406 F. Dellmann on Atmospheric Electricity. 



and by means of the transferrer put in communication with the 

 measuring instrument* ; sphere and transferrer were then put 

 on one side, the angle of deflection a measured, and, with the 

 exact time t of charging, noted. After some minutes a precisely 

 similar observation was made, and the angle of deflection b, with 

 the corresponding time t, noted. My Table II. (of which more 

 hereafter), where the quantities and angles stand side by side, 

 gave the quantities u and /3 which correspond to the deflections a 

 and b. The required loss is given by the expression. 

 («-/3)2 X 5 X 100 _ («-/3)1000 

 12(« + /3)Ui-0 ~U{ci + ^){t^-ty 

 where t^ — t represents minutes. Where it appeared advisable, an 

 observation with the psychrometer was combined with the elec- 

 tric measui'ements. 



The first series of measurements, where the intervals amounted 

 to 3| minutes, gave at once the following numbers correspond- 

 ing to the losses per cent, per 25 seconds : 

 2-02, 2-38, 2-34, 1-85, 1-60, 1-19, 1-04, 101, 1-00, 097. 

 Here, therefore, a pretty regular diminution in the loss of elec- 

 tricity is unmistakeable ; some of the following series showed 

 less regularity. In the loss here under consideration four pieces of 

 shell-lac are concerned, — two on the carrier of the sphere, one on 

 the transferrer, and one on the measuring instrument ; of these, 

 however, the ring around the carrier of the sphere being most 

 important was principally kept in view ; the effect of the shell- 

 lac on the electrometer was known and very small. In the fol- 

 lowing measurements, however, the influence of the two remain- 

 ing pieces of shell-lac was not entirely neglected ; that of the 

 shell-lac on the transferrer was specially tested several times, 

 by charging the instrument from the transferrer alone before 

 charging it from the sphere. The ring on the carrier was mo- 

 dified in several ways — by rubbing and drying it, by diminishing 

 its mass, and by changing its material ; but none of these expe- 

 dients produced a proper regularity in the phsenomena, though 

 drying the shell-lac ring was the most eff'ective in bringing about 

 regularity. After exposing the ring for more than an hour to 

 the current of warm air rising from a stove, the per-centages of 

 loss for 25 seconds were, — 



1-00 (10), 071 (13), 0-68 (15), 0-68 (15), 0-60 (18) 



0-68 (17), 0-67 (16), 0-68 (17), 0-74 (17), 070 (17), 



where the numbers in brackets denote the intervals in minutes. 



This notation will be retained ; and it is well also to remark that 



the measurements of each series were always made with one and 



* See Phil. Mag. Ser. 4. vol. xv. p. 460. 



