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.

152 MESSRS GLAZEBROOK AND LODGE, EXPERIMENTS ON THE OSCILLATORY
It appears that the greatest difference between two results is '0014 in a total of
4-6500, or less than 1 part in 3000.
It will be noticed also that the agreement is very decidedly improved by the
temperature corrections of the last column. Thus the value of the coefficient of selfinduction has been determined to an accuracy which requires that the temperature of
the various coils used should be known to a fraction of a degree.
The value given, 4-6493, is in legal quadrants; i.e., the resistance of a column of
mercury 106 cm. long has been taken as 109 C.G.S. units. To reduce it to "Henry's"
or "International Quadrants" it must be multiplied by the ratio 106/106-3. We then
find as the value of the coefficient of self-induction of the coil
4-6362 Quadrants,
or 4-6362 x 10lO0u centimetres.
SELF-INDUCTION OF EACH HALF OF THE COIL.
Since the coil was wound in two parts and one of the parts occasionally used
alone, it was thought well to find the coefficients for the two parts separately, and to
check the result by observing also the value when they were arranged so that the
mutual induction of the two opposed the self-induction. Let L1, L2 be the two coefficients of self-induction of the two parts, M the coefficient of mutual induction between
the parts, L' the coefficient of self-induction of the whole with the two parts opposed.
Then L = L1 + L2 + 2M,
L' = L + L2 - 2M
= 2(L + L2)-L.
Thus L = 2 (L, + L) - L'.
The coefficients are all small and the probable errors of the measures are greater
than those in the direct measurement.
The following values, however, were obtained:
L = 1-405 Quadrants for semi-coil marked A.
L, = 1393.. " " B.
L' = 0-963
Whence L = 4'633 legal quadrants; agreeing fairly well with the true result.