278 Scott and Brebner . — On the Anatomy 
that the production of external secondary bast is at first very 
slight. When the stem is two years old the outer boundary 
of the wood becomes uneven, and its depressions are filled 
by little groups of phloem, which subsequently become the 
f phloem-islands V The cambium remains in contact with the 
wood during their development, and hence the phloem-islands 
must be produced on its exterior side. The intercepted 
portions of the wood are reunited by means of divisions in 
the cells of the pericycle lying outside the newly formed 
phloem group, and thus the latter becomes imbedded. Herail 
compares this process with that in Bignonia , in which the 
phloem-plates are also formed by increased centripetal and 
diminished centrifugal activity of the cambium at certain 
points, though in Bignonia the phloem does not become 
enclosed in the wood. The fact that regular radial rows of 
cells, continuous with those of the wood, can only be traced 
in the inner part of the phloem-islands, is used by Herail 
as an argument in favour of their centripetal development. 
He concludes by pointing out that Strychnos conforms to 
the * general law * that a cambium only forms one kind 
of tissue, wood or bast, on each of its faces. He gives 
two figures, one of which (Fig. 24) accurately represents a 
developing phloem-island, showing clearly the continuity of 
the radial series of cells inwards, but not outwards. The 
other figure will have to be referred to again later on. The 
main point is, that Herail’s account directly contradicts that 
of de Bary. According to the former the phloem-islands are 
formed centripetally, and thus belong, like other phloem, to the 
bast tissues, while, as we have seen, de Bary believed that they 
are formed centrifugally, and are thus a constituent of the wood. 
We will now proceed to describe our own observations, 
deferring all comparative consideration of parallel cases to the 
end of the paper. It may be mentioned that the essential 
part of our work, in so far as it relates to the development of 
1 This name is so appropriate for the isolated phloem-groups imbedded in the 
xylem, as they appear in transverse sections, that it will be adopted throughout 
this paper. 
