TRANSPIEATION AND LEAF- WATER 3 



observation, except that the pieces were carried into an adjoining 

 building, where they remained a minute or two during weighing. 

 The time error introduced amounted at most to about three min- 

 utes, but this error practically disappears by making the burette 

 reading at the conclusion of the weighing. 



In the tables following, the first readings, namely those made 

 precisely on the hour have been omitted, since their introduction 

 will serve no useful purpose. They have, however, been care- 

 fully compared with the readings following the weighings, in order 

 to detect errors. No such errors, which would most likely have 

 occurred during periods of rapid transpiration, sufficient to dis- 

 turb the validity of the data which are here presented, have been 

 observed. In plotting the graphs (fig. 1) it was assumed that 

 the intervals between the readings (with conjoint weighings) 

 are equal. It is probable that this assumption is correct within 

 3 per cent. 



Regarding the use of the potometer for the ocotillo, it emerges 

 that the differences between the lowest and highest rates, deter- 

 mined by volume, are less than those determined by weight (in 

 the ratio of 27 to 28 for piece 1, approximately). This is seen at 

 a glance by comparing the course of the corresponding graphs in 

 the figure. For the purpose intended in my earlier work, the 

 error due to the use of the potometer in the study of this plant is, 

 therefore, on the right side of the balance sheet, and my inferences, 

 based on the data earlier published, unless invalidated by other 

 evidence, are true. 



It appears further that the error, due to the use of the potome- 

 ter, relatively to the total loss of water during a part of the night, 

 amounts to 50 per cent, that is, about one-half the small volume 

 of water taken up by the plant, when the rate of transpiration is 

 extremely low, is retained within it. This may be regarded as 

 due chiefly to the damping effect of the closed or nearly closed 

 stomata or of other physical conditions in the leaf, or to the less 

 favorable conditions for evaporation. That the rate of transpira- 

 tion during the night, as compared with that of evaporation from 

 three atmometers (the data for one of them being given in table 

 VI) is more uniform indicates that there is some factor at work in 



