226 [Dec. 12, 



particularly in the cyanide of potassium solution, by not employing 

 too great an excess of hydrocyanic acid, and^electrolysing the mixture 

 with mercury electrodes as soon as it is made. From the various 

 results obtained, it is evident that the sounds only occur under very 

 special, limited, and probably complex conditions. 



19. The presence of certain impurities in the electrolyte did not 

 prevent the vibrations or sounds ; for instance, chloride of ammo- 

 nium, nitrate of ammonia, or bromide of potassium, added in mode- 

 rate quantities to the cyanide of potassium solution, produced no con- 

 spicuous effect. 



20. After the current has been passed a long time in one direction 

 and is then reversed, the crispations and sounds do not occur at the 

 cathode for a little time, partly because a layer of liquid containing 

 an excess of cyanide of mercury has been formed at the surface of 

 that electrode, and a layer of liquid with a deficiency of cyanide of 

 mercury has been formed at the other electrode by the previous 

 electrolytic action ; diffusion of these layers by stirring the liquid 

 hastens the reproduction of the crispations and sounds. 



21. By repeating the phonetic experiment many times with 

 unguarded wires (4), the cyanide solution gradually loses its power 

 of producing the sounds, probably from acquiring an excess of 

 cyanide of mercury by electrolysis, hydrogen being evolved and 

 mercury replacing it : the loss of power does not result from an 

 alteration in the electrodes, because with the same mercury and a new 

 solution the power was restored, whilst with fresh mercury and the 

 same liquid the power was not restored. The mercury remains 

 unaltered, except absorbing a trace of alkali metal, which is readily 

 removed by washing, and it is therefore not necessary to take fresh 

 mercury for fresh experiments. 



Influence of Mechanical circumstances and Temperature. 



22. In some solutions (or with a feeble battery), where the crispa- 

 tions did not spontaneously occur, a slight blow with the end of a 

 glass rod against the negative ring induced them to commence, and 

 also a similar blow sometimes stopped them. Frequently, also, when 

 the vibrations have stopped, stirring the liquid reproduces them by 

 mixing the layers of altered liquid that accumulate about the elec- 

 trodes. 



