48 REPORT — 1871. 



resistance ; but the measurement should be made by the use of a eondeuser, or, 

 infinitely better, by my " Potentiometer," which, -n-ith a Thomson's reflecting 

 galvanometer, readily measures to the millionth part of a Daniell's cell, or very 

 much less if required. 



Notice of and Ohservations with a New Dip-circle. 

 By J. P. JoTTLE, LL.D., F.R.S., ^r. 



The method of suspension of the needle, which formed the principal feature of 

 the new instrument, was explained. The increased facilities of observation had 

 enabled the author to ti-ace the diurnal variation of inclination with greater 

 accuracy than he believed had hitherto been done. At Manchester, about the 

 summer solstice, the greatest inclination was found to occur at 21>> 40™ local time, 

 and the range extended to 5'. The simultaneous variation of horizontal intensity 

 was such as to indicate that the total intensity was very nearly a constant quantity. 



On Tlif.rmo-eJectriclty. By Professor Tait. 



It results from Thomson's investigations, founded on the beautiful discoveries of 

 Peltier and Gumming, that the graphic representation of the electromotive force of 

 a thermo-electric circuit, in tenns of temperatures as abscissae, is a cm've .symme- 

 trical about a vertical axis. This I have found to be, within the limits of experi- 

 mental error, a parabola in each one of a very extensive series of investigations 

 which I have made with wires of every metal I could procure. To veiify this 

 result with gTeat exactness, and at the same time to extend the trial to temperatures 

 beyond the range of a mercurial thermometer, I made a graphic representation, in 

 which the abscissaj were the successive indications of one circuit, the ordinates 

 those of another, the temperatures being the same in both. It is easy to see that 

 if the separate circuits give parabolas (as above ) in terms of temperatm'e, this pro- 

 cess also should lead to a parabola, the axis, however, being no longer vertical. 

 This severe test was well borne, even to temperatm"es approaching a didl red heat. 

 Unfortimately, it is difficult to procure wires of the more infusible metals, with the 

 exception of platinum and palladium, so that I have not yet been able to push this 

 test to very high temperatures. I hope, however, with the kind assistance of 

 M. H. Sainte-Claire DeviUe, to have wares of nickel and cobalt, with which to test 

 the parabolic law through a very wide range. 



Parabolas being similar figures, it is easy to adjust the resistances in any two 

 circuits so as to make their parabolas (in terms of temperature) equal. When this 

 is done, if the neutral points be different, it is obvious that by making them act in 

 opposite directions on a difterential galvanometer we shall have deflections directly 

 proportional to the temperature-difierences of the junctions. 



It is a curious residt of this investigation, that, supposing the parabolic law to be 

 true, the Peltier effect is also expressed by a parabolic function of temperature, 

 vanishing at absolute zero. 



I was led to this inquiiy by a hypothetical application of the Dissipation of 

 Energy to what Thomson calls the electric convection of heat, and my result is 

 verified (within the range of my experiments), that the specific heat of electricity 

 is directly proportional to the absolute temperature. It is scarcely necessary to 

 point out that the above results appear to promise a very simple solution of' the 

 problem of measuring high temperatm-es, such as those of fiu-naces, the melting- 

 points of rocks, &c. 



071 a Method of Testing Submerged Electric Cables. By C. F. Yaeley. 



On a Neiv Key for the Morse Printing Telegraph. By Ch. Y. Zenger, Pro- 

 fessor of Natural Philosophy at the Polytechnic School in Prague. 



I had devised in 1868 a new automatic key to work the Morse telegraph. 

 It produced three marks, viz. a point, a short line, and a long line. It con- 



