Transmission of Electric Waves along Wires. 379 



incident waves was determined by Bjerknes's method to be 

 8*5 m. 



For the interference phenomena, observed in order to 

 detect the phase change on reflexion, the distance ED was 

 varied by half a metre from 0*75 m. to about 8'0 m. so as 

 to furnish nearly two wave-lengths of the interference 

 curve. 



The methods of taking the readings and their results will 

 be sufficiently understood from the following tables and 

 accompanying curves. The three condensers used were 

 chosen by reference to the curve in fig. 5 as likely to yield 

 typical results. 



57. Large Condenser. — Disks 15 centim. radius and placed 

 1 centim. apart. 



(i.) Transmitted System. On taking seven readings with- 

 out condenser on alternated with six readings with condenser 

 on, the values were respectively of the order 



d = 26'8 and d' = 2-6 ~] 

 whence | . a . 



y=^- =*097 nearly J 



With such small readings as those here obtained for d' it 

 cannot be professed that they are free from errors due to 

 small variations of the zero, which is never quite steady and 

 is always read before each throw is taken throughout all the 

 experimental work. 



(ii) Reflected System. Two sets of 41 readings each were 

 taken for this arrangement. The first gave 



meantf 6 

 meany= — -= = 1*862 (47) 



The second, in which the behaviour of sparks and electro- 

 meter was more regular, gave 



mean ys= ^^ = 1-986 (48) 



This second set of readings is given in Table II. (p. 380). 



(iii.) Interference System. Four sets of readings were 

 taken with this arrangement of the apparatus. The last of 

 these is given in Table III., and the points corresponding to 

 all of them are plotted in fig. 6, a curve is then drawn through 

 the mean position to represent as fairly as may be the result 

 of all the readings. 



