476 



MEMOIBI OP THE AMERICAN ACADEMY 



Gaugain despairs of realizing this essential condition upon any long telegraphic line, 

 and be therefore concludes that it is difficult, if not impossible, to verify these laws on 

 such a scale of magnitude. Guillemin has made many experiments upon the tele- 

 graphic lines in Fin nee. In Gaugain's experiments upon short, imperfect conductors, 

 it appeared that the rate of propagation was no greater with strong batteries than 

 with weak ones ; and this has been the general verdict in experiments upon air and 

 cable lines of telegraph. But Guillemin found that when the number of elements in 

 his battery was doubled, the transmission time was diminished about ten per cent. 

 To ascertain the rate of transmission, he employed lines varying in length from 280 

 kilometres ( 174 miles) to 1,004 kilometres (624 miles). The rate of transmission was 

 much greater than in the simple proportion of the length, but decidedly less than in 

 proportion to the square of the length. With 60 elements of Bunsen the transmis- 

 sion time over 354 miles of iron wire 4 mm (or .157 of an inch) in diameter was only 

 .020 of one second. Therefore the electrical disturbance travelled over this particular 

 distance at the rate of 17,710 miles in one second. 1 It should be understood that 

 generally the transmission time has been ascertained by the arrival of the current in 

 sufficient force to affect a galvanometer ; but in Gaugain's experiments the time re- 

 quired for the conductor to acquire a stated tension was determined by means of a 

 delicate gold-leaf electroscope. The resistance per knot of the Anglo-Atlantic cable 

 is less than that of the cable between Duxbury and St. Pierre, and very much less 

 than that of the cable between St. Pierre and Brest, partly on account of difference 

 of dimensions and partly because of differences of pressure and temperature where 

 they are laid. Nevertheless, the total insulation of the Anglo-Atlantic cable is 

 about 1,316,000 ohms, and that of the whole length between Brest and Duxbury 

 1,300,000. I infer, therefore, that little or no allowance need be made, in consequence 

 of any large difference of leakage, in the comparison which has already been presented 

 of the transmission times over these two long cable-lines. 



We may hope that the longitude campaigns of the United States Coast Survey may 

 have been useful, not only in determining differences of longitude, but also in throw- 

 ing light on the delicate problem of the transmission of electrical disturbances. As 

 we have pursued this discussion we have felt our obligation to Professor Winlock of 

 the Harvard College Observatory, and to Mr. Dean and Mr. Goodfellow, and their 

 assistants in the United States Coast Survey service, for the intelligence and energy 

 with which the observations were planned and executed. But the best observations, 

 however numerous, will not bring out a satisfactory result, unless they are skilfully 



Chim 



t 



4 



