Dielectric Constants by means of Hertz Oscillations. 179 



denser CC, this being the same condenser as I used for the 

 previous experiments. The electrometer-deflexions were read 

 by the aid of a telescope. 



This method is surprisingly sensitive ; the adjustments are 

 very easily made, and moreover transient disturbances are 

 easily recognized by the suddenness with which the effect 

 appears ; the raising of a hand, or a person passing, causes a 

 momentary deflexion of the electrometer-needle *. 



After the adjustment bad been made with the plate xx ] be- 

 the condenser-disks, xri was removed and then, without any 

 alteration whatever in EE', the plates CC were made to 

 approach each other, until the electrometer-needle returned 

 to zero. 



In the following table the measurements have all been 

 collected together. The numbers give the distance apart of 

 the condenser-plates in eentim., when the capacity has been 

 kept the same, as described above. 





2 ©W 



Glass 

 (0-8797 

 eentim.). 



Petroleum 



+ two glass 



plates (1-9272 



+0-4338 



eentim.). 



Water + two 



glass plates 



(1-9342 



+0-4338 



eentim.). 



Two glass 



plates (0-4338 



eentim.). 



With the dielectric . 

 With air 



2-1610 

 1-5785 



1-6090 

 1-0249 



2-6278 

 1-7240 



20325 

 1-1285 



3-6590 

 1-8120 



3-6560 

 1-8240 



4-2100 

 1-3410 



4-0035 

 1-1200 



1-8834 

 1-4401 



1-9240 

 1-4859 





Difference 



0-5825 



05841 



! 



0-9038 



0-9040 



1-8470 



1-8320 



2-8690 



2 8835 



0-4433 



0-4381 



We thus obtain for the dielectric constant of 



Plate Glass. " Solinglas." Ebonite. Petroleum. Water. 



5-34 5-09 2-81 2'35 ao 



These numbers are, apart from errors of the method, more 

 exact than those of the previous section ; the separate 

 measurements by the next method show the least agreement. 



* It should be noticed in passing, that if capacities are connected to C 

 or E' the alteration in the adj ustment is considerable ; if, on the contrary, 

 a coil of wire be inserted in the conductor leading from C or E' to the 

 electrometer-quadrants the effect is smaller. The self-induction of the in- 

 duction-coil is itself very large, the capacities on the contrary comparatively 

 small. Of the disturbance introduced in consequence of damping the 

 oscillations by causing the condenser-plates to approach each other, I shall 

 speak later. 



