FROM THE MAGNETISM OF THEIR BALANCES, 359 



The best material, in this case, to be substituted for steel, 

 would probably be platina, or an alloy of platina. As the ex- 

 pansion of platina with heat appears to be less than that of al- 

 most any other metal, being nearly one-third less than that of 

 steel, and less by above one-half than that of brass *, I should 

 imagine that it would be better suited for effecting the com- 

 pensation than steel. 



For accomplishing the second object, I should recommend 

 that the flat surface of the balance be the last part that is 

 finished, and that it be ground and polished in the plane of 

 the magnetic equator. From various experiments with differ- 

 ent ferruginous substances, I found that no friction, however 

 severe, produced magnetism in flat plates or slender bars, when 

 the friction was endured in the plane of the magnetic equa- 

 tor ; but that, on the contrary, such substances, especially iron 

 and soft steel, when hammered, bent, twisted, scowered, filed, 

 or polished in the plane of the magnetic equator, were depriv- 

 ed of any small quantity of magnetism that they might have 

 previously acquired f ; whereas by a treatment precisely si- 

 milar, excepting as to position, ferruginous bodies were inva- 

 riably rendered magnetic. I have made some experiments on 

 chronometer balances, with the view of removing their polari- 

 ties ; but although a sensible diminution of their magnetism 

 occurred after grinding them in the plane of the magnetic 

 equator, or striking them in the same plane with a small 

 smooth-faced hammer, while resting upon a hard, flat substance, 

 yet 1 have not had leisure to accomplish their neutralisation so 



vol. ix. p. ii. z z effectually 



* Henry's Chemistry : Table of Expansion of Solids by Heat. 



-f- See Transactions of the Royal Society of Edinburgh for 1881 ; " Descrip- 

 " tion of a Magnetimeter," &c. Propositions 4, 5, 6, 10, &c. 



