146 PROCEEDINGS OF THE AMERICAN ACADEMY. 



method at 0° is affected by the same source of error, so that the com- 

 parison is a fair one. 



The only alternative to this inference, — the supposition that the po- 

 rous cup introduces a new complication causing too low a weight, — is 

 difficult to disprove ; but, on the other hand, the possible nature of such a 

 complication is not easily conceivable. 



Recurring to the first inference, an hypothesis supported by many 

 facts, we admit it to be probable that the porous cup does not wholly 

 prevent the admixture of the hypothetical complex ions with the liquid 

 around the cathode ; for even if diffusion were wholly prevented, some of 

 these ions might migrate with the current. Both diffusion and migration 

 are increased in rapidity by a rise in temperature, hence even with the 

 porous cup the hot solution gives a somewhat higher value. 



It is a matter of great interest to study the effect of the low value 

 obtained by the new method upon the electrical atomic weight of copper, 

 upon the electrolytic value assigned to the ampere, and upon the elec- 

 trical result for Joule's equivalent. 



We concluded, at the close of the section devoted to copper, that the 

 atomic weight of this metal must lie between the limits 63.547 and 

 63.563, if the common silver voltameter is correct in its verdicts. We 

 have now shown that the results of this voltameter are in all probability 

 too high by 0.081 per cent; hence the conclusion is that the true atomic 

 weight of copper must lie between 68.598 and 63.615. Since the 

 elaborate chemical investigation of copper carried out some years ago in 

 this Laboratory yielded the value 63.604, we have here a remarkable con- 

 firmation of the results of the " porous cup voltameter." 



In this connection, it is worth while to call attention to the reasons 

 why copper placed in an ice-cold solution of argentic nitrate causes the 

 deposition of the true amount of silver.* Of this automatic voltaic cell 

 the anode is copper ; hence the complications which arise around a silver 

 anode do not exist. On the other hand, the cathode is silver; hence the 

 complications which take place around a copper cathode do not exist. 

 The trace of cuprous salt undoubtedly formed at the anode does not de- 

 compose the very cold ions of nitric acid, but sooner or later deposits 

 its equivalent of silver. The action is so slow at the low temperature 

 that the current density where the silver is deposited is very small, hence 

 there is no danger of the deposition of a weighable amount of copper 



* Richards, These Proceedings, 22, 342, 23, 177 (1887). 



