INDUCED RHYTHM IN TRANSPIRATION. 65 



upon the question of transpiration than stoma tal response. Regarded in 

 this light they become open to the objection, equally applicable to the major- 

 ity of Curtis's experiments (though not to those above mentioned), that 

 the difference in normal morning and afternoon illumination introduces a 

 source of error. Nor is the element of fatigue considered at all by either. I 

 must say further that I believe Burgerstein is quite mistaken in his evaluation 

 of the evidence above considered (loc. cit., p. 153) and appears to have over- 

 looked that which we shall now consider. 



The data which Curtis (loc. cit., p. 367) derived from plants kept in constant 

 darkness are, however, distinctly pertinent. I quote his conclusions as 

 formulated by himself : 



Occasionally there is apparent rhythm corresponding to the daily periodicity, but on 

 the other hand it not frequently happens that there may be an irregular decline from the 

 start, or the curve may be characterized by a rise or fall without reference to the period 

 of the day. 



It would appear that Curtis was expecting, on a priori grounds, a rhythmic 

 maximum to fall at or about the hour at which the normal daily maximum 

 occurs. This is what Unger believed to occur, but as Burgerstein has pointed 

 out (loc. cit., p. 151), his experiments were not carried out with sufficient 

 rigor. That at present unexplainable fluctuations of transpiration rate 

 occur even under conditions as constant as we have been able to get them is 

 a matter of general observation, and that for some plants some maxima will 

 fall at or about the hour of the normal maximum is, so far as our understand- 

 ing at present goes, a matter of chance, and must be so regarded. A few 

 instances in different plants of a time correspondence of this kind may or may 

 not be properly regarded as rhythm. From my own experience, however, it 

 seems more logical to look for a maximum nearer to the hour at which the 

 stimulus which leads to the increased transpiration response of daytime is 

 normally applied, namely, at or somewhat later than at sunrise.* This I 

 had satisfied myself to be, in fact, the case, and on examining Curtis's data 

 I found that he too had noted on several occasions that pronounced maxima 

 may be found early in the morning, between 5 and 9 o'clock, an observation 

 which Curtis seemed to have attached but little importance to, but which in 

 the light of the evidence now to be presented appears to be of great impor- 

 tance as bearing on the problem under discussion. 



This evidence, as will be manifest upon examination, possesses a remarkable 

 degree of uniformity. The manner by which it was obtained has been fully 

 described (see p. 18). 



*The fact that Fonquicria splendens produces a crop of leaves in December and January 

 should enable us to test this view, for, since the sun rises at this time about 2 hours later 

 than in the summer, the maximum of induced rhythm in the winter leaves should be pushed 

 forward accordingly. 



