October 28, 1 S97J 



NATURE 



625 



ating X-rays wherever they impinged, behaving in fact as kathode 

 rays. (4) That in X-ray tubes, as usually exhausted, without 

 excessive precautions for the drying of the gases, and the com- 

 plete removal of residual hydrogen from the electrodes, the 

 residual gas was in most cases hydrogen or water vapour. 



In order to test the behaviour of other gases as compared with 

 argon, similar experiments, in the same tubes, were made with 

 dry air, with hydrogen, with oxygen, and with water vapour. 



With dry oxygen and nitrogen, the absorption of the gas was 

 very rapid at a pressure of one-tenth of a millimetre, if the i 

 electrodes were sufficiently heated. Although hydrogen was not j 

 observable and was presumably absent, the blackening of the tubes j 

 was very slight, and a much greater power could be applied 1 

 than in the case of argon, without melting the electrodes. 



With water vapour under the same conditions an X-ray vacuum ] 

 could not be obtained (owing probably to the slowness of diffu- | 

 sioni, unless the tube were considerably heated, either with a j 

 flame or by means of an excessive current. On allowing the 

 tubes to cool under these circumstances, the vacuum improved 

 very greatly, owing to absorption by the surface of the glass, 

 and the discharge often refused to pass. Under steady con- 

 ditions of running at a low temperature, there was no clear 

 evidence of absorption of the water vapour, in spite of the drying 

 tube on the pump. 



With carefully dried hydrogen, under the same conditions, 

 the process of exhausting the lubes with the mechanical pump 

 was extremely rapid as compared with the other gases, owing to 

 the greater velocity of difiiision of the lighter gas. With the 

 smaller tubes, ten or twenty strokes were sufficient to give 

 brilliant X-rays, starling in each case with a pressure of half a 

 millimetre to a millimetre. There was no marked absorption at 

 any stage of the vacuum, and no trace of sputtering of the 

 electrodes. We expected to find aome evidence of absorpticm 

 by the electrodes or the platinum antikathode, but it is possible 

 that these became saturated with gas very rapidly at an early 

 stage, and ceased to absorb gas at an X-ray vacuum. We con- 

 cluded that hydrogen was the most suitable gas to use in X-ray 

 tubes, but it is possible that helium, being also a very light gas, 

 might be equally good, if its inert or monatomic character does 

 not lead to the disintegration of the electrodes in the same 

 manner as in the case of argon. 



If the great resistance to the passage of the discharge from 

 the kathode to the gas in the case of argon, is dependent upon 

 the monatomic nature of the gas, it might be expected that 

 similar phenomena would be observed in the case of mercury, j 

 Some mercury vacuum lubes were therefore made in the form of 

 inverted U tubes. The electrodes were liquid surfaces of 

 mercury in each limb, to which connection was made by short j 

 pieces of platinum wire, which did not project above the surface j 

 of the mercury. These tubes were exhausted and boiled with j 

 an alternating discharge passing, until more than half the j 

 mercury had distilled over. They then presumably contained 

 only mercury vapour. When cool, the two-inch spark discharge 

 refused to pass at first, but if the tube were tilted for a moment, 

 so as to expose the platinum wire, it appeared that sufficient gas 

 was liberated to enable the discharge to pass without any 

 difficulty. The tubes showed only the mercury spectrum. In 

 the high resistance state, immediately after boiling, the kathode 

 limb, with a larger spark coil, showed brilliant fluorescence and 

 feeble X-rays. We concluded from these experiments that a 

 very small trace of another gas was sufficient enormously to 

 reduce the resistance of a mercury vapour tube, and that if the 

 vapour could be obtained quite pure, it would possibly not con- j 

 duct at all. [ 



To verify more accurately the conditions of vacuum at which j 

 these phenomena occurred, the whole apparatus was subsequently | 

 connected to an automatic Sprengel mercury pump to which a 

 McLeod gauge wa* attached. The pump and all its connections 

 were carefully tested for leakage, and the drying tube was filled 

 with fresh PjOj. We had found in previous experiments of a 

 similar character made two or three years previously, that 

 sulphuric acid, however carefully prej^ared, gave off" appreciable 

 quantities of water vapour, which would have been quite 

 sufficient to vitiate these results. 



Using a large coil and a slow mercury break, to avoid over- 

 healing the tubes, we found that fairly efficient X-rays were 

 obtained in most of the tubes at an average vacuum of •006 milli- 

 metre, if the tul)es were exhausted in the ordinary way without j 

 taking special pains to remove the hydrogen. The H lines 

 always showed faintly in the kathode light before this vacuum 



was reached. After using an alternating dicharge to heat the 

 electrodes, and carefully washing out the hydrogen as far as 

 po-sible with argon, we found that the pres-ure corresponding 

 to an X-ray vacuum gradually increased up to -030 millimetre. 

 Before letting the argon into the tube it was allowed to remain 

 ten or fifteen minutes in contact with the fresh PjOg On 

 omitting this precaution and admitting the argon direct from a 

 bulb containing an old sample of PjOj, which was beginning to 

 deliquesce on the surface, it was necessary to raise the vacuum 

 to "015 millimetre before X-rays were produced. On the other 

 hand, the sudden addition of dry argon at this stage up to a 

 pressure of '029, produced no change in the appearance of the 

 tube. It is probable that we never succeeded entirely in elin* 

 inating the residual hydrogen, but we concluded from these and 

 similar experiments that the presence of the argon by itself had 

 little, if any, effect on the production of X-rays, since the 

 amount present in the tube could be varied within wide limits. 



We next endeavoured to ascertain at what degree of vacuum 

 the apparent absorption of the argon previously observed could 

 be reproduced. For this purpose we used two tubes of the 

 double-focus pattern, and an alternating discharge. Taking the 

 first tube slightly damp from the blowpipe, we exhausted it to 

 one-fifth millimetre vacuum with the mechanical pump. The dis- 

 charge was then turned on and adjusted to heat the tube and 

 electrodes as much as possible with safety, and the pump was not 

 further worked. Under these conditions the remaining water 

 vapour was rapidly expelled and absorbed by the P2O5, the tube 

 soon showed a brilliant hydrogen spectrum followed by green 

 fluorescence, the antikathode became red hot, then cooled, and 

 in fifteen minutes the discharge (20,000 volts) refused to pass. 

 The tube was not appreciably blackened. On connecting to the 

 direct current discharge, it gave brilliant X-rays. This case is 

 interesting as showing that a very good vacuum may be obtained 

 in these cases by simple absorption. 



Dry argon was then admitted into the tube up to a pressure 

 of 0160 millimetre. At this pressure, in tubes three inches in 

 diameter, with the direct or alternating current, the tube was 

 filled with blue light, and gave a spectrum which was verified 

 to be that of blue argon, without any visible trace of hydrogen 

 or nitrogen. After running the direct and alternating current 

 through the tube for half an hour, the tube became very black, 

 but there was no change in the pressure as measured by the 

 gauge. It should be remarked that in measuring these high 

 vacua, the pump was usually stopped to allow time for the 

 equalisation of the pressure throughout the apparatus. The 

 alternating current in the primary of the coil was only 2"5 amperes 

 with argon, whereas 4 amperes had been used with air. The 

 latter current, if used with argon, would have melted up the 

 kathodes. Finding no absorption at this pressure, the pump 

 was started to run very slowly, and the same alternating dis- 

 charge was continued. The sputtering of the electrodes rapidly 

 increased, and at a vacuum of about one-tenth of a millimetre, 

 the upper electrode suddenly melted off. Another tube was 

 then tried, but met with a similar fate at the same degree of 

 vacuum. The failure was so sudden that it was difficult to 

 control the current in time. If the argon is actually absorbed, 

 it is clear that the phenomenon depends upon very special con- 

 ditions of temperature and discharge. It is possibly that the 

 absorption is only apparent, and corresponds to a very sudden 

 increase of resistance to the discharge at a particular degree of 

 vacuum, such as occurs in an ordinary X-ray tube when the 

 boundary of the dark space reaches the antikathode. With 

 greater care, it may be possible to decide this point, but an 

 unfortunate accident to our water mains prevented further 

 investigations at the time. It is clear, however, that the 

 behaviour of argon is peculiar, and it seems probable that most 

 of the ordinary kathode ray phenomena are due to residual 

 hydrogen. 



THE ORIGIN OF THE EUROPEAN FAUNA} 

 T HAVE endeavoured to show how the present fauna of 

 -'■ Europe originated. For that purpose it was found advisable 

 to commence the inquiry by the study of the past and present 

 fauna of an island. The British Islands, and in particular 

 Ireland, seemed to me most suitable for that object. 



1 Summary of a paper by Dr. R. F. Scharf, read before the Royal Irish 

 Academy. (Reprinted from the Proceedings of the Academy, 3 ser. vol. iv. 

 No. 3, 1897.) 



NO. 1 46 I, VOL. 56] 



