494 
Thoday . — On the Organization of Growth and 
Figs. 5 and 6 , are photographs of the larger section; they show the 
hollowness of the pith and the lines of strain in the outer still intact 
part of it. 
Incidentally the structure of the upper part of the stem is the better 
adapted to tangential growth, since the independence of the leaf-trace 
bundles through greater distances means a corresponding continuity of 
medullary rays and less frequent cross-connexions between the tangentially 
growing sectors. Frequent anastomosis would obviously hinder the tangential 
growth of the anastomosing strands, for the fibrous cross-connexions would 
offer greater resistance to the separation of the strands than the yielding 
ray parenchyma. Radial segmentation of the woody cylinder is indeed 
a sine qua non . Active tangential growth in the cambial region outside 
a continuous zone of wood is hardly conceivable. 
Anastomosis of phloem strands at the node and elsewhere. The 
independence of the leaf-trace bundles for a long distance entails no 
serious disadvantage in relation to the conduction of water, since the supply 
comes from below and serves the one leaf to which it is directly transmitted. 
For the translocation of food, however, which comes from the leaf and is 
required mainly above, the path would be greatly lengthened. It is 
therefore interesting to find that a short circuit is established at the node, 
by lateral union of phloem associated with the incoming leaf-trace and the 
phloem of adjoining minor bundles. 
The details require further elucidation in the petiole on the lines of the 
present study of the stem, but the following brief outline, together with 
Text-fig. 5 , will suffice in the present connexion. In the petiole of a large 
leaf there are a number of small bundles as well as the three principal ones. 
The former appear to anastomose in rather intricate fashion and one or two 
larger ones unite with the two lateral bundles. In the base the remainder 
cluster round the principal bundles as these diverge. Many of the smallest 
consist, even in the case of a mature leaf, of phloem only, and in the others 
the xylem dies out in the nodal region. Thus, just above the level at which 
the trace bundle enters the vascular zone of the stem, there are associated 
with it a number of small strands of phloem, which encircle the xylem 
(Text-fig. 5 (6)). 
At the same level cambium is continuous across the rays. In the case 
figured the two small bundles are the branches of a small bundle which has 
forked just above. At the approach of the leaf-trace bundle these branches 
diverge, the cambium ring bulges inwards and divides to admit the xylem 
of the incoming bundle. Meanwhile the phloem strands that encircle the 
latter take their place as they arrive, along with a few other phloem strands 
already present, outside the cambium, and follow it when it divides. The 
free ends of the divided ring then gradually withdraw into the undistorted 
parts adjoining, and the associated phloem strands crowd together, along 
