189 
out of a descending bark-sap, but out of a cell-development 
of the cambium already existing in the stock, and having es- 
sentially the same characters. The formation of new wood 
of the nature of the graft has always been taken for granted, 
in order to prove the descent of the bark-sap; but we find 
that this wood does not partake of the nature of the graft, and 
that it must, therefore, be formed independently of any de- 
scending juices.’ ‘These being the views held by the best 
authorities on the matter at present, I shall now detail my 
experiments, and show how far they bear on either. 
‘* My predecessor, Mr. Niven, had been conducting some 
physiological experiments before he left the Botanic Gardens, 
the results of which are already before the Public. I consider, 
however, it only just on my part towards him, that I shall 
here state my principal experiment to be founded on one he 
had commenced, though we do not appear to have been 
aiming to attain similar objects. He had cut several trees 
more or less through their boles in various ways, one of 
them a large horse-chesnut tree, then four feet in circum- 
ference, and now four feet nine inches. At three feet from 
the surface of the ground, two deep incisions had been made 
through the stem, crossing each other at right angles, and 
reaching the circumference on each side (Fig. 1). The tree was 
thus left growing on four separate pillars of wood, alburnum 
and bark, but no results, that I am aware of, were deducible 
from this experiment when I commenced the following. Seeing 
that it afforded an excellent example for observing the growth 
of woody matter, as it would form to fill up the perforations 
through the stem, I examined the portion of the tree where 
it was cut, and found that the heart wood was completely 
dead, and beginning to decay, at both the upper and lower 
lips of the cut. It, therefore, could render no assistance 
whatever for the phenomena of life being carried on through 
its medium. ‘The ascent of the sap and formation of wood 
must then have depended altogether on the functions of the 
