It would be unnecessary and tedious to describe all the experiments which led me to adopt the above arrangement, but there are one or two of interest that it may be as well to mention. If jet (a) in fig. 1 is used alone, the branches of the three-way piece Z being put into connexion with the air-bag and the gas supply respectively, a common blowpipeflame can easily be obtained from (a), and this will be found to sing, but not nearly so loudly as the flame already spoken of the blue tip of the inner cone can be seen shooting in and out as it emits the different sounds. If now a thin sheet of platinum-foil be brought into this part of the flame, the sound is very markedly reinforced, the tones given out by the hot platinum being very "round," smooth, and pleasant to the ear, though, as I have already said, they are not loud. I attribute this reinforcement to the rapid expansion and contraction of the platinum in response to the variation in the heat of the flame caused by the sound-waves from the phonograph, and not to simple reflexion of these waves from the platinum surface. The well-known surface action of the metal may also play some part in the phenomenon, although there is no reinforcement, if, whilst the platinum is in position and sounding, the blowpipe-flame is blown out: the platinum, under these circumstances, remains at a yellow heat, from the combustion of the unignited gas and air on its surface, but there is no sound until the flame is rekindled.

Finally, the action of the flame in reinforcing sounds originally of equal loudness but of different pitch deserves notice. This can be generally stated as follows:-The smaller the flame the more rapid the current of air or of gas, or of both; or the more jet (b) is made to encroach on jet (a) in fig. 2 (b), the higher is the pitch to which the flame most readily responds and reinforces, and vice versa: but it is very remarkable how wide a range of pitch the tones may have which are reinforced at one and the same time, so that it is possible and easy in the case of a phonograph record in which the high-pitched notes have been recorded too strongly, to lessen their loudness, or, indeed, to eliminate them altogether without seriously interfering with the other notes on the other hand, if it is the base which is too prominent, this may be reduced, whilst additional strength is added to the treble.

I do not think that it is too much to say that these facts*

*Note.-Since the reading of this paper, the writer has been much indebted to Mr. Chichester Bell for calling his attention to a communication of Mr. Bell's, read before the Royal Society, and published in the Transactions' of that Society, 1886, Part 2, and entitled "The Sympathetic Vibrations of Jets." It appears that as early as 1866 Kundt obtained musical tones by the impact of two flames, and of an air-jet

place at our disposal a method of reinforcing sounds possessing some special advantages, and may prove capable of increasing faint sounds so that many can hear them at once, to a degree beyond that attained by any other method at present known. Eton, Oct. 1903.

XXX. The Photographic Action of Radium Rays. By S. SKINNER, M.A., University Demonstrator in Experimental Physics at the Cavendish Laboratory, Cambridge *.

IT

T is well known that a photographic plate by exposure to Radium rays is affected in such a way that the plate develops similarly to its development after exposure to light. The following experiments are an attempt to answer the question: Are the actions the same? As far as can be seen by eye, plates exposed to the two agents appear on development to be darkened in a similar way, and we may conclude that the final result of the actions and of the development is the same. I have attempted to obtain information of the intermediate stages, and some of the experiments described below appear to indicate that only slight differences occur in these early stages. I shall describe the experiments in the order in which they may be best understood, although this was not the order in which they were made.

A plate, inclosed in two envelopes of paper, one red and the other black, was placed under a capsule containing 10 milligrammes of radium bromide at a distance of 1 centimetre from the plate for periods of time varying from half a minute to 48 hours. A different part of the plate was

against a flame. Barrett and Tyndall were apparently the first to notice that a sensitive flame sometimes reproduces the tones by which it is affected. In 1875 Decharme found that carbon dioxide blown through a jet against a flame gave a feeble effect, and attributed this to the decomposition of the CO-a chemical explanation. He also found that pure oxygen gave only a feeble effect in comparison with air, and that nitrogen gave no effect at all. Mr. Bell himself, by boring a small hole in a telephone-plate, and forcing through it, at a gentle pressure, a stream of air which impinged on a small flame, reinforced the otherwise inaudible sounds of the telephone till they could be distinctly heard over a small room. The explanation of the reinforcement, -founded by Mr. Bell on a long series of delicate and beautiful experiments, including the measurement of the relative pressures at different points in and near the jets,- -seems to be a purely physical one. actual experiments as described in the present paper are, however, new, so far as the author can ascertain-though Mr. Bell has experimented on the use of flames in connexion with the graphophone, and took out a patent thereon in 1886.

*Communicated by the Physical Society: read January 22, 1904.

The

presented to the radium for each period of time. It was then developed in alkaline hydroquinone for three minutes, and the intensities of the developed images were compared by means of a commercial densitometer. A similar experiment was also made with 50 milligrammes of radium bromide. For most of the experiments a red label rapid Ilford plate was used, and in a few a Lumière' plate.

The general result of these experiments showed that the intensity of the developed image increased rapidly to a maximum value, then decreased rapidly, and finally the intensity decreased very slowly until a stage was reached in which there was practically no dark image formed on development. The time of exposure required to reach the maximum depended on the mass of the radium bromide used, that with 50 mgr. being earlier than with 10 mgr. In fact, after 40 hours' exposure the effect on the plate might be compared to the complete reversal of image obtained when a plate is very much over-exposed to a bright light. The following curves refer to Ilford red label rapid plates, I. exposed respectively to 50 mgr. and II. to 10 mgr. (fig. 1). With this class of plate the maximum was reached

With Lumière plates the hours, see fig. 2 (p. 290).

in 15 minutes with 10 mgr. of radium bromide, and in 7 or 8 minutes with the 50 mgr. maximum was reached in about 2 To verify the result that an exposure greater than a certain critical value will produce a less dense image, the following experiment was made. A slit was cut in a thick sheet of lead and a covered photographic plate was placed below it. It was then exposed to radium rays for 10 minutes. Then the lead plate was removed, and the plate was exposed for a further period of 10 minutes. In this way the parts opposite the open slit received altogether 20 minutes'

exposure, whilst the protected parts were exposed only for half that time. On development the slit appeared as a light line crossing a dark field. This result follows at once from the diagram (fig. 1), for if OA represents the 10 minutes' exposure the density will be AM, and for a time exposure OB the density is BN, a value less than AM.

Here it will be recognized that the action on the plate is like that of light, for an overexposure produces a fainter image. The action, however, is more under control as its rate is slower.

From the arrangement of the radium* I am inclined to think that the a-rays were all absorbed before reaching the plate, and that we have to consider the action as coming from the B- and y-rays alone, these two sets of rays having much more penetrating properties than those of the a group.

The next experiments were made with the view of finding out whether a plate exposed to the light of an electric spark would undergo any reversal if it was afterwards exposed to radium rays. In fact, I tried to obtain a reversal effect similar to that got by Clayden, by exposing a plate first to electric sparks and then to faint gas-light, only I substituted radium rays for the after-exposure to faint light. Clayden showed that it was easy to reverse the images of bright sparks by exposing the plate subsequently to a gas-flame for a few seconds. In this way he explained the reversed images

The radium bromide was contained in an ebonite capsule covered with a thin mica plate. The rays passed through the mica.

of lightning-flashes which are obtained when several images are taken on one plate, or when images are obtained of lightning-flashes with a bright cloud behind them.

In my experiments an Ilford plate (red label) was exposed in a camera to a series of Wimshurst machine sparks, the plate being shifted between each spark. Then the plate was placed in its envelopes, and a part of it was protected by a thick leaden screen. The rest of the plate was exposed to the radium. This was repeated for different periods of radium exposure. On developing these plates, it was found that in those which had been exposed to the radium for only short periods, gradually increasing in length, there was a progressive elimination of the spark image, but the density of the parts corresponding to the radium exposure was not so great as that of the parts of the plate only exposed to the spark, although the radium action had been able to obliterate completely the spark effect. On the other hand, when those plates which had been exposed longer, or had been exposed to the greater mass of radium, were developed, I found that reversed images of the sparks appeared. These images were not so wide as the spark images. In another case, with still further exposure, I observed the radium reversing its own image, and across the radium reversed part of the plate there appeared a faint dark image of the spark. This might be described as a re-reversal of the spark.

The conditions of these experiments were such that probably only the B- and y-rays were acting. In the Philosophical Magazine for November, 1903, a paper appears by R. W. Wood, on the subject of photographic reversal, and in it he says that he has been unable to reverse a sparkimage by Röntgen rays, which are supposed to be similar, if not identical with the y radium rays. If this is the case we must ascribe the reversal, which I have described as produced by radium, to the action of the B-rays alone. This point, however, needs further investigation with screens capable of cutting off each class of rays.

In Wood's paper he gives what may be called a reversal order. It is :It is Pressure-marks, x-rays, Light-shock, Lamplight. In this order each successive influence is able to reverse those before it. He also states that Becquerel rays will reverse pressure-marks, and that lamplight will reverse the effect of Becquerel rays. The Becquerel rays which he used were probably those from a compound of uranium. My experiments show that Light-shock (from an electric spark) may be reversed by Radium rays, and consequently