582 



NA TURE 



[April 23, 1896 



the position of this point with respect to the tube varied 

 with the shape and character of the tube used to pro- 

 duce the rays. When the bulb was pear-shaped, with 

 the negative electrode in the narrow part of the tube, 

 the point was near the place where the bulb began to 

 widen out ; while with bulbs containing metallic plates 

 to reflect the rays, this origin of the rays was on the 

 metal plate whatever might be the position of the nega- 

 tive electrode. But in no case did the origin of 

 the rays fall on the part of the tube nearest to the air 

 space through which the leakage was measured, though 

 this was the part of the tube which was most brightly 

 phosphorescent. The origin, however, in these experi- 

 ments was always at a place where some kathode rays 

 struck against a solid obstacle. In these experiments 

 the rays came through a small hole, so that the want of 

 coincidence between the origin of the rays and the part 

 of the tube where the phosphorescence was brightest, 

 could hardly be explained by supposing that the direction 

 of emission of these rays is practically almost confined 

 to the normal to the phosphorescent surface, so that the 

 apparent source of the rays would be the locus of inter- 

 section of the normals to this surface. 



The results of the various investigations as to the source 

 of the rays seem to show that this is most commonly at 

 a place where some kathode rays strike against a solid 

 obstacle ; they also prove that there are regions in the 

 tube where, though many kathode rays are stopped, very 

 few Rontgen rays are produced. The experiment of 

 Rowland previously mentioned, as well as the one by 

 Lord Blythswood, where the rays were produced at the 

 negative terminal of a Wimshurst machine without a 

 bulb at all, suggest that there are other sources of these 

 rays besides places where solids are bombarded by 

 kathode rays. Judging from analogy with the behaviour 

 of an ordinary discharge tube giving a luminous dis- 

 charge, there seems nothing unreasonable in the idea 

 that these rays may sometimes originate in the gas itself 

 by the splitting up of its molecules under the electric 

 discharge. When the discharge is luminous there is, as 

 Dr. Schusterhas pointed out, generally a peculiar spectrum 

 emitted by the negative glow outside the negative dark 

 space ; there is thus in this region, under the com- 

 paratively feeble electric field which can exist when the 

 discharge is luminous, some process going on which is 

 favourable to the emission of radiation. Now, when the 

 tube is emitting Rontgen rays the dark space reaches to 

 the walls of the tube, and the intensity of the electric 

 field is enormously greater than when the discharge is 

 luminous. (It is most interesting to connect a voltmeter 

 to the terminals of a tube while it is still on the pump, 

 and observe the way the readings go up when the tube 

 is getting into a fit state to emit Rontgen rays.) It would 

 seem likely that under this very intense electric field the 

 gas in the tube would be more thoroughly split up than 

 under the feeble field which accompanies the luminous 

 discharge, and that this . finer subdivision of the gas 

 would cause the emission of radiations of much smaller 

 wave-length. It would be interesting to investigate 

 whether the nature of the gas in the bulb has any 

 effect upon the character of the Rontgen rays emitted 

 by it ; it would, however, be necessary to use other 

 pumps as well as mercury ones to try this experiment. 



The existence of a definite metallic reflection, which 

 is considerable at grazing incidences, has been proved 

 by Dr. Joly ; while Lord Blythswood and other ex- 

 perimenters have published accounts of investigations 

 which confirm Rontgen's observation of a copious diffuse 

 return of rays from a solid obstacle. The definite metallic 

 reflection gives us hope of arriving at the wave-lengths 

 of these rays by interference experiments with m.etallic 

 mirrors inclined at a small angle, though before this ex- 

 periment can be successfully attempted it will be 

 necessary to study more closely the want of homogeneity 



NO. T382, VOL. 53] 



in the Rontgen rays, and to devise methods of producing 

 monochromatic rays From an interference experiment 

 with negative results, M. Sagnac concludes that the 

 Rontgen rays he was using must have had a wave-length 

 less than 4 X io~^ centimetres. 



A very remarkable experiment has been made by M. 

 Lafay, in which effects are produced similar to those 

 which would occur if under certain circumstances the 

 Rontgen rays were absorbed, and rays having the pro- 

 perty of being deflected by a magnet emitted. The 

 experiment is as follows : a photograph of a needle is 

 taken through a thin piece of silver-foil ; the Rontgen 

 rays, after passing through the foil, traverse a magnetic 

 field. If the foil is not electrified, the magnetic field 

 produces no effect on the position of the shadow of the 

 needle ; when, however, the foil is electrified (by a source 

 quite independent of the one used to work the coil), the 

 shadow is deflected when the magnetic field is on, the 

 direction of deflection is reversed when the sign of the 

 electrification of the foil is reversed. Only a preliminary 

 account of this experiment has yet been published, and 

 in this it is not explicitly stated whether or no the width 

 of the shadow cast by the needle is affected by the 

 magnetic field. The point is an important one in the 

 interpretation of this experiment, for the Rontgen rays, 

 falling on the charged silver foil, will discharge its elec- 

 trification and produce electric currents in the air ; 

 these currents, probably in this experiment flowing 

 somewhat in the direction of the rays, would, when 

 they passed through the photographic plate, prob- 

 ably affect it. Now these currents will be deflected 

 by a magnet, the magnet driving some of them on to 

 the part of the photographic plate previously occu- 

 pied by the shadow of the needle. They would thus 

 encroach on one side of the shadow ; they would not, 

 however, affect the other side unless the Rontgen rays 

 were stopped by the charged plate; as though these 

 currents can obliterate a shadow, they could not produce 

 one where the Rontgen rays are present. Thus the 

 effect of the currents would be to cut a piece off the 

 shadow, and the piece would be cut off from one side or 

 the other, according as the silver-foil was positively or 

 negatively electrified. These currents, however, could 

 not, with the presence of the Rontgen rays, explain a 

 simultaneous shifting of both sides of the shadow of the 

 needle. 



Rontgen's discovery of the close connection between 

 the absorption of these rays and the atomic weight of the 

 absorber, has been confirmed and extended Isy many 

 observers. This is a most interesting result, and it may 

 be remarked is what would occur on Front's hypothesis 

 of the constitution of the elements, if each little prim- 

 ordial atom furnished its quota to the absorption of 

 these rays. 



The rate of leakage of electricity through different gases 

 under the influence of these rays has been measured by 

 Mr. McClelland and the writer. We find that in general 

 the rate of leakage increases with the atomic weight of 

 the gas, although there are exceptions to this rule. The 

 rapidity with which electrification leaks through the 

 halogens is very remarkable ; it is interesting to find that 

 the gas through which the rate of leak was greatest was 

 mercury vapour, although the ordinary electric discharge 

 only passes through this gas with great difficulty. When 

 leakage takes place between two platinum plates in a gas 

 exposed to these rays, the plates show strong polarisation. 



The connection between the rate of leak through a gas 

 and the potential difference is very remarkable. When the 

 potential difference does not exceed two or three volts, the 

 rate of leak is proportional to the potential difference. 

 Then increasing the potential difference, we arrive at a 

 stage where the rate of leak increases more rapidly than 

 the potential difference. Increasing still further the 

 potential difference, we soon arrive at a stage where the 



