Films of Sodium and Potassium, 105 



■continuous, while from the inner surface one might expect 

 phenomena analogous to those exhibited by the " freak" 

 gratings with very deep grooves, which I have described, and 

 which have never been explained dynamically. 



Fig-. 1. 



While it is true that the greater part of the energy in the 

 region of the dark spectrum-band, seen by internal reflexion, 

 is removed by absorption, it should be mentioned that a small 

 proportion is scattered. For example, if the film which 

 reflects the golden-yellow light is examined in a narrow 

 pencil of sun-light, it will be observed that a faint violet 

 light is scattered. The scattered violet light is, however, 

 only a very minute proportion of the total violet light removed, 

 as can be shown by illuminating the film with a beam of 

 pure violet light from a monochromator, and comparing the 

 diffuse radiation with that shown by white paper. It can, 

 in fact, be nearly matched by a piece of black paper. It 

 seems, then, as if the cavities between the crystals acted as 

 resonators for a very restricted region of frequencies, 

 absorbing most of the energy in this region, but at the same 

 time permitting the escape of a very small proportion in the 

 form of diffuse radiation. 



Determinations were made with the photometer of the 

 percentages reflected, absorbed, and transmitted for different 

 regions of the spectrum, by films of various thickness, both 

 for internal and external reflexion. 



The arrangement of the apparatus for internal reflexion 

 is shown by fig. 2. 



The light from an arc passed out through a hole in a box 

 to a revolving disk provided with an open sector of variable 

 width, graduated to read percentages, and kindly loaned to 

 me by Dr. Pfund. The beam passed by the open sector 

 illuminated a portion of a sheet of white paper, in front of 

 which the sodium bulb was mounted as shown. The sector, 

 when wide open, embraced an angle of 180° and was gradua- 

 ted from to 100, and when wide open, passed 50 per cent, of 

 the incident radiation. The light which passed by the edge of 

 the disk illuminated a strip about 5 cm. wide along the edge 



