of Electrical Decomposition. 



205 



of platinum, the spark recombiiics a portion of it, and the num- 

 ber of sparks, the size of the bubble preceding the spark, and 

 the quantity recombined, would vary indefinitely in each different 

 experiment, and even in the course of the same experiment. 



I now give a selection of experiments from my note-book with 

 the volumes and analyses of the gases evolved. The decompo- 

 sition was arrested in each case when the larger volume of gas 

 had reached to 0'3 c. i. The analyses were made by an eudio- 

 meter, which I have formerly described*, consisting of a loop of 

 fine platinum wire sealed into a graduated glass tube, and heated 

 or ignited by the voltaic battery : — 



1st. The gas from the positive pole contracted to 0*4 of its 

 original volume ; the residue, mixed with half its volume of oxy- 

 gen, detonated, leaving a minute residue, apparently nitrogen. 

 The gas from negative pole, 2-3 times the volume of the positive 

 contracted to 0*15 of its volume: the residue contracted very 

 slightly on the addition of hydrogen . 



2nd. Gas from positive, 2*5 volumes contracted to 0*7 of its 

 volume. 



From negative, 1 volume contracted to 0*4; the residues 

 mixed contracted again to 0*7; and this residue, mixed with 

 half its volume of oxygen, detonated, leaving only a very minute 

 bubble. 



3rd. Instead of two coated wires, one coated wire was em- 

 ployed, and the opposite terminal consisted of a strip of platinum 

 foil, \ inch width and 4 inches length, exposing consequently, 

 reckoning both sides, 2 square inches of surface. 



This strip was attached to a platinum wire, which was sealed 

 into a glass tube, the extremity of the platinum foil being ^ an 

 inch from the orifice of the tube, so that no gas could escape. 

 (See fig. 2.) 



Fig. 2. 



* Phil. Mag. Aug. 1841. Phil. Trans. 1847, p. h 



