148 J. H. Pratt — Capillary Electrometer. 



through the tube to cleanse the surface ; and the bubbles of 

 gas, if any, were driven from the tube. The arrangement of 

 the instrument was the same as before, and the readings were 

 taken with great care. 



Table III.— April 1, 1881. 



E. M. F. 



Deflect. 



E. M. F. 



Deflect. 



E.M. F. 



Deflect. 



1-20 Dan. 







1-32 Dan. 



•25 mm 



1 '44 Dan. 



1-30 mm 



1-21 







1-34 



•35 



1-50 



2-30 



1-22 



•05 mra 



1-36 



•45 



1-60 



6-70 



1-24 



•10 



1-38 



■55 



1-T0 



12-90 



1-26 



•15 



1-40 



•60 



1-80 



21-20 



1-30 



•20 



1-42 



1-20 







Since in a Daniell's cell the replacement of one gram-equiva- 

 lent of zinc evolves 24,200 cal. of heat ; and, when one gram of 

 hydrogen and 8 grams of oxygen unite to form, one gram- 

 equivalent of water, 32,462 cal. of heat are evolved,* therefore a 

 cell of potential fUff Dan. (equals 1*34 Dan.) would be just 

 sufficient to decompose water. For about this potential (i - 34 

 Dan.) it might be supposed that conduction through the elec- 

 trometer would take place. Molecular changes which accom- 

 pany electrolytic action may, very probably, slightly precede it 

 and cause variations in the electrical conditions which allow a 

 current to pass. If, moreover, there were any leakage in the 

 instrument, this would be evidenced by a current through the 

 galvanometer, when the potential was below 1*34 Dan. De- 

 fective insulation would cause leakage, and conduction would 

 begin at a somewhat lower potential. It will be seen by re- 

 ference to the table that conduction begins not far below the 

 electrolytic limit, i.e., at 1*22 Daniell ; and that, at first slowly 

 and then rapidly, it increases with the potential. 



When the mercury was treated with oxygen polarization, 

 conduction began at very low potentials. An E. M. F. of 0*01 

 D. gives a permanent deflection of 6 mm ; and other higher 

 potentials gave higher deflections; which, however, seem to 

 follow no uniform law. This, like the former experiments 

 with oxygen polarization, shows that such polarization cannot 

 be too carefully avoided. 



III. To determine the capacity of the electrometer. — Three 

 methods of making this determination were tried. The nature 

 of the instrument makes it impossible to secure rigorously 

 accurate results. However, several not inconsistent values of 

 the instrument's capacity were secured, and the final determina- 

 tions should be regarded as fair approximations. By the first 

 method a condenser of one microfarad capacity was charged 

 with electricity at a known potential (V), and the charge was 

 * Daniell's Physics, p. 609. London, 1885. 



