LUMINOSITY OF FLAMES CONTAINING SALT VAPOURS. 
71 
nearer together. The increase in the current was about 25 per cent, with potentials 
up to 1800 volts. With 2700 volts the increase was larger, varying from 50 to 300 per 
cent, with different salts. This indicates that between 1800 and 2700 volts the 
positive ions began to move down the flame. If the salt beads were put into the 
flame about 1 cm. or 2 cm. below the upper grating, instead of close to it, an arc 
started with 2700 volts. Luminous salt vapour could be seen along the whole length 
of the arc. 
The potential gradient below the beads of salt with potentials between 1800 and 
2700 volts was about 300 volts per centimetre. Since the velocity of the flame was 
300 cm. per second, this shows that the velocity of the positive ions was about 
1 cm. per second for one volt per centimetre. The velocity was about the same for all 
salts, but could not be estimated accurately because the changes in the current due to 
putting in the beads increased gradually with the potential difference. It appears 
that the maximum velocity of the positive salt ions due to 1 volt per centimetre is not 
much greater than 1 cm. per second. 
In the earlier experiments by the writer, # it was found that a bead of salt below 
the upper electrode began to increase the current at a P.D. of about 100 volts. 
LusbyI obtained a similar result and also found that the salt diminished the potential 
gradient below the bead. These results appear to have been due to a small amount 
of salt getting into the flame below the bead. This was prevented, in the experiments 
just described, by the chimney attached to the upper electrode. In the absence of 
the chimney, it was found that putting a bead below the upper electrode increased 
the current and diminished the potential gradient even with small potential differences- 
The salt vapour from the bead gets into the lower parts of the flame partly by 
circulating round with the air surrounding the flame inside the apparatus and partly 
by escaping into the air in the room and then entering the burner along with the air 
supply. The chimney entirely stopped both these processes. There was a strong 
draught up the chimney so that the salt vapour from the bead went straight up the 
chimney and did not get into the air surrounding the flame. 
It appears that the earlier attempts to measure the velocity of the positive ions of 
salt vapours in flames gave values much too high. AndradeJ has recently obtained 
values of about 3 cm. per second for 1 volt per centimetre for the positive ions of 
strontium salts. The difference between this and the value of about 1 cm. per second 
just obtained may be due partly to differences between Andrade’s flames and mine. 
Also there may have been a few positive ions, even in my flames, with velocities 
greater than 1 cm. per second. 
When a large increase in the current is produced by putting in the bead an arc 
forms as already mentioned. The potential gradient necessary to start an arc was 
* ‘Phil. Trans.,’ A, vol. 192, p. 499, 1899. 
t ‘ Proc. Camb. Phil. Soc.,’ vol. xvi., PI. 1, 1910. 
J Loc . at . 
VOL. CCXVI.-A. 
L 
