PERCENTAGE OF BEAM TRANSMITTED LOGARITHMIC SCALE)

cent. A slight creep of the zero (mentioned in the original paper) was also observed, but did not seriously affect the results, and could be allowed for.

The logarithm of the observed ratio was plotted as a function of the thickness of the aluminium, giving an X-ray absorption curve for the metal (fig. 3). As the tube was operated much below its critical voltage for characteristic radiation, the resulting line shows a slight curvature, due to

the beam being heterogeneous. The points, however, lie on a good line, especially considering that only one value was taken for each point and that the whole experiment did not occupy more than about twenty minutes. The scale was calibrated by a method described later, and found to be linear within the degree of accuracy obtainable in this experiment.

It is interesting to note that, if positive ionization currents are used (i. e. if the high-tension electrode of the chamber is

positive), the deflexions due to ionization currents, and those due to high-frequency surges picked up by the lead from the grid to the collector, are in opposite directions, a fact which is useful in deciding whether the amplifier is sufficiently well screened or not.

The effect of substituting a gas-tube supplied from the same coil for the Coolidge was next investigated. It was found that the output of the former was far too variable to give a steady deflexion, fluctuations of up to 100 per cent. being obtained. For work with such a tube some form of "integrating" indicator, such as an insulated electrometer etc., is evidently preferable to the valve bridge. The latter, however, might be useful for demonstrating the variation in the output of a gas-tube.

In the above tests, while the X-ray tube was close (about 2 feet) to the valve bridge, and in the open, the induction coil was several feet away. It is possible that, if the latter were situated near the amplifier, slight unsteadiness of the zero might be occasioned from the magnetic field of the coil, despite the most careful screening. Fortunately, however, in most X-ray installations it does not matter where the coil or transformer is situated, and it should not be difficult to arrange that it is several feet from the amplifier.

Photo-electric Effects.

Tests were carried out to ascertain whether the amplifier was suitable for measuring photo-electric currents. As a rough qualitative test, the apparatus and ionization chamber was set up as previously described, and a polished zinc plate. attached to the collector of the chamber, the other electrode of the chamber being, as before, maintained at a potential of plus 60 volts by connexion to the high-tension battery supplying the anode currents. On allowing a beam of ultraviolet light from a mercury arc to enter the chamber, a large galvanometer deflexion was obtained, showing that the amplifier can be used to demonstrate the phenomenon of photoelectric emission.

For quantitative tests a potassium photo-electric cell was employed. The anode, or collector of the cell, was connected to one of the grids of the amplifier, and a suitable potential applied to the cathode from a separate battery, through a safety resistance. The cell was enclosed in a metal case, which was connected to the screening-case of the amplifier, and also to the guard-ring of the photo-electric cell. In this particular case the battery supplying the

potential to the photo-electric cell was outside the screeningcase. If, however, trouble is experienced from highfrequency surges in the laboratory, it might be advisable to enclose it in the case also.

Light from a 60-watt metal filament lamp was admitted to the cell through a narrow slit in the outer screening-case. On gradually closing up the slit, by decreasing its length by means of a metal wedge placed across it and plotting the observed galvanometer deflexion as a function of the area of the slit remaining exposed, a curve was obtained which was practically a straight line, showing that the relation between the galvanometer deflexion and the current to the grid was approximately linear for small currents. This is the method of calibrating the scale referred to in the previous section.

Using this simple apparatus, it was possible to locate the positions of X-ray spectral lines on a photographic negative by moving the latter slowly across the slit and observing the galvanometer deflexion. This suggests that the amplifier, in conjunction with a comparatively low-sensitivity galvanometer, may be used with advantage to replace the electrometer, or high-sensitivity galvanometer, usually employed for photometer work in connexion with photo-electric cells. The advantages to be gained in such a case are: (1) high insulation of the leads etc. is unnecessary, (2) readings could probably be taken more quickly, and (3) the amplifier is much more portable and easier to set up than an electrometer.

It should be observed, however, that with such an arrangement, in taking zero readings, the grids (and hence the "collectors") must not be earthed. Instead, to obtain the zero reading the beam of light entering the cell should be cut off by means of a shutter, as described in the previous section in connexion with X-rays. It is also advisable, in choosing a galvanometer for the bridge, to see that it has a fairly high damping factor, as this will tend to mask any slight unsteadiness of the zero.

In conclusion, the writer desires to express his sincere thanks to Dr. E. A. Owen for offering facilities for carrying on this investigation at the Physics Laboratory of the University College of North Wales, Bangor, and also for the valuable suggestions made by him and his interest in the work. Further, he wishes to record his thanks to the various research students of the department for their assistance in carrying out tests of the amplifier with their X-ray and photo-electric apparatus.

XXXIV. Notes on Active Nitrogen. By ARTHUR EDWARD RUARK, Ph.D., Mellon Institute of Industrial Research, University of Pittsburgh, and Gulf Oil Companies

R ECENTLY Okubo and Hamada † have published a paper on spectra excited by active nitrogen when it comes in contact with metallic vapours. Some of their results do not agree with those previously obtained by Ruark, Foote, Rudnick, and Chenault ‡, and it seems of interest to examine the causes of this discrepancy. Okubo and Hamada used a discharge-tube which they describe as being similar to those used by Strutt and Fowler § and by Mulliken . The pressure-range in which these tubes were operated is not stated, but it may be inferred that it was of the order of several tenths of a millimetre, since this is the pressure which ordinarily gives the brightest afterglow when the Geissler tube is used to produce active nitrogen. On the other hand, Ruark, Foote, Rudnick, and Chenault used an electrodeless ring-discharge in a pyrex sphere about 30 cm. in diameter, because this discharge can be run at much lower pressures, of the order of 01 mm. (see p. 19 of our paper). Under these conditions, with a pressure ten to thirty times smaller than that which gives the best results in the ordinary discharge-tube, secondary effects are minimised. It is felt that this difference of pressure explains why we recorded fewer lines of thallium than Okubo and Hamada, and that it accounts, at least in part, for the fact that we obtained only the resonance line of cadmium at 3261 Å., and did not observe any spectral lines of sodium under the conditions. described.

Because of possible secondary effects in the discharge used by Okubo and Hamada, such as collisions of the second kind between metal atoms and excited nitrogen molecules, it seems reasonable to say that their observation of the lines 23P-2 P' of magnesium does not prove conclusively that two electrons can be displaced simultaneously to higher energy levels by the primary process which gives rise to metallic spectra at much lower pressures.

1

The writer and his colleagues recorded the mercury line 23P1-63D1, which has an excitation potential of 100 volts. On the other hand, Okubo and Hamada state that they could *Communicated by the Author.

+ Phil. Mag v. p. 372 (1928).

1 J. O. S. A. & R. S. I. xiv. p. 17 (1927).

§ Proc. Roy. Soc. lxxxv. p. 377 (1911), and lxxxvi. p. 108 (1911). Phys. Rev. xxxvi. p. 1 (1925), and previous papers.

not obtain any lines coming from levels higher than 4 D, with an excitation potential of 9.51 volts, even with exposures of 100 hours, although they were able to obtain the lines 2 P-4 D in four hours. It is probable that this discrepancy also is caused by the difference in pressure, although it is difficult to construct a detailed explanation in the present state of our knowledge.

The second positive bands of nitrogen were observed in some of our afterglow spectra. Okubo and Hamada have doubted our results, stating: "It may be questioned whether their sectored disks operated satisfactorily and perfectly cut out the direct discharge or not." This objection is invalid. Careful tests were always made to be sure that the sectored disks did cut out the direct discharge. It was easy to attain this condition since the disk was run slowly. Frequently it was so arranged that it made one revolution in five to ten seconds, and it was never run faster than one revolution per second.

Okubo and Hamada mention the existence of surface fluorescence of metals placed in the afterglow tube. This has been discussed by E. P. Lewis and others. The writer has observed a surface fluorescence of magnesium in active nitrogen as well as in mixtures of active nitrogen and helium. The distribution of the glow was not uniform, and it is believed to be associated with the presence of magnesium oxide or hydride on the surface. Small particles of magnesium in contact with active nitrogen glow with a bright white light. Presumably they become incandescent, although this cannot be stated with certainty.

Following Sponer, it is often assumed that active nitrogen owes its properties to the presence of neutral unexcited nitrogen atoms which combine in triple collisions with other molecules and atoms, exciting them by virtue of the heat of association which is transferred to them. Various theories may be advanced as to the subsequent history of the excited entities. It may be worth while to point out that a study of the absorption spectrum of activated nitrogen in the region where the absorption lines of the neutral nitrogen atom lie would provide evidence as to the presence of such atoms. The writer is not in a position to undertake such a test, and hopes this note will call it to the attention of someone equipped with suitable spectrographs.

It is a pleasure to thank Mr. Philip Rudnick for helpful comments.

Pittsburgh, Pa.,

May 8, 1928.