"42 BOTANICAL GAZETTE [JANUARY 
leaves, stems, and roots. The transpiration of the leaves and the 
absorption by the roots sets up a difference in pressure at the two 
ends of the plant. The principal factor, however, in the ascent 
of sap is the osmotic power of the wood parenchyma, which tends 
to keep the quantity of water constant. He thinks that transpira- _ 4 
tion simply serves to accelerate the rate of flow, but does not cause : 
it (‘Le méchanisme des mouvements de la séve est donc le méme 
dans un arbre haut de 100 métres que dans une herbe de quelque 
centimétres, dans une tige verticale que dans une tige horizontale”). 
Drxon’s theory of sap-flow is perhaps the most thoroughly in 
accord with the experimental facts as they stand today. He holds _ 
that the elevation of water even in the highest trees is due to a 
tensile stress being set up in the tracheae of the leaves when water 
is abstracted from them by the evaporation from the leaves, which — 
stress is transmitted to the water in the larger vessels. The water 4 
adheres to the walls of the vessels, so that the stress is resisted. 
The water ‘‘hangs” in the vessels “by virtue of its cohesion.” 
This theory has been objected to on the ground that the water — 
columns are not continuous, and that air bubbles and vapor are — 
present in the vessels. | 
system. It will be seen that this theory eliminates the action 0 
the living cells of the stem. Drxon cites the classic experiments of 
STRASBURGER (31) in which an oak 22 m. high was sawed througt” 
and placed in picric acid for three days until the liquid had risen 
3m. The tree was then placed in fuchsin, which rose 18 m. in 
the dead stem in 8 days. Drxon believes that this experimen! 
proves conclusively the uselessness of the living cells of the stem im 
sap-flow. 
Preliminary experiments a 
It is plain that the question as to the effect of killing a section 
of a stem on the character and quantity of the sap-flow is still 
unsettled as to many particulars. There is no question that, as 
UrspruNG admits, such stems conduct a certain amount of watel 
for a considerable period of time. Experiments conducted on 
quantitative basis are essential in such a problem, and the follow 
ing studies have been carried on from this standpoint. Afte 
