789.  In the preceding cases, except the first, the water is believed to be inactive; but to avoid any ambiguity arising from its presence, I sought for substances from which it should be absent altogether; and, taking advantage of the law of conduction already developed (380. &c.), I soon found abundance, amongst which protochloride of tin was first subjected to decomposition in the following manner.  A piece of platina wire had one extremity coiled up into a small knob, and, having been carefully weighed, was sealed hermetically into a piece of bottle-glass tube, so that the knob should be at the bottom of the tube within (fig. 68.).  The tube was suspended by a piece of platina wire, so that the heat of a spirit-lamp could be applied to it.  Recently fused protochloride of tin was introduced in sufficient quantity to occupy, when melted, about one-half of the tube; the wire of the tube was connected with a volta-electrometer (711.), which was itself connected with the negative end of a voltaic battery; and a platina wire connected with the positive end of the same battery was dipped into the fused chloride in the tube; being however so bent, that it could not by any shake of the hand or apparatus touch the negative electrode at the bottom of the vessel.  The whole arrangement is delineated in fig. 69.

790.  Under these circumstances the chloride of tin was decomposed:  the chlorine evolved at the positive electrode formed bichloride of tin (779.), which passed away in fumes, and the tin evolved at the negative electrode combined with the platina, forming an alloy, fusible at the temperature to which the tube was subjected, and therefore never occasioning metallic communication through the decomposing chloride.  When the experiment had been continued so long as to yield a reasonable quantity of gas in the volta-electrometer, the battery connexion was broken, the positive electrode removed, and the tube and remaining chloride allowed to cool.  When cold, the tube was broken open, the rest of the chloride and the glass being easily separable from the platina wire and its button of alloy.  The latter when washed was then reweighed, and the increase gave the weight of the tin reduced.

791.  I will give the particular results of one experiment, in illustration of the mode adopted in this and others, the results of which I shall have occasion to quote.  The negative electrode weighed at first 20 grains; after the experiment, it, with its button of alloy, weighed 23.2 grains.  The tin evolved by the electric current at the cathode:  weighed therefore 3.2 grains.  The quantity of oxygen and hydrogen collected in the volta-electrometer = 3.85 cubic inches.  As 100 cubic inches of oxygen and hydrogen, in the proportions to form water, may be considered as weighing 12.92 grains, the 3.85 cubic inches would weigh 0.49742 of a grain; that being, therefore, the weight of water decomposed by the same electric current as was able to decompose such weight