1902] RISE OF THE TRANSPIRATION STREAM 165 



there when the apparatus was set up, but still so little that it 

 was in small isolated flat bubbles. 



While no experiment with any kind of an artificial tube can 

 have the force of a demonstration as to what occurs in a real tree, 

 the behavior of this apparatus was like that in trees as we know 

 it, not only in the general achievement of the elevation of water 

 to a still undefined height, but also in enough known details to 

 make it most probable that the same fundamental physical prin- 

 ciples are operative in both cases. But as already said, an 

 analysis showing the elementary factors by which the water is 

 raised has baffled me, even in my own apparatus. 



The motive idea in setting up the apparatus — that the cohe- 

 sion of the water in the tube must be responsible for any suction 

 exerted more than lO"" below the top — was certainly a mistake. 

 This is shown by the readings of the manometers higher up, and 

 also by the continued rise of the mercury at the bottom, after 

 suction at the top ceased, January 18. Both of these observa- 

 tions indicate that the suction developed in the tube is not as 

 great as one atmosphere; that a difference in pressure of less ' 

 than one atmosphere between top and bottom will lift water 

 much more than \o^ under the peculiar conditions here present. 

 But how? It might be that when water is drawn from the top 

 of the tube it rarefies the air there, and that the suction so aris- 

 ing is propagated downward as largely as possible by the air, 

 and that the bubbles and as little water as possible separating 

 em are thus put in motion. In this way the suction necessary 

 to lift the water should be only equal to that of a column of 

 water equal to the total height of the indices separating the 

 bubbles. But I did not observe that the bubbles moved, and my 

 doubt as to their doing so is very much strengthened by the 

 final observation that the rise of the eosin solution did not drive 

 all the air before it. The tube did become more full of air 

 in the upper part ; but I think this was chiefly due to greater 

 suction there, this causing the air present to occupy more 

 space, and hasteninsf diffusion inward through the rubber con- 

 nections ; and two of the " leaves " used leaked. It may be taken 



th 



