1070 
great accuracy. Of all observations on iron the mean is 8.86°/, for 
a variation of temperature between 10° and 100°. 
According to the theory of the molecular magnetic field of Werss 
the spontaneous magnetisation of the ‘elementary crystals’ which 
Weiss regards as the building stones of the iron decreases with a 
rise of the temperature. The law of this decrease is given by the 
formula 
M 30 M ] 
ED 
M, T M 
where J/ is the magnetisation at the temperature 7’, M, that at the 
absolute zero-point and @ the Curt-point. In this formula the mutual 
action between the different elementary crystals has been neglected. 
If by means of it we calculate the change between 10° and 100°, 
we obtain, taking for the Curim-point 757° C., 4,2 °/,, what agrees 
rather well with the experimentally found variation of 3,9 °/, between 
10° and 100°. 
Also for steel the above described magnetometric method was 
used. To render possible however the investigation of different parts 
of one and the same magnet a ballistic method was used. A flat coil 
placed round the experimental magnet could be shifted with regard to 
the latter over a small distance. In order to make the deviation 
of the galvanometer not too great, it was for the greater part 
compensated by means of a second coil equal to the first one and 
mechanically connected with it, so that it followed its motion, 
moving itself however in the opposite direction; within this second 
coil a second magnet was placed. If now the first magnet was 
heated, the ballistic deviation changed and from this the influence 
of the heating could be investigated. By bringing respectively different 
parts of the experimental magnet within the first coil, compensating 
each time by means of the second magnet, the change of the mag- 
netisation could be determined for the different parts. 
_For steel the results of AsHwortH were confirmed. In the first 
place for different kinds of steel very different values of « were 
found; the highest value for a very old steel magnet viz., nearly 
0.0015, for other kinds of steel « was smaller, while piano-strings 
even gave a negative value. 
By the mentioned ballistic method was found however, that the 
value of « was not the same for all parts of the magnet. Near the 
ends « proved to be larger in the positive sense than more towards 
the middle, Between 20° and 100° the change for the above men- 
