Magnetic Characters of Metak, ^c, 88 



found to be more higly magnetic than the old half-crown of 

 William and Mary (29.) 



33. This aingular result has cost me much thought and a 

 great deal of trouble. The crucible employed was quite clean, 

 having never been used before ; and its contents during the 

 time it was in the fire were the silver and copper, and a mix- 

 ture of pulverized charcoal and common salt. Similar pieces 

 of charcoal and slices of the same quality of salt have been 

 tested, but no magnetic action could be detected in either. 

 Whence, then, this almost unexpected magnetism in the 

 metallic alloy ? Fragments of a broken crucible similar to 

 that used were found to be slightly magnetic, probably from 

 a portion of iron in the clay of which it was formed : but the 

 magnetic action in the fragment of the crucible was not 

 nearly so great as that displayed by the alloy. Moreover, 

 the pure portion of the silver that entered the alloy had pre- 

 viously been fused in a similar crucible (one of the same nest), 

 and with a similar mixture of pulverized charcoal and salt, 

 and yet shewed no trace whatever of magnetic action. 

 Hence, it could hardly be imagined that the magnetism dis- 

 played by the alloy was due to iron derived from the crucible. 



34. After subjecting a portion of the alloy (32) to dilute 

 nitric acid, and finding no iron in the solution, the surface of 

 the metal was washed in clean water and thrown into dilute 

 sulphuric acid, which, upon the principles of electro-chemis- 

 try, took up a portion of the copper ; and, had iron been pre- 

 sent, would have taken it up also. But no indication of that 

 metal appeared by the test of gall-nut, nor by that of ferro- 

 cyanuret of potassium ; but the formation of ferrocyanuret 

 of copper was manifested in an eminent degree, though, pre- 

 viously to the introduction of the ferrocyanuret of potassium, 

 scarcely any colour was perceptible in the liquor.* 



35. From the facts above described (32, 33, 34), — and at 

 present I rest on no other data, — I am inclined to think that 



* Another portion of the alloy ha^ since bepn analysed, with similar results. 



