iqiq] weaver &• MOGENSEN— TRANSPIRATION 413 



the early autumn losses (September 20 to October 11) and the 

 initial leaf area. Also a comparison of the graph of mean tempera- 

 ture for the several periods (fig. 15), with the graphs showing the 

 march of transpiration, shows that they have a striking resemblance. 

 The mean temperature was obtained from the thermograph records, 

 and is given on a time-humidity basis. The means for the several 

 periods respectively were obtained by drawing a horizontal line 

 through the weekly record sheet in such a manner that the total 

 area included by the graph above this line was equal to the total 

 area below the line. The areas were determined by the aid of a 

 planimeter. In this interpretation both temperature and time 

 factors are taken into consideration. Temperature, indeed, seems 

 to be the controlling factor, for no such correlation between tran- 

 spiration and humidity (here the graph is inverted) is discernible 

 after October 5; that is, transpiration dropped ofif at this time, 

 following the descending temperature graph in spite of a decreasing 

 humidity. 



These two factors, however, together with wind velocity, are 

 well summed up in evaporimeter readings. These were obtained 

 by Livingston's porous cup atmometers and show an increase or 

 decrease very similar to the transpiration graphs. They are as 



follows : 



September 20-24 .... 25.1 cc. average daily loss 



September 24-28 .... 11. 8 cc. 



September 2 8-October 6 . . 22.2 cc. 



Readings were discontinued on October 11 because of freezing 

 weather at night. Unfortunately no instrument has yet been 

 devised for satisfactorily measuring winter evaporation losses. 



The irregularity in the graphs, especially in the fourth interval, 

 is due to the non-uniform drying out of the leaves. Up to October 

 5 the leaves of all the dicotyledonous plants were green and appar- 

 ently functioning normally. By October 11 a few of the older 

 basal leaves were beginning to turn brown. While some dropped 

 at once, many held on after they were dry. This irregularity in 

 defoKation defeated an attempt to base the daily losses upon the 

 actual average leaf area during any given period, although an exact 



