PERMANENT WILTING IN PLANTS 95 



the open on August 10, in the shelter on August 12, and in the 

 chamber in the shelter on August i:^. 



The numerical data pertaining to this series are given in table 

 II, the colunm headings of which are self-explanatory. It will 

 be noted that the data are serially arranged according to the 

 hourly evaporation rates, beginning with the lowest. Each item 

 refers to a single culture. Where two or more cultures wilted 

 under the same conditions and at the same time, the results have 

 been a\'eraged and these averages are also shown in the table. 



It is clear from table II that Caldwell's main general conclusion 

 is definitely substantiated, and that with increasing evaporation 

 rates for the period of wilting, the moisture remaining in the soil 

 at permanent wilting also increases. That the data from individ- 

 ual cultures does not always agree with this generalization is 

 not surprising (for there must be considerable experimental 

 error involved with methods as crude as those here employed, 

 error brought about by unknown conditions either within or 

 without the plant), but if the data referring to single cultures 

 and to averages of only two are neglected, there are then no 

 discrepancies between the actual facts and the requirements of 

 the generalization. 



To determine what may be the limits of the atmospheric evap- 

 orating power, within which this factor may be considered as 

 without important effect upon the magnitude of the soil moisture 

 residue, the data of residues and evaporation rates have been 

 plotted to form a graph (fig. 1), in which abscissas represent 

 evaporation rates and ordinates represent soil moisture residues. 

 In order to indicate, in a general manner, the relative weights 

 to be attached to the various points of the graph, those referring 

 to the averages of the larger numbers of cultures wilting at about 

 the same time and under about the same conditions are denoted 

 by heavy circles. The range of variation of the group from which 

 each of these averages has been derived is shown by a vertical 

 line; the ordinate of the upper extremity of this line represents 

 the maximum and that for the lower end the minimum moisture 

 residue as actually encountered in the group. Other averages are 

 shown by hght circles, and data from single cultures (where no 



