416 



Mr. F. Jenkin on the Retardation of 



Table T. 



Times, in ten-thousandths 1 

 of a second J 



17 

 0-5 

 1-9 



1-6 



10-6 



50 

 2-5 



94 



22 

 23 



85 

 11-5 



41 



41 

 21 



127 

 18-5 



70 



6-5 

 19-5 



146 

 22 



83 



8-5 

 17 



175 

 24-2 

 91 



10-6 

 16-5 



26-5 



Deviations, in degrees 



Percentage of maximum 1 

 deviation J 



Times expressed in func-1 

 tion of a according to > 

 theoretical curve J 



Value of a obtained by di-"j 

 viding the numbers in 1 

 the first line by those in [ 





The meaning of the figures in the Table will be made clearer 

 by an example, thus : 127 ten-thousandths of a second after the 

 signal had been sent, the received current caused a deflection 

 of 18°'5 in the galvanometer, showing that the current had 

 reached 70 per cent, of its maximum strength. The fourth 

 line expresses the fact that, according to Professor Thomson's 

 theory, this fraction of the maximum strength corresponds to a 

 period of 6*5 a ; and the last line gives the corresponding value 

 of a (i. e. the quotient of 127 by 6*5). 



If M. Guillemin's results were in exact accordance with 

 theory, the value of a would be constant from whatever part of 

 the observed curve it might be calculated. We find, however, 

 that the value is least of all at first, rapidly increases, and again 

 slightly decreases. The small value of a obtained from the 

 shortest period of time is possibly due to the want of propor- 

 tionality between the galvanometer-deflections and the observed 

 current. Still more probably it is due to the difficulty of 

 making an exact observation of so small a deflection as half a 

 degree. We may therefore reject the first result from our cal- 

 culations. Between one or two degrees and twenty, there is 

 often a pretty good agreement between the deflections shown by 

 the common detector and the current passing through it. We 

 find, therefore, as was to be expected, that in passing from 10 

 per cent, to 70 per cent, of the maximum current, the value of a 

 varies little. Between 70 and 90 per cent, the values of a again 

 diminish, showing a more rapid increase than would be due to 

 theory. We may therefore say that the observed curve is some- 

 what flatter than the calculated one, resembling in this respect 

 the curves obtained by the writer in experiments on a submarine 

 cable, and due possibly to a similar cause, viz. varying insulation 

 caused on the land lines by the polarization of the moist film on 

 the insulators — and possibly to electromagnetic induction produ- 

 cing currents in the neighbouring wires. 



