OF ARTS AND SCIKNCES. 323 



A long half-inch bar was magnetized at one end by a (primary) 

 coil connected with a battery. A small sliding secondary, of about 

 15 turns, was clamped at a short distance along the bar, and some- 

 what further off was placed a larger one, of about 1 ')0 turns. 



The two coils were connected (in series) with a galvanometer so 

 as to give opposed induction effects, and their relative distance was 

 varied so that these effects should neutralize each other as nearly as 

 possible. 



It was found, as expected, that when the latter adjustment was 

 made as accurately as possible, the spot of light thrown by the mir- 

 ror of the galvanometer still made invariably a sudden motion, as if 

 under the influence of the nearer of the two coils, followed by a 

 sudden stop, due supposably to the somewhat tardier action of the 

 more distant. 



The interval of time between the two effects was estimated by vai;i- 

 ous methods sufficiently obvious ; the most successful being the actual 

 determination of how rapidly the circuit must be closed and broken in 

 order to reproduce, with a single secondary, the same effects as before. 

 The various distances between the two secondaries, divided by the 

 corresponding intervals of time, gave, of course, certain values of the 

 velocity of propagation. These all lay within a few feet of 30 per 

 second, for distances of not over one foot, beginning at 2 inches 

 from the primary. How nearly these earlier estimates agree witli the 

 results of later investigation may be found below. 



The method of measm-ement, however, is open to serious criticism, 

 since the first sudden motion of the spot of light might be due to 

 the small counter-current which ordinarily precedes the arrival of the 

 normal current, and the magnitude of the counter-current would 

 doubtless be greater in the larger of the two coils ; so that the first 

 sudden motion might really be due to the action of the more distant, 

 rather than to that of the nearer coil. The experiments were, there- 

 fore, discontinued. 



It was not until the autumn of 1880 that the experiments were 

 again resumed, no new method less open to criticism having sug- 

 gested itself in the mean time. I should, however, mention that cer- 

 tain experiments of Professor Trowbridge at Newport (which will 

 doubtless be published in due time) had added both a confirmation of 

 previous results and a renewed interest in the continuation of tlie 

 investigation. 



For the sake of continuity, the earlier experiments were repeated ; 

 then varied by the substitution of one secondary and two primaries 



