exhibited by certain Non-metallic Substances. 399 



doubly absorbing, doubly metallic, and doubly fluorescent. By 

 the last expression I mean that the fluorescence, which the cry- 

 stals generally exhibit in an eminent degree, is related to direc- 

 tions fixed relatively to the crystal, and to the azimuths of the 

 planes of polarization of the incident and emitted rays. 



M. Haidinger has expressed the relation between the surface 

 and substance colours of bodies by saying that they are comple- 

 mentary. This expression was probably not intended to be 

 rigorously exact; and that it cannot be so, is shown by the fol- 

 lowing simple consideration. The tint of the light transmitted 

 across a stratum of a given substance almost always, if not 

 always, varies more or less according to the thickness of the 

 stratum. Now one and the same tint, namely, that of the 

 reflected light, cannot be rigorously complementary to an infinite 

 variety of tints of different shades, namely, those of the light 

 transmitted across strata of different thickness. In most cases, 

 indeed, the variation of tint may not be so great as to prevent 

 us from regarding the reflected and transmitted tints in a general 

 sense as complementary. But as media exist (for example, salts 

 of sesquioxide of chromium, solutions of chlorophyll) which 

 change their tint in a remarkable manner according to the 

 thickness of the stratum through which the light has to pass, it 

 is probable that instances may yet be observed in which M. Hai- 

 dinger's law would appear at first sight to be violated, although 

 in reality, when understood in the proper sense, it would be 

 found to be obeyed. As the existence of surface-colour seems 

 necessarily to imply a very intense absorption of those rays which 

 are reflected according to the laws which belong to metals, it 

 follows that it is in the very thinnest crystals or films of those 

 which it is commonly practically possible to procure, that the 

 transmitted tint is to be sought which is most properly com- 

 plementary to the tint of the reflected light. 



I will here mention another instance of the connexion between 

 metallic reflexion and intense absorption. I choose this instance 

 because a different explanation from that which I am about to 

 offer has been given of a certain phenomenon observed in the 

 substance. The instance I allude to is specular iron. As it is 

 already known that various metallic oxides and sulphurets possess 

 the optical properties of metals, there is nothing new in bringing 

 forward this particular mineral as a substance of that kind. It 

 is to the chromatic variation of the metallicity that I wish to 

 direct attention. If light polarized at an azimuth of about 45° 

 be reflected from a scale of this substance at about the polarizing 

 angle, and the reflected light be viewed through a plate of cal- 

 careous spar and a NicoPs prism, it will be found, by using dif- 

 ferent absorbing media in succession, that the change of phase, as 



