112 



SCIENCE PROGRESS 



properties appear to be inherent in certain crystalline masses, 

 not themselves the magnetic units but which contain, as one of 

 their structural elements, molecular groups that are magnetic at 

 proper temperatures. 



The magnetic quality of any particular specimen depends 

 much more upon its thermal history than upon its composition, 



Fig. I. — Explanation of curves. All curves refer to the same sample. 



29'6 per cent. Mn, 8'2 per cent. Al, 6'2 per cent. Cu, o'^ per cent. Si. Curve 1, best conditions after 

 heating to 125" during thirteen hours; 2, after eight hours near 200° and five hours between 275° 

 and 300" (another heating of five hours at 100" and two hours at 200^ produced no further 

 change); 3, five hours at 2oo°-22 5^ ; 4, annealed at 225° aftei* heating above transformation point ; 

 5, after nine and a half hours at 250° ; 6, after fourteen hours at 29o°-35o°. 



although the maximum attainable intensity of magnetisation — 

 i.e. the saturation value after most favourable heat treatment — 

 depends upon the composition. As is well known, iron loses its 

 magnetic properties when heated to 785° C. and regains them 

 upon being cooled to a slightly lower temperature. The Heusler 

 alloys undergo a similar transformation at temperatures which 



