THE DISTURBANCE OF ISOMAG NETICS. 



159 



Observed Length 

 of Ma<rnet. 



7.0053 

 6.9994 

 6.9973 

 6.9990 



In the reduction of öcale to 0",C. the expansion coefficient of brass 

 was assumed to be 0.0000189, and that of steel, 0.0000106. From the 

 mean of these observations, the length of the magnet is 6.9979 cm. 

 The leno-th has thus somewhat diminislied since the former deter- 



o 



mination, by which it was found to be 7.002-1 cm. 



The breadth of the magnet was measured at three points; i.e., at 

 the ends and at the middle. The m;ignet was placed perpendicular 

 to the metre scale and between the two prisms before mentioned. 

 The same process of measurement as used for the length, gave 0.8066 

 cm. at 0"C. for the mean breadth of the bar or its thickness between 

 the face with the mark, and the one parallel to it. The breadth 

 of the other faces is of minor importance, and has not been mea- 

 sured. 



In weighing the magnet a Sartorius' balance was used. Care 

 was taken not to place the magnet near any of the iron used in 

 various parts of the balance, by not putting it on the scale pan, but 

 suspending it from this l)y a slender thread, bringing it about 50 cm. 

 below the balance case and about the same distance above the floor. 

 The magnet was shielded from air currents by placing it in a nearly 

 closed glass vessel. The mass of the magnet, as determined by a set 

 of weights from Sartorius, was 35.0561 grm., and 35.0633 grm. by 

 the set of weights formerly used. This discrepancy was cleared 

 up by comparing these weights with those made of quartz by the 



