776 Mr. S. G.Lusby'ontlie" 



all doubt by Gold * and H. A. Wilson f that in high tempe- 

 rature flames the negative ion is a pure electron. Within 

 the last few years, however, the attention of physicists has 

 been transferred to positive carriers of electricity. The 

 present paper gives an account of some measurements made 

 on the mobility of the positive ion in a Bunsen flame whose 

 temperature could be varied within fairly wide limits. Obser- 

 vations were made at intervals of about 200° on all salts whose 

 conductivity was high enough for accurate measurement. 



The first determinations of the velocities of flame ions were 

 made by H. A. Wilson J, whose results may be summarised 

 thus :— In a Bunsen flame at 2000° C. all salts of all metals 

 of the alkali and the alkaline earth groups produce a positive 

 ion whose mobility is 62, and a negative ion whose mobility 

 is 1030 cm. a second under an electric force of 1 volt per 

 centimetre ; in hot air at 1000° 0. positive ions of the former 

 group have a mobility of 7*2, those of the latter group 3*8, 

 whilst both groups give a negative ion whose mobility is 26. 

 It is now known that these results are inaccurate owing to 

 serious errors introduced in the determination of the potential 

 gradient in the flame. In the negative measurements a 

 uniform gradient was assumed, whilst in the positive mea- 

 surements the gradient was plotted, but only in the pure 

 flame, whereas the salt bead introduced in the experiment 

 very appreciably alters the distribution of potential in the 

 flame. An indirect determination of mobility was made by 

 Marx § from considerations of the potential gradient in a 

 flame in which salt vapours were sprayed. In order to make 

 his experiments amenable to mathematical treatment, Marx 

 was obliged to assume that all the ionization took place in 

 the flame, and none at the surface of the electrodes. The 

 values of the positive mobility thus obtained were very 

 variable, ranging from 120 to 340, with a doubtful mean 

 of 250. This method is very interesting from the theoretical 

 point of view, but is too indirect to give accurate results. 

 The only other measurements which concern us are those of 

 Moreau ||, who simplified matters by assuming a uniform fall 

 of potential in the flame, thus deducing a mobility of 80 at 

 2000° C. Thus all three of the results quoted suffer by 

 reason of incorrect values given to the potential gradient in 

 the flame. The order of the errors introduced is illustrated 



* Gold, Proc. Roy. Soc. Ixxix. A. (1907). 

 t Wilson, Proc. Eoy. Soc. lxxxii. A. (1909). 

 t Wilson, Phil. Trans. A. (1899). 

 § Marx, Ann. der Phi/s. ii. p. 790 (1900). 

 || Moreau, Ann. de Chim. etPhys. 1908. 



