AUernating Currents to Calibration of Capacity-boxes. 711 



thickuess of 2*5 cms. The lower part of each long limb was 

 surrounded by a wooden frame 12' 7 cms. long, on which 

 were wound 370 turns of No. 20 copper wire. A brass 

 cross-piece supported from the poles of the electromagnet 

 carried an insulated torsion-head, from which was suspended 

 the moving coil, which swung between the poles of the 

 electromagnet. 



We fnlly anticipated that the moA^able coil (especially if 

 wound on a copper frame) w^ould exhibit great axial stability, 

 and consequentl}^ very little sensitiveness. To overcome this 

 defect we originally tried supporting the coil by a crossed 

 bifilar so as to partially neutralize the axial stability by super- 

 posing gravitational instability, but the result was far from 

 successful. We then tried the effect of increasing the accuracy 

 of reading without attempting to reduce the axial stability by 

 attaching a pointer to the coil. Horizontal rigidity was 

 imparted to the pointer by making it of a light aluminium 

 wire 30 cms. long, bent in the middle at a very acute angle 

 so as to form a V. This was lightly latticed across the 

 middle to increase the rigidity. The two free ends of the V 

 were attached in a horizontal plane to the upper part of the 

 coil, and the w^eight suitably counterpoised. The acute end 

 of the V, to which w^as attached a suitable mark, could, be 

 examined by a microscope, magnifying 16 diameters, placed 

 with its axis horizontal, although we should have preferred 

 a vertical arrangement of axis if that had been equally easy 

 to attain. After considerable practice with this method of 

 reading the position of a movable coil, we much prefer it for 

 laboratory purposes to the ordinary mirror- and- scale arrange- 

 ment. For rough work the pointer can be viewed directly 

 in full daylight without the microscope, the latter only being- 

 used for the final adjustment of the balancing resistances. 

 However, the necessity for using this method of indication 

 has now disappeared, for we have by accident discovered a 

 method of adjusting the sensitiveness with the utmost ease 

 and convenience. If the moving coil is wrapped on a closed 

 copper frame of appropriate breadth, this frame will posse.ss 

 axial instability in the alternating field (see expt. 5 above), 

 whereas the movable coil, if short-circuited, would possess 

 immense axial stability (see expt. 7); so that if we shunt 

 the movable coil through a variable resistance, in all cases 

 many times its own, we can get any degree of stability 

 desired. We are thus in a position to employ the mirror 

 method of reading the deflexions if we so desire. We 

 are, however, ourselves convinced that the mirror method is 

 a mistake in many cases where the movable system is heavy 



