F. Dellmanu on Atmospheric Electricity. 413 



From these one might imagine that somewhat too high a 

 value had been given in the Tables to the larger angles ; but the 

 day before the following numbers had been obtained: — 

 2-92(64), 2-91(80), 2-98(96), 2-96 (112), 3-00 (128), 

 2-97(144), 3-01(160), 295(176), 2-95(192), 3-01(208), 

 2-99 (224), 3-03 (240), 3-09 (256), 3-06 (240), 3-04 (224), 

 3-01(208), 3-00(192), 3-01(176), 3-06(160), 3-04(144), 

 3-05 (128), 3-01 (112), 3-06 (96), 2-91 (80), 2-99 (-64), 

 3-12(48), 3 12(32), 3-00(16). 



The angle of deflection for 256 elements was 55°, that for 16 

 elements 3°. 



The Tables by means of which our observations here, on atmo- 

 spheric electricity, are calculated range with tolerable certainty 

 from 2° to 88°. In fact the torsions from 5° to 85° could be 

 measured with a good agreement amongst the results ; still for 

 these extremes it appeared advisable to adopt a particular ex- 

 pedient. Two electrometers which stood close to one another 

 were charged with the same quantity of electricity, one immedi- 

 ately after the other, that is to say, ndth the same pole-wire of 

 the battery. The ratio of sensitiveness of the two instruments 

 was thus ascertained with tolerable exactitude by a series of 

 measurements whose angles fell within the limits of the Tables 

 already calculated, the mean of many calculated values for a 

 single element being taken. Greater and less numbers of ele- 

 ments were now taken, which with the more sensitive instrument 

 gave deflections up to 88°, and with the other under 10°. The 

 tabular value of the one was then multiplied or divided by the 

 quotient of sensitiveness, when the product or quotient ought to 

 give the tabular value for the measured angle of the other. 

 Lastly, the mean of many measurements was adopted. This 

 method would not have been practicable had the instrument 

 under the vessel not had almost free space. 



For the instruction of those who may wish to form Tables for 

 their own instruments, it may be well to notice that the former 

 assertion*, that the arm of the balance must be straight through- 

 out, is no longer applicable when the angle read off" is always on 

 the same side of the arm as that which was taken when mea- 

 suring the torsions on which the Tables were based. The great- 

 est difficulty attending the construction of the apparatus thus 

 disappears. Care must of course also be taken that the direc- 

 tion of the arm does not change very much ; for which purpose 

 it is better to have a brass arm, since the brass wire does not 

 enter the lamp, and therefore remains harder. All the above 



* PoggendorfF's Annalen, vol. Ixxxvi. ]». 530. 



