Electric Waves passing through Telegraph Lines. 331 



periments about 300,000 megohms. The capacity of the accu- 

 mulator introduces an error, because it requires to be supplied 

 with as much electricity as a mile of submarine cable, or as 50 

 to 100 miles of land line. The effect which it has on the form 

 of the wave is therefore the effect of an error of position of less 

 than 1 mile in the cable. Now in a length of 200 or 300 miles 

 this effect is of very little practical importance and may be 

 neglected ; but if we have to measure contours of curves or 

 speeds of waves in short cables or land lines, the accumulator 

 employed should have a proportionally small capacity. 



When two batteries are used (that is, the waves are observed 

 by reversed currents), the second battery should be introduced 

 at the same instant that the first battery-circuit is interrupted, 

 not to allow time for the charged state of the cable to fall. 

 This was first done with two batteries of ten cells and five cells 

 respectively connected in opposite directions, so that when the 

 current passed through them all, the resultant electromotive 

 force was 10 — 5 = 5 cells in one direction. The reversal was 

 made by simply shortcircuiting the ten cells and allowing the 

 five others to operate alone in the reverse direction. This plan 

 was not without objection; and an arrangement was therefore 

 made by which the interval between the removal of one bat- 

 tery and the introduction of the other was reduced to a mini- 

 mum by means of a spring following the contact, and arrested 

 just before the reversal, by which the interval was made less 

 than 0-0005 second. 



The time-contact apparatus may be used so as approximately 

 to imitate any two successive currents given in the ordinary 

 ways of telegraphing ; and the speeds and contours of the waves 

 under such conditions can then be ascertained. In these ex- 

 periments, as in actual telegraphing, we can leave out of con- 

 sideration the absorption of electricity into the dielectric ma- 

 terial. Practically we have to deal only with what may be 

 termed the "free charge" of the cable, or that which is free 

 to be quickly discharged. The electricity which takes time 

 to creep into and out of the material does not influence per- 

 ceptibly the rate of signalling or the contour of waves. 



In conclusion, I must thank Mr. Willoughby Smith for 

 affording me opportunities of practically illustrating this 

 method, and Mr. May for his kindness and assistance in the 

 matter. 



25 Cumberland Terrace, Kegent's Park, 

 September 12, 187G. 



