CAUSES OF VARIATION IN DETERMINATIONS. 53 



of the system accompanied the wilting of the indicator plant. This 

 is not due to the loss of water from the indicator plants alone, for 

 the record shows that a steady loss of water took place from the soil 

 from October 11 to October 14. On October 15, however, the 

 system was found balanced. On October 16 the plant end of the 

 system was up, indicating the loss of water from the plant tissues, 

 and the squash was found wilting. The squash was cut off below the 

 wax and the pot resealed, weighed, and again balanced on October 17. 

 After five hours the plant end of the system was again up, showing 

 the simultaneous loss of water from the growing tissues of the cactus. 

 In other words, the cactus loses water from its storage tissues at the 

 time the squash plants wilt, which means that the cactus is initially 

 no more effective than the squash in reducing the moisture content 

 of the soil. The two plants therefore have the same wilting coefficient. 

 Precisely similar results were obtained for cactus and Kubanka 

 wheat, and with the olive and wheat. The results afford no ground for 

 believing that either the cactus or the olive can reduce the moisture 

 content of the soil below the point at which other plants wilt without 

 simultaneously reducing the water content of its aerial tissues. 



CAUSES OF VARIATION IN WTLTING-COEFFICIENT 

 DETERMINATIONS. 



In making determinations of the wilting coefhcent, variations of 

 greater or less magnitude occur. These variations are found even 

 though a single strain is used as an indicator, and when individual varia- 

 tion is minimized by growing five plants or more of the same strain in 

 each pot. The causes for such variation may be treated under two 

 headings: (1) Causes for determinations having a value less than 

 normal and (2) causes for determinations having a value greater 

 than normal. 



CAUSES FOR DETERMINATIONS HAVING A VALUE LESS THAN NORMAL. 



This condition is generally due to direct evaporation from the 

 soil. It is evident that the moisture evaporated directly into the 

 air may reduce the moisture content of the soil on the surface or 

 outer portions of the soil mass far below that to which it could be 

 reduced by the plant at the time of wilting. In our experiments 

 this source of error has been avoided by using impervious pots and a 

 wax seal. In the experiments of earlier investigators, however, 

 this must have been a constant source of error. In field determina- 

 tions this constitutes an almost insurmountable source of error or 

 uncertainty, since it is difficult to determine to what extent direct 

 loss of water from the soil has continued in the upper layers while 

 the plant was drawing upon some deeper source of moisture. 



230 



