very small Intervals of Time. 



341 



the discharging resistance r while charging the accumulator ; 

 and therefore a very much lower value of r may be employed 

 than in using the preceding method, without fear of error. By 

 way of illustration, B is a jrj„ 2 



block of metal or anvil 

 against which the contact, a, 

 is pressed in order to charge 

 the accumulator, A, by the 

 battery E. The light ham 

 mer, b (weighing about 1 

 ounce), is then suddenly 

 struck against B and allowed 

 immediately to rebound off 

 again. During the interval 

 which elapses between the 

 contact by b and its inter- 

 ruption, or what may be 

 termed the electrical dura- 

 tion of the blow, the accumulator leaks through r ; and the 

 remainder is measured on the galvanometer (G) by depressing 

 the key (K) as before. The same accumulator was used in 

 these experiments as in the previous ones. 



Excursion of light -point. 



r. 



Duration of flow, 

 calculated 



(Mog s 2), 

 t. 



Initial charge, 

 C. 



Bemainder, 

 c. 



308 

 308 

 308 

 308 

 307 

 306 

 306 

 307 



265 

 264 

 269 

 265 

 261 

 260 

 262 

 267 



ohms. 

 1000 

 1000 

 1000 

 1000 

 1000 

 1000 

 1000 

 1000 



second. 

 0-000050 

 000051 

 0000045 

 0000050 

 0000054 

 0000054 

 0-000052 

 0000047 







Mean . . . 



0000050 



It now remains to be shown : — 



1. That the results obtained by this method agree amongst 

 themselves, and 



2. That the intervals of time calculated by the formula agree 

 with the observed time. 



The observations made with the hammer and anvil (fig. 2) 

 were sufficiently uniform to suggest that, by discharging the 

 accumulator gradually by a series of blows of as nearly equal 

 force as possible, the time of each blow might be regarded as 



