the effect of moving electrified bodies still unanswered, at least satisfactorily; for example, the motion of a dielectric in a uniform field; also the reverse of the above experiment. These problems are also to be attacked.

It is a source of great satisfaction to me that the above investigation was practically completed before the death of Prof. Rowland. His advice and criticism throughout the work were invaluable.

To Prof. Ames also I am indebted for much valuable advice. I also take this opportunity of thanking Mr. N. E. Gilbert for his valuable assistance in the early part of the work, and Messrs. J. T. Barrett and Jas. Barnes for their help in taking the final readings.

Physical Laboratory, Johns Hopkins University,

June 10, 1901.

XV. The Thermal Properties of Isopentane compared with those of Normal Pentane. By J. ROSE-INNES, M.A., B.Sc., and SYDNEY YOUNG, D.Sc., F.R.S.*

IN

[Plate IV.]

Na paper read before the Physical Society in December 1898, attention was drawn to the great interest attaching to the comparison of the isothermals of two isomeric substances (Phil. Mag. xlvii. p. 366). An experimental investigation of the relations between the temperatures, pressures, and volumes of isopentane had been carried out by one of us some time previously; and in the paper above referred to the results of a similar investigation into the thermal properties of normal pentane were given. A preliminary comparison was instituted between the thermal data belonging to the two substances respectively, and the conclusions provisionally arrived at were stated (loc. cit. pp. 366-367). We have since undertaken a more exhaustive examination of the experimental results, and the conclusions now reached form the subject of our present paper.

pv

If we could treat normal pentane and isopentane as perfect gases, we should be able to write pv=RT, where all the letters have their usual signification. În practice it is found that po changes as p diminishes along any isothermal, in such a way that the limiting value of pv for p=0 is equal to RT; but any actual value of pv lies below RT within the limits of temperature of the experiments. The quantity RT-pe, which

*Communicated by the Physical Society: read May 10, 1901.

should be zero if Boyle's law were correct, may be considered as measuring the error of Boyle's law, and we may term it "the departure from Boyle's law" for the particular volume. and temperature in question. It then appears that the departure from Boyle's law in the case of isopentane bears a constant ratio to the departure for normal pentane at the same volume and temperature*. This simple law is found to yield numerical results so nearly in accordance with the experimental facts that the error incurred by assuming the law as entirely correct is comparable with the errors of experiment.

We will next proceed to describe our method of testing the law. Take any volume and temperature, v and T say; and let Pa and Pi denote the pressures of normal pentane and of isopentane respectively at the said volume and temperature. Then the departure from Boyle's law for normal pentane is RT-P., while the departure from Boyle's law for isopentane ᎡᎢ is RT-pr. The proposed law tells us that

By examining the experimental data in the neighbourhood of the critical point of normal pentane, it was concluded that A might conveniently be put equal to '9463. By the last formula it was then possible to calculate pv for any volume and temperature whenever pay for the same volume and temperature was known. In this way a large number of values of pv were found for the isothermals 280° C., 240° C., 200° C., 160° C., 120° C. These were plotted in a diagram against v, and the dots so drawn distinguished by means of a cross. The values of pe found by direct experiment for the same temperatures were also plotted against v, and the resulting dots distinguished by means of a small circle. As both sets of dots were very numerous, and occurred at small intervals of -, we were well able to judge by the eye whether the curve determined by the one set might be considered as

The connexion between chemical problems and the departure from Boyle's law has also been investigated by M. Daniel Berthelot. (See Comptes Rendus, 1898, cxxvi. pp. 954, 1030, & 1415.) His immediate object, however, was different from ours. He confined himself to studying how the departures from Boyle's law might be allowed for in deducing the true molecular weights of substances by physical methods; and he nowhere pays special attention to isomeric bodies.

Phil. Mag. S. 6. Vol. 2. No. 8. Aug. 1901.

P

coinciding with the curve determined by the second set. The diagram is reproduced on Pl. IV. Only the volumes above 3-4 were taken into consideration, for reasons stated in our former paper (loc. cit. p. 364).

It will be seen that on the whole the two sets of dots agree very well. In places, however, there is a discrepancy between the two sets: e. g., at small volumes on the isothermal of 240° C. the calculated values of pv differ from the experimental values by slightly over 1 per cent. This error is partly to be accounted for by our not having chosen the most suitable value of λ; by choosing a rather smaller value we could easily diminish the error spoken of. But we cannot diminish the error indefinitely in this way, since we are hampered by the condition that we must not introduce more serious error into fresh places by our changes in λ. Hence there will necessarily be some small outstanding differences between the two sets of dots whatever value of λ be chosen. We have not thought it worth while to draw a fresh diagram with a more suitable value of A, as a slight correction in the diagram corresponding to a slight change in λ is easily allowed for by

the eye.

As to whether the outstanding differences referred to above indicate the necessity for a correction in the law we have proposed, or whether they can be considered as due to errors of experiment, we are unable to decide with any confidence at present. But it seems certain that by far the greatest part of the difference in the behaviour of normal pentane and of isopentane can be attributed to the action of the law we have mentioned, We are therefore confirmed in the conclusion stated by us in a former paper" that the difference of pressure between two isomeric substances at the same temperature and volume involves the same power of the density as the first deviation from Boyle's law, i. e. the second power" (Phil. Mag. xlviii. p. 214).

XVI. Dependence of the Current through Conducting Gases on the Direction of the Electric Field. By E. RUTHERFORD, M.A., Macdonald Professor of Physics, McGill University,

Montreal*

IN

N the conduction of gases under the influence of Röntgen and Becquerel rays, it has generally been considered that the magnitude of the current between the electrodes is independent of the direction of the electric field, except in * Communicated by Prof. J. J. Thomson.

the case of potential-differences of the order of one volt*, when the contact-differences of potential between the electrodes may cause inequalities in the currents. In the general case, however, of gas-conduction, I have found that there is, in most cases, a difference between the values of the current with the reversal of the electric field. This difference is due to the inequality of the velocities of the positive and negative ions.

It is only in certain special cases that the current is independent of the direction of the electric field.

These cases are:

(1) When the ionization of the gas is symmetrical with regard to the electrodes.

(2) When the electric field is so great that the current is a maximum, i. e. when all the ions reach the electrodes before recombination occurs.

(3) When the number of ions present is so small that their movement between the charged electrodes does not appreciably disturb the potential-gradient.

(4) When the positive and negative ions have equal velocities. În all other cases the positive and negative currents are unequal in value, the difference depending upon the distribution and intensity of the ionization, the distance apart and shape of the electrodes, and the potential-difference applied.

In many of the experimental arrangements of previous observers, by means of which the equality of the current in the two directions was observed, one or more of the above conditions was fulfilled.

The essential conditions for obtaining unequal currents

are:

(1) Ionization unsymmetrical with regard to the electrodes. (2) Disturbance of the potential-gradient due to the movement of the ions in the electric field.

(3) Unequal velocity of the ions.

A few simple cases will now be considered, in which the difference between positive and negative currents is strongly

marked.

The difference between the currents in air is most readily shown when the gas is dry, and when the ionization between the electrodes is powerful and confined mainly to the surface of one electrode. This condition can readily be fulfilled by allowing a thin stratum of strong Röntgen-rays to pass between two parallel plate-electrodes, but nearer one plate than the other, or by using a powerful radio-active substance, * E. Rutherford, Phil. Mag. Jan. 1899.

like radium, where the ionizing-power due to the radiation is mainly confined to within a few centimetres of the active surface. Some experiments with Röntgen-rays will be given later, but, on account of the inconstancy of a Röntgen-tube, it was found more convenient and accurate to experiment with radio-active substances.

Fig. 1 shows the general arrangement of the experiment with radio-active substances.

Two parallel circular insulated plates of lead A and B,

19 centims. in diameter, were fixed horizontally inside a tin vessel C. The lower plate was covered with a layer of tinfoil on which a thin layer of radium was lightly sprinkled. The upper plate was also covered with tinfoil and rigidly attached to the lid of the vessel. The central portion D of the upper plate was separated from the outer by a narrow air-gap and insulated from it, so that the outer served as a guard-ring. The electrometer was connected to the central plate in the usual manner; the guard-ring and vessel were connected to earth; the lower plate to one terminal of a large battery of small accumulators, the other terminal of which was to earth. The crossed lines in the figure represent insulators. The current between the plates was measured by noting the deflexion of the electrometer-needle in a given time.

The quadrants were separated, and the connexion with the