Films of Sodium and Potassium. 107 



The preliminary observations were made by the eye, in 

 combination with monochromatic ray-filters, but the tinal 

 readings, from which the more accurate curves were plotted, 

 were taken with a ITilger spectroscope, arranged as a mono- 

 chromatic telescope, i. e. with a lens in front of the first slit 

 (Lord Rayleigh/s method), the eye being applied to the 

 second slit. Both slits should be well open, to secure good 

 definition of the glass sliver. Transmission readings were 

 obtained by mounting the bulb between the eye and the 

 portion of the paper illuminated by the light passed by the 

 absorbing gelatine, i. e. a little to the left of its position in 

 fig. 2, and mounting the sliver of" silvered glass in front of 

 it at such an angle as to reflect the light from that part of 

 the paper illuminated by light passed by the disk. In this 

 case the eye was at X. 



Six sets of curves, obtained with films of increasing- 

 thickness, are reproduced in fig. 3, dotted lines corre- 

 sponding to incidence on the inner or concave side of the 

 film, unbroken lines to incidence on the outer surface. The 

 absorption values were represented by subtracting the sum 

 of the reflexion and transmission percentages from 100. 



The transmission values are independent of the direction 

 of the ray, consequently a single curve represents both 

 conditions of incidence. 



On the other hand, the absorption curves are very different. 

 Take, for example, series 5 (purple film). In the case of 

 incidence on the outer surface, we have 85 per cent, reflected 

 for all values of \. Of the 15 per cent, which enters the 

 film, about one half is removed by absorption in the blue, 

 and somewhat more in the red, consequently the film is 

 fairly opaque, and bluish green in colour. 



In the case of incidence on the inner surface, however, 

 the reflexion falls to 6 per cent, for \ = 5100 (dotted curve 

 Kef.). Since this wave-length does not appear in excess in 

 the transmitted light, we must assign a value of over 85 

 per cent, for the absorption in this case.- This absorption, 

 I believe, results from the action of the spaces between the 

 crystals as resonators or selective radiation traps. 



The absorption which occurs in the case of incidence on 

 the outer surface may be of a different type, though both 

 types probably occur simultaneously to a greater or less 

 degree. 



We will now examine in detail the phenomena which occur 

 as the thickness of the film is gradually increased. 



In the case of exceedingly thin deposits, i.e. consisting of 



