410 F. Dellmann on Atmospheric Electricity. 



b. The Measuring Instrument. 



The measui'ing instrunieut is the one before described, the 

 construction of which was proved to be excellent in these as in 

 all other measurements*. Nevertheless a slight modification 

 was deemed advisable ; the small strip is now quite rectilinear, 

 and no longer possesses a slit. The arm of the balance is also 

 rectilinear, and therefore hangs over the strip. 



The calculation of new Tables appeared necessary for the fol- 

 lowing reasons : first, there is some doubt attending the bending 

 of the strip, for the amount of curvature is not exactly ascertain- 

 able, and the same calculating Tables cannot be applied to differ- 

 ent curvatures ; secondly, the observations on the battery, calcu- 

 lated according to Kohlrausch's Table II., did not agree well with 

 each other, and the discrepancy before alluded tof must arise 

 from the fact that my strip has not the same curvature as that 

 of Kohlrausch's ; lastly, the results to be communicated to the 

 Royal Meteorological Institution must necessarily have as trust- 

 worthy a foundation as possible. 



In order to calculate the new Tables, three methods presented 

 themselves, — that of Kohlrausch by torsion-measurements, that 

 of Quetelet by division of the same quantity of electricity over 

 well-constructed spheres, and that, not hitherto employed, of 

 measuring the tension of the pile. The torsion-measurements 

 had before furnished me with imperfect results, for a reason 

 to be immediately mentioned. I had three spheres exceed- 

 ingly well constructed by Reimann of Berlin under my own 

 superintendence; two of them had equal radii double that of 

 the third, but on account of the loss of electricity I could not 

 employ them. The third method was hopefully employed, but 

 did not lead to satisfactory results. The tension of the pile is 

 not sufficiently constant in winter, and I neither could nor would 

 wait until summer. In the meanwhile repeated and cai'eful 

 measurements of the tension of the pile led to the detection of 

 the defect which the electrometer then possessed. It was fre- 

 quently noticed that, with small angles of deflection, the arm of 

 the balance suspended in the slit is far too apt to come into 

 contact with the edges of the latter. Electricity then passes 

 from the arm to the strip, and the angle is smaller than that 

 corresponding to the charge. Now since this defect (which may, 

 of course, be in a great measure avoided by slowly displacing 

 the upper index) might often occm* in measurements which must 

 be frequently, and therefore quickly made, the arrangement was 

 necessarily faulty. Moreover I hoped with a perfectly straight 

 strip to be able to prepare Tables of general applicability ; but, as 



* Poggcnfloiff's/4MwaZeM,vol.lxxxvi.p.527; alsoPlnl.Mag.vol. xv.p.4()0. 

 t Poggemloiff's Annalen, vol. Ixxxvi. p. 539. 



