260 H. S. Carhart — Surface Transmission 



pacity of the principal conductor and of the coefficient of self- 

 induction. This inference, so far as it relates to the capacity 

 of the principal conductor, was confirmed by some experiments 

 of Helmholtz in 1869.* 



The bearing of the duration of an oscillation on the reversal 

 of the superficial magnetism is that the mean intensity of the 

 current during the first period is inversely proportional to the 

 length of the period, since the whole quantit}^ of electricity on 

 the principal conductor is discharged during this short interval. 

 Then the magnetic moments, being nearly proportional to the 

 current strength, so long as the saturation point is not ap- 

 proached, will therefore vary inversely as the periods of an 

 oscillation, or inversely as the square root of the coefficients of 

 self-induction, in the two branches of the discharger, even 

 though the quantity discharged through the two branches is 

 the same. The magnetic impulse of both the direct and re- 

 verse oscillatory discharge is greater the smaller the coefficient 

 of self-induction in the discharger. 



In the experiments thus far described no care was taken to' 

 make the coefficient of self-induction in the two branches equal 

 to each other. The inner branch was of copper wire longer 

 than the glass tubes, and coiled into a long helix with the con- 

 secutive turns far enough apart to prevent a spark from leaping 

 across. This disposition of the wire increased largely its self- 

 induction. The time of an oscillation through it was therefore 

 greater than in the outer branch, and the magnetic effect of the 

 weaker current less, while the whole quantity discharged 

 through the two branches was the same. Not only then was 

 the mean magnetic moment of B greater than that of A, but 

 the effect of the return discharge was. sufficient to produce a 

 reversal of magnetism in B near the surface. 



These conclusions are fully confirmed by experiments made 

 with a new arrangement of apparatus, so devised that the 

 coefficients of self-induction in the two branches should be as 

 nearly equal as possible ; while at the same time their capaci- 

 ties were made more nearly equal than before, in order to elim- 

 inate any possible effect on the period of an oscillation arising 

 from dissimilarity in this respect. 



Twof large tubes, 22 mm in diameter, were covered with tin- 

 foil for a combined length of 265 cm . Two other tubes 15 mm in 

 diameter, were also covered for a combined length of 296 cm , 

 and the smaller tubes were placed inside the larger ones. One 

 of the magnetizing spirals was made to connect the outer tubes 

 in series, and the other, the inner ones, the other ends of the 

 tin-foil conductors being joined by means of copper clamps and 



* Wissensch-iftliehe Abhandlungen, " Ueber Elektrische Oscillationen," p. 531. 

 f Tbese experiments were made after the reading of trie paper at Ann Arbor. 



