220 Mr. H. H. Lester: Determination of Work Function 



There is a temperature term in the expression for <p, viz. 



-(6\— #o) j but this term contributes at most only about 



7 per cent, to the value of <£. 



If the Richardson equation be written in the form 



log i- s log = -b.^ + A, 



or 



where 



X = -6Y + A, 



X = log i— ol°g^ an d Y = 7j, 



then the current-temperature curve becomes a straight line 

 whose slope is — b. Knowing <j>, b may be obtained from 

 equation (7). A single determination of i for a known 

 temperature will fix the position of the line. Such a 

 determination may be made by observing i for a known 

 temperature, say the melting-point of platinum. The 

 curve once fixed may be used to determine higher tempe- 

 ratures. The method would give relative temperatures 

 with great accuracy, and would give absolute values with an 

 accuracy depending upon the accuracy of the determination 

 of the melting-point of platinum. It has the advantage 

 that it is independent of radiation formulae. If, as seems 

 probable, <fr is very nearly constant for all metals, it would 

 not be necessary to determine <£ in applying the method. 

 In regard to the constancy of 0, it may be pointed out 

 that all the determinations of the cooling effect so far 

 made have given identical values of (f> within the limits 

 of experimental error. However, the metals tested have 

 been at one end of the voltaic series where large differ- 

 ences would not be expected. Richardson and Cooke 

 determined from the heating effect for much more 

 electro-positive metals. Their measurements were not 

 accurate, but the values found were greater than, instead 

 of less than, the values found for cj) from the cooling effect. 

 It has been pointed out that their measurements probably 

 contained an error, due to the ionized film at the cold 

 surface. If suitable allowance be made for this error, the 

 values found for <j> for these metals fall in with the values 

 found from the heating effect at about 4*5 volts. The point 

 is important, and will be investigated further. 



