LAWS AND WARREN. — HYSTERESIS. 491 



1892, a descriptioD of a "Differential Wattmeter." This instrument 

 gives the difference of the mean values of two definite integrals repre- 

 senting the primary and secondary energy in a transformer ; that is, 

 the losses. The Watt-balance measures the ratio of the mean values 

 of two definite integrals. It can therefore be used to measure the 

 efiiciency of a convertor, giving the result by a single reading. 



The Watt-balance consists of two electrodynamometer Wattmeters, 

 one above the other. The movable coils are rigidly connected in our 

 instrument by a spiral-wound paper tube, and consist of 100 turns 

 of silk-covered German silver wire No. 32 B. & S. gage. The 

 resistance, including leads, is 103 ohms; the diameter of coils, 2 cm. 

 In series with the movable coils are two adjustable resistances, r^ and 

 ^2, as shown in Figure 2. The stationary coils are arranged in pairs ; 

 each is 2.3 cm. in mean radius, and contains 130 turns of No. 12 

 cotton-covered wire. They are movable along their axes, so that the 

 factor of the instrument may be adjusted. The vertical distance 

 between the two dynamometers is 30 cm. The movable parts are 

 suspended by a silk thread inside a narrow paper tube to prevent 

 disturbance by air currents, and directive force is given by four spiral 

 springs, made of copper wire 0.005 cm. in diameter. These springs 

 serve as leading in wires for the movable coils. The instrument is 

 read by a telescope and scale. 



In Figure 2 is shown the scheme of connections for a hysteresis test. 

 At the left the slide resistances for controlling the currents are shown. 

 By the proper insertion of the connection, G. either a direct or an 

 alternating current could be sent through the remainder of the appar- 

 atus, which in the main circuit consisted of an electrodynamometer, 

 one pair of stationary coils, the specimen under test, a known resist- 

 ance, R^ of German silver strip immersed in kerosene oil to keep it 

 at a constant temperature, and the second pair of fixed coils. At one 

 terminal of ^ a conn,ection was made by means of a short piece of 

 stout wire and a key, K; beyond the key the circuit branched, pass- 

 ing through the adjustable resistances r-^ and r,, the proper movable 

 coils, and returning one lead to the external terminal of the specimen, 

 and the other to the corresponding terminal of R. The resistance r^ 

 was some hundreds of ohms, seldom falling below 300, and was more 

 frequently in the neighborhood of 1 ,000 ; rg was kept at a constant 

 value, 976 ohms. The resistances r^ and r^ were ordinary plug boxes, 

 and the assumption was made throughout the work that the resistances 

 of circuits r^ and r^, were so high in comparison with their inductances 

 that no correction factor need be applied to the indications of the 



