RICHARDS, COLLINS, AND HEIMROD. — COPPER AND SILVER. 139 



Lord Rayleigh did not neglect this source of error, and by some very 

 ingenious qualitative experiments showed that it could not be large. It 

 seemed to us w^orth while, however, to analyze quantitatively the precip- 

 itated silver, and the following table recounts this series of trials, in which 

 the silver was dissolved, and precipitated by the purest hydrobromic 

 acid with great care.* 



The result of this table somewhat surprised us; for few crystalline 

 precipitates contain so little included mother liquor as 0.0] per cent. 

 While the analytical work is not perfect, (for the result of Experiment 33 

 can only be ascribed to accident,) one is forced to conclude that the pre- 

 cipitate is very nearly pure silver. Evidently the electrical method of 

 precipitation insures a more compact structure than is possible when the 

 precipitation does not involve outside electromotive force. One of us 

 found three times as much water in crystallized silver prepared at 0° in 

 another way.f The source of error in the voltameter is apparently more 

 subtle than such a merely mechanical cause as inclusion. 



Further experimentation upon such silver deposits showed that, after 

 having been properly washed, they neither gained nor lost in weight 

 upon continued digestion with water or with neutral solutions of ar- 



* For the method, consult recent papers on atomic vveiglits from this Labo- 

 ratory. 



t These Proceedings, 23, 177 (1887). 



