4S4 SCIENCE PROGRESS 



formula the value p = y6$.io 2Z . The value of the e/mc was 

 deduced from the same experiments to be 2*04. io 7 , and the 

 tolerable agreement with the values found by more direct 

 methods serves as a measure of the degree of accuracy attained. 

 This value of p gives N = 3.10 14 as the number of electrons per 

 unit volume in a flame coloured by sodium vapour. 



Another mode of experiment, devised by Macaluso and 

 Corbino, was used by Geiger, 1 depending upon the displacements 

 of interference bands, and values of p were obtained varying 

 with the wave length between the limits for 



Sodium r63.io 2S to 4'83.io 33 



Potassium .... 0.82.10" to 2 ro.io 93 

 Lithium 5*2. 10" 



values which are of the same order of magnitude as the one 

 above. Now Hallo (Dissert p. 92) estimated the number of 

 molecules present per unit volume in his experiments, and 

 concluded that only a small fraction of these molecules are, at 

 any instant, concerned in the emission or absorption of light, 

 and that accordingly the mere presence of a sodium molecule in 

 the flame is not a sufficient condition for its taking part in 

 radiation or absorption : to do so, it must necessarily be in some 

 special state. The exact nature of that state is unknown, but 

 only a small fraction of the molecules are in it at any given 

 instant. 



If now this result be accepted it would appear as though 

 the values of N required to account for the observed pressure 

 displacement are possibly rather large ; if, on the other hand, 

 these are decreased the calculated value of dX/A, will become 

 smaller than the experimental. There is the further difficulty 

 that if this conclusion is true there is no reason why the number 

 of electrons emitting radiation of any given wave length should 

 increase proportionately to the density, as it has above been 

 assumed to do. Too much stress must not, however, be laid on 

 these objections. Hallo's result is by no means conclusive, and 

 the conditions existing in the arc are so different from those in 

 a flame coloured with sodium vapour, and so little is known 

 about the exact nature of these conditions, that it cannot be said 

 with certainty whether the numerical agreement between theory 

 and experiment is good or otherwise, and to draw premature 



1 Ann. der Phys. 23, p. 758, 1907 ; 24. p. 597, 1907. 



