3 50 PHENOMENA, ATOMS, AND MOLECULES 



posite surface of Cs and bare tungsten as is the case at lower ^'s. At still 

 higher values of 6 the already crowded layer tends to be compressed and the 

 adsorbed atoms are given in efifect smaller atomic volumes. In general, 

 smaller atomic volume is accompanied by lower electron emission. This 

 may explain the decrease in Vg at values of d^o.'j. 



Since the values of contact potential from v^ and Vp agree so well below 

 6 = 0.50, the values from Va have been used with Eq. (19} to construct 

 plots of Ve vs. i/T down to low values of 6. Fig. 15 shows a family of curves 

 calculated for a series of values of \ka. The curves were found for ^>o.5 

 by using the values of Vc taken from the smooth curve (Fig. 14) through 

 the points calculated from Vg. The highest (412° K.) corresponds to a 

 pressure of 8.8 baryes ; the lowest (237° K.) to 6.7 X 10"^ baryes. The 

 curves at higher pressures may be compared with older unpublished data 

 of Kingdon taken in the region of very low ^ (o — 3 percent), where the 

 caesium emission approaches that of pure tungsten (yw)- This is of 

 particular interest because it may show whether the equations and con- 

 stants obtained at the pressures of the present experiments can be used 

 to calculate electron emissions at pressures many thousand times greater. 



The comparison showed a difference of about 20 percent in the ratio 

 Ve/vw, the calculated values of Ve/v„, being lower. This is a satisfactory 

 agreement considering the difficulty of making experiments at high 

 pressures. 



In connection with this comparison with the data of Kingdon, it is 

 desired to correct certain equations and statements in the first paper by 

 Langmuir and Kingdon ^ on the Thermionic Effects Caused by Vapors of 

 Alkali Metals. In this work no scale of 6 was available and an equation 

 (Eq. (10), page (71)) from which had been eliminated was derived. 

 This equation was to be applied in the region of very low 6. This equation 

 in corrected form should be, 



{pjv^.y^"= \n{vjv:)^{c/d)va ■ (21) 



derived from the two approximate equations, valid for low values of B, 



\n{va/e)=a+aie\ \n{ue/e)=cd. 



The equations representing v^ and v^ in the present experiments have 

 been put in this form and the constants determined. 



ai = 23,100/r-f6.8; In d==62-32,380/T 



a = 62-32380/T\ aiA = 0.188 + 5.53X IQ-^r. 



c=123,000/r. 

 * I. Langmuir and K. H. Kingdon. Proc. of the Royal Soc, A, 707, 61 (1925)- 



