498 SPECIAL METHODS OF DISCOVERY. 



ing any arsenic in it. He then requested Roloff to repeat 

 his experiments. This he did, and now perceived that 

 the precipitate, which lie had taken for orpiment, was not 

 so in reality, but owed its existence to the presence of 

 another metallic oxide, and probably new. Specimens of 

 this oxide of zinc, and of the yellow precipitate, were sent 

 to Stromeyer for examination, who readily recognised the 

 presence of cadmium, and was able to extract from it a 

 considerable quantity of that metal.' l 



After having discovered the anomalous molecular 

 movements in red-hot iron during the process of cooling 

 from a red heat, 2 I was informed by a machinist that the 

 best temperature for shrinking iron hoops on metal wheels 

 is at a very low red heat, and that if a higher temperature 

 is employed the hoops are liable to burst from some un- 

 explained cause ; the cause is probably connected in some 

 way with the molecular change I have referred to. The 

 pleasant odour emitted by good cast steel during the process 

 of hammering, and the odour of alkalies in soap-works, 

 are unexplained phenomena worthy of examination. 



d. By the investigation of exceptional cases. This is 

 one of the most important methods of research, because it 

 leads to the discovery of greater laws and more correct 

 principles than those already known ; the exceptional 

 cases themselves being usually discovered only by means of 

 extensive research. 3 An exception to a general principle 

 indicates the existence of a wider law, and the necessity of 

 a new definition, which will include both the ordinary cases 

 and the exceptional ones. For instance, there are two 

 exceptions viz., copper and zinc to the statement that 

 the magnetic capacity of different elementary bodies in- 

 creases in a given volume with the number of atoms they 



1 History of Chemistry, vol. ii. p. 220. 

 2 See p. 519. 8 See Chapter XXL 



