53 EEPORT — 1872. 



and r is tlie temperature, and consequently that in the case of ec^uilibiium of tem- 

 perature, /■. c. where v^r = 0, the known results, 



where 6 is the dilation at any point, p the density, and n the displacement parallel 

 to the axis of a; are still true ; in fact r appears as part of the potential of external 

 forces. 



In the case of a spherical shell, the interior of which is maintained at one con- 

 stant temperature and the exterior at another, it is found that the stresses are 

 independent of the thickness of the envelope, and that the g-reater liability of thick 

 vessels to break where the temperature is maintained different in the two surfaces 

 is due to the fact that the thickness of the vessel makes it possible to maintain a 

 greater difference between the surfaces, whilst the general temperatures of the 

 media within and without remain the same as for the thin vessel. In fact the 

 greater safety of a thin vessel lies in its greater conducting-power, and not in the 

 mechanical properties of its form. 



The principle of superposition may be applied by integration to the case of the 

 solidification of a fluid sphere in shells beginning at the outside, the effect of the 

 solidification of each infinitely thin shell being calculated, and the strains produced 

 by each added together as an integral. 



Electhicitt and Magnetism. 



On Double Neutral Points in Thermoelectric Currents. 

 By Prof. P. G. Tait, F.R.S.E. 



On the Use of Electromagnetic instead of Electrostatic Induction in Cable- 

 Si(/naUinff. By G. K. "Winter, F.R.A.S., Tehyraiih Enyineer, Madras 

 Railway. 



The experiments on this subject were made by the author in ignorance of the 

 contents of Mr. Varley's specification of 1802. The sending-apparatus being the 

 same as that now in use, the currents from the cable were made to pass through a 

 long, fine, primary wire of an induction-coil, and the induced currents in the 

 secondary wire were used for worldng the receiving-instrument. In long sub- 

 marine cables the receiving- instrument was a Thomson's galvanometer ; but the 

 siphon recorder might also be used. The signals obtained in this way were 

 steadier, and the elements of the letters more distinctly formed than with the 

 condenser or electrostatic method. On short submarine cables and land-lines, 

 on which Morse's instruments are used, this method, though requiring more battery- 

 power than that now in use, and necessitating the use of a polarized relay, would 

 almost entirely ])revent the delays caused by earth-currents during magnetic storms ; 

 and on long cables this method, while, equally with the condenser method, render- 

 ing earth-currents harmless as far as signalling is concerned, would, besides, cause 

 the cable to be only dynamically instead of statically charged by them, and the 

 danger of damage to the insulator to be at least halved thereby. 



Since the paper was read, the author has been informed by Mr. C. F. Varley 

 that the induction-coil was tried by him as early as 1861, on the Dunwich and Zan- 

 voort cable. In 1802 he tried it on the Dunwich and Zanvoort and the Lowestoft 

 and Zanvoort cables, in one circuit of about 1000 nauts ; also in 1865 upon the 

 Atlantic cable, on board the ' Great Eastern,' besides manj' times on his artificial 

 cable. 



