39° 



NA TURE 



[February 22, 1906 



xerting a pressure on the ions against the translation 

 resulting from radiation ; besides this force an electro- 

 magnetic force — of second order of the ratio of velocity of 

 translation to velocity of light — may arise from the moved 

 charges of the ions and act on the vibrating electrons. 

 The experimental research of the light of Kanal-strahlen 

 emitted normally to their direction has given the following 

 results. The observations have been made on hvdrogen ; 

 ■the velocity of the Kanal-strahlen was o'q.io' and i'2.io* 

 cm. sec.- 1 . The spectrograms were taken with a prism- 

 spectrograph and with a concave grating of i metre 

 radius. 



The total radiation of the line spectrum (Ha, H,8, . . .) 

 is partially polarised, and the electrical vibrations parallel 

 to the direction of translation have a greater intensity than 

 the vibrations at right-angles to the direction of trans- 

 lation. The difference of intensities is very small. 



The lines of hydrogen (when observed normal to the 

 Kanal-strahlen) are displaced towards the red, when com- 

 pared with the lines emitted by the slow ions in the 

 negative glow. The displacement seems to be proportional 

 to the wave-length, and also proportional to the square 

 ■of velocity. The displacement of the centre of H/3 is 

 approximately 0-8 Angstrom unit for a velocity of i - 2.io 3 

 cm. sec. - '. 



Besides this displacement there is observed a broadening 

 of the lines ; it seems also to be proportional to the square 

 of velocity, and to increase somewhat with decrease of 

 wave-length. The observations as to the splitting up into 

 components of the broadened line, and also as to the 

 polarisation of its edges, are not concordant enough in the 

 different spectrograms, and are therefore not ready for 

 publication. J. Stark. 



Gdttingen, January 6. 



Inversion-point of the Joule-Kelvin Effect. 



In discussing the Joule-Kelvin effect for a fluid like 

 hydrogen, which shows an inversion point above which 

 Treating takes place on free expansion, it is usuall} assumed 

 that this point is unique. Thus, for example, Olszewski 

 lias fixed it experimentally at -8o"-5 C. An examination 

 of the consequences of any of the usually assumed .^na- 

 tions of state (such as Van der Waals's or Dieterici's) 

 easil) reveals the fact that it must in reality be a function 

 of the pressures to which the gas is subjected. But this 

 is not all. If these consequences are examined for the 

 inversion point corresponding to an infinitesimal change 

 in pressure, it is seen that all the equations of state (which 

 at the same time indicate a critical point) demand that 

 there shall be two inversion points (if any) for nnv given 

 pressure, and that, moreover, for suffii ientl} high pressures 

 no inversion point will exist. Different equations ol state, 

 while unanimous in the above respects, indicate verj 

 different temperatures at which inversion should occur. 1 

 desire to point_ out, therefore, that a complete determin- 

 ation of the inversion points corresponding to various 

 pressures affords an exceedingly sensitive means ol dis 

 criminating between characteristic equations and oi 

 indicating the direction in which these require modification. 



I his matter is discussed in detail in a paper shortly to 

 '" published. Alfred W. Porter. 



I nivei -;i | < oil. ge, \V.< , Februarj 19. 



A Definition of Temperature. 

 A BODY containing heat is in a condition from which ii 

 tends I,, release itself (by radiating or conducting 

 heat), and this tendenc} only ceases when the bodv has 

 passed inn , heatless condition. The temperature of a 

 body is the measure of its tendency at any instant to 

 recover this heatless state (cf. Maxwell, "Theory of 

 Heat," iuih ed., p. 32). This suggests a mechanical 

 analogy; a body containing heat is analogous to an elastic 

 medium in a stale of strain, from which it tends to release 

 itself in virtue of its institutional forces; the magnitude 

 oi the restitutional force when a bodv is in a given strained 

 condition measures its tendency to release itself from that 

 tram, and so is analogous to the temperature of a bod} 

 when in a given thermal condition. The quantity of work 



NO. 1895, v OL 7$] 



stored up in producing this strained condition, and which 

 can be given out again when the body returns to its un- 

 strained condition, is analogous to the quantity of heat the 

 body contains when at a given temperature ; it is quite 

 easy to show that we can completely represent the thermal 

 condition of a body by means of a model consisting merely 

 of an elastic rod subjected to a tension. A temperature, 

 therefore, is analogous to a tension or pressure. We are 

 now in a position to give a real physical meaning to the 

 " temperature " of a body, and so enable it to be measured 

 in absolute units like a mass or a length. Let us take 

 a molecular body devoid of all heat motion and plunge it 

 into a medium the temperature of which is T. Then the 

 medium will exert an intermittent pressure or force on 

 the molecules, thus setting them into motion and generating 

 heat motion in the body. It can easily be shown that this 

 force cannot be infinite, or a cold body placed in a hot 

 medium would instantly acquire the temperature of the 

 medium, whereas it always takes a definite time to do so. 



The maximum force which the medium exerts on a 

 molecule at rest when placed therein is the numerical 

 value of its temperature. Hence we arrive at the follow- 

 ing definition of temperature : — 



.1 molecule at rest when placed in a medium possessing 

 temperature is subjected to an intermittent pressure; the 

 greatest value of this pressure is the correct measure of 

 the temperature of the medium m the neighbourhood of the 

 molecule. Another method of stating the same thing is to 

 say that the greatest force required to hold a molecule at 

 rest when placed in a medium is the measure of the 

 temperature of the medium. .Still another statement is to 

 say that the temperature of a medium is the magnitude 

 of the force tending to drive heat motion into an absolutely 

 cold body placed therein. A temperature, therefore, 

 should be measured as a pressure in dynes per sq. cm. 

 All the ordinary laws of thermodynamics, the flowing of 

 heat from bodies of higher to bodies of lower temperature, 

 Waterston's hypothesis, &c, follow quite simply as a con- 

 sequence of this definition, as the reader can doubtless 

 work out for himself. Geoffrey Martin. 



Kiel, February 10. 



Chinese Names of Colours. 

 I\ reply to the letter of Mr. Alfred H. Crook contained 

 in your issue of January 11, I would say that it is possible 

 that the explanation of the Chinese colour-name is to be 

 found in the violet coloured halo which is very commonly 

 noticed by Alpine climbers surrounding moving objects. 

 Dr. Ellis attributes it, I believe, to fatigue of the eye (see 

 discussion in Nature, May, 1897). 



Reginald A. Fessenden. 



In your issue of January 11, Mr. Alfred H. Crook, of 

 Hong Kong, asks why the Chinese should call a bright 

 purple (almost a mauve) " snow green," and he adds that 

 the term " green " is sometimes applied to the colour of 

 the sky, which I take to mean blue. The following is a 

 possible explanation : — 



One of the commonest places in nature to find purplish 

 hues is in shadows, and shadows on the snow, when the 

 sky is clear, are decidedly purple. If purple is to be 

 classified among the colours, it will go with the blues, 

 hence " snow green " as meaning " snow blue " would 

 111.1 be such a misnomer as might at first sight appear. 



Pittsburg, Pa., February 7. Alfred Sang. 



Sounding Stones. 



Mr Aii kin Tingle (January 4, p. 222) and Mr. Carus- 

 Wilson (January 11, p. 246) may be interested to know 

 that at the caves of Ellora, mar Aurangabad, one of the 

 pillars in the rock-rut temples lias the same property of 

 sounding tinder a blow. 



The pillar is a massive one close to, or part of, the 

 doorway Lading to an inner shrine, and if struck with the 

 clenched fist emits .1 deep note 



s " far as I recollect, this property was confined to a 

 portion of the pillar. W. G. Barnett. 



Poona, January 29. 



