On a Method of Studying Transpiration. 279 



humidity. This conclusion, which is a physical necessity, does not seem to 

 have been definitely proved or represented diagrammatically. 



In some cases {e.g. figs. 2 and 6) the line of dots {i.e. the transpiration 

 curve) is not straight — the change in rate of transpiration lags behind the 

 change in i/r — for reasons not yet clear. 



(2) The second characteristic of the diagrams is that the diagonal does 

 not pass through the point of intersection of the axes — or, in other words, 

 transpiration is not zero in saturated air. I have not hitherto seen this 

 graphically represented as the result of experiment, although it might have 

 been foretold. The fact that transpiration occurs in saturated air, and that 

 it is due to the production of heat in plant-respiration was first made clear 

 by Sachs*, who proposed that the fact should be utilised as a means of 

 measuring the " Eigenwarme " of plants.f We shall see later that the 

 diagram (fig. 9) may perhaps be applied to the same end. The position of 



30 













C 



) 



O 20 

 V. 



OS 



£ 10 



F 







1 









J05 



100 



|90 







80 





70 



G A B C 



Relative Humidity. 

 Fig. 9. 



the point G varies in different cases. In the earlier experiments, I estimated 

 AG = 7, but I now consider 5 a more reasonable average. The construction 

 here given has been used throughout my work for the rough reduction of 

 transpiration-rates to a common degree of relative humidity. Thus, 

 supposing that in fig. 9 the transpiration-rates DC and EB have been 

 obtained under different conditions of illumination, it is clear that we 

 cannot estimate the effect of such conditions until the amounts have been 

 corrected for the differences in relative humidity. 



* < Sitzb. K. Akad. Wier,' 1857, vol. 26, p. 326. 



t See Sachs, ' Physiologie Exp6rimentale, ! 1868, p. 249 (the French translation of his 

 book on plant physiology). 



