SPA Mr F. Darwin and Mr R. W. Phillips, [Nov. £3, 



5*6 seconds 



5-8 „ 



6-0 „ 



5-8 „ 



6-0 „ 



5-8 „ 



5-8 „ 



In our experiments we usually employed Cambridge water- 

 supply water, which is fairly free from impurities, in some cases 

 however distilled water was used. The success of the method 

 depends on there being no leaks in the apparatus; it is therefore 

 essential that the india-rubber corks should be good, and the liga- 

 tures where the india-rubber tube is attached either to the glass or 

 the branch should be carefully made. For this purpose either 

 copper wire or (what is better) india-rubber thread was em- 

 ployed. 



Advantages and disadvantages of the method. 



The apparatus is rapidly put together so that readings can be 

 taken within one minute after the branch has been cut. It is so 

 simple that it is not likely to get out of order, and the places of 

 possible leaks are reduced to the lowest possible number. Each 

 reading takes only a few seconds, so that a number of observations 

 may be made in a short space of time. It is easy to take obser- 

 vations without in any way disturbing the plant, and this is of 

 importance since it has been shown that the shaking caused by 

 such a disturbance as is involved in weighing a transpiring plant 

 influences the transpiration*. 



Owing to the sinking of the level in the vessels c and e, as the 

 branch absorbs water, the conditions do not remain absolutely 

 identical from hour to hour, but since these sources of error are 

 extremely small, and their effect is spread over a considerable 

 interval, they may be considered as of no moment. 



Changes in the temperature of the water, or in the pressure of 



, the surrounding air, must theoretically disturb the reading of the 



instrument. If the water in the Y tube is increasing in "bulk 



owing to increase in temperature the passage of water up the tube 



b must be delayed. 



If it be remembered how slowly the temperature of the 17 cc. 

 contained in the poto meter would be changed in the course of 

 experiments conducted in ordinary rooms, and when it is further 

 considered that the bubble serving as index often moves at the 

 rate of 12 mm. per second, it will be seen that no serious eri'or 



* Baranetzky. Bot. Zeitung, 1872. 



