342 
CHEMISTRY: RICHARDS AND HALL 
other. The weight thus removed was subtracted from the weight of 
the fused salt. Before filtration the boat had been carefully lifted out 
and thoroughly washed; and the filtration was collected in the pre- 
cipitating flask. 
A supposedly equivalent weight of silver was next dissolved in pure 
nitric acid mixed with an equal bulk of water in a flask with a chain of 
bulbs ground in to the neck, a steam bath being used to heat the solu- 
tion. More water was then added and the temperature raised until a 
few seconds' gentle boiling had removed nitrous fumes. The solution 
having been made up to a liter, it was added very slowly with constant 
agitation to the lead chloride solution. Actinic light was henceforth 
rigorously excluded from the precipitated silver chloride. The pre- 
cipitate was coagulated by violent shaking, usually for fifteen minutes, 
and allowed to stand at least twenty-four hours, with further occasional 
shakings, before samples were removed for tests. The end point was 
found by the customary method by means of the nephelometer, and 
tested for several days to be certain that it did not change. 
Vacuum corrections were applied as usual, and all customary pre- 
cautions were taken. The results follow. 
TABLE 1 
0 Atomic Weights 
ORDINARY LEAD CARNOTITE LEAD 
A B 
207.187 206.426 206.406 
207.187 206.409 206.422 
206.431 206.399 
Average 207.187 206.422 206.409 
This table includes all the analyses of these materials which were 
completed ; any which had become the object of suspicion were rejected 
at an early stage. 
Obviously, if the fractional crystallizations produced any change in 
the relative concentrations of radium G and lead, this change is very 
small. The means of the two series of determinations of isotopfc lead 
differ only by about 6 parts in 100,000. This is certainly within the 
bounds of possible experimental error, though the low 'average errors' 
of the two sets (0.008 unit in each case) tempt one to assign a meaning 
to the observed difference. At any rate, we are safe in concluding that 
assuming the atomic weights of 'radium G' and lead to be 206.0 and 
207.2, respectively, no change in concentration greater than 13/1200 
or 1.1% has been obtained by nine hundred crystalHzations. Even 
supposing that a real difference in solubility exists, it is evident that 
