plotted against tan 6.

These latter are straight lines as ex

e

plained above, and give at once the values of a and for

E

each frequency. It is also interesting, having obtained the

values of a and to split the "bismuth E.M.F."

e

E

,

up

into

two components, one 90° and the other 180° behind the current, for the former component may be considered as due to something in the nature of self-induction, and the latter to a real rise in the resistance of the bismuth.

e

e

Table I. shows the values of a, E, E cos a (the resistance

e

E'

component), and sin a (the self-induction component) for

E

Phil. Mag. S. 6. Vol. 2. No. 9. Sept. 1901.

40

The results of these experiments, although they do not indicate what really takes place in the bismuth, show that, whatever it is, it may be represented by an alternating E.M.F.

set up in the bismuth, the magnitude and phase of which for frequencies between 10 and 60 per sec. is very nearly given by the equations

60

On studying Curve I. (fig. 2) it can at once be seen that these conclusions explain the results obtained by M. Sadovsky. For, on closing the galvanometer-arm of a bridge, which has been balanced for steady currents during the 60° of rising current, a current represented by the area ADBC would pass through the galvanometer; this would appear to be due to a rise in the resistance of the bismuth. On closing the galvanometer-arin for the next 60° a current, represented by the area CDE, would pass through the galvanometer; this current would be less in value, but of the same sign as the previous one: hence, again, the resistance of the bismuth would appear to have been raised. This, then, accounts for the apparent fact that

R,>R>R..

But when the galvanometer-arm is closed during the next 60°, the current through the galvanometer is in the opposite direction, for it is represented by the area EFG; thus Rwould appear less than R..

These experiments were made in the New Physical Laboratory of the Owens College, Manchester: where, in order to complete the research, I hope to be able to investigate the relations between the magnitude and phase of the "bismuth E.M.F." and the field-strength; also to investigate the whole. effect at the temperature of liquid air.

XXVIII. Note on the Spark-discharge.

By SIEGFR. GUGGENHEIMER, Ph.D.*

THE HE interest in the phenomena accompanying the sparkdischarge in gases has been revived recently by a discussion between Mr. Swyngedauwt and Prof. Warburgt. The complete discordance between the opinions of these authors led me to undertake the experiments described below with the view eventually to decide the question at issue. may at once say that the result of my experiments and theoretical considerations is to confirm the views of Prof. Warburg.

I

1. The fact that the spark-potential is independent of the nature of the radiation employed to shorten the time of retardation (Warburg's Verzögerung) made it seem probable that this potential depends upon the momentary state, i. e. upon the degree of ionization of the gas. Therefore it was to be expected that the "Verzögerung" would also be destroyed, or at least shortened, if, instead of using direct radiation, we introduce a sufficient number of ions in the space containing the sparking system. The experiments confirmed this expectation.

2. In the sides of a brass tube 20 cm. long and 3 cm. interior diameter were fitted two ebonite plugs facing one another. Through each of these plugs passed a brass wire terminated inside the tube by a brass ball of 7 mm. diameter.

*Communicated by Prof. J. J. Thomson, F.R.S.

R. Swyngedauw, Journ. de Phys. ix. p. 488 (1900); Bichat & Swyngedauw, Rapports of the Paris Congress, iii. p. 164 (1900).

E. Warburg, Verhandlungen der Deutsch. Phys. Gesell. ii. p. 212

(1900).

Opposite these electrodes was an aluminium window over which a brass cover could be placed. One end of the tube was connected to a U-tube filled with calcium chloride and thence to a wash-bottle containing concentrated sulphuric acid. The other end of the tube was connected by means of a glass tube (from 8 mm. to 2 cm. diameter and 10 cm. length) to a second brass tube of 20 cm. length, and of the same diameter

as the first one. This second tube was connected at the other end by means of a U-tube containing calcium chloride to a blowpipe. In the side of this brass tube was an opening 8 cm. long and 2 cm. broad, which was closed by thin aluminium-foil.

One part of the electric circuit comprised the one pole of a Wimshurst machine, the inner coating of a leyden-jar, one of the spherical electrodes, and the indicating portion of a Braun electrometer; the other part of the circuit consisted of the other pole of the Wimshurst, the outer coating of the leyden-jar, the other sphere, and the cage of the electrometer, and was generally put to earth.

The observations were carried out in the following order:(1) The discharge-potential was measured without exposing the sparking system to radiation, and without introducing ions into the tube A. Then (2) from an X-ray bulb placed near the aluminium window of the tube B, strong X-rays entered B, and the ions thus produced were driven by a strong current of air into the tube A, and the potential was measured whilst this operation went on. (3) A third measurement was then made whilst X-rays fell directly through the aluminium window in A upon the electrodes. The current of air was also blown through the apparatus during the operations 1 and 3. Sheets of lead protected A when В was exposed to radiation.

I give below the results of two particularly striking series of experiments.