OF ARTS AND SCIENCES. 327 



Btrong alternating current, had I not been informed by Professor Trow- 

 bridge that experiments at Newport had ah-eady placed the matter 

 be3'ond all doubt. 



It had been supposed that only a powerful battery (ten or more 

 Bunsen cells) wouKl be able to give distinctly measurable results for 

 the remote portions of the bar, whereas it now appeared that the con- 

 ditions of success in these experiments depend upon the insufficiency 

 of each current to produce a condition of magnetic saturation within 

 the limited time of its action. The use of a single cell obviated all 

 these difficulties, and was found to give sufficiently precise results for 

 all portions of the bar. 



In order to determine the several positions of the sliding-piece 

 which corresponded to the simultaneous action of the two commuta- 

 tors, the shaft was set revolving very slowly, and the contact-piece 

 was moved until the spot of light thrown by the galvanometer, though 

 feeling every pulse, show^ed no constant deflection to right or left. A 

 fixed pointer then covered, in each case, a certain division of a scale 

 attached to the sliding-piece, the number of which being noted served 

 as a zero for the calculation of angular measurements. The branches 

 of the commutator making not quite one and one half turns, there 

 were at most three points which satisfied this condition, one of which 

 was taken as 0°, the next as 180°, and the third as 360°. Interme- 

 diate inclinations of the two commutators were measured by simple 

 interpolation. Owing to the wearing away of the centres which sup- 

 ported the shaft, and their occasional readjustment, these three points 

 had frequently to be redetermined. 



The result of a long series of experiments * upon a half-inch rod, 

 some 53 inches long, was to show that the phase, in the portions of 

 the bar not very remote from the primary, was invariably later than 

 within the primary itself ; and that this retardation increased up to a 

 certain distance (not under 10 nor over 20 inches), to a maximum ; 

 and then diminished again, so that the furthest end of the bar did not 

 differ in phase from the primary. 



The phase of the primary was itself retarded more and more as 

 the speed of the commutator increased, until for the highest velocities 

 (100 or more reversals per second) it became fully equal to 90°, 

 while the maximum retardation found anywhere in the bar never 

 exceeded 127° or 130°; and was, accordhigly, never more than 37** 

 or 40° later than the primary. 



* See Table for Phase Retardation, appended. 



