248 M. Viard on the Electro-chemical Deportment of Oocygen. 



their execution. It is necessary, in order to avoid oxidation at 

 the exterior surface of the electrolyte, to cover the exterior parts 

 of the stems to the plates in both glass tubes with gum -lac ; first, 

 to fill the apparatus with the liquid deprived of air by ebullition, 

 in order to be assured of the equality of the plates before admit- 

 ting oxygen ; and lastly, in all observations to avoid the effects 

 of agitation as much as possible. 



In order that my researches might include all the metals, I 

 wished to examine the influence of oxygen on the one hand upon 

 potassium, and on the other upon platinum. 



As potassium possesses the property of decomposing water, it 

 was of course impossible to employ it in its pure state: the 

 amalgam of potassium alone could be operated upon. 



To the bottom of two broad glass tubes, and where the glass 

 was thin, two platinum wires were soldered, the upper parts of 

 which were curved back in a circle. An amalgam of potassium 

 was formed with distilled mercury and a very small quantity of 

 potassium ; one-half of the amalgam was then poured into each 

 tube. The two tubes were first filled with boiling water, the 

 two wires were joined to the galvanometer, and the two liquids 

 were united by means of a kind of siphon, full of water, but 

 closed at its two extremities by membranes. At first only a 

 feeble current was observable ; but when the one tube contained 

 boiled water and the other aerated, it could again be recognized 

 that the presence of oxygen rendered the amalgam negative. 

 The experiment also succeeded when the water was replaced by 

 marine salt. 



Platinum led to the same conclusion ; but the first experi- 

 ments made with the same apparatus which had served for zinc, 

 iron, and copper, having caused some incertitude, much larger 

 plates of platinum were operated with, and, at the same time, all 

 contact between the platinum plates and the membranes was 

 avoided. The surface of these plates contained 20 square centi- 

 metres, and they were connected with the galvanometer by means 

 of extremely fine platinum wires. The plates and the wires were 

 submitted to the action of heated nitric acid, of concentrated 

 sulphuric acid, and of dilute sulphuric acid, and lastly, they were 

 washed several times in water. Two pairs of bottles were used 

 large enough completely to enclose the plates. In the first pair, 

 both glasses contained the same electrolyte at the same degree 

 of oxidation, boiled or aerated ; in the second pair, the one ves- 

 sel contained boiled liquid, the other aerated. When the two 

 plates were plunged into the first, and a communication was 

 established between the two liquids by means of a large siphon 

 terminated by membranes, or rather by means of a small basin, 

 bearing two vertical tubes closed by membranes (fig. 4), the 



