346 On a Method of Measuring very small Intervals of Time. 



In employing this system I have found that there are certain 

 precautions which it is necessary to keep in view. In the first 

 place, it is desirable, in taking the readings of the excursions 

 of the needle, to so adjust the resistance-wire that the value of 

 c is not too nearly the same as that of the initial charge C. 

 When they are nearly of the same value, a very slight error of 

 observation makes a material error in the result, because the 

 difference of their logarithms may increase or decrease much 

 more rapidly than the numbers themselves. On the other 

 hand, the proportion between C and c should not be too great ; 

 in other words, the accumulator should not be too empty when 

 its reduced charge is observed. The electricity whicn flows 

 from a battery into an accumulator (assumed to be perfectly 

 insulated) is employed in two ways : (1) it suddenly charges 

 the coatings inductively ; and (2) it gradually polarizes or 

 charges the dielectric material. When the charged coatings are 

 connected through a small resistance, the coatings are first 

 discharged suddenly, and the " diselectrification " or discharge 

 of the dielectric (which commences at the same instant of time) 

 goes on gradually. We should therefore be prepared to find 

 a small quantity of electricity still in the accumulator long- 

 after the coatings have been discharged. This is more obser- 

 vable when the discharge takes place suddenly than when it 

 takes place gradually (that is, through a high resistance), 

 because in the latter case the diselectrification-current, which 

 is very small in comparison with the discharge-current, has 

 time to creep out with it unobserved. 



This, however, has an effect only when the accumulator 

 becomes very empty and the readings of c are so small as to 

 come within its disturbing influence. By selecting the dis- 

 charging-resistance of such a value that the measured remainder 

 from the accumulator is never less than one third or one fourth 

 of the initial charge, all fear of error from this cause may be 

 dismissed. 



There are numerous applications in which the determination 

 of time-intervals of very small duration may be found of prac- 

 tical value ; and I think that this system may be depended upon. 



In conclusion I must acknowledge the valuable assistance 

 which I have had, both in arranging the apparatus and in 

 making the observations, from Mr. M'Eniry. 



25 Cumberland Terrace, N.W., 

 February 1876. 



