EVAPORATION OF ATOMS 337 



the leads at bath temperature. Below, ^ 750° K., an increase in filament 

 temperature due to radiation received from filament B (1200°) sets in 



The lead cooling correction was calculated '^ and it was shown that 

 though the temperature of the 3 cm long central part of the filament was 

 reduced, the distribution remained uniform to <i°. The lead loss correc- 

 tion was —3° at 550° K. and —20° at 430° K., the lowest temperature 

 used in measurements of atom evaporation. 



The radiation correction was determined in two ways. First the change 

 in resistance of the filament A was observed with B hot and cold, at various 

 temperatures of A. Second, the current required to produce a given electron 

 emission (|Xa constant) when filament B was hot and cold was observed. 

 By comparison the true temperature of A with B hot was obtained. The 

 radiation correction amounted to -|-3° at 750° K. and +50° at 430° K. 



V. DETERMINATION OF o^i 



Various low pressures ([Xg. = 10^^ to 10^^) of caesium were established 

 and then filament A was maintained at a low temperature Ti until trial 

 showed that 04 had reached its limiting steady value fixed by the balance 

 between evaporation and condensation. In successive experiments, as T\ 

 was progressively lowered, it w^as found that 0.1, as given by the 2-filament 

 method, increased at first rather rapidly and then very slowly until finally 

 when Ti was reduced below about 325° K. no further increase in 0.4 

 occurred. This quite definite limiting value which was observed in numer- 

 ous experiments was presumed to be the value a^i corresponding to a com- 

 plete monatomic film. These experiments gave a^i = (4.80 ± 0.05) X 10^* 

 atoms cm"^. 



Intervals at higher temperatures (.—'400° K.) to allow possible favor- 

 able rearrangements by migration did not change a^i. Only by cooling the 

 filament (by immersion of leads in liquid air) below the temperature (bath 

 temperature) corresponding to saturated caesium vapor were other values 

 of a observed. These were greater than 4.8 X io~^^ and increased very 

 rapidly as the filament was cooled below bath temperature. If the filament 

 was now heated only slightly above bath temperature, a^i = 4.8 X lo"^'* 

 was again obtained. 



In these experiments no detailed study was made of the conditions 

 which give values of o^ slightly lower than a^i. In later experiments, 

 Section XII, the actual slow variation of o with T and [Aa in this region 

 was recorded and further justification is given for regarding this limiting 

 value as that corresponding to a complete monatomic film. 



® I. Langmuir, S. MacLane and K. B. Blodgett, Phys. Rev., 35, 478 (1930). 



