256 BOTANICAL GAZETTE focroper 
does not modify their distribution; and as the roots occur in the 
same way whether the piece is erect, horizontal, or inverted, gravity 
cannot be a determining factor. The other conception is that of 
specific formative substances, which in this case, moving toward the 
base, would accumulate at the lowest part and incite the formation 
of roots there. As this theory is to be discussed later, only two 
cases will be mentioned here to show its inefficiency. If in Salix, 
instead of cutting the stem off just above a bud, the bud is cut entirely 
away from the stem, it starts to develop just the same; or, as in 
jig. 7, where a notch is cut at the base of the stem, there can be no 
accumulation of these substances further up where the roots actually 
occur. It is not impossible that the upper parts, especially the 
leaves, exert some influence on the formation of roots on the stem. 
If so, it is not through the transpiration current, for the roots develop 
as well when transpiration is entirely checked as when it is quite 
active. Also, in fig. 9 the leaves are transpiring and a current must 
be passing in through the end that does not produce roots. The 
removal of all leaves greatly decreases the vigor of the roots and 
also the number that are formed, but the same result is obtained by 
placing the leafy part in the dark, or in an atmosphere free from 
carbon dioxid. The fact that pieces of the internode produce roots 
indicates that the leaves are not necessary; their influence probably 
lies in keeping the stem better nourished. 
On the other hand, the evidence seems to point to the second 
line of explanation, namely, that just as the growing shoots seem to 
exert a retarding influence on the buds below them, so the growing 
roots exert an influence which inhibits cells, otherwise able to do ” 
from forming roots. If, as just mentioned, the stem is cut off at its 
lower end and placed in water, and at the same time a portion of the 
stem higher up is surrounded with water, we get roots at both places. 
When these are growing well, if we cut off the lower part of the stem 
having the roots on and then submerge this end again, new rools 
soon appear at this point, although roots are growing vigorously . 
few centimeters above. In no case could I ever get 4 retarding 
influence of the roots to pass down the stem. 
The problem to be solved in regeneration seems t - 
much the growth of parts following a removal (the causes here 4 
o be not 5° 
