with special reference to solutions of its salts. 561 
the solution was saturated at its boiling point. The residue 
poured off from the crystals after cooling was taken as the first 
fraction. The process of solution and crystallisation was repeated 
in the same manner 21 times, and the crystals of the last operation 
were taken as the second fraction. The most careful experiments 
failed to show any difference between the first and last fractions, 
though a difference of 1 °/ o could have been detected with cer- 
tainty. 
(3) Aluminium hydroxide, barium sulphate, silver chloride, 
and copper sulphide were precipitated from aluminium sulphate, 
barium chloride, silver nitrate and copper sulphate respectively 
mixed with separate portions of a strong solution of potassium 
sulphate. No activity similar to that of potassium could be 
detected in the precipitates. The potassium sulphate remaining 
was not examined, for it was thought that the impurities introduced 
in the processes of precipitation would render any result that might 
be obtained devoid of value. 
(4) A current of three amperes was passed for four hours 
between electrodes of platinum wire immersed in about 5 litres of 
a strong solution of potassium chloride. Immediately after the 
operation the electrodes were taken out, rinsed and examined for 
activity. No activity whatever could be observed. 
These four methods include most of those that have been used 
for the separation of radioactive products. They all yielded negative 
results. For the present, therefore, I have abandoned as hopeless 
the attempt to effect a separation, though I do not regard it as 
certain that others with better opportunities and greater chemical 
skill may not obtain a more favourable result. Even if potassium 
is intrinsically radioactive, it is contrary to all precedent that only 
one member of a radioactive series should show activity. 
II. The activity and absorption of solutions. 
§ 6. While endeavouring to find some preparation of a 
potassium compound which would show an activity differing 
markedly from the normal, it occurred to me to try the activity 
of such a compound in the state of solution. It was discovered at 
once that the activity of a solution was considerably less than that 
which would be predicted from the amount of potassium which it 
contains: the variation is shown in Table II, which gives a few of 
the many observations made upon solutions. 
