LUMINOSITY OF FLAMES CONTAINING SALT VAPOURS. 
69 
small amounts and large potentials the fall of potential was nearly uniform, but with 
potassium it nearly all occurred near the lower negative electrode. 
It was found that if the tube, A A' (fig. l), carrying the lower grating was removed 
and the burner itself used as the lower electrode then the puffs of salt were deflected 
downwards when the burner was charged negatively. The deflexion was about 
0‘5 cm. for 1000 volts in the case of all salts. 
This gives 
7 _ 100 x 3 x Q‘5 
1000 
= 0'15 cm. per second per volt per centimetre. 
Charging the burner negatively when the tube, AA', was removed depressed the 
inner cones of the flame and caused the flame to broaden out slightly just above the 
burner. Evidently the electric field retarded the upward motion of the flame and so 
produced the downward deflexion of the puffs of salt vapour. When the burner was 
charged positively no deflexion could be detected. The tube, AA', and platinum 
grating were therefore used to shield the inner cones from the electric field and to 
allow the puffs of salt to get well mixed with the flame before entering the field. 
The tube, AA', also served to greatly steady the flame. The flame is not affected by 
the field when the tube, A A', and lower grating are used because it is practically 
uniform throughout the distance between the upper and lower gratings. 
Since no deflexion of the puffs was obtained when using the lower grating and the 
tube, AA', it appears that the luminous salt vapours in flames do not really move 
relatively to the flame in an electric field. 1 think the motion observed by Lenard, 
Ebert and Andrade was due to the deflexion of the flame which is difficult to 
estimate and allow for. 
Andrade found k = 0T6 cm. per second per volt per centimetre, which is about 
twenty times the upper limit given by the experiments just described, but agrees 
nearly with the value found when the lower part of the flame was not protected from 
the electric field. When the potential difference between the two gratings was increased 
sufficiently an arc formed accompanied by a great increase in the current. The arc 
could be seen to form a luminous streak down the flame, and the electrodes got very 
hot at each end of the arc and fused if the arc was allowed to continue for long. 
When an arc was formed the puffs of salt vapour became more brightly luminous and 
the coloration due to the salt could be seen, although faintly, all along the track of 
the arc. The position of the puffs was not changed when an arc formed. It appears, 
therefore, that positive ions move down the flame from the puffs when an arc is 
formed and form a small amount of luminous salt vapour in the arc. I think some 
positive ions are probably deflected downwards from the puffs, even when the field is 
not strong enough to arc, but they are not numerous enough to form visible vapour. 
In the arc the temperature is higher so that a smaller amount of salt vapour becomes 
perceptibly luminous. 
