20 



THE PHYSIOLOGY OF STOMATA. 



as the relative rates of transpiration in successive periods are concerned. 

 Whether the absolute amount of water transpired may safely be measured in 

 this way can not be settled until comparative results based upon experiments 

 with rooted plants are available. It will be seen, however, that my object, 

 namely, to record sudden changes of rate in transpiration, was attained. It 

 may be noted in this connection that of 38 experiments conducted in batteries, 

 only one had to be disregarded. 



One obvious objection had to be met, namely, that the rate of transpiration 

 might not be equal to the rate at which the water leaves the potometer. This 

 objection was answered first by comparative weighing and absorption experi- 

 ments and also by the comparison of the synchronous experiments of a battery. 

 The former discovers the absolute error and also its distribution, and therefore 

 whether it affects the character of the curve or not ; the latter by comparison, 

 point by point, gives the probability of error in any one or more of the series. 

 The data in tables 10 to 12 illustrate the former for ocotillo. 



Table 10. — Ocotillo — Error due to use 0} potometer, June 19, 1904. — Two pieces. 



*8.6 per cent of water drawn from potometer was absorbed by plant. 

 1 10 per cent of water drawn from potometer was absorbed by plant. 



In piece No. 1, table io, 8.6 percent of the water drawn from the potometer was absorbed 

 by the plant, while in the second piece, the amount retained was 10 per cent. 



In table 1 1, of all the water retained (viz, 3.7 per cent of the whole amount 

 absorbed) 92.6 per cent was retained during the first 8.5 hours, the rest dur- 

 ing the last 11 hours. The hourly readings taken showed an abnormally 

 high rate between 9 p. m. and 3 a. m., a behavior out of harmony with nearly 

 all my experiments. Even then the total error is small, and if distributed 



Table ii. — Ocotillc- 



-Error due to use 0} a potometer, July 18-19, 1904; 

 continual darkness. 



during the entire period would be quite negligible for any of my own pur- 

 poses. Under the conditions of this experiment the last 10 hours of the 

 curve show the rate of loss of water by the plant. With very few exceptions 



