j58 PHILOSOPHICAL TRANSACTIONS. [aNNO 1786. 



drew one end of it over the surface of the stone in various directions. This done, 

 the piece of steel was weighed again, and was found to have lost so small a part 

 of its weight as not to be discernible by that pair of scales ; yet the Turkey-stone, 

 which had acquired only that small quantity of steel, affected the magnetic needle 

 very sensibly. Chemistry seems not to afford any means by which so small a 

 quantity of iron may be decisively detected in a body that weighs 1 oz. Hence it 

 follows, that though no iron is to be discovered in a body by chemical methods, 

 yet it should not be concluded, that the said body, if it affect the magnetic needle, 

 does not owe its magnetism to some small quantity of iron concealed in its 

 substance. 



The most of my experiments are relative to the properties of brass ; and they 

 seem to prove that this compound metal, which is often magnetic, does not owe 

 its magnetism to iron, but to some particular configuration of its component par- 

 ticles, occasioned by the usual method of hardening it, which is by hammering. 

 In some specimens of brass, and especially in that which has often passed from 

 the work- shop to the furnace, and from the latter to the former, there are some- 

 times pieces of iron sensible not only to the magnet, or the chemical analysis, but 

 even to the sight, which render the brass strongly magnetic. But the brass 

 generally used in my experiment swas such as, when quite soft, had no sensi- 

 ble degree of magetism. 



Examination of the Magnetical Properties of Brass. — A few years ago, being 

 intent on making some magnetic experiments, in which brass was concerned, I 

 used to examine first whether the pieces of brass had any magnetism or not, and 

 rejected those pieces which had an evident degree of that power. In the course 

 of those experiments I remember to have observed, that those pieces of brass 

 which had been hammered were generally magnetic, and much more so than 

 others ; in consequence of which I made no use of hammered brass in those 

 experiments. But lately, having ordered a theodolite at a philosophical instru- 

 ment shop, I particularly enjoined the workmen to try the brass, both soft and 

 hammered, before they worked it, and to make no use of that which had any 

 magnetism. They found that hammered brass, even such as before the hammer- 

 ing had no magnetism, could afterwards disturb the magnetic needle very sensi- 

 bly. These observations induced me to make the following experiments. Mr. C. 

 here describes 6 experiments on this metal, from which he draws the following 

 conclusions : 



1 st. That most brass becomes magnetic by hammering, and loses the mag* 

 netism by annealing or softening in the fire. 2dly, That the acquired magnetism 

 is not owing to particles of iron or steel imparted to the brass by the tools 

 employed. 3dly, Those pieces of brass which have that property, retain it with- 

 out any diminution after a great number of repeated trials, viz. after having been 



