[ 902 ; 



CI. Intelligence and Miscellaneous Articles. 



THE DIFFUSION OF ALKALI SALT VAPOURS IN FLAMES. 



The Rice Institute, 

 Houston, 

 Texas, U.S.A. 



Oct. 25, 1912. 



To the Editors of the Philosophical Magazine. 

 Gentlemen, — 



IN reply to Mr. A. Becker's letter published in the Philosophical 

 Magazine for October 1912, I beg to say that I very much 

 regret that I was not acquainted with his previous work on the 

 diffusion of alkali salt vapours in flames when I wrote my paper 

 on that subject which appeared in your July number. Unfortu- 

 nately the publications of the Heidelberg Academy of Sciences are 

 not accessible to me. As to my measurement of the velocity of 

 the flame gases, it was not my intention to claim any novelty. 

 Similar methods had previously been used by various people, and 

 as I only required a rough, estimate of the velocity I considered 

 the rough method I adopted sufficient for the purpose. I did not 

 suppose anyone would desire to be known as the inventor of the 

 method I used. 



I do not consider methods depending on the force exerted on a 

 small ball by the stream of gas as likely to be exact, since they 

 must require a knowledge of the coefficient of viscosity of the 

 flame gases, a quantity which can scarcely be fouud with any 

 exactness. I think that the very considerable divergence between 

 the values found for such quantities as ionic velocities and co- 

 efficients of diffusion in flames by different observers is partly due 

 to differences between the flames used. These quantities un- 

 doubtedly vary very much in different parts of the same flame, and 

 large differences must occur in different flames. I think it quite 

 possible that Mr. Becker's flames have been colder than mine to 

 an extent sufficient to account for the smaller value he obtained 

 for the coefficient of diffusion of sodium vapour. In my experi- 

 ments I also used lithium, potassium, rubidium, and caesium 

 vapours, and obtained consistent results with all these metals. 

 The possible errors in such experiments are certainly very large, 

 and I have no desire to claim greater accuracy for my work than 

 for that of others. 



Tours very truly, 



Harold A. Wilson. 



