734 Dr. S. P. Thompson and Mr. E. W. Moss on the 



having similar axial ratios. More recently, Riborg Mann 

 obtained a series of values slightly higher than those obtained 

 by Du Bois, who has accepted them as more correct than his 

 own figures. E wing's observations ranged over rods the 

 lengths o£ which varied from 300 diameters down to 50 

 diameters. Du Bois' results go from a dimension-ratio of 

 1000 down to one of 10 ; those of Riborg Mann from one of 

 300 down to 5. The magnitude of the outstanding dis- 

 crepancies may be indicated by stating the values found by 

 different observers for the self-demagnetizing factor J\ 7 for 

 cylinders having a dimension-ratio of 50. For rods of this 

 proportion Du Bois found _/V = 0*0162 ; Riborg Mann 

 iY T = 0*01825. For the ellipsoid of revolution having the 

 same axial ratio of 50, Du Bois and Riborg Mann agree in 

 assigning the value 0"0181, and presumably the true value 

 for the cylinder is less than that figure. Greater discrepancy 

 is found for shorter cylinders. For a dimension-ratio of 

 10 Du Bois gives JST = 0*2160, while Riborg Mann gives 

 JN = 0-25500. 



To clear up, if possible, such discrepancies a research 

 was undertaken in the laboratory at the Technical College, 

 Finsbury. 



The bars used were cut from two long rods of best Swedish 

 iron carefully annealed, and for comparison a ring was forged 

 from the same material. To each and all of the rods the 

 same diameter was given, namely, 1*128 cm., in order that 

 each might have a cross-section of precisely 1 sq. cm. After 

 being turned down to approximate size they were annealed, 

 and then finally turned to the precise size required. 



The magnetizing coil used to magnetize the rods was a 

 long coil wound on a brass tube 91*4 cm. in length and 

 4*75 cm. in external diameter. It was carefully overwound 

 with 5800 turns of wire of No. 20 s.w.G., in seven layers. 

 With this coil a very uniform field could be produced of any 

 desired intensity up to ^f=255. The uniformity of the 

 field between the ends of this coil was tested by means of a 

 short coil of somewhat smaller diameter, wound on a turned 

 bobbin of hard fibre, of a size fitted to slide inside the brass 

 tube. The wires of this smaller coil were connected with a 

 ballistic galvanometer, the throw of which was observed 

 when the current in the long magnetizing coil was reversed. 

 The field was found to be sensibly uniform for a length of 

 60 cm.; while the longest specimen of iron was only 40 cm. 

 There was therefore no need to apply any corrections for 

 non-uniformity of field. 



The ballistic method was also used for determining; the 



