Transparent Bodies by the Action of Magnetism. 431 



diamagnetic oxides can result from the union of magnetic metals 

 with a magnetic gas, oxygen. However, we know so Uttle as 

 yet about the true cause of magnetism, that I do not consider 

 myself entitled on this account alone to speak positively. 



To recapitulate. All diamagnetic substances which do not 

 contain any magnetic metal have a positive rotary power. This 

 is not the case with substances containing magnetic metals; and 

 according to all the observations which have hitherto been made, 

 magnetic metals may be classed in three divisions, having for 

 their types, iron, nickel, and manganese respectively. The 

 magneto-rotary power of all compounds of iron is negative, with 

 the exception of the ferrocyanides, in which, as is known, the 

 properties of the iron are completely disguised. The magneto- 

 rotary power of all compounds of nickel is positive. With iron 

 are to be placed titanium, cerium, lanthanium, and probably also 

 chromium and uranium. With nickel are to be placed cobalt 

 and molybdenum. Manganese represents an intermediate type, 

 the magneto-rotary powers of its compounds being sometimes 

 positive, sometimes negative. It is possible that chromium and 

 uranium belong rather to the manganese group than to that of 

 iron. 



Moreover, no relation seems to exist between the positive or 

 negative magneto-rotary power and any property whatever of 

 the metals. It is not the magnitude of the magnetic power 

 which determines the separation of the magnetic metals into 

 the three preceding classes ; because iron and nickel, the most 

 strongly magnetic of all the metals, are the types of two opposite 

 classes. Chemical analogy is equally unavailable as a rule. If 

 we see without surprise cobalt grouped with nickel, chromium 

 with iron, lanthanium with cerium, and manganese serving as a 

 transition between these two opposite classes, we must be asto- 

 nished to see titanium and aluminium separated so completely 

 from tin and iron respectively. 



Another hypothesis concerning the anomalous nature of the 

 phsenomena was suggested to me by the old experiments of M. 

 PlUcker. It will be remembered that this physicist obtained 

 mixtures of magnetic and diamagnetic bodies capable of being 

 repelled by the poles of an electro-magnet of certain strength, 

 and attracted by those of a weaker one. He thence inferred 

 that magnetic attraction varied with the intensity of the electro- 

 magnet, according to a law different from that of diamagnetism*. 

 It occurred to me that something similar might occur with the 

 magneto-rotary power; that, for instance, the magneto-rotary 

 power of nickel salts might be positive with% certain strength of 



♦ M. Pliicker .siil)sequcntly recognized tlic inaccuracy of this interpreta- 

 tion of his experiments. 



