THE NATIONAL PHYSICAL LABORATORY. 137 



the Reichsanstalt. Specialized schools for the training of young 

 mechanics in the scientific side of their calling have heen formed and 

 now the majority of the leading firms retain in their permanent service 

 one or more trained mathematicians or physicists. In this way again 

 the importance of science to industry is recognized. I have thus noted 

 very briefly some of the ways in which science has become identified 

 with trade in Germany, and have indicated some of the investigations 

 by which the staff of the Eeichsanstalt and others have advanced manu- 

 factures and commerce. 



Let us turn now to the other side, to some of the problems which 

 remain unsolved, to the work which our laboratorv is to do and bv 

 doing which it Avill realize the aims of its founders. The microscopic 

 examination of metals was begun by Sorby in 1864. Since that date 

 many distinguished experimenters, Andrews, Arnold, Ewing, Martens, 

 Osmond, Eoberts- Austen, Stead and others have added much to our 

 knowledge. I am indebted to Sir W. Eoberts- Austen for the slides 

 vdiich I am about to show you to illustrate some of the points arrived at. 

 Professor Ewing a year ago laid before the Eoyal Institution the results 

 of the experiment of Mr.Eosenhain and himself. This microscopic work 

 has revealed to us the fact that steel must be regarded as a crystallized 

 igneous rock. Moreover, it is capable at temperatures far below its 

 melting point of altering its structure completely, and its mechanical 

 and magnetic properties are intimately related to its structure. The 

 chemical constitution of the steel may be unaltered, the amounts of 

 carbon, silicon, manganese, etc., in the different forms remain the same, 

 but the structure changes, and with it the properties of the steel. 

 1'he figure on page 136 represents sections of the same steel ^lolishod 

 rnd etched after various treatments. 



Section of Bad Rail. 



The steel is a highly carbonized form, containing 1.5 per cent, of 

 carbon. If it be cooled down from the liquid state, the temperature 

 being read by the deflexion of a galvanometer needle in circuit with a 



