Nov. 30, 1876] 



NATURE 



J 03 



the electrification along the cable may he represented bv 

 a harmonic curve, with single maximum in the middle 

 and zero at each end. Now let both ends 

 be connected to earth. Then the time that 

 the harmonic electrification takes to subside 

 to three-fourths of its initial value is de- 

 noted hy a. Curve I. coincides so nearly 

 with the Ime, o X, at first, as to indicate 

 that there is no sensible current until the 

 interval of time corresponding to a has 

 elapsed ; although, strictly speaking, the 

 commencement of effect at the remote end 

 is instantaneous. After the interval, a, the 

 current rapidly increases in strength. When 

 an interval of time equal to 5a has elapsed 

 from the first application of the battery, the 

 current will be about half its ultimate 

 strength. After \oa, the current will have 

 attained to nearly its full strength, that 

 the further increase will be scarcely sensible. 

 Theoretically the full strength is not reached 

 until an infinite time has elapsed. 



Fig. 4, Curve II., shows the effect at the 

 remote end of applying a battery during a 

 time equal to \a, and then putting the cable 

 to earth. It will be observed that a current 

 gradually diminishing in strength continues 

 to flow out of the cable for a considerable 

 time after the battery has been disconnected. 

 It is this after-effect which interferes so seri- 

 ously with the working of submarine cables. 



The Automatic Curb Sender was designed by Sir 

 William Thomson and Prof. Fleeming Jenkin for the 

 purpose of diminishing the effects 

 of the retardation due to the elec- 

 trostatic capacity of submarine 

 cables. This was accomplished by 

 making each signal consist of two 

 currents, the second being of oppo- 

 site name to the first and of a 

 shorter duration. The latter, or 

 curbing current, hastens the neu- 

 tralisation of the cable after the 

 first current has passed, and tends 

 to do away with the effecis of the 

 first current. For example, let one 

 «nd of the cable be connected to 

 one pole ot a battery for an interval 

 of time equal to 4^:, and then let 

 it be connected to the opposite 

 pole for a time 3«. The effect at 

 the remote end of this latter current, 

 if it had occurred alone, would be 

 represented by Curve III. of Fig. 

 4. The joint effect of the two cur- 

 rents is represented by Curve IV., 

 whose ordi nates are the algebraic 

 sum of the ordinates of Curves II. 

 and III. The quick return of 

 Curve IV. to the zero line OX, as 

 compared with Curve II., shows the 

 advantage of curb sending. 



The curve traced out by the 

 point of the siphon represents the 

 curve of arrival, and the effect of 

 curb sending is to give a sharp outline to the signals and 

 to bring the point of the siphon back 01 nearly back to 



the zero line between each signal. It is , 



necessary for the success of curb sending ? o o o o 

 that the spaces for the signal and curb I o o o o o 

 currents should be perfectly correct, and this ) ' o 



can only be tffected by means of automatic ' 



mechanism. 



Fig. 5 is an engraving from a photograph'of the auto- 

 matic sender. The message to be sent is punched on a 



strip of paper in right and left hale^ which correspond to 

 the dot and dash of the Morse alphabet, and a central 



Fig. 4. 



row of holes is punched for the purpose of carrying the 

 paper through the machine. The paper is then put into 



the machine and drawn along at a uniform speed by 

 means of clock-work. The paper passes through the 



ooooooooo 



O 

 o o o o 

 O 



0090a 



Fig. 6. 



tube u and underneath two prickers placed so as to cor- 

 respond to the right and left holes in the paper. When 



