﻿Electricity in Flames containing Metallic Vapours. 881 



It is interesting to compare the positive carriers in the 

 flame with the canal rays in a vacuum-tube, which are also 

 charged atoms. In both cases we have the constant alterna- 

 tion of the positive and neutral states *. Similarly, we have 

 found in flames a result already known for canal rays, namely, 

 that there exist metallic atoms in flames, in comparatively 

 small numbers, which are negatively charged. These are 

 analogous to the negatively charged atoms found in canal 

 rays by Wien f, whose energy was much less than that of 

 the positive ones. 



In the following we give a critical resume of previous 

 work on the subject. H. A. Wilson, in the work already 



quoted (p. 865), obtained the value 60 — '- / as the 



* u ' sec./ cm. 



velocity of the positive metallic carrier, a value about 

 20 times as large as that found here for the carriers of the 

 second kind . It is difficult, however, to determine exactly 

 the smallest electric force which suffices to drive the carriers 

 against the current of the flame-gases. Wilson measured 

 the distribution of potential in the flame ivithout the presence 

 of metallic salt ; he drew the potential curve and assumed 

 that the smallest gradient here shown was sufficient to give 

 the carriers the velocity of the flame. This smallest gradient 

 lies, however, close by the positive upper electrode, in the 

 place occupied by the metallic vapour in the experiment itself, 

 and, since the carriers which come into account here proceed, 

 as we think to have demonstrated, from the* edge of the 

 luminous vapour (i. e. in this case from the lower boundary 

 of the vapour), and not from the upper electrode, this smallest 

 gradient occurs in a region where the carriers are not 

 necessarily driven down against the flame. 



Moreau % imeasured the velocity of the positive carriers 

 with two vertical electrodes, one standing in metal vapour, 

 the other in free flame, and assumed that the potential 

 gradient was approximately uniform between the electrons. 

 This is far from being true (as Lusby has shown, for instance, 

 in work to be quoted immediately); it is not to be wondered 



at that his value, 80 / , d 



sec./ cm. 



oes 



not ao-ree with mine. 



* W. Wien, Ann. der Physik [4] xxvii. 1903, p. 1025. The result 

 was first obtained by Lenard for flames {Ami. der Physik [4] ix. 1902, 

 p. 649) and found a<?ain for canal rays by Wien, cf. W. Wien, Ann. der 

 Physik [4] xxx. p. 368 (1909). 



t W. Wien, Ann. der Physik [4] xxxiii. p. 871 (1910). 



} G. Moreau, Ann. de Chimie et de Physique [7] xxx. p. 39 (1903). 



Phil. Mag. S. 6. Vol. 23. No. 138. June 1912. 3 M 



