Polarization by Diffusion of Light. 209 



being insufficient, combustion does not destroy the lampblack 

 in proportion to its production, the particles accumulate and 

 become sufficiently abundant to render visible the trace of the 

 pencil of solar light. 



2nd. When the combustion is complete the light is very in- 

 tense, it dazzles the eye; and it is easy to comprehend how dif- 

 ficult it must be to perceive the polarization of the very slight 

 proportion of reflected light. We can convince ourselves of this 

 by endeavouring to observe through a gas-flame the polarization 

 which accompanies the illumination of a liquid for example, 

 though it is so easily verified under the usual conditions. 



I have also studied the diffusion of light by any solid surface 

 covered with lampblack deposited from the flame of essence of 

 turpentine or of benzine. When a pencil of sunlight falls on this 

 surface in a dark chamber, a very notable proportion of light is 

 seen to be diffused in all directions ; the incident pencil is marked 

 on the black by a white spot of considerable splendour, much 

 more brilliant than that which is obtained upon a smoky flame : 

 the reflecting particles are, in fact, much more abundant. What- 

 ever be the angle and the plane of incidence of the luminous 

 peucil relative to the blackened surface, the diffused light is seen 

 to be polarized in the plane of vision. The maximum corresponds 

 to a visual angle of 90°, though there the polarization is not 

 complete. When looked at through an analyzer which is rotated, 

 the spot augments or diminishes in brightness, but remains 

 always distinctly visible. 



If in place of natural light we employ a pencil of solar rays 

 polarized, for instance, in a vertical plane, the light diffused by 

 the lampblack presents in all directions the same phenomena of 

 partial polarization as the trace which is produced in an illumi- 

 nated gas or liquid. In a horizontal plane of vision, and with 

 a visual angle of 90°, we have the minimum of brightness, but 

 not complete extinction as this takes place with a smoky flame 

 or, in general, with gases or liquids holding in suspension fine 

 particles in relatively little abundance. The residual light is 

 neutral and shows no trace of polarization ; but as soon as the 

 angle of vision becomes greater or less than 50° partial polariza- 

 tion reappears immediately. 



I explain this residue of neutral light by multiple reflections 

 from the minute rugosities produced by the particles of lamp- 

 black on the diffusing surface. I do not think it can be attributed 

 to a fluorescence, for the following reasons : — 



1st. When this residual light is analyzed with the spectro- 

 scope, the principal lines of the solar spectrum are recognized in 

 it; therefore there is no change of refran gib ility*. 



* This proof is perhaps not quite conclusive ; for under the influence of 

 ■Phil. Mag. S. 4. Vol. 47. No. 311. March 1874. P 



