351 Specific Inductive Capacity of Electrolytes. 



small second term, we have for air, 



B = 0*52 dyne. 



It does not appear as yet what is the strength of the field, 

 or the difference of potential between the electrodes. Our 

 formula for the potential due to two parallel wires is 



V = logR'-logR"=log|i; 



this shows that along the axis of Y where B/ = R", V = 0. At 

 a (figure 3) R' = 037 cm. and R" = 2-07 cm., whence V = 

 + 1*72 electrostatic units, or +516 volts. The force of 0*52 

 dyne calculated above is therefore for a difference of potential 

 of J 032 volts between the electrodes. For 100 volts, since 

 the force is inversely as the square of the difference of poten- 

 tial, it would be ( n \ =106*5 times less ; or 



B = 0-0049 dyne for lOO volts. 



The torsion of the suspending wire was determined and 

 found to require 0*0040 dyne for each centimetre of deflexion 

 on the scale. As 100 volts gave 1*50 cm. with carbon cylin- 

 ders in air (for this position of the scale), this would corre- 

 spond to 0*0060 dyne. Thus the measured force is nearly of 

 the same magnitude as the calculated, but about 20 per cent, 

 greater. This is probably due in large part to the effect of 

 the ends of the cylinders, which tend to increase the force, 

 and of which no account was taken in the calculation. More- 

 over, the electrodes being short, the field would not be quite 

 uniform along the axis of Z as was assumed ; this would be 

 strictly true only for infinitely long electrodes. 



The agreement is, however, sufficient to serve as a verifica- 

 tion of the formula which, although only an approximation, 

 is yet of a good deal of interest. For the purpose of deter- 

 mining the specific inductive capacities of liquids, we have 

 only to compare the forces on the same cylinders in the 

 several liquids, and the error due to the ends of the cylinders 

 has no effect. 



The experiments were carried out last year at the Physical 

 Laboratory of the Johns Hopkins University, and this paper, 

 together with the one already referred to, constituted a Thesis 

 offered for the degree of Doctor of Philosophy. I have to 

 thank Prof. Rowland for the solution given in the fourth part 

 of the present paper. 



Wesleyau University, 

 Middletown, Conn., July 1802. 



