of Electrical Discharges. 257 



It is evident that the experiment is not conclusive, for the 

 arrangement described constitutes a divided circuit in which 

 one branch, consisting of the gas-pipe, presents much less resist- 

 ance than the interior copper wire. The greater part of the 

 discharge should accordingly pass along the gas-pipe, and the 

 magnetization of the interior needle would be so slight as to 

 escape detection unless a delicate method of examination were 

 employed. 



.The matter seemed to be of sufficient importance to justify 

 an experimental inquiry. The investigation has been more 

 extended than was anticipated, because it has met with greater 

 difficulties than were at first apparent; but the results obtained 

 are of considerable interest, and the conclusion of the whole 

 matter settles the main question. 



The following arrangement of apparatus was adopted. Two 

 glass tubes about 15 mm in diameter were covered with tin-foil 

 for a length of about one meter each. About the tin-foil at 

 each end was secured tightly a copper clamp. Two similar 

 magnetizing spirals were made by winding twenty turns of 

 gutta-percha-covered wire round small glass tubes. One of 

 the spirals served to connect the two tin-foil coatings in series 

 by having its terminals soldered to the copper clamps. The 

 other formed the middle of a circuit extending through the 

 inside of the tubes. The two conductors were then joined at 

 the remote ends of the glass tubes so as to form a divided 

 circuit, — one of the tin-foil and connecting helix, the other of 

 the wire passing through the inside of the tubes and containing 

 the other helix. The two branches were adjusted to very accurate 

 equality of resistance. Through this compound circuit a Leydeu 

 jar was discharged ; and the method finally employed to 

 measure the relative portions of the charge passing through 

 the two branches was to determine the magnetic moment im- 

 parted to two similar rods of steel, l-8 mm in diameter and 6 om 

 long, tempered glass-hard, one inserted in each spiral. The 

 Leyden jar of about 700 square centimeters surface was charged 

 by a Holtz machine with fifteen sparks from two to three 

 centimeters in length. The magnet in the inner branch of the 

 divided circuit will be called A ; the one in the outer, B„ 



What Hughes calls " the evident magnetism " of A was in 

 every case greater than that of B after a single discharge of the 

 jar. Upon sending repeated discharges through the circuit, the 

 ratio of the two magnetic moments gradually diminished and 

 the moments themselves approached a maximum, as is shown 

 in fig. 1, where the ordinates denote magnetic moments and 

 the abscissas discharges of the Leyden jar. The moments 

 in arbitrary units are simply fractional parts of the deflection 



Am. Jour. Sci.— Third Series, Vol. XXXT, No. 184, April, 1886. 

 17 



