350 



Messrs. G. F. C. Searle and T. G. Bedford. 



This method of determining X + Y has lately been used successfully 

 at the Cavendish Laboratory by Mr. R L. Wills in the case of 

 specimens of large section. In the authors' experiments X was 

 generally negligible. 



As the corrections X and Y depend upon dC/dt it is necessary that 

 the primary current should change only gradually. By inserting a 

 choking coil of great self-induction in the primary circuit, and by 

 using a special key to cause the reversal of the current, this end is 

 satisfactorily attained. 



The authors have made many comparisons between the values of 

 W found by their method and those calculated from the areas of 

 cyclic B-H curves obtained by a ballistic galvanometer, and have 

 found satisfactory agreement. 



§ 3. By using a ballistic galvanometer in addition to the dynamo- 

 meter, the two authors were able to make simultaneous observations of 

 the range of the magnetic induction ± B and of the energy dissipated 

 in each cycle. The range of the magnetic force ± H was also 

 observed. 



It was found that the cyclic B-H curve is not always divided into 

 two parts of equal area by the line H = 0. The effect is well 

 marked in the case of an iron wire freshly annealed, and sometimes 

 does not disappear in spite of many reversals. 



When the magnetic force is reversed many times both B and W 

 decrease. The effect is most apparent in soft iron freshly annealed, 

 and subjected to a small magnetic force. Thus when the limits of H 

 were ± 2 -5, in the first cycle after the annealing, B = 2220 and W 

 = 598. In the forty-first cycle B - 1840, W - 433. 



§ 4. When an iron wire is stretched by a variable load, and is put 

 through cycles with the limits ± H , the first application of the 

 tension results in an increase in both B and W. As the tension 

 increases, B and W reach maxima and then decrease. The effect is 

 more marked when H is small than when it is large. Thus with a 

 wire of section 0-00708 cm. 2 a load of 16 kilos, raised B from 1233 to 

 5870 and W from 494 to 3820, with H = 4-524. 



A series of experiments was made upon the effects of torsion. 

 When H is kept constant, as the torsion increases there is a large 

 decrease in both B and W. Thus in the case of a soft iron wire when 

 H = 3'0, by torsion within the elastic limit B was brought down 

 from 2280 to 1070 and W from 907 to 276. Further, both B and 

 W exhibit hysteresis with respect to the torsion. 



Experiments were also made in which the torsion was gradually 

 increased till the wire broke. In other experiments the authors 

 studied the influence of permanent torsional set upon the effects of 

 cycles of torsion. They also examined the development of a cyclic 

 state, for cycles of torsion, after initial permanent torsional set. 



