966 Transactions of the American Institute. 



can be taken only as approximation to truth, which will be more or 

 less remote as our materials are more 'or less impure. 



60. In the investigation of these laws, the writer has made a great 

 number of tests with water voltameters, and those of copper, silver 

 and gold, the results of which were at first very discouraging ; no two 

 being found to agree in their equivalents. 



61. But finally two copper baths, one of sulphate of copper, and 

 one of nitrate of copper, were tried. The plates to be used were first 

 coated with reguline copper, by deposition from solutions in which 

 anodes were used of the purest copper to be obtained. Now, by using 

 one of such plates as anode and another as cathode, and occasionally 

 reversing the current until the solutions became so entirely neutral 

 that the weight of metal lost from the anode, and that gained upon 

 the cathode, were equal ; and the transport in the two voltameters 

 were also equal ; these were, therefore, taken as giving correctly 

 equivalent proportions for copper. 



62. But in the electrolysis of acidulated water by the same current, 

 it was found that no such voltameter as is described by Faraday or 

 other authors would develop a measure of corrected gas sufficient to 

 amount to the same equivalent — a circumstance evincing clearly the 

 fact, that between electrodes of any considerable size a part of the 

 current is conducted without decomposing the water. 



63. But by fixing two pieces of No. 20 plantinum wire across a 

 glass tube of an interior diameter of four-tenths inch, and a length 

 of say six inches (the tube being broken and the wires put across and 

 melted in), a voltameter was at length constructed which gives, in 

 same circuit, a volume of gas precisely equivalent to the copper 

 deposed in the copper voltameters, and from this true multipliers are 

 obtained. 



64. In common practice every operator can easily obtain for him- 

 self the constant multiplier, which, for his own galvanometer and bath, 

 will- determine the amount of work performed by the current with 

 as much accuracy as is attainable by weights and measures. 



65. To do this let him take a few articles (such as present large sur- 

 faces in proportion to their weights are best), and accurately weigh 

 them ; then let them be placed in the bath and remain a suitable time, 

 which must be accurately noted together with the mean deflection of 

 the galvanometer. Now, let them be accurately weighed again, and 

 the weight per minute estimated in grains ; dividing the number of 

 grains by the tangent of the mean deflection gives the constant 

 sought. 



