Chfervations on the fuppofed Magnetic Property of Ntciel. 289 



Its oxide u-lien held in folutioii by acid nienfttua. Indeed the magnet will, in advantageous 

 circumilances, be influenced by the fmalleft particle of metallic iron adhering to the fur- 

 face of any body, without regard to the aflual quantity of the fubftance to which it is 

 attached ; while the efficacy of chymical tefts is, in fome meafure, bounded by the pro- 

 portion of heterogeneous elements. In one cafe the a£lion is independent and abfolute ; 

 in the other, relative and limited. Thus a lump of copper, of whatever fize, being ftruck 

 with an iron hammer upon an anvil, fometimes acquires the power of influencing the 

 magnet ; but I believe it would be rather difficult to difcover, by reagents, the iron that 

 confers on it that power ; and which perhaps does not amount to a millionth part of the 

 whole mafs. 



As far as I have obferved, the following cautions alone are neceflary in thus treating 

 the oxide of nickel by ammonia. 



1. All oxides of iron, except the red, are foluble, more or lefs, in ammonia; therefore 

 the iron muft, by the means of nitric acid, be brought to its full ftatc of oxidation, before 

 precipitation and redlfTolution. 



2. The ammonia muft be perfedlly cauftic ; for carbonate of ammonia difTolves red 

 oxide of iron. And, 



3. It may be expedient to allow fome precipitate of nickel to be formed at the bottom, 

 before the ammoniacal liquor is drawn off for evaporation ; as, by that time, we may be 

 fure, that if the two firft conditions have been obferved, all the oxide of iron has previoufly 

 fallen to the bottom. 



Although I have not directly performed a fimilar operation upon cobalt, or obtained 

 that metal unattraflable, I have little doubt but that fuch a quality is intirely owing to iron; 

 and that the property of magnetifm muft be again reftrided to its former limits. 



III. 



On the Mechanlfm of the E-^e. By Thomas Young, M.D. F.R.S. 

 ( Continued from page 26^. J 



HESE confiderations will explain the Inconfiftency of different obfervatlons on the rc- 

 fra£Hve power of the cryftalllne ; and, in particular, how the rcfradtion which I formerly 

 calculated, from meafuring the focal length of the lens,* is fo much greater than that 

 which is determined by other means. But, for direft experiments, Dr. WoUafton's method 

 is exceedingly accurate; 



* Phil. Tranf. for 1793, p. 174. 

 Vol. v.— November J 801. Pp When 



