230 PROCEEDINGS OP THE AMERICAN ACADEMY. 



boxes L and N. These were 10 cm. square and 4 cm. thick, and 

 mounted upon the wooden bar E by insulating supports. Within the 

 boxes were wooden bobbins fixed to a hard rubber axle, and each 

 capable of holding 10 m. of the largest wire experimented upon. In 

 the front of each box was a small opening for the passage of the 

 wire, but, to assure a firmer contact between the wires and the boxes, 

 a brass block was soldered on the inner side of the front and a bind- 

 ing screw passed in from the outer side of the box. The bar E was 

 fastened to a wooden support resting upon the car, which ran on a 

 wooden track extending the entire length of the room. The car car- 

 ried a brake, so that the wires could be drawn taut, and the wooden 

 screw held the axle from turning. With this arrangement the length 

 of the wires could be varied at will, while the end capacities would 

 remain constant. The end capacities are not a feature desirable for 

 their own sake, as they destroy the perfect simplicity of the plain rec- 

 tangle and seem to detract somewhat from the sharpness of the maxi- 

 mum ; but the gain in convenience, and the possibility of obtaining a 

 large number of observations whose average values can be used, quite 

 overbalance these considerations in most cases where the apparatus 

 may be applied. 



In the early part of the investigation the " Foucault " mterruptor 

 was used, but it was extremely irregular in its action, and caused 

 endless annoyance. It ran at an ever varying rate, and required 

 repeated adjustment and constant attention. To remedy at least 

 some of these defects, an interrupter actuated by a small electric motor 

 was devised.* The results obtained from this motor-interruptor were 

 so satisfactory that a detailed description is added. 



A Porter's motor. No. 1 (Fig. 6, M), was used to produce the mo- 

 tion. This was actuated by the current from two storage cells, and it 

 ran at a fairly constant speed. The armature of the motor was wound 

 in three sections, and was thus free from dead points, giving it the 

 great advantage for the present purpose that it could be set in motion 

 simply by closing the circuit, making it possible to control it from the 

 observer's station. The motor was geared to the two-crank shaft K 

 by means of a wheel and pinion. The wheel and pinion liad the 

 ratio of 144 to 24 in the following investigation, but the motor could 

 slide on the brass bed-plate so that pinions of other sizes were avail- 

 able. The speed of the shaft K was about 750 revolutions per min- 



*This and the other apparatus especially prepared for this investigation, and 

 requiring nmch technical skill in its construction, was made by the mechanician 

 of the laboratory. 



