INssociation and Electrical Conductivity. 45.3 
apparatus. The air was drawn through a tube much narrower 
than that shown in fig. 9, and a strong field created between 
it and an electrode along the centre. We were unable to 
stop by this means the formation of fog. Bubbling through 
several inches depth of various liquids also failed to stop it, 
as did also a tight plug of cotton-wool when placed near the 
water-bottle. When the plug was placed near the apparatus, 
the fog was greatly reduced. The air could be allowed to 
remain in the connecting tubing for several minutes before 
admission to the water-bottle without greatly impairing its 
efficiency as a fog producer. ‘The introduction of iodine 
vapour into the bottle, brought about by sucking air over 
boiling iodine, caused no fog. 
These fog experiments seem to show that the ions are bodies 
of considerable size, while those on their velocity would, if 
trustworthy, show that they are similar whether they come 
from the chloride, iodide, or bromide, and would thus preclude 
the halogens themselves. Their source would thus be narrowed 
down to either the zine or, perhaps, the oxygen of the air. 
The former of these views would correspond to the ejection 
of a rays from radioactive substances, the fact of the chemical 
dissociation of the compound, bromide, iodide, or chloride 
into its elements rendering this ejection possible. The latter 
of the two explanations would proceed on the lines marked 
by Prof. J. J. Thomson * to explain the power of oxygen to 
change a negative into a positive electrification, in the gas 
near an incandescent solid. Experiments to test these views 
are now in progress. 
The energy required to ionize a gram-molecule of zinc- 
chloride vapour may be calculated, if we take the curve in 
fig. 4 and assume that the negative ions are ejected from the 
ZnCl, vapour. This quantity, g, in calories, is given by the 
formula fT 
m, and nz being the small currents, in arbitrary units, at 
absolute temperatures 0, and @,, where 6,—0, is only a small 
interval of temperature. Fora mean temperature of 335° C. 
the result is 165,000 calories. Taking the charges on the 
ions to be the same as in electrolysis, an estimate may be 
formed of the potential-difference, V volts, through which 
the charge must fall freely in order to gain energy enough 
to ionize a molecule. Since it requires 96,000 coulombs to 
* ‘Conduction in Gases.’ 
+ H. A. Wilson, Phil. Trans. exevii. A. p. 429 (1901). 
