June i6, 1898] 



NATURE 



167 



ing the way in which an expert telegraph clerk reads siphon- 

 recorder signals on a long cable, that it might be possible to 

 analyse waves without the supposition of a resonating apparatus. 

 The clerk interprets not so much the motions to one side or 

 other of the zero-line, as the rate of change of velocity, i.e. 

 the acceleration of the siphon. This had been recognised in 

 the design of those relays for long cables, where the 

 lever makes contact when the received current exceeds a 

 certain value, and breaks contact when the current falls below 

 a certain minimum. Messrs. Siemens had adopted a relay in 

 which the lever was carried on the suspended coil o( a 

 D'Arsonval galvanometer by a pivot with a small amount 

 of friction. If contact was made, the coil could, nevertheless, 

 continue its motion in a given direction. If that direction 

 altered, contact was immediately broken, and the lever passed 

 over to the opposite stop, thereby reversing the local 

 circuit. It was possible that, in the process of hearing, 

 something akin to this took place, the ear behaving as a 

 mechanism responsive, not by resonance to the complete waves, 

 but by its sensitiveness to changes of direction of the received 

 impulses. Dr. S. P. Thompson thought that a mechanism 

 similar to the relay described by Prof. Ayrton was contained in 

 the telautograph of Elisha Gray ; it was a "Prony" mechanism. 

 In the acoustical problem the ear was probably sensitive to 

 abrupt changes of shape in the waves as well as to reversals. 

 In the case of mistuned octaves, something is heard that suggests 

 "revolving" in the ear, indicating a cyclic change. In this 

 regard it was necessary to take into account the phase-relations 

 as well as the relative intensities of the component tones. — Mr. 

 E. H, Barton then read a paper on the attenuation of electric 

 waves along a line of negligible leakage. It forms a sequel to a 

 paper communicated to the Physical Society and printed in 

 their Proceedings of December 1897 and January 1898. Shortly 

 after the publication of the earlier results, Mr. Oliver Heaviside 

 drew attention to Lord Rayleigh's high-frequency formula for 

 the " effective resistance " of wires to alternating currents, and 

 suggested that the formula might be approximately applicable 

 to the case ; but he thought the experimental value of the 

 attenuation would be considerably higher than the one derived 

 from calculations. Mr. Barton here repeats the work, with 

 special precautions as to the mode of insulating the parallel 

 copper wires through which the wave-train proceeds. The 

 value of the attenuation constant deduced from these experi- 

 ments is 0000013. By applying Lord Rayleigh's formula for 

 the effective-resistance of the circuit, and using this value in 

 Mr. Heaviside's expression for the attenuation, the calculated 

 constant is 00000062. To account for the discrepancy, the 

 author points out that the effective-resistance formula was 

 originally developed for a wire placed at a considerable 

 distance from other parts of the circuit, and for currents 

 following the harmonic law. Whereas, in the experiments 

 the conditions are (i) wires i'5 mm. diameter, only 8 

 centim. apart, and (2) the waves are propagated in the form 

 of a damped train, with the large end leading ; they are 

 extinguished after ten or a dozen vibrations. Mr. Oliver 

 Heaviside (communicated) pointed out that, as there was human 

 interest in error, it might be worth mentioning that at first it 

 was supposed the previous experiments of Dr. Barton made the 

 index of the attenuation factor to be six times that of the long- 

 wave theory for simple periodic waves. And it was hard to 

 account for so large a discrepancy. The discovery of an error 

 in the figures, reduced the result from six to two. The small 

 depth of the surface-layer of effective conduction, and the dis- 

 tance apart of the wires, seemed now to make it improbable that 

 Dr. Barton's first reason (i) was adequate to account for the 

 doubling of resistances. The second (2) was of course a sub- 

 stantial reason for increased resistance. A third one, Mr. 

 Heaviside suggested, was the external resistance at the boundary 



)f the waves. A combination of the second and third reasons, 

 with a little of the first, might account for most of the extra 

 attenuation observed, and, if more was wanted, one could "try 

 ihe K.R. law." Mr. Appleyard said it was rather to be regretted 

 that, in all the experiments, the distance between the wires had 

 been the same, i.e. 8 cms. By taking a few different values (l) 

 might have been checked. Lord Rayleigh's formula for the 

 effective-resistance, involved the square-root of the magnetic 

 permeability of the wires. The author had, throughout, used 

 copper, a paramagnetic metal, and had assumed ^ = i. It 

 would be of advantage to try other metals. Mr. Barton, in 



reply, said he would make further experiments with the two 



NO. 1494, VOL. 58] 



conductors at different distances apart, and he would also try 

 iron wires. With iron, the thickness of the surface-layer of the 

 effective conductor was about one-thirteenth that of copper. Iron 

 should therefore giveagreater value of the attenuation than copper. 

 — Mr. A. Griffiths then read a paper on diffusive convection, a 

 phenomenon analogous to caloric convection. The differences of 

 density that produce convection-currents are not due to changes 

 of temperature, but to variations in the quantity of dissolved 

 substance per unit volume. The author has devised an appa- 

 ratus consisting of a vessel divided horizontally by a diaphragm, 

 through which pass two vertical tubes of unequal lengths. A 

 solution of copper-sulphate, maintained at constant strength, is 

 placed in the lower compartment. The upper compartment is 

 filled with water. Diffusion takes place up the tulies. One 

 tube is 4 cm. long ; the other is 4 05 cm. The tops of the 

 tubes are exactly at the same level. Up the longer tube, and 

 down the shorter, diffusive convection occurs at the rate of 5 

 cm. per year. This flow increases the quantity of copper- 

 sulphate transmitted by the long tube by about 2 per cent., 

 and diminishes that transmitted by the shorter tube by about 

 the same amount. Consequently, the resultant increase due to 

 the motion is only a fraction of i per cent. To detect the flow, 

 the author employs a second piece of apparatus, in which the 

 upper ends of the tubes are separated by a capillary, containing 

 coloured liquid. By this means the motion is considerably 

 magnified. Dr. S. P. Thompson asked whether, in a case 

 where a large tube was used in determining the velocity, the 

 viscosity of the liquid would not play a very much less part 

 than with narrow tubes. Mr. Griffiths explained that viscosity 

 was not important until very small tubes were considered, e.g. 

 those of the order O'ooi mm. diameter. — The President pro- 

 posed votes of thanks to the authors, and to Dr. Max Meyer 

 for lending the Society his model. — The meeting then adjourned 

 until June 24. 



Edinburgh. 

 Mathematical Society, May 13.— Mr. J. B. Clark, Presi- 

 dent in the chair. — The following papers were read : — On the 

 second solutions of Lame's equation, by Mr. Lawrence Craw- 

 ford (communicated by Mr. J. W. Butters) ; on the insolation of 

 a sun of sensible magnitude, by Mr. A. Ritchie Scott ; the 

 singular solutions of a certain differential equation of the second 

 order, by Mr. Hugh Mitchell. 



Paris. 

 Academy of Sciences, June 6. — M. Wolf in the chair. — 

 New photographic studies of the surface of the moon, by MM. 

 Loewy and Puiseux. A discussion of the data contained in the 

 third part of the photographic atlas of the moon. — On a new 

 absolute electrodynamometer, by M. Marcel Deprez. In the 

 system described, the forces due to the action of the current 

 are simple algebraic functions, rigorously and without approxi- 

 mation, of the dimensions of the fixed and movable circuits. — 

 On a new constituent of the atmosphere, by MM. William 

 Ramsay and Morris W. Travers (see Nature, p. 127). M. 

 Berthelot observed that the green ray of krypton coincided 

 almost exactly with the bright green line of the aurora borealis. 

 He suggested the name eosiuin for the new element. — On the 

 propagation and deformation of the tidal wave which ascends 

 rivers, by M. Partiot. The curve of the experimental results 

 obtained on the Gironde and Garonne are compared with five 

 formulae ; of these, that suggested by M. Boussinesq agrees best 

 with the experiments. — On surfaces of total constant curvature, 

 by M. C. Guichard. — On the systems of differential equations 

 which satisfy the quadruply periodic functions of the second 

 species, by M. Martin Krause. — On discontinuous functions 

 which are allied to continuous functions, by M. R. Baire. — On 

 the determination of the order of interference fringes, by MM. 

 A. Perot and Ch. Fabry. — On the rotatory power of quartz in 

 the infra-red, by M. R. Dongier. A comparison of the ex- 

 perimental results with those calculated from a formula given by 

 M. Carvallo. — On the discharge of a Leyden jar, by M. R. 

 Swygedauw. — Comparison of the Hertzian field in air and in 

 oil, by M. Albert Turpain. In a resonator kept in a plane-per- 

 pendicular to the direction of the wires the wave-lengths vary 

 with the nature of the dielectric ; if the resonator is in the same 

 planes as the wires, the wave-lengths are independent of the nature 

 of the dielectric— On resonators, by M. Oudin. The resonator 

 now used consists of a solenoid of bare copper wire wound 

 round a cylinder of paraffined wood, the high frequency current 

 being produced by the arrangements of Hertz, of Tesla, or ot 



