Dr Davy on Atmospheric Electricity. 443 



This result was obtained during stormy and showery weather, 

 with a sirocco wind. 



Experiments were tried with a view to the decomposition of 

 some metallic solutions, such as the nitrate of silver, muriate of 

 platinum, and sulphate of copper, using wires of different metals 

 not capable of acting on them chemically, but no satisfactory re- 

 sult was witnessed, excepting once with a solution of sulphate of 

 copper and platina wires, and during a thunder storm, when a 

 slight, just perceptible, precipitate of copper took place on each 

 of the wires immersed. 



Occasionally, for comparison, the electro-chemical experiments 

 were interrupted, and the conductors were connected with a gal- 

 vanometer, and also with a spiral wire, containing a sewing 

 needle, placed at about right angles to the magnetic meridian. 

 The magnetic effects witnessed were very inconsiderable. On 

 one occasion only, when the atmosphere was in a highly electri- 

 cal state — lightning frequent, near and vivid — was the galvano- 

 meter very distinctly affected. The deviation of the needle was 

 from 8° to 10°. I may mention a particular instance, in which 

 the absence of magnetic effect to me appeared remarkable. It 

 was on the 20th December, when I happened to be present du- 

 ring a shower of hail without thunder. For two minutes that I 

 attentively watched the galvanometer, I could not observe the 

 slightest movement of the needle. I then rapidly connected the 

 conductors with the platina wires immersed in the compound of 

 the iodide of potassium and starch, — now, in less than half a mi- 

 nute, there was a considerable precipitation of iodine on the wire 

 communicating with the inferior conductor. 



On the needle in the spiral the effect was even less than on 

 the galvanometer. Sometimes it appeared on trial to have ac- 

 quired a very feeble magnetic power, but which it soon lost on 

 continuing the experiment. It may be worthy of mention, that 

 in the instance in which the galvanometer was affected, as be- 



VOL. XIII. PART II. 3 L 



