208 PROCEEDINGS OF THE AMERICAN ACADEMY. 



This paper gives the results of some tests made at low excitations, 

 ■which are interesting because of the great susceptibility which the 

 rings showed in fields less than two gausses, and compares the magnetic 

 behavior of this metal vdth that of a rin^ of the so-called " American 

 Ingot Iron," which I obtained through the kindness of Dr. P. W. 

 Bridgman. This well-known iron, which was made by the American 

 Rolling Mill Company of Middletown, Ohio, seems to be perfectly 

 homogeneous, and, according to the makers, contains less than 0.03 per 

 cent of impurities all told. 



All the rings were very accurately made by Mr. G. W. Thompson, 

 tbe mechanician of the Jefferson Laboratory. The external diameters 

 of the Norway iron rings were 5.000 cm. and 4.996 cm. respectively; 

 their thicknesses were 0.250 cm. and 0.254 cm., and their breadths 

 were 1.2204 cm. and 1.210 cm. The measurements were made with 

 the help of Zeiss Comparator No. 3196 and a set of auxiliary gauges. 

 After each ring had been measured, a coil of very fine double-silk-^ 

 covered copper wire was wound on the metal in a single layer and then 

 baked in shellac. Over this was wound, usually in two layers, the 

 exciting coil of well-insulated wire nearly one millimeter in diameter. 

 The ballistic galvanometers were of the moving coil type, and had 

 periods amply long enough ^ for the work. The fine coil on the ring 

 was always in simple circuit with the galvanometer and the second- 

 ary coil of a standard of self inductance tested by the Bureau of 

 Standards. 



The maximum value of the permeability (5480) which I obtained 

 for the first ring tested seemed so high that at first I suspected that 

 there was some error in the determination, so I changed the galvanom- 

 eter, and then took off the coils and wound on new ones with different 

 numbers of turns ; but when the result was unchanged and the second 

 ring gave values for the ordinates of the HB diagram which were prac- 

 tically indistinguishable from those obtained from the first ring, there 

 seemed to be no doubt that the work had been accurately done. The 

 two rings lay side by side in the original bar, and both must have had 

 nearly the same discontinuities. Table I, founded upon several hundred 

 separate determinations, gives values of the permeability of the metal 

 obtained from 35 different excitations of the first ring and 25 of the 

 second. A ripg of very pure annealed iron from the Armstrong Works 

 at Elswick gave in the hand^ of Wilson the same maximum value of 

 the permeability as the rings just mentioned ; but apart from the re- 

 ports of some tests upon thin pieces of electrolytically deposited iron, 



? Peirce, These Proceedings, 44, 1909 (283). 



