October 17, 1907J 



NATURE 



631 



•curriculum followed is also given, and it will be of interest 

 to both engineers and otliers to note that the course at 

 Karlsruhe includes political economy and labour problems. 

 There is also a greater tendency in Germany for the 

 various branches of engineering to be treated by professors 

 who are also engaged in practical work. On another 

 point the institution at Karlsruhe differs from the colleges 

 in this country, namely, the students themselves. There 

 appear to be no student organisations, and the sociability 

 so characteristic of an English college is almost entirely 

 wanting. 



SOCIETIES AND ACADEMIES. 

 London. 



Royal Society, June 27. — " The Annealing of Copper, with 

 Special Reference to Dilatation." By Prof. T. Turner 

 and D. M. Levy. Communicated by Prof. J. U. Poynting, 

 F.R..S. 



The authors have employed a special form of cxtenso- 

 meter, in conjunction with a Le Chatelier pyrometer, in 

 order to trace the changes in the length of metallic rods 

 during the process of annealing, and have thus obtained 

 continuous curves connecting dilatation and temperature. 

 The rods, which were :|-inch in square section and 

 35 inches long, were uniformly and regularly heated in a 

 gas-fired furnace. The chief feature of the apparatus was 

 the use of water-cooled copper lubes for connecting the 

 rod with the extensometer, a plug of non-conducting 

 material being used to prevent any cooling of the end of 

 the rod by the water. By this method the whole of the 

 rod was in the furnace and uniformly heated, while the 

 rest of the system was maintained at a constant tempera- 

 ture. The water-cooled tube attached to one end of the 

 rod was firmly clamped, the other being free to move. 

 On to this tube was screwed a brightly polished brass 

 disc, against which pressed a finely rounded projection 

 attached to the short arm of a bell-crank lever, which, 

 traversing a scale, indicated the expansion of the rod. 



The scale was divided into millimetres, each millimetre 

 representing i/1200-inch expansion, corresponding lo a 

 magnification of 48: i. The suitability and delicacy of the 

 apparatus were ascertained by testing bars of wrought iron 

 and steel ; from the former a regular line was obtained, 

 while the latter gave a curve showing a marked change 

 of volume at the critical point (about 690° C), thus agree- 

 ing with Le Chatclier's results obtained by an entirely 

 different method. Hard-drawn copper bars gave a perfectly 

 regular line, similar to that obtained with wrought iron ; 

 annealed copper also gave a straight line. Experiments 

 on rods of brass of different composition, on gun-metal, 

 and on phosphor bronze gave similar results. 



It thus appears that the change from the hard, elastic 

 condition of worked copper and copper alloys to that of 

 extremely soft metal is not accompanied by any alteration 

 in length. On the other hand, it is known that allotropic 

 changes in an element, such as occur in pure iron at 

 about 880°, or such chemical constitutional changes as 

 occur in iron-carbon alloys at critical temperatures, are 

 accompanied by marked alterations in volume, and the 

 authors therefore conclude that the changes brought about 

 by mechanical work, or by annealing of worked metals, 

 produce only internal re-arrangement of the metallic 

 molecules, but are of a different order from the chemical 

 and physical changes, such as are correctly regarded as 

 allotropic. 



" Experiments on a New Kathode Dark Space 

 in Helium and Hydrogen." By F. W. Aston. Com- 

 municated by Prof. J. H. Poynting, F.R.S. 



This paper is a description of a new dark space, close 

 up to the kathode and inside the Crookes dark space, dis- 

 covered by the author while investigating the length of 

 the latter phenomenon in helium, and later found to be 

 exhibited in a less marked degree in hydrogen. The 

 length of the new dark space, which under measurable 

 conditions varies from 02 mm. to i-o mm., is almost 

 unatTe<ted by the pressure of the gas, but varies roughly 

 with the inverse square root of the current density. 



Careful observations show that the fall of potential across 

 the ijfii' dark space is constant for the same gas under 



NO. 1 98 1, VOL. 76] 



all observed conditions, and is in helium jo volts, in 

 hydrogen 15 volts. 'Ihe phenomenon may be accounted 

 for by the supposition that the energy required to ionise a 

 molecule of helium is a definite quantity, and that an 

 electron liberated from the surface of the kathode virtually 

 at rest must fall freely through a definite potential in 

 order to acquire that energy, so that the new dark space 

 may be regarded as the distance through which the 

 electrons fall in order to attain sufficient energy to ionise 

 the gas by collision with its molecules. 



The intense blackness of the new dark space in pure 

 helium bears out this theory, by which also the following 

 effect was predicted ; — Since the behaviour of electrons 

 liberated from molecules of the g.is by collisions should 

 he the same as that of those derived from the kathode, if 

 the ionisation just beyond a potential distance from the 

 kathode of 30 volts is sufl^iciently concentrated there will 

 be a further maximum of ionisation— and therefore of 

 light — just beyond a potential distance of Co volts, another 

 beyond 90 volts, and so on, each getting less definite than 

 the previous one, so that the light in the Crookes dark 

 space should be striated in appearance. By suitable adjust- 

 ment of conditions, several successive striations can be 

 seen and photographed in helium. Coinbining the poten- 

 tial differences obtained from the length of the new dark 

 space with the accepted values of e and m for the electron, 

 the following values of the energy required to ionise and 

 the velocity of the ionising electron are obtained : — 

 Hydrogen ... 1 7 x lO"" ergs 2-25 x io« cm. per sec. 



Helium 3-4x10-" „ 32 x 10' „ 



Not the slightest indication of the phenomenon has yet 

 been observed in any other gas. 



Received July 5.—" The Dispersion of Double Refrac- 

 tion in Relation to Crystal Structure." By Dr. T. H. 

 Havelock. Communicated by Prof. J. Larmor, Sec.R.S. 



In this paper the object is to consider to what extent 

 it is possible to regard double refraction, whether pro- 

 duced artificially or Occurring in natural crystals, as due 

 simply to an iEolotropic distribution of similar particles. 

 The ordinary theory of double refraction amounts to con- 

 sidering the medium as a collection of crystal molecules in 

 cubical order, all the varieties of dispersion being postu- 

 lated of the single particle ; and similarly in artihcial 

 double refraction, the a;olotropy is assumed to originate 

 wholly in the individual molecule, and to be effected 

 through the quasi-elastic force under which the polarisa- 

 tion electrons are supposed to vibrate. On the other hand, 

 a theory which confines the effect to a re-arrangement of 

 the molecules in space will express the result by a modifi- 

 cation of the effective electric force operative at each 

 particle. From this point of view the author develops a 

 theory of the optical properties of a homogeneous assem- 

 blage of isotropic particles. If the medium behaves like a 

 uniaxal crystal, it is found that the double refraction, 

 equal to the difference between the two principal indices, 

 is proportional to (n„=-i)=/"o. where n. is a mean value 

 measuring the refractive index of a medium composed of 

 the same particles in the same density, but arranged in 

 simple cubical order. , , , . , , ^■ 



This gives a law of dispersion of the double refraction, 

 provided the dispersion of the mean index n, is known. 

 The formula is applied first to the double refraction pro- 

 duced artificially in simple isotropic media. Experimental 

 data are available for the dispersion of the double refrac- 

 tion produced in carbon disulphide by an electric field, and 

 these results are found to agree well with the formula 

 given above. <r . • 



In the case of natural crystals, where the etlect is 

 larger, the theory indicates that the quantity 



(»,= -!)-"-(«,= - I)-' 



should be independent of the wave-length, «, and n, being 

 the ordinary and extraordinary indices of the crystal. This 

 relation is 'found to hold very well for quartz over a wide 

 range, and gives in this case the equation 



n,^ — n,^ = o-oi44i(i«,^— i)(n2^— i)' 



It follows, further, that the double refraction decreases 



in absolute value with the mean index n„, that is, it 



deci eases in general with increasing wave-length. This 



is the rule of dispersion in most actual crystals, but there 



