53 
1920-21.] Interruption of Endodermis. 
not appear at first in even typical radial rows. The external cambium 
spread gradually round the sections, the cells dividing by periclinal walls, 
and their arrangement becoming regular and broader, while the endodermis 
remained undisturbed. A few sections further up showed cambium 
appearing in parts of the pericycle also, and from this point upwards 
the internal cambium caused breaking of the endodermal ring in the 
manner already described and figured. For a considerable number of 
sections the pericyclic cambium showed only a slight development of 
secondary tissue, and the external cambium none at all, but about f of an 
inch from the attachment of the root on the stem, there appeared the 
beginnings of two small secondary bundles — typical xylem elements 
surrounding central phloem — on the outside of the endodermis, which was 
continuous below them. The following sections showed rapid development 
of secondary tissue both from 
external and internal cam- 
bium. In consequence of the 
activity of the latter, single 
cells or groups of two or more 
endodermal cells were forced 
to the outside of the second- 
ary tissue, or were left lying 
between the bundles in the 
midst of this tissue. This follows from the division and growth of the 
cambial cells on either side of an endodermal group being quicker than 
that of the cells immediately beneath it. The last sections made, about \ 
inch from the end of the root, showed internal cambium and development 
of secondary tissue to a more or less degree practically all round, but the 
thickening was much more strongly developed on that side of the root 
which showed both internal and external cambial activity. The endodermis 
appeared almost completely broken up into patches of cells, and even into 
isolated cells (Photo II). 
The .second root examined was of slightly greater diameter than the 
first throughout all the series of sections, and its tissues had become much 
more sclerotic and woody, even the pericycle being comparatively thick- 
walled. The only thin-walled cells apparent were those of the cortex, and 
of the cambium to the outside of the secondary tissue. The development 
of cambium first began here in the cortex just outside the endodermis, and 
spread round, giving rise to secondary bundles to a greater or lesser degree 
in several isolated places, the endodermis being complete below the bundles 
(fig. 7). When a considerable amount of external secondary tissue had 
