434 
Dl;. TTAIJOLI) A. WILSON OX THE 
A solution containing 20 grainmes of NaOl in a litre was sprayed for one hour, and 
the amount of NaCl in the phig determined h}^ washing out with water and titrating 
against standard AgNOg solution. That the plug stopped all the spray was shown 
by passing the escaping air into a Bunsen flame when no coloration could be 
detected. With a shorter plug of glass-wool all the spray was not stopped. In 
this way it was found that 4’7 milligrammes of NaCl were stopped by the plug in 
one hour, this number being the mean of several concordant determinations. 
In the later experiments a new sprayer was used, which sprayed 2'2 times as 
mucli salt as tlie old one. The i-atio of the amounts of salt spraved by the two 
s])rayers was got Ijy comparing the conductivities due to 1 per cent. KCl solutions 
when sprayed by the two sprayers under similar conditions at 1200 °, and with an 
E. M.F. of 800 volts. Under tliese conditions the current is proportional to the 
electrochemical equivalent of tlie salt })assing through the tube, as will be shown 
below. The amount of KI passing through the tube during the above measurements 
was therefore 1'46 X i0“® milligramme per second. 
To electrolyse this amount of KI in a solution retpiires a current of 
I'lG X 10“’" X 39 
1G6 X (j'403 
= 8'5 X 10 ^ ampere. 
'the current obtained at 1400° with 800 volts was 7 X 10“'^ ampere, which is 
tlierefore nearly ecpial to the current recjuired to electrolyse the KI used in a 
sohttion. 
It is prol:)able tliat at high temperatures the KI is converted into KOH by the 
watei' vapoui' present. The reaction l>eing 
KI + HoO = KOH + HI. 
Tlie view that all salts of the alkali metals are converted into hydrates at a high 
tem|)erature in the presence of water vapour was used by Aruhexius to explain his 
results on the electric conductivity of flames, and this view is also in accord with the 
results of the investigation published in 1809, “ On the Electric Oonductivity and 
Luminosity of Flames containing Va])orised Salts.According to this view the 
falling off in the current above 900° might lie explained as being due to KOH being 
less ionized than KI at this temperature. 
Thus up to about 900° the conductivity may be supposed entirely due to KI, but 
at this tenqierature KOH and HI begin to lie formed, and conduct less well than KI, 
so that the current falls off. As the temperature rises, however, the KOH begins to 
dissociate into ions, so that at 1200° the current attains a maximum value when it is 
C( >nipletely electrolysed. 
The energy required to ionize a gramme molecular weight of KI can he calculated 
in the same way as was done for air by using tlie values of the current obtained at 
* ‘ Phil. Trans.,’ A, vol. 193, 1899. 
