292 MAGNETISM 



as T rises ; and B the constant diamagnetistD of th< n>titn- 



ent. As T increases x approaches H. 



Curie also studied the behaviour of iron, magnetite, and nickel 

 when carried to high temperatures. In the case of iron he 

 experimented in succession on three \\iiv-, each about 1 rm. lon^, 

 and of respective diameters 0*002, O'OU, and 0'035 cm. These 

 were enclosed in glass or platinum tube* to protect them from 

 oxidation and arranged on the tor-inn rod nlon.i: Mf, 1 

 The temperature was raised from 20 C. to 1360 ( .. and tin- 

 field from 25 to 1300. 



About 750 ('. tlic value of I, the intensity of magnetisation, 

 began to be nearly independent of the field and tell rapidly. >n- 

 tinuing to fall to 950 C. Then it ;lv con-taut, though 



falling slightly to 1280 C. Then it rose llightly and ftf 



fell again. Apparently at a sufficiently high temperate I 



as for feebly paramagnetic hod I MID the < \JM i munN In 



obtained the following roults for iron : 



Tetnperatmv 



<>:> B I KXXX) 



20 1300 



1000 25 to 1300 



Thus at 1000 C. the value of x i- independent of the lield : 

 is nearly the same as for air at 20.* 



Similar results were obtained \\ith magnetite and nirkrl. 



Wills' experiment.! In this i pillar 



plate of the Mih.stai.i ^ as liiin_u t'lmn oiu- arm 



balance. Its lower edge was hori/ontal and \\a- l>. t \\een the poles 

 of an electro-magnet as represcn t ed in filiation and plan in 1'ig. 221. 



If A is the area of i -ro^-si ction of the plate indicatttl in 

 in (b\ the forcv on the plate in direction ./ upuanU i> 



F= / A 



where the limits for II are the field at the lower t dg- and that at 

 the upper. Hut the field at the upper edge is negligible in 

 comparison with that at the lower, and we may put the force 

 measured bv the balance 



where H is the field between the poles. i tMire this a 



"dummy" plate of plaster of Paris with tinfoil strip round its 



* For details of the work on iron at high temperature?, Curie's paper (Joe. ett.> 

 should be consulted. 



t Phil. Mag., xlv. (1898). p. 432. 



