1906.] on Hoiv to Improve TeUphomj. 211 



the relatively high self-induction of these instruments has prevented 

 their general application. Practically all the instruments which are 

 at present being applied to the measure of high-frequency currents are 

 thermal instruments, that is to say, they depend for their action on 

 the heating produced by the current when it flows through a suitable 

 small high-resistance conductor. These instruments may be broadly 

 divided into three classes, according to whether the rise in tempera- 

 ture of the conductor and consequently the current is measured by 

 (1) the expansion of the conductor ; (2) the change in its resistance ; 

 (3) the E.M.F. of a thermocouple either forming part of or near to 

 the heated conductor.* 



The first method, viz., the use of the expansion of the conductor 

 as a measure of the current, has not up to the present lent itself to 

 the production of very sensitive instruments. The second and third 

 methods above have each given instruments of high sensibility such 

 as the " barretter " employed by Dr. Kennelly, t and the thermo- 

 galvanometer. | From the point of view of ultimate sensibility there 

 is very little choice between these two instruments, but the simplicity 

 and ease of standardisation of the thermo-galvanometer make it the 

 more convenient in practice. 



Some very interesting results have been obtained by Dr. H. V. 

 Hayes § on the attenuation of the current through cables and long 

 overhead lines, and on the improvement that can be obtained by 

 adding self-induction to the line. These experimental results amply 

 bear out the theoretical conclusions of Heaviside as to the great 

 advantage of increasing the self-induction, or " loading " the line for 

 long-distance transmission. The great importance of avoiding re- 

 flection of the current at the terminal apparatus, and the means of 

 reducing it by the use of a " terminal taper " is also very clearly 

 shown. It is greatly to be hoped that these investigations will be 

 actively pursued, and a satisfactory design of loading coil will be 

 developed, as increasing the self-induction of the circuit gives great 

 promise of successfully increasing the length (now limited to about 

 50 miles) of subterranean or submarine cable through which telephony 

 can be commercially accomplished. 



So far the methods of measurement dealt with only give the root- 

 mean-squared or heating value of the current. To investigate the 

 distortion in the sound pattern when translated into a varying electric 

 current as it flows along the line and through the different pieces of 

 apparatus, we require to be able to record the current at every instant 

 and also at two or more points in the circuit. This can be easily 

 accomplished by means of an oscillograph, and the sound patterns 

 given at the commencement of this discourse were thus recorded. 



* See further, L'Electricien, xxxi,, p. 145, 1906. 



t Internat. Else. Congress, St. Louis, 1904. 



j Internat. Elec. Congress, St. Louis, 1904. 



$ Proc. Phys. Soc, xix., p. 91 ; and also Phil. Mag., 1904, 



