54 M. J.-L. Soret on Polarization 



not go so far, but continue to attribute it to the general fact of 

 the reflection produced at the surface of separation between two 

 unequally refracting media. Each particle of carbon, although 

 very minute, is composed of a great number of molecules > it 

 forms a minute body reflecting light, only, as its dimensions 

 are very small, there is no annihilation by interference of the 

 rays emitted in directions different from those determined by the 

 ordinary laws of reflection : there is not specular reflection, but 

 diffusion in all directions. 



Let us look at some consequences of these two interpreta- 

 tions. Suppose a liquid having no power of illumination ; then 

 introduce some very line particles of a solid body also having 

 no power of illumination. If M. Lallemand's theory is accurate, 

 the liquid, on ceasing to be homogeneous, will not receive the 

 property of being illuminated, since neither of its constituents 

 possesses it ; while, according to my view, the trace of a lumi- 

 nous pencil must be marked in the liquid and present the usual 

 phenomena of polarization, except in the quite exceptional case 

 of the solid and liquid having the same index of refraction. 

 The experiment is hardly realizable absolutely ; but an approxi- 

 mation is possible. Water purified to the highest degree attain- 

 able, and containing only very few particles in suspension, has 

 very little illumination-power ; and calcspar has no sensible 

 capacity of illumination, the trace of a pencil manifesting itself 

 in the interior only by the slight red fluorescence of this body. 

 Now, on suspending finely pulverized spar in water, and 

 then filtering the liquor to separate the particles of too great a 

 volume, I have found the power of illumination become much 

 superior to that of the water alone*. 



On the other hand, if it is the molecules of the body them- 

 selves that produce the diffusion, it seems to me that the inten- 

 sity of its manifestation should not depend on the refrangibility 

 of the medium in which we operate. A smoked surface, for in- 

 stance, whether it be in the air or in a more or less refracting 

 liquid, should emit sensibly the same quantity of light. Now 

 this is not what takes place; for it is easy to prove that the dif- 

 fusing-power of lampblack is less intense in a liquid than in the 



* I cannot, however, give this experiment as entirely conclusive. The 

 trace becomes incontestably more evident, and the polarization more pro- 

 nounced ; but the polarization is not complete, and there remains a residue 

 of neutral light analogous to that which would be produced by an action of 

 fluorescence. The colour of this residue is greenish, while the spar has a 

 red fluorescence ; besides, the proportion of solid particles is so trifling 

 that it would be difficult to attribute a sensible fluorescence to them. 

 Further, the trace presents very pronounced differences of colour, accord- 

 ing to the angle under which it is viewed — a phenomenon manifested in 

 other cases, and to which 1 purpose, one day, to return. 



