186 PROCEEDINGS OF THE AMERICAN ACADEMY. 



TABLE I. 



and similar experiments, Barus and Strouhal made out a table connect- 

 ing s and a which they subsequently found to fit other kinds of steel 

 pretty well. Some of their results are given in Table I. 



If corresponding values of s and a be used as coordinates, a fairly 

 smooth curve results, and the mean values of 79 for s and 0.0013 for 

 a which Barus and Strouhal got for three pieces of cast iron which they 

 tested, yield a point which seems to lie closely enough upon the pro- 

 longation of this curve. It appears also that the values of .« and a 

 which Matthiessen, Vogt, and Benoit obtained for different kinds of 

 wrought iron agree numerically with the values for steel ; and some 

 persons have thought that it is possible to determine the position of 

 any piece of iron or steel in the scale of mechanical hardness, without 

 any knowledge of the percentage of combined carbon, by finding s alone. 



For bar magnets or for simple bent magnets, fine tool steel, or better, 

 some of the kinds of special magnet steel, serve very well, but if a 

 permanent magnet is required of such a shape that the steel has to be 

 heated red hot a number of times during the process of forging and 

 before it is made glass-hard, irregular temper thus introduced into the 

 material often shows itself in the presence of irregular magnetization 

 when the magnet is finally charged, and this sometimes makes the 

 magnet worthless. For this and other reasons, some makers of elec- 

 trical instruments are now using chilled cast iron for such magnets, 

 and these have usually proved to be satisfactory. They are cheap, 

 they can be made quite as strong as tool steel magnets of the same 

 dimensions, they are very permanent after they have once been aged, 



