Polarization by Diffusion of Light. 211 



tion reappears in the contrary direction ; that is, the light is 

 polarized in the plane of vision. 



The fumes obtained on agitating some drops of nitric or chlor- 

 hydric acid in a moistened balloon, or chlorhydric acid and am- 

 monia, give also anomalous polarization for an angle of vision 

 of 90°. 



On the contrary, on clouds and fogs illuminated by the sun I 

 have never observed polarization in a plane perpendicular to the 

 plane of vision; the diffused light is either almost neutral or 

 polarized normally. With a light mist the polarization is very 

 marked. The white bow is strongly polarized in the plane of 

 vision. 



When water is pulverized by making a very thin jet strike 

 forcibly against a solid body, a vapour is obtained composed of 

 minute drops. The light diffused by it is partially polarized in 

 the plane of vision for a visual angle of 90°. Towards 100° a 

 neutral point is observed*; for greater angles the polarization 

 is inverse, by refraction. If, starting from 90°, the angle of 

 vision is diminished, the polarization remains normal ; it is very 

 strong in a direction corresponding approximately to the position 

 of the first two rainbows. I say approximately, because the two 

 rainbows, which are usually so readily visible, disappear almost 

 completely in these conditions of extreme minuteness of the glo- 

 bules of water ; they only manifest themselves by two maxima 

 of light sensibly white, between which a red coloration is ob- 

 served. The other colours cannot be distinguished; it seems 

 that they have spread out and mingled, and that the reds of the 

 two bows have united with each other. This remarkable fact is, 

 I think, thus explained : — because on extremely minute surfaces 

 reflection and refraction do not follow the laws of geometrical 

 optics ; as Fresnel has demonstrated, diffraction is produced. 



I will not add more ; the examples just cited suffice to show 

 the considerable influence exercised by the state in which the 

 particles are found. I defer the publication of the other facts 

 which I have hitherto been able to observe, or which I may 

 ascertain in pursuing this study on diffusion. 



* The position of neutral points in general is not very constant; it varies, 

 as Tyndall has shown, with the circumstances of magnitude and abundance 

 of the particles. 



P2 



