Permanent Magnetism, 313 



The variation in the resistance of the steel under different 

 values of the magnetic induction has not yet been regarded. 



The change in the air-resistance for unit radius, from about 

 •45 in the case of the thick disks to *30 in the case of the long 

 bar, shows that the length of the bar has some influence on 

 this quantity. It is proposed to examine further the law of 

 change. 



The temperature during the work was always high and 

 tolerably constant. So far its variation has not been taken 

 into account. In the early morning, before the room is warm, 

 an increased moment is observed. The permanent magnet as 

 it now stands in the bifilar has a deflection of about 14°, read- 

 ing during the day 73 c, 7 (zero 88° nearly). I have gone in 

 early on several mornings and found the reading to be about 

 72°'4. It returns to its former position as soon as the room 

 is warm. 



Heat therefore diminishes the magnetism*. It remains to 

 be determined whether the magnetomotive force or the resist- 

 ance is affected. In the latter case we should have a curious 

 analogy to the behaviour of electrical conductors, the resist- 

 ance increasing with the temperature. 



So long as the permanent magnet was kept free from mag- 

 netic influences, its mean condition did not sensibly change. 

 I wished to make it take a higher moment, and tried magne- 

 tizing it on the dynamo. It took a higher moment, but soon 

 lost it again. I had reason subsequently to think that one of 

 the pieces was reversed in this process, wdiich threw doubt on 

 the first direct determination of permeability. I then remag- 

 netized it carefully on the dynamo. The moment soon fell 

 again to about its original value; it was slightly increased by 

 the first action of the currents used in the second determina- 

 tion of permeability, but did not change further during the 

 determination, and still remains the same. 



Calculation of Experiments. 

 Constants. 

 t ~ i • i ' > distance of points of suspension. 



/4K _ 

 h had the values I 37 ) inches, height of suspension. 



X 34^ 

 m = 41 oz. = 1162*3 gnu,, ) . -,, 

 <7=981, j we, S ht - 



* This conclusion is opposed to that of Meyer (Wiedemann's A?malen, 

 xviii. p. 246). His conclusion is founded on the difference between expe- 

 riments in September and January. I measure the difference directly, and 

 see the change repeat itself daily. 



