372 



BOTANICAL GAZETTE 



[may 



up of at least three component curves. The general relations of 

 these to one another in the composite curve are shown graphically 

 in fig. 2, which has been somewhat exaggerated, especially in respect 

 of the first curve, for the sake of clearness. The effect of the initial 

 rapid intake is to throw the main part of the curve upward from 

 the base line. Careful examination showed that it was not possible 

 to find a parabolic curve that would follow the data at any tempera- 

 ture. The problem then was to find an empirical formula or 

 equation or such a combination of equations as would very closely 

 approximate the given data of observation. This was necessary 



e 



Fig. 2. — Curves showing composite nature of moisture intake curves in Xanthium: 

 first curve exaggerated ; oa, initial intake; be, main curve; de, approaching saturation. 



measunn 



could be substituted for the uncertainties of the graphic method 

 used by Brown and Worley. The only proof we have that any 

 equation or group of equations is adapted to such use lies in a 

 comparison of the experimental data with corresponding values 

 computed from the equation under consideration. As it is impos- 

 sible to avoid slight irregularities in obtaining data, the equation 

 must be so determined as to distribute the irregularities rather 



from 



laws of chance variation. 



made 



approximate 



l discovered which 

 the series of data 



