54:8 Mr. F. B. Young on the Field at 



Tests were applied for the purpose of ascertaining whether 

 the capacity o£ the connecting wire was affected by the dis- 

 placement of the sphere. With a thin plate of paraffin-wax 

 and tinfoil a condenser was made whose capacity was ap- 

 proximately equal to that of the proof -plane system, so that, 

 when added to that system, it reduced the deflexions to about 

 a half of their original magnitude. The values of a were 

 then found as before for two or three values of 6. If the 

 capacity of the proof-plane system is constant the values so 

 obtained should agree with those previously found ; if the 

 values differ slightly a = a 1 (l + 2cr 2 — a^ approximately, 

 where a 1 and <7 2 are the apparent values of a without and 

 with the condenser respectively. It was found that the 

 results agreed well within the limits of experimental error 

 except for the position of the proof-plane most remote from 

 the centre, where a change of capacity of nearly 2 per cent, 

 seemed to be indicated. In this position the wire and its 

 covering were in a somewhat strained position, so that a slight 

 change of capacity was possible. The corrected value for the 

 value of 6 in question does not, however, fit so well into the 

 density curve as the uncorrected value, and the difference 

 may well be accounted for by the diminution of the deflexions 

 which rendered the values of cr 2 liable to errors twice as great 

 as those of o-j. The test, indeed, cannot be regarded as giving 

 more than a general indication that no progressive rise or fall 

 of capacity occurs as 6 is increased ; the smoothness of the 

 curve and the consistency of the four series may be taken as 

 evidence that no irregular fluctuations occurred which were 

 not of negligible magnitude. The question of constancy of 

 capacity was important as it probably constituted the weakest 

 link in the chain of experimental proof. 



In order to integrate I o- 2 sin 26 ,d6 a second curve was 



J 



developed from curve A by plotting the value of a 2 sin 29 

 with 6. The area of this curve, which gave the required 

 integral, was found to be 0'85. The ordinates of curve A / 

 are proportional to the contributions of the annular elements 

 of the hemispherical surface to the pull P for the correspond- 

 ing values of 0, so that the curve serves to show the relative 

 importance of these contributions. The part of the surface 

 which contributes most to the pull is that for which 6 



approximates to j. Hence the assumption of uniform 



density will give an approximate result only in so far as the 



