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.

172 MESSRS GLAZEBROOK AND LODGE, EXPERIMENTS ON THE OSCILLATORY
Now the resistance of our wire circuit is only 200 ohms as ordinarily measured,
but it is well known that under rapid oscillations the resistance of a conductor is
increased by reason of the extra peripheral distribution of the current. The spark gap
has also a certain resistance which it is not easy exactly to estimate.
Some observations were made with a condenser discharging through several known
circuits and the same air gap, in order to study the damping and make an
estimate of what the resistance of the spark was. These indicate that for feeble
discharge a spark resistance is high, while for powerfil discharge it may be quite
low. With our feeble spark it is undoubtedly large, and quite eclipses the resistance
of the wire part of the circuit, though it does not amount to anything like the critical
resistance at which the discharge ceases to be oscillatory; but it cannot be considered
as constant, and its complete specification will be difficult.
With regard to the throttling by reason of rapid alternation, it must be observed,
lst, that the alternations were not excessively rapid, always comparable to 1000 per second;
and 2nd, that the wire on the coil was copper and very thin.
The coil had a mean diameter of 38 centims. and consisted of 3493 turns of copper
wire, half a millimetre in diameter. At 1000 alternations per second uniform distribution
of current through such a wire would hardly be departed from, and neither the resistance or the self-induction would be greatly different from their ordinary values.
It is important to note that no correction to self-induction is necessary, for even
with infinite rapidity of oscillation, when all the current flows by the periphery, the
value of the self-induction would not be greatly disturbed; though the throttling resistance
would then be enormous. The reason why the self-induction is not very dependent on
distribution in a thin wire is that it is only the space inside the wire which ceases
to be magnetised by a peripheral distribution, and this is small in comparison with all
the space outside.
(iii) SEI,F-INDUCTION OF LEADING WIRES.
The self-induction of the leading wires between condenser, coil and spark-gap, was
about 100 metres; but as the self-induction of the whole circuit was considerably more
than an earth quadrant this is entirely insignificant.
(iv) EFFECT OF LEAKAGE.
The insulation resistance between the two halves of the coil was measured and
found to be 20 megohms; hence leakage during a discharge was practically non-existent.
(v) EFFECT OF WAVE LENGTH.
The electric oscillations have not been assumed quick enough to give waves comparable in length with the circuit, else different parts of the circuit would be in different
phases, and some complications would result.