280 Scott and Brebner . — On Internal Phloem 
As the activity of the cambium goes on, the isolated protoxy- 
lem-elements are pushed further and further outwards, 
accompanying the external phloem 1 . Often this proceeds 
until a large amount of secondary wood has been formed, as 
in the root shown in Fig. 9, where two protoxylem-elements 
are seen outside the cambium. Sooner or later, however, the 
protoxylem is restored to the main body of the wood, from 
which the cambium had severed it. Sometimes this happens 
very soon (Fig. 11), sometimes very late (Fig. 9), and it takes 
place very irregularly. Thus in Fig. 9, between the two 
isolated protoxylem-elements is a third, which has long been 
enclosed in secondary wood. It is difficult to distinguish, but 
is probably the element marked with letter a. The enclosure 
of the protoxylem in the. wood happens as follows : the cam- 
bial divisions within it cease, the cells lying just to the outside 
begin to divide and complete the cambial ring ; new wood is 
formed and the protoxylem-element is now left behind, im- 
bedded in secondary wood (cf. Fig. 11). Ultimately all these 
elements are thus regained by the wood. As is often the 
case in roots, the protoxylem-vessels are not spiral but reticu- 
late. Their thickening is sufficiently different from that of 
the pitted vessels of the secondary xylem for them to be 
recognised long after their enclosure in the latter. Since we 
made these observations we have found the same mode of 
development in the root of an Impatiens , in which the 
protoxylem is also cut off by the cambium from the wood. 
We traced carefully the development of the secondary 
phloem-islands of the root. Here, as in the stem, they are 
formed centrifugally, on the inner side of the cambium. A 
cell cut off from the latter divides two or three times by 
tangential or inclined walls. The products of division are the 
mother-cells of sieve-tubes. They then divide again to form 
their companion-cells and the phloem-strand is complete 
(see Fig. 12). The development can be studied with advan- 
tage in radial longitudinal sections, which also leave no 
1 It is a curious coincidence, though certainly nothing more, that the same root 
should show phloem-strands in the xylem and xylem-strands in the phloem. 
