Memoirs presented to the Cambridge philosophical society on the occasion of the jubilee of Sir George Gabriel Stokes, bart., Hon. LL. D., Hon. SC. D., Lucasian professor.

184 MESSRS GLAZEBROOK AND LODGE, EXPERIMENTS ON THE OSCILLATORY
Hence replaced them by pieces of ebonite all cut out of the same sheet; each
piece 7 millimetres square; 52 pieces end to end, measured in vernier callipers,
occupied 10'24 inches.
So the thickness of each distance piece was
*1989 inch or -5001 centim.
With 5 of these between each plate, one at each corner and one in the middle,
the plates were once more clamped up, connexions carefully made, and an experiment
begun.
The plates stood vertically on a pair of sharp-edged ebonite wedges, at a height
of 1 inch above the floor of the frame, which was tinfoiled to make it definite. The
sides and top were at first open, so that the edges of the plates were then free; but
afterwards in order to keep the inside air dry for all the best experiments, the box
was panelled in. The distances of the wood panelling from the plates were as follows:
From edge of plates to wall of case......... 5'15 centim.
". ".. roof.........8 8.".. ". floor..........25
A simple wooden X formed the front and back at a distance from the outer flat
of the plates, 4'0 on one face and 3'2 on the other.
ESTIMATE OF CAPACITY OF CONDENSER.
The method of correcting for the edges of a thin plate is given in Maxwell,
vol. I. ~ 293. A term has to be added to the linear dimensions as if an extra strip
of a certain breadth were put on all round a uniformly charged plate.
This extra breadth, on account of the extra density at terininations of thin parallel
planes, is
-oge 2,
'n.
where "b" is the distance between the plates. But the plates are thick and square
edged, and a further correction has to be made for their thickness; Maxwell's further
7r/3
correction for thickness /, cos b, assumes the edges to be rounded and is therefore
inapplicable, but acoustic analogies suggest the addition to the dimensions of each plate
of a quarter of its thickness, to represent the effect of the edges themselves. (Cf. Rayleigh,
Sound, vol. iI. ~~ 307 and 314.)
The total correction is thus
22b + '25/B = '11 + 18 = '29 centim.
to be added all round.