122 PROCEEDINGS OF THE AMERICAN ACADEMY. 



If the lines of magnetic induction in a slender homogeneous cylinder, 

 made of perfectly soft iron, are known to be straight and parallel to the 

 generating lines of the cylinder, we may infer that the induction 

 vector — which in this case must be solenoidal and lamellar in the 

 metal, — has the same intensity throughout the space considered. 

 If, moreover, the lines of force in the air about the cylinder and near 

 it on all sides, seem to be straight, we may believe, since the tangential 

 components of the magnetic force and the normal component of the 

 induction are continuous at the surface of the iron, that the lines of 

 force and induction in the metal are straight and parallel to the lines 

 of the cylinder and to the lines just outside the metal in the air. If, 

 therefore, by means of a test coil of very fine insulated wire wound 

 tightly around the cylinder, and a somewhat larger coaxial coil which 

 does not extend into any portion of the air where the lines of force 

 are not straight, we determine B in the metal and // in the closely 

 surrounding space, the ratio of the two may be supposed to give the 

 value of the permeability in the iron. This is the theory that under- 

 lies one form of the " Isthmus Method" of measuring the value of the 

 magnetization vector in the metal at high excitations. If the results 

 are to be satisfactory, great care must be taken to make sure that the 

 magnetic lines just outside the isthmus are really straight in the 

 region to he used, and the dimensions of the test coils must be deter- 

 mined with the aid of trustworthy comparators with great accuracy. 

 The larger test coil must be mounted upon some sort of support, if 

 it is to keep its form unchanged, and the choice of material for a spool 

 is very narrow. No brass or copper that I have ever tried is unmag- 

 netic in very strong fields; paraffine wax and ebonite are often para- 

 magnetic and introduce errors into the readings. Silk insulation for 

 the wire of which the test coils are made is inadmissible without 

 careful examination and even shellack, when dried from an alcoholic 

 solution, is almost always strongly magnetic. 



The form of bobbin used by Ewing and Low requires a fresh outer 

 test coil for each specimen, but the little rods inserted at the ends 

 into holes in the pole pieces, as in the work of Gumlich, or the shorter 

 rods butted against the faces of the pole pieces, as in the work of Had- 

 field and Hopkinson, do not have this disadvantage. 



If the lines of force in the air about the isthmus are practically 

 straight for one excitation they often cease to be so when the intensity 

 of the field is much increased. If, with soft iron pole pieces the lines 

 are parallel for a soft iron bobbin, they may not be even approxi- 

 mately so for a bobbin of fairly hard steel, as I have found to my cost 

 in a somewhat large experience. 



