686 lectuhe lv. 



are still of great utility in assisting us to generalise, and to retain in memory, 

 a number of particular facts wliicli would otherwise be insulated. The doc- 

 trine of the circulation of streams of the magnetic fluid has been justly and 

 universally abandoned, and some other theories, much more ingenious and 

 more probable, for instance that of Mr. Prevost, appear to be too compli- 

 cated, and too little supported by facts, to require much of our attention. 



The distinction between conductors and nonconductors is, with respect to 

 the electric fluid, irregular and intricate: but in magnetism, the softness or 

 hardness of the iron or steel constitutes the only difference. Heat, as soft- 

 ening iron, must consequently render it a conductor; even the heat of boil- 

 ing water affects it in a certain degree, although it can scarcely be supposed 

 to alter its temper; but the effect of a moderate heat is not so considerable in 

 magnetism as in electricity. A strong degree of heat appears, from the expe- 

 riments of Gilbert, and of Mr. Cavallo, to destroy completely all magnetic 

 action. 



It is perfectly certain that magnetic effects are produced by quantities of iron 

 incapable of being detected either by their weight or by any chemical tests. 

 Mr. Cavallo found that a few particles of steel, adhering to a hone, on which 

 the point of a needle was slightly rubbed, imparted to it magnetic properties; 

 and Mr. Coulomb has observed that there are scarcely any bodies in nature 

 which do not exhibit some marks of being subjected to the influence of mag- 

 netism, although its force is always proportional to the quantity of iron which 

 they contain, as far as that quantity can be ascertained; a single grain being 

 sufficient to make 20 pounds of another metal sensibly magnetic. A combi- 

 nation with a large proportion of oxygen deprives iron of the whole or the 

 greater part of its magnetic properties; finery cinder is still considerably 

 magnetic, but the more perfect oxids and the salts of iron only in a slight de- 

 gree; it is also said that antimony renders iron incapable of being attracted 

 by the magnet. Nickel, when freed from arsenic and from cobalt, is decid- 

 edly magnetic, and the more so as it contains less iron. Some of the older 

 chemists supposed nickel to be a compound metal containing iron, and we may 

 still venture to assume this opinion as a magnetical hypothesis. There is in- 

 deed no way of demonstrating that it is impossible for two substances to be so 

 united as to be incapable of separation by the art of the chemist; had nickel 



4 



