Diffuse Reflexion of Light on Matt Surfaces. 201 



I. Experimental Part. 



Preparing the Plates. — Powders loosely sifted and having 

 a level surface would undoubtedly- answer the postulate of 

 the theory best. This would, however, necessitate a horizontal 

 position of the surfaces, which renders the investigations very 

 laborious. It is therefore advisable to have the substances 

 in the shape of plane plates, so that the surface to be experi- 

 mented upon can be brought into a vertical position. Cements 

 such as gum &c. have been shown to change the optical 

 properties of the powders in a high degree. Casting (plaster 

 of Paris) also causes more or less normal reflexion, which 

 must by all means be avoided. The best method, as I have 

 found, is to compress the powders by means of an hydraulic 

 press and a steel stamping-iron. With suitable support at 

 the base and a pressure of from 4 to 20 tons, the necessary 

 solidity for delicate treatment is obtained. No normal 

 reflexion can be detected on the plates, as the following 

 results will show ; even the size of the powdery particles is 

 not changed. Thus the chief source of such errors as all 

 investigators have hitherto fallen into is fully avoided. 



The Polarization of the Scattered Light had to be dealt with 

 first in order to decide whether a polarization-photometer 

 could be used. In contrast io media with very small 

 suspended particles such as solutions of colloids &c, the 

 scattered light reflected by powders is not polarized, as 

 Provostaye and Desains * have shown. The linear dimensions 

 compared with the length of the reflected waves are in the first 

 case of the same order or even smaller than the waves, whilst 

 the powder particles have an average diameter of O002 mm. 

 The size is evidently of the same importance here as it is in 

 respect to the colour (Rayleigh's theory) of the light of the 

 sky. The above-mentioned plates have been studied with a 

 " Haidinger'sche Lupe " and no polarization was found ; even 

 polarized incident light was reflected depolarized (with 

 exception of the limit angles e = 80° and i = 60° or 80°). 

 This shows the perfection of the plates and permits the 

 employment of a polarization-photometer, to the use of which 

 Messerschmitt was not entitled as his substances polarized 

 the light. 



Ihe Disposition of the Experiments must be such that 

 parallel light can fall under any angle of incidence (i) on the 

 surface of the substance, and that the reflected scattered 

 light can be observed under any angle of emission (e) by the 



* F. de la Provostaye et P. Desains, Ann. chim. et phi/s. (3) xxxiv, 

 pp. 196, 216 (1852). 



