June 2, 1898] 



NATURE 



19 



iepends upon the fact that aluminium expands 2'6 times as 

 much as platinum-iridium for the same increment of temperature. 

 The author therefore employs, like Fizeau and Benoit, a tripod 

 of platinum-iridium, and places upon its transverse table, through 

 which pass the three screws, a disc of aluminium whose thick- 

 ness is i/2'6ths of the length of the screws. The space between 

 the lower surface of the glass plate which is laid upon the upper 

 ends of the screws to assist in producing the interference, and 

 the upper surface of the aluminium, then remains constant for 

 all temperatures under observation, and if a crystal is laid upon 

 the aluminium compensator the whole amount of its expansion 

 by rise of temperature is available for measurement by the 

 interference method. Hence the method is no longer a merely 



hitive one, aftbrding the difference of expansion between the 



ipod and the substance investigated, but affords directly 

 aiisolute measurements of the expansion. 



The results of numerous determinations of the expansion of 

 the platinum-iridium of the tripod are given, carried out with 

 the surface of the tripod table and the cover-wedge separated 

 at the long interval of 12 mm., by the aid of green mercury 

 light. The mean value is very similar to that of Benoit, and is 

 o = io"9(86oo + 4-56/), 



The result of several determinations in red hydrogen light of 

 the expansion of the pure aluminium used for the series of com- 

 pensators, carried out by the Fizeau relative method with a 

 block 12 mm. thick, is 



a = io"*(2204 + 2- 1 2/). 

 Similar determinations for the black glass of the crystal- 

 covering plates afford the value : 



a = lO~9(7257 + 10-4/), 

 In a subsequent memoir the author intends to present the 

 results of determinations of the expansion of the sulphates and 

 selenates of potassium, rubidium, and ccesium. 



Physical Society, May 27.— Mr. Shelford Bid well, Presi- 

 dent, in the chair. — A paper by Messrs. Edwin Edser and C. P. 

 I'lUller, on a simple method of reducing prismatic spectra, was 

 ul by Mr. Edser. The production of interference-bands in a 

 iitinuous spectrum is capable of furnishing a reference-spec- 

 !ui, which can be employed to determine the wave-lengths 

 rresponding to the bright lines in a spectrum of a metal or of a 

 ^. The authors discuss various methods by which such bands 

 n be formed. In their final experiments, an air-film between 

 1 plane parallel glass plates is inserted in front of the slit of 

 spectrometer, in the path of the incident light. Owing to 

 c interference of the direct ray with that twice internally 

 fleeted, bright bands separated by dark intervals are observed 

 the spectrum ; these bright bands correspond to a series of 

 iitlerent waves, whose lengths are easily determined for the 

 whole series, when two of them are known. The bands are 

 nnich improved by partial silvering of the two internal 

 1 faces of the glass. It has been found that ordinary plate- 

 iss, if well chosen, is good enough for all these experiments, 

 order to adjust for parallelism, a spot of light, or the filament 

 a glow-lamp, is viewed through the silvered surfaces. A long 

 in of images is generally visible ; these must be brought into 

 incidence. If now a sodium flame is looked at through the 

 111, interference bands are seen. These bands must be adjusted 

 l^ressure, to be as broad as possible. An arc-lamp is used 

 1 illuminating the collimator slit. The authors exhibited the 

 ..pi)aratus, and showed photographs of spectra-scales with the 

 appropriate wave-lengths, calibrated upon them by this method. 

 ["he results there obtained were read from the spectrometer to 0*4 

 a tenth-metre, with an ordinary pocket-lens. A simple 

 iphic method enables wave-lengths, corresponding to a great 

 inber of spectral lines, easily to be determined by inspection. 

 e phase-changes introduced by the silver do not affect the final 

 -ult. Prof. Thrclfall congratulated the authors on their discovery 

 a method that would greatly reduce the labour of calibrating 

 cctra, and at the same time give such accurate results. Prof. Boys 

 id the simplicity of the apparatus added greatly to the value of 

 method. It would seem to him better if the slit were some- 

 w contrived within the film-space. All want of definition due 

 rays falling at different angles upon the collimator object- 

 i-,s would thus be avoided, and only a small part of the glass 

 lies, i.e. the slit, would require to be strictly parallel planes. 

 le limit of accuracy in the authors' method depended upon the 

 Himator, not upon the optical perfection of the silvering of the 

 lates. Mr. Butler pointed out that previous methods had 



NO. 1492, VOL. 58] 



always required experienced spectroscopists for mapping-out 

 results. In the new method that work could easily be done by 

 an assistant. Mr. Edser said that by putting the two plates 

 immediately in front of the slit only a very small part of the 

 glass is concerned in the action, light coming through at an 

 angle would not reach the lens in the collimator.— Prof. Boys, 

 Vice-President, then took the chair, and Mr. Cam.pbell Swinton 

 read a paper on some further experiments on the circulation of 

 the residual gaseous matter in Crookes' tubes. In the discussion 

 that followed the former paper on this subject, at the Physical 

 Society on March 25, 1898, Mr. Appleyard had suggested that, 

 in tracing the cause of the rotation of the exploring mill, it 

 would lead to simpler results if the vanes were made of some 

 light conducting substance, for it was probable that mica intro- 

 duced complications by retaining the charges. Prof. Boys then 

 pointed out that the mica might be gilded. Such a tube has 

 now been made by Mr. Wolff. With the gilded mica vanes so 

 placed as to be outside the kathode stream, the mill behaves in 

 a manner similar to the non-conducting insulated mill. It shows 

 a greater tendency to as.sume a position of stability, due to 

 electrostatic induction ; this renders it somewhat troublesome in 

 starting, but, when once under way, the mill rotates always when 

 excited. Occasionally, when starting, a few reverse revolutions 

 are observed ; these are probably due to electrostatic influence 

 and momentum, and also possibly to eddy currents in the residual 

 gaseous matter. But it is found, in all cases, that rotation in the 

 direction that indicates a stream of residual gaseous matter from 

 anode to kathode, follows the reversal immediately after one or 

 two oscillations. An electrometer connected to the mill through 

 the pivot and needle-point, shows the vanes to be always electri- 

 fied positively. The results are confirmed by a second tube with 

 oblique vanes. The author concludes that at very high exhaus- 

 tions there exists a molecular or atomic stream from anode to 

 kathode, which carries a positive charge, and travels at high 

 velocity outside the opposite kathode stream. Mr. J. Quick 

 asked what was the minimum degree of exhaustion required to 

 produce these results. Prof. Boys said that the experiment 

 gave some amount of probability to the truth of Mr. Campbell 

 Swinton's hypothesis, but it did not altogether prove the 

 mechanical theory of rotation to be correct. He was glad that 

 the chance suggestion at the last discussion had led to such in- 

 teresting experiments being continued. Prof. Threlfall men- 

 tioned that Boettger had devised a method for gilding mica, by 

 a chemical process, that was much to be preferred to ordinary 

 gilding. Mr. Campbell Swinton said it was necessary to exhaust 

 the tubes as completely as possible to a point where it was only 

 just possible for any discharge at all to pass through them. If 

 the rotation was due to electrification, there must still be some 

 mechanical process whereby the charges get to the vanes — a 

 stream of residual gas satisfied that condition. — The \'ice- 

 President proposed votes of thanks, and the meeting adjourned 

 until June 10. 



Paris. 

 Academy of Sciences, May 23.— M. Wolf in the chain- 

 Notice on the late M. Souillart, Correspondant in the Section of 

 Astronomy, by M. O. Callandreau. — Some remarks on the 

 periods of double integrals, and on cycles of two dimensions in 

 algebraic surfaces, by M. Emile Picard.— New researches on 

 the reaction between pyrogallol and oxygen in presence of 

 alkalies, by M. Berthelot. The reaction depends upon the 

 nature of the alkali employed. The amounts of oxygen 

 absorbed, and carbon monoxide evolved, were measured and 

 the oxidation products studied. — Chronophotography applied to 

 the study of muscular action in locomotion, by M. Marey. The 

 paper is accompanied by four plates, illustrating the methods 

 used. After a set of photographs of the living animal has been 

 obtained, its skeleton is prepared, and these photographed 

 upon the same scale. From these the curves of change of 

 length of each muscle can be deduced. — The origin of the 

 vertebrates, by M. Edmond Perrier. — On minimum surfaces, by 

 M. C. Guichard. — On systems of differential equations which 

 satisfy quadruply periodic functions of the second species, by M. 

 Martm Krause. — On the determination of the terminal curves of 

 spirals, by MM. C. E. (juillaume and J. Pettavel. A mechanical 

 method for determining Phillips' curve for spiral balance springs. 

 — On a new method of determining the mechanical equivalent 

 of heat, by MM. J. B. Bailie and C. Fery. A cylinder of copper 

 is fixed in a rotating magnetic field, and the heating effect 

 measured. The moment of the couple required to keep the 

 sphere at rest and the velocity of the field can be exactly 



