LONDON, EDINBURGH, AND DUBLIN

PHILOSOPHICAL MAGAZINE

AND

JOURNAL OF SCIENCE.

[FOURTH SERIES.]

MAY 1875.

XXXVIII. On Spectra of Gases. By M. EUGEN GOLDSTEIN *.

WÜLLNER has recently published some experiments from

which he concludes that the appearance of different spectra of one and the same gas does not depend on the temperature.

It is known that the complete spectra of the gases have hitherto only been produced by means of electrical discharges; and we are accustomed to distinguish two different kinds of discharges, the continuous discharge and the disruptive discharge.

Discharges which we call continuous show in a mirror rotating about thirty times a second the image of a continuous field of light. I will not decide whether this discharge is really continuous, or whether it consists of sparks following each other so rapidly that the rotating mirror cannot resolve them into separate images, at least not with the velocities which we can conveniently give to the mirror. This latter opinion, which Professor Helmholtz considers to be the more probable one, is suggested by the fact that just those means by which we can shorten the time elapsing between the different discharges, lead at the end to the so-called continuous discharge. The temperature is doubtless the greater the more electricity passes in a single discharge. It is therefore higher in discharges following each other slowly than in the continuous discharge. The measurements made by Riess with the electrical thermometer agree with this.

According to Wüllner, the two forms of spectra which are called by Plücker and Hittorf spectra of the first and second *Communicated by Dr. Arthur Schuster, from Poggendorff's Annalen, vol. cliv. pp. 128–149.

Phil. Mag. S. 4. Vol. 49. No. 326. May 1875. 2 A

order, or band- and line-spectra, are closely related to the modes of discharge. He believes he has established by a series of experiments" that the line-spectrum of a gas only appears in the disruptive discharge, the band-spectrum only in the continuous discharge through the gas." After these facts have been established, Wüllner tries to explain the spectra of different orders by means of the different thicknesses of the luminous layer in the two modes of discharge. In the disruptive discharge (according to Wüllner) only a few molecules of the gas emit light; in the continuous discharge, on the other hand, nearly the whole mass of the gas contained in the spectral tube is luminous. If we accept the principles deduced by Zöllner, the gas-spectrum can only consist of a few luminous lines in the first case, while the thick layer of the continuous discharge will show light of all wave-lengths which the gas can emit at the temperature in question. I shall now give the results of observations and experiments which were made in reference to the above assertions.

The close relation between the two modes of discharge and the two kinds of spectra which seemed to be established by Wüllner's experiments, became first doubtful to me in working with a tube filled with air under low pressure and_with a Leyden jar inserted between one of its electrodes and the induction-coil. The image of the discharge in a rotating mirror showed, besides continuous fields, groups of images of the tube not wider than the tube itself. ✨

The observations were at first made with a mirror rotating about four times a second, but were repeated with a velocity about twenty-five times as great. The spectroscope showed only the band-spectrum of nitrogen. Geissler's tube filled with hydrogen, on the other hand, the capillary portion of which showed the characteristic red colour when the spark was passing, gave only widened fields in the rotating mirror, the colour of which was everywhere of the same crimson colour. The spectrum consisted of the well-known lines of hydrogen. A number of other tubes filled with hydrogen gave the same result. Both experiments contradict directly the law given by Wüllner. In order to arrive at a definite conclusion, a close investigation seemed to be ne

cessary.

During the course of my experiments I found it often necessary to introduce into the circuit a layer of air, which was sometimes replaced by a second tube filled with air under high pres In these cases the tube filled with air under low pressure always shows the band-spectrum; the spark in air, as in the second spectral tube, shows the line-spectrum. If we take the views of Wüllner as proved, we must assume

sures.

that the discharge can be disruptive in one part of the circuit and continuous in the other part. We must therefore investigate the question whether the discharges go through adjacent. parts of the same circuit in the same rhythm or not, in order to be able to decide whether Wüllner's hypothesis can be correct.

If we take an analogy in the flow of an incompressible liquid, the isochronism of the discharge in different parts of the same circuit is self-evident. The experimental investigation gave the following results.

The rotating mirror showed that if the exhausted tube was at the same distance as the spark, both images were of the same width. If the discharge through air under atmospheric pressure begins, as shown by Lissajous, by a spark followed by the aureola, the image of the spectral tube begins with a luminous band, the width of which remains constant, whatever the velocity of the mirror may be.

When the discharge in the air consisted of several partial discharges, the same number of narrow images of the tube were seen at corresponding distances.

The image of the discharge through a vacuum-tube is, when the remaining circuit is metallic, very complicated, consisting of fields of different brightness, &c. Two or more tubes put into the same circuit, even if the gas in the different tubes is chemically different (e. g. nitrogen and carbonic acid), always give exactly the same image, showing in every detail the same groups of illuminated fields and the same variation in brightness. The exact coincidence of rhythm and intensity in the same discharge, according to these experiments, is evident.

A tube filled with carbonic oxide which is in my possession allows only sparks to pass; every discharge of the inductioncoil is subdivided into a greater or smaller number of sparks. This tube was inserted in the circuit together with other tubes containing nitrogen. These nitrogen-tubes separately allowed a continuous discharge to pass; but when the carbonic-oxide tube was introduced, sparks only passed, the same number as in the carbonic-oxide tube.

A Leyden jar gives very strong sparks; the effect of the induction-spark (which allows only a small quantity of electricity to pass) will only show quantitative differences compared with the discharge from a battery of Leyden jars, and will be about the same as the discharge from a small jar.

If the electricity goes through the whole circuit in the same rhythm, the introduction of a small spark through air will have the same effect as a condenser which is feebly charged.

The introduction of a condenser changes the reddish colour of the positive light seen in narrow tubes into blue. The introduction of a spark will give the same result.

2 A 2