118 PROCEEDINGS OF THE AMERICAN ACADEMY. 



1873, and Stefan, 1400, in the following year. In 1881, however, 

 Fromme got the value 1737 for one specimen, and in 1884, Weber, 

 exposing a long rod in a solenoid to a field which had an intensity of 

 only 900 gausses before the iron was introduced, found the correspond- 

 ing value of / to be 1700. 



In 1887, Messrs. Ewing and Low introduced a new and most in- 

 genious method for experimenting upon slender isthmuses of iron and 

 steel under ^'ery high excitations and showed that different specimens 

 of soft iron often behaved very differently in very strong fields. For 

 one brand of fine Swedish iron, they found the final value of / to be 

 only 1620, while for a certain kind of Bessemer steel, the value I^ 

 was as high as 1770. 



Du Bois published in 1890 the results of a series of experiments 

 upon iron in very intense fields the strengths of which he had deter- 

 mined by optical means. In order to obtain, for each brand of mate- 

 rial, the Kerr's constant which he needed, he first examined an 

 ellipsoidal test piece of the material in much weaker fields, in a 

 solenoid. In a typical case, the soft iron ellipsoid of revolution was 

 18 centimeters long and 6 millimeters in diameter at the centre. 

 The solenoid was 30 centimeters long and consisted of 1080 turns of 

 insulated wire wound in twelve layers of about 4 centimeters mean 

 radius. The field intensity at a point 9 centimeters from the centre 

 of the solenoid was a})out 6% less than at the centre and this intro- 

 duced a correction into the formula for /. The moment acquired by 

 the bar when it was under excitation was determined, after the effect 

 of the current in the solenoid had been compensated for, by the indi- 

 cations of a magnetometer in Gauss's A Position with respect to the 

 specimen. The correction for the ends of the ellipsoid was made by 

 the use of the formula H' = H — 0.052 /. We shall find it con- 

 venient to refer to these results later on in this paper and some of 

 them appear in Tal)le I. 



TABLE I. 



H'. I. 



600 1680 



800 1695 



1000 1703 



1300 1712 



In the stronger fields the results were not so regular, for the speci- 

 men w^as magnetized between the poles of a powerful electromagnet 



