338 Mr F. Darwin and Mr R. W. Phillips, [Nov. 23, 



cavital theory of water-transmission must be condemned. It is 

 therefore of importance to test the validity of his results. The 

 Essay consists of several sections — we shall first deal experimen- 

 tally with one of these. 



Dufour made a series of experiments in which two incisions 

 were made on two opposite sides of a branch and at a short 

 distance apart, as shown in fig. 2 or 7. If the incisions reach the 

 centre of the branch the continuity of all vessels must be broken. 

 Branches thus treated were tested as regards their power of trans- 

 mitting the water of transpiration, and also as to their power of 

 transmitting a current of water forced in under pressure. 



He was able to show that a transpiring branch which had been 

 " doubly sawn*" can transmit a sufficient amount of water to 

 prevent the withering of the leaves ; but that when the branch is 

 cut off above and below the double saw-cuts and it is attempted to 

 force water through the branch it is found that the current cannot 

 pass the double cuts. 



From this it is argued that the transpiration-stream which can 

 pass the double cuts must be of a different nature to the filtration- 

 stream (the stream travelling in the cavities of the elements 

 under pressure) which cannot pass this part of the branch. That 

 is to say, that the argument leads up to the belief that the stream 

 of water travelling in a transpiring branch does not travel in the 

 cavities of the wood elements. That it travels as water of imbibi- 

 tion in the cell-walls seems the only other theory which suggests 

 itself. 



We believe that Dufour is wrong in two points : 



(i) He has not realised how great is the obstruction to the 

 transpiration-current produced by double-sawing. 



(ii) He has exaggerated the difficulty of forcing water through 

 a doubly-sawn branch. We believe when the transpiration-stream 

 and the filtration-stream are properly estimated, that they will be 

 seen to be equally transmissible through a doubly-sawn branch. 



We proceed to the experimental proof of these statements. 



§ 4. I. Estimation of the transpiration-stream in doubly-sawn 



branches. 



In the paper already referred to-f- it has been shown that in 

 the case of Portugal laurel, double-sawing may produce a great 

 effect. The fall in the rate of absorption was from 100 to 3*2+. 



* i.e. sawn as shown in figs 2 and 7. 



+ Francis Darwin in Nature, May 1, 1884. 



X The cuts were ^ inch apart. 



