122 Prof. H. A. Wilson on the Diffusion of 



bead of sodium sulphate which was supported on a wire 

 attached to one prong of a vibrating tuning-fork. The fork 

 also carried two small screens between which the flame could 

 be seen once in each vibration of the fork. The salt vapour 

 from the vibrating bead when observed through the screens 

 had a wavy outline, and the wave-length was measured. The 

 wave-length multiplied by the number of vibrations per sec. 

 gives the upward velocity of the flame. In this way the 

 velocity was found to be about 200 cm. per sec. 



The experiments do not admit of much accuracy because it 

 is difficult to set the platinum wires on the rather hazy 

 boundary of the region emitting light. After some practice, 

 however, independent values found for K did not differ by 

 more than about 10 per cent. 



In addition to the region emitting light which surrounds 

 the bead, there is a separate region which forms a sheath 

 covering the outside of the flame. This extends from the top 

 of the flame down to a few centims. above the bead. This 

 outer sheath obscures the upper part of the inner region 

 and makes the observations more difficult. Very often the 

 inner region is not the same colour as the sheath. The 

 ca?sium and lithium salts available contained some sodium, 

 and the bright coloration due to it made it difficult to set the 

 wires on the edge of the coloration. In the case of the lithium 

 salt the red colour was outside the sodium yellow, so that it 

 did not matter much, but the caesium coloration was badly 

 obscured. 



The following table gives the values found in centims. for 

 the distance from the bead to a point on the widest part of 

 the inner region emitting light (r m ) and the width of the 

 widest part (2y m ). Each number is the mean of several 

 determinations. 





r in . 



2y m . 



K. 



Lithium 



... 2-4 



1-6 



14-5 



Sodium 



... 2-0 



1-3 



11-5 



Potassium! . . . 



... 2-0 



0-85 



47 



Rubidium ... 



... 1-9 



0-80 



4-4 



Caesium ...... 



... 1-9 



0-80 



44 



It appears that caesium, rubidium, and potassium salt 

 vapours diffuse at nearly equal rates, while sodium and 

 lithium diffuse more quickly. The velocity of the positive 

 ions of alkali salt vapours in flames appears to be about 

 60 cm. per sec. for 1 volt per cm.* at high temperatures. 



* H. A. Wilson. Phil. Trans. A. vol. cxcii. p. 499 (1899). 



