300 BOTANICAL GAZETTE [OCTOBER 
in moist air. Leaves of the latter type (fig. 12) never formed shoots 
in their notches, and at best only a beginning of root formation was 
noticeable, which never led to roots longer than 1mm. The 
bud on the stem opposite the leaf grew out into a vigorous shoot (S), 
and at the lowest point of the callus of the stem, roots (R) were 
formed. 
The other type of leaves with only a piece of cortex attached 
(fig. 11) pommed always numerous long roots and quite frequently 
shoots, while the cortex formed roots 
(R) at the basal end and small leaves (L) 
from the axillary bud of the leaf. The 
drawing was made after 5 weeks. The 
growth of the roots and shoots at 
the cortex did not inhibit the growth of 
the notches, since in our assumption the 
suction which the cortex produced was 
not as strong as if the piece of stem had 
been complete. 
This furnishes us, therefore, with a 
method of retarding the velocity of the 
flow away from the notches of the leaf 
without interrupting it completely (as 
in a completely isolated leaf) or with- 
out making the flow too strong, as is the case when a whole 
piece of stem (without the opposite leaf) is attached (as in fig. 12). 
This difference should find expression in the relative velocity of 
growth in the notches if the proper experiments were made. Figs. 
13, 14, and 15 represent such a series, 3 weeks after the experiment 
was begun. The leaves were suspended in moist air. The com- 
pletely isolated leaf (fig. 13) has formed 3 shoots of more than 1 cm. 
in length and roots of considerable size. The second leaf with a 
piece of cortex attached (fig. 14) has roots and only one short shoot 
in the notches, but has a considerable shoot and roots in the cortex. 
The flow of sap from the leaf to the cortex was not strong enough to 
completely prevent the growth of the notches, but only to retard it. 
The third leaf (fig. 15) has a piece of complete stem attached and 
in this leaf no growth has taken place, and in all probability none 
Fic. 15 
