F. Dellmann on Atmospheric Electricity. 409 



manifested itself; it was only after 4 hours, however, that this 

 loss increased threefold. Electricity, then, clearly disperses 

 drops of water. 



The influence of change of air on the electric loss was next 

 examined, and it was always found that the loss was unaltered 

 by can-ying the sphere about, provided that it remained in air of 

 the same constitution. When carried about in a room wherein 

 the air was mixed with the moister air of the kitchen, the loss 

 was a little increased, but it would certainly have also been in- 

 creased had the sphere remained stationary. 



Positive electricity, too, appears to suffer less loss than nega- 

 tive electricity,— a fact easily explained, since the electricity of the 

 air is generally positive. 



So far the new shell-lac ring has approved itself by givino- the 

 following series of losses after having been moistened and rubbed 

 dry with clean linen : — 



0-53, 0-51, 0-50, 0-51, 052, 0-53, 055, 056, 056. 

 The whole series extended over 3 hours, and the intervals were 

 almost equal. It will only be necessary to test the ring occa- 

 sionally, and to replace it by a new one whenever it can no 

 longer be restored by treating it with alcohol or heating it before 

 a lamp j the latter expedient generally succeeds perfectly. 



Apart from the intrinsic interest of many of the results, this 

 communication on the electric loss will be sufficiently justified 

 by its principal object,--to prove that the insulating power of 

 the described apparatus is sufficient when properly experimented 

 with; for the necessity has been shown of frequently rubbin"- 

 the shell-lac with clean linen in order to free it from moisture 

 and dust. But to do this, the apparatus must have the pre- 

 cise form which it possesses ; the sphere with its carrier must 

 admit of being easily lifted out of the box in order to reach the 

 foot, — although the latter certainly little disturbed the rcu- 

 larity of the foregoing measurements, since it was always made 

 of yellow shell-lac. The lac on the transferrer, as well as on the 

 electrometer, was also of the yellow description. In the last 

 series we may assume with tolerable safety that the loss, whose 

 mean value was 0-53, was caused by the air alone ; as also in the 

 last series but one, where the mean loss was nearly the same 

 (0-555). As this loss amounts to about i per cent, with a 

 moisture of 75, and as the latter oscillates between 50 and 100 

 the approximate correction to be applied to the results of my 

 observations on atmospheric electricity, on the assumption of a 

 proportionality between moisture and electric loss, would vary 

 between -^h and iJs, and consequently remain within the limits 

 of errors of observation; on this account such correction appears 

 superfluous. 



