560 Mr Campbell, The Radioactivity of Potassium, 
Since the value of X in these experiments is not less than 10, the 
activity of a layer 0'5 cm. thick will not differ from that of an 
infinite layer by more than 1^ °/ o , and the apparent activity of 
materials containing the same proportion of potassium should be 
inversely proportional to the value of X/p. 
To test the view that the variation in the calculated values of 
the activity of K 2 0 is due to variations in this quantity and not 
in the activity of the potassium, a sample of the sulphate was pre- 
pared from the nitrate which gives too low a value for K 2 0. The 
activity of this sample is given in Table I, No. 8, and agrees well 
with that of the other samples. Potassium sulphate Nos. 5 and 6 
were prepared from the specimen of kainite shown in Table I, and 
No. 7 from the “muriate of potash.” Potassium chloride No. 3 
was prepared from potassium sulphate No. 2 by precipitation with 
barium chloride. It appears then that, though the activity is not 
always proportional to the content of potassium, it is always the 
same when the apparently abnormal compound is turned into 
sulphate or chloride. The explanation by means of the absorption 
coefficient seems more reasonable than the only alternative, that 
the activity actually varies with the state of chemical combination, 
for no such variation has been observed in any other case of radio- 
activity. It is, moreover, supported by the observations on solutions 
described in the later part of this paper. 
It may be noted that, while the value of K.,0 deduced from 
the iodide is lower than that deduced from the sulphate in these 
experiments, it is too high in those of a former paper. This 
circumstance accords well with the hypothesis just suggested, for 
in the former experiments the layers were only 0T5 cm. thick. If 
the value of X is greater for the iodide than for the sulphate, 
a layer of this thickness would be more nearly “ infinitely thick ” 
for the former than for the latter salt. In thin layers the iodide 
may appear more active, in thick layers less active than the 
sulphate. 
§ 5. Since no evidence could be found of a separation from 
potassium of any part of its activity in salts obtained by the 
methods described, an endeavour was made to effect a separation 
by the means employed in the study of the decomposition pro- 
ducts of the elements recognised as radioactive. 
(1) 400 grammes of potassium sulphate were heated to a dull 
redness for eight hours and to the highest temperature attainable 
in a blow-pipe flame for 15 minutes. No change was produced in 
the activity measured after cooling. 
(2) In the former paper quoted above an account was given 
of an attempt to separate the activity by re-crystallisation. This 
attempt has been repeated with yet greater care. 1000 grammes 
of the sulphate were dissolved in such a quantity of water that 
