( 424 ) 
the top) and the middle piece, II, evidently meet their want of con- 
nection sooner and perhaps better by means of the neo-formation of 
rootlets, than by strengthening the existing but feeble communica- 
tion with the old ones. *) 
The third part of the leaf, UL, (which consequently is in direct 
communication with the base) never shows any inclination to the 
formation of rootlets, obviously because the communication has 
remained unimpaired here. 
If we must assume that the stream-bundle in the uninjured leat 
is regulated by a basally directed impulse, then, when the “reversion” 
has succeeded, the newly formed current in the middle piece must be 
directed by an opposite impulse, or, to speak more correctly, by the 
same impulse, after it has, so to say, been reflected by the eross-wound. 
That this current in fact behaves in this manner, follows at once 
from the fact that the new current is first visible below in the middle 
piece and is gradually prolonged upward. 
A still more convincing proof of this can be given by a further 
experimental operation : if namely these new currents are interrupted 
in I and II by a small cross-wound (as in fig. 1) one sees the thrust 
in I occur above, in II on the other hand below the wound, and 
the currents take their way as is indicated by the dotted line in the 
figure. This is a proof that these two adjacent pieces behave oppositely. 
Though we finally often succeeded in bringing about the “reversion” 
in the middle piece, vet this reversion is very incomplete, as I infer 
from the following observation. In one of the leaves with a double- 
hooked wound a prolification had been formed above in the middle piece, 
while the complete reversion was being brought about; the new 
leaflet lay a little sideways of the current. Proceeding from this 
leaflet a little bundle of three currents had developed. One of them 
proceeded along the lower side of the upper eross-wound into the 
third part of the leaf, after having joined the main current coming 
from below. The other two, however, took their way straight down- 
ward as if the connection with the base of the leaf were still exactly 
as before the lesion. Hence one of the currents, when coming forth 
from the leallet, obeyed the action of the reflected impulse, whereas 
the other two experienced no influence. In that place of the middle 
piece the old basipetal impulse must consequently have been preserved. 
A similar case, occurring in another experiment, will be men- 
tioned later. 
1) Above the upper cross wound also rootlets are sometimes formed, although 
only when the top-part is large and so powerful enough, or when prolifications 
occur on it. 
