180 PHENOMENA, ATOMS, AND MOLECULES 



The choice of Ta = 300, at the lowest pressures and the tabulated 

 values at 0.207 nim-. was based upon calculations similar to those previously 

 employed in obtaining the temperature drop around a wire at low pressures. 

 Thus we may calculate cpa ~ (Pi according to (12), substituting, however, 

 Wc + Wd in place of Wp. The data for the calculation, together with the 

 results, are given in Table X. The values of Wc + Wd were 



taken directly from Table II. In this calculation it is assumed that the 

 hydrogen atoms leaving the filament all recombine to form molecules 

 within a distance X from the filament and that the heat thus liberated is 

 carried to the bulb by ordinary heat conduction. This assumption, of 

 course, cannot correspond exactly with the facts, but would lead to too 

 high a value for T^. It is therefore not surprising that the value T^ = 760 

 found at 0.207 mm. should be higher than the value To = 545 previously 

 calculated by Equation 46 from the data of Table VII. Taking the value 

 Ta = 545 as being correct and making a proportionate change in the 

 corresponding values of cpa — ^>l, we obtain the values of T^ given in Table 

 IX. At the lower pressures the values of To are so close to 300° that it 

 is not worth while to take into account the departures from this temperature. 



Before discussing the values of K of Table IX, we shall first attempt to 

 calculate K from the data at pressures of i.i and 4.4 mm. 



At these higher pressures there is more uncertainty as to the value that 

 should be assigned to To. We may, however, adopt a method similar to 

 that we employed to calculate To at a pressure of 0.207 mm. In the latter 

 case at very high temperatures Wd became independent of T showing 

 that the denominator in Equation 31 was practically zero. This led to 

 Equation 40, from which we determined C and then by (33) we solved 

 for To. At the pressures above 0.207 mm., however, Wd does not reach 

 a constant value at the highest temperatures, although it shows a distinct 

 tendency to do so. This is readily seen from Fig. 2, where log Wd has 

 been plotted against i/T for the data at pressures up to 4.4 mm. At the 

 three lower pressures the curve becomes horizontal at the highest tem- 

 peratures, but at I.I mm. the temperatures reached were not high enough 

 to give the maximum value of Wd- From the degree of curvature or the 



