and there in it occur isolated elements over which 
Heterodera has had no inhibiting influence (Plate V. fig. 2). 
They are more or less completely lignified. Where the 
unthickened tissue joins the sheath there is frequently a 
gradual transition from such isolated cells to a broad 
continuous band of sclerenchyma. In other cases the line 
of demarcation is sharp. 
As regards the final result of the attack, it has already 
been remarked that in extreme cases the entire central 
portion of the root is killed below the pomt of the 
worm’s lodgment. In many instances—where the attack 
has been severe—the cells of the tissue a little above 
the parasite have been thickened or filled up. But as 
this condition is frequently observed in old normal roots 
