1876.] cases of Electromotive Force. 185 



The author supposed these effects to be due to electrolytic action 

 through the glass, not suspecting the true cause, excepting in V. But 

 Sir William Thomson pointed out to him that the rate of development 

 of the charge was greater than could occur through a substance of the 

 low conductivity of the most conducting glass, and that the circuit must 

 have been completed by conduction through a film of moisture on the 

 surface of the glass. The next two experiments prove this to be the 

 case. 



VI. Drops of strong sulphuric acid and of caustic potash were placed 

 on a sheet of common window-glass, previously carefully cleaned, but ex- 

 posed to the air of the room. Platinum electrodes, dipping into each drop, 

 communicated with the electrometer. 



1. Drops half an inch apart, E. E. == 47. 



2. The drops were connected by a thin trail of alkali drawn from the 

 drop of alkali ; E. E. = 105. 



3. Drops 5 inches apart, two minutes being allowed for the charge 

 to develop ; E. E. = 12. 



4. A trail was drawn halfway from one drop to the other; E. E.=31. 



5. The trail of liquid was continued till but |- inch of clear glass 

 separated the liquids ; E. E. = 43. 



6. The connecting trail was completed from one drop to the other ; E. 

 E. = 70. It was observed that the potash trail had dried up, leaving a line 

 of alkali between the drops. 



VII. A dry chip of deal 6 inches long was split at each end, and a platinum 

 wire let into each slit ; the two wires were moistened with sulphuric 

 acid and potash respectively at the points of contact with the wood ; 

 E. E. = 43. 



VIII. Clean platinum wires were let into slits in a second dry chip of deal 

 12 inches long ; these were connected for twenty minutes with the poles of a 

 battery of twoDaniell's elements, and then detached and connected through 

 the reversing-key with the electrometer. As was expected an electromo- 

 tive force opposite to that of the battery was observed, at first amounting 

 to 33 divisions. 



These experiments show that imperfect insulation, such as glass exposed 

 to the air or wood, may cause errors in electrical experiments, not merely 

 by leakage, but by introducing unknown electromotive forces, arising either 

 from the imperfect insulators connecting different liquids, or from elec- 

 trolytic polarization after a current has for some time been creeping through 

 or over the surface of the insulators. 



Several experiments were then tried on the direct action of liquids on 

 liquids ; two only are given here, because determinations have been made 

 by other methods by Becquerel and others. 



IX. In a previous experiment a plug of moist sand had been rammed into 

 the bend of a U-tube, and strong sulphuric acid and caustic potash poured 

 into the limbs. When this tube was washed out, it was found that a plug 



