Ice as an Electrolyte* 115 



three legs in the inside of a wooden tub F F. Resting on 

 the bottom of the copper box, but separated from it by three 

 small pieces of glass (a) {b) (c), is a copper disk, d H, 13 

 centims. in diameter, to which is fastened a copper strip J K. 

 The box is closed bj a very tightly fitting cover furnished 

 with two openings — one, L M, to allow J K to pass through 

 without touching the cover, the other for the admission of a 

 thermometer. Distilled water having been introduced into 

 the copper box to a sufficient depth to cover the disk G H, 

 the temperature could be lowered by the introduction of a 

 freezing-mixture of snow and salt inside the tub above and 

 below the copper box, great care being taken that none of the 

 mixture fell into the box through the tube L M, which was 

 necessarily left open. 



I. Preliminary experiments. — The current from one Meidinger 

 cell was passed through the ice and a reflecting-galvanometer, 

 the space between the copper plates being 2 millims. Keeping 

 the temperature nearly constant (it varying between — 18°-2 

 and — 17°"2 C), we found that in one hour the galvanometer- 

 readings increased from 46*8 to 168-8, corresponding to a di- 

 minution in resistance per cubic centimetre of the ice from 

 354 megohms to ^^ megohms. The cause of this may have 

 been that, with such thin glass separating the copper disk from 

 the box, some tilting may have occurred at freezing, so that 

 the coppers were not perfectly parallel ; or it may have been 

 due to a very little of the salt water of the freezing-mixture 

 having found its way into the copper box by passing between 

 the box and the cover. The box was therefore opened and 

 cleaned, fresh distilled water put in, and the cover cemented 

 to the box by "cap-cement" to avoid the possibility of the 

 salt water entering the box. 



II. The current from eighty-seven porous Daniell's cells 

 joined in series was passed through the ice, the temperature 

 being kept very nearly constant at — 8°-0 C. The galva- 

 nometer-reading fell regularly from 28*5 to 12*73 in twenty- 

 seven minutes, being equivalent to an apparent rise in the 

 resistance per cubic centimetre of the ice from 3767 megohms 

 to 8443 megohms. The temperature remaining constant, the 

 reading fell to 7*20 in about two hours, corresponding with a 

 specific resistance of 17,310 megohms. Disconnecting the 

 battery and joining the coppers on the two sides of the ice 

 through the galvanometer, a discharge-current, which rapidly 

 diminished, was obtained, as was to be expected. 



Further experiments showing the difficulty of maintaining 

 the copper disk approximately parallel to the bottom of the 



12 



