PERMANENT WILTING IN PLANTS 85 



that an evaporation rate during wilting, of 0.3 cc. per hour from 

 the standard atmometer was still somewhat too high for such 

 agreement. Of course this discrepancy means either that some 

 environmental condition besides the nature of the soil (as shown 

 by its water holding power) and the evaporating power of the 

 air must have been effective in determining the soil moisture 

 residuum of one or both of the two series of experiments so far 

 recorded, or else that the plants used by Caldwell at Tucson 

 were in some way different in their physical nature from those used 

 by the Washington experimenters. 



With these considerations in view, it becomes even more 

 highly de&irable than is indicated by Caldwell's statement 

 (loc. cit., page 50), that it be determined whether, with decreasing 

 evaporating power of the air, the limit assumed by the last- 

 named \\Titer may be actually attainable. Beyond such a limit, 

 if it be found, any further decrease in the evaporating power 

 of the air should have no marked effect upon the magnitude of 

 the soil moisture content at permanent wilting; the latter having 

 thus become practically a constant, as is indicated for the con- 

 ditions employed by Briggs and Shantz. The study to be reported 

 here was planned to determine whether Caldwell's physiological 

 Hmit might indeed be determined by simple methods and, if so, 

 to approximate its magnitude. 



The experiments of the present work were performed during 

 the "months of July, August and September, 1913, at the Desert 

 Laboratory of the Carnegie Institution, at Tucson, Arizona, 

 where atmospheric conditions are very favorable for experiments 

 involving the wilting of plants. It is a pleasure to acknowledge 

 here the writers' obligation to the Department of Botanical 

 Research of the Carnegie Institution of Washington, for the 

 facilities of the Desert Laboratory during the summer just men- 

 tioned and for financial support in carrying out the study. 



METHODS 



The natural soils used in these experiments were two, a coarse 

 sand and a clay loam, the former from a streamway north of the 



