576 
MESSES. GIBSON AND BAKCLAY ON MEASUBEMENTS OF 
The mean of these values, and the previously measured volume of the containing 
spherical surface, give 4‘5107 centimetres as the radius of the inner sphere. 
The capacity of a spherical condenser is calculated by the formula 
Capacity = -^-, 
where r' is the radius of the outer, and r the radius of the inner sphere. 
In this case r'=4‘857 centimetres and r=4 - 5107 centimetres. Hence the capacity 
is equal to 63‘264 centimetres. The specific inductive capacity of the vulcanite of the 
pins which support the inner sphere being greater than that of air, causes an increase 
of the whole capacity of about T9 centimetre. The hole in the top causes a diminution 
of -085 centimetre, and the electrode passing through it an increase of T5 centimetre. 
The actual value of the condenser is therefore 63‘519 centimetres. 
To reduce this to scale-divisions of the sliding condenser, the value of one division in 
absolute measure is calculated by the formula for a cylindrical condenser, 
Capacity =g- — — p 
log — 
r 
where r' is the radius of the outer surface, r the radius of the inner surface, and l the 
length of the condenser whose capacity is to be calculated. 
In this case ^ = 2-4837 centimetres, r — 1T515 centimetre, and l =-£ 0 of an inch or 
•063499 centimetre. Here r' was determined from a measurement of the volume of 
water contained by the tube, the length of which was accurately measured. To deter- 
mine r, the circumference of the core was measured by winding fine wire round it, and 
measuring the length of a certain number of turns, the necessary corrections being made 
for the thickness of the wire and its spiral arrangement. The value for electrostatic 
capacity of one scale-division is therefore - 0413 centimetre; and hence the spherical 
condenser is equal in capacity to 1538 scale-divisions. 
[Direct electrical measurements taken subsequently on a sliding condenser of greater 
range gave 1607 scale-divisions as the value of the spherical condenser. This is pro- 
bably greater than the actual value, while that derived from calculation is too small by 
a quantity due to the action of parts whose capacity could not be numerically deter- 
mined. The mean of these values, 1572 scale-divisions, may therefore be taken as the 
value of this condenser.] 
When the capacity of the spherical condenser was measured electrically while in 
connexion with the side p of the platymeter, the reading obtained on the scale of the 
sliding condenser was 211. When the connexions were reversed, that is, when the 
spherical condenser was connected with the side p' of the platymeter, the reading 
obtained was 183. The difference of the two readings thus obtained shows some 
inequality of the sides of the platymeter. 
Now to find the true reading in any case from two such readings, suppose the sidep 
to be equal to n Xp 1 - If no deflection of the needle takes place on the distribution of 
