303 
a small shrub with rather small leaves which, in my opinion, is 
not S. arborescens, T. And., but, if anything, is nearer S. Wal- 
lichit, Hook., £. & Thoms. Of the other numbers of Wilson’s and 
S. melanocarpa is described as ‘‘0°3-1 m. altus,’’ but in brackets 
ded “‘ planta yunnanensis interdum ad 5 m. secun- 
dum Cl. Henry.’? The statement in brackets refers to the 
S. arborescens. It seems clear t e, now, that the Sikkim 
plant is not the type melanocarpa, but should continue to bear 
Anderson’s name , In my opinion, Rehder and Wil- 
XXXI.—NATURAL GRAFTING OF BRANCHES 
AND ROOTS. 
W. Davumore. 
(With Plates.) 
Natural grafting of branches and roots is of common occur- 
are sometimes very curious, not only by reason of the compli- 
1 eres or roots of in- 
dividual trees, but also by the joining together of large branches 
and even trunks of different trees of the same species, 
Naturally-grafted branches are fairly common on beech, oak, 
olly, lime, willow, yew, and Scots pine, whilst they may also 
be noted on many other trees. The commonest form of natural 
Another fairly common method of natural grafting of branches 
