PEIRCE. — MAGNETIZATION IN IRON. 119 



and the fields were far from uniform. The final value of I which 

 du Bois obtained lay somewhere between 1700 and 1750. 



A paper by Roessler in the Elektrotechnische Zeitschrift for 1893 

 describes some experiments very like those made by du Bois with the 

 solenoid mentioned above. Roessler's solenoid was 1 meter long and 

 consisted of 16 layers of wire 3 millimeters in diameter. The mean 

 radius of the solenoid was about 5.5 centimeters and the field at a 

 point on the axis 25 centimeters from the centre Avas about 1% less 

 than at the centre itself. The test piece was an ellipsoid 50 centime- 

 ters long and 1 centimeter in diameter. The results which Roessler 

 obtained for a certain specimen of so called "soft iron" are given in 

 Table II. 



TABLE II. 



H'. I. 



414 1645 



481 1663 



516 1670 



587 1675 



741 1679 



777 1681 



The values of I^ published in 1896 l)y E. T. Jones, who magnetized 

 a short, slender wire of the material to be tested between the conical 

 pole pieces of an electromagnet of the du Bois form, ranged as high as 

 1818; and the results of the joint work of du Bois and Jones, printed 

 in 1899, gave values of I^ between 1780 and 1850. 



Weiss, in 1907 and 1909, experimented upon small ellipsoids of 

 revolution made of iron, nickel, and cobalt, placed symmetrically 

 between the flat pole pieces of a powerful electromagnet. Each 

 ellipsoid was about 9 millimeters long and 3.5 millimeters in diameter. 

 The gap between the pole pieces was about 6 centimeters long and the 

 diameter of the magnet core was 15 centimeters. An excitation of 

 94000 ampere turns corresponded to a field of about 9000 gausses in 

 the gap centre. The small ellipsoid was suddenly drawn out of the 

 field through a hole in the axis of one of the pole pieces and the flux 

 change in a test solenoid outside the iron was determined. Weiss's 

 values of I^ were 1731 and 1706. 



Gumlich in 1909 made a series of extremely accurate determina- 

 tions of the final value of / in soft iron by the Isthmus Method, using 

 an electromagnet of the du Bois form, which was furnished with two 

 soft pole pieces fastened together with the isthmus between them and 



