276 Peii'ce — Permeabilities and Reluctivities for Steel. 



long solenoid, the permeability of a soft, homogeneous rod four 

 or five hundred diameters long, for excitations up to H = 400. 

 Here the value of I is probably between 95 per cent and 

 97 per cent of the final value ( 1^ ), which can be found by the 

 use of the Isthmus Method. If, then, any other method . of 

 measuring permeability be used on the specimen between 

 H — 300 and i/ = 2000 and if this method yields the proper 

 values of I at both ends of the interval, it is comparatively easy 

 to judge, from a graphical representation of all the observa- 

 tions, whether the short interval corresponding to from 3 

 to 5 per cent of 1^ is properly bridged. The principal 

 difficulty with this procedure is that several isthmus specimens 

 of a metal and. several testing coils must be used before one can 

 be satisfied that the resulting value of 7^ is correct to within 

 1 per cent ; for, as Ewing has shown, the separate results of a 

 series of determinations of 7^ by the Isthmus Method may 

 differ from the mean on either side by as much as 4 per cent. 

 As a matter of fact, in all the materials I have tested the final 

 value of /obtained by the Isthmus Method does not differ by so 

 much as 1 per cent from the final value as obtained by the other 

 method I have used. This second method, however, gives in 

 any case a series of determinations of 1^ which do not as a 

 rule range over so much as 1/3 per cent of the mean, while 

 the Isthmus Method in my hands is much less satisfactory in 

 this respect. 



The Bessemer and the compressed steel were procured in 

 specially long pieces from which lengths of about 450 centi- 

 meters were cut, and these, under the severest test which I 

 could conveniently apply, seemed to be practically uniform 

 throughout. The Norway iron, on the contrary, was not every- 

 where of the same temper and could not be used satisfactorily 

 until it had been carefully annealed by Mr. George W. Thomp- 

 son, the mechanician of the Jefferson Laboratory, who has 

 had much experience in work of this kind. Upon each of these 

 rods at its center a test coil of fine wire was w T ound by Mr. 

 John Coulson, who has made all the test coils and has helped 

 me in every part of the work, and then the coil and its leads 

 were carefully covered by pieces of rubber and rubber tape to 

 make the whole waterproof. The rod thus prepared was 

 placed inside a horizontal solenoid nearly 500 centimeters long, 

 placed perpendicular to the meridian. This solenoid was 

 uniformly wound with 20,904 turns of well insulated wire and 

 a stream of tap water could be kept running through the bore 

 around the rod to prevent any appreciable rise of temperature. 

 The rod was demagnetized in situ by means of a long series 

 of currents in the solenoid, alternating in direction and grad- 

 ually decreasing in intensity, and then a curve of ascending 



