Films of Sodium and Potassium. 101 



constricted until it circumscribed closely the image reflected 

 from the sodium mirror, preventing all other scattered or 

 reflected rajs from entering the spectrograph. The film 

 was then cleaned off by brushing the bulb with a small 

 pointed flame, taking care to clean off the front wall as well. 

 A series of exposures of different exposure times was then 

 made, moving the plate-holder up one notch between ex- 

 posures. This gives us a record of the reflecting power of 

 the quartz bulb. A small pad of cotton, wet with liquid 

 air, was now pressed gently against the outer wall of the 

 bulb and a deposit of sodium made at low temperature by 

 brushing the flame against the sodium deposits on the sides 

 and top of the bulb. This deposit gives a white specular 

 reflexion like silver. A similar series of spectrograms was 

 then taken with much shorter exposure times. Similar 

 experiments were made with potassium. During the ex- 

 posures the pad of cotton was kept w r et with liquid air for, as 

 has been said, the films break up into granules at room 

 •temperature. Records were also made of the transmission 

 of the films by placing the bulb between the slit and the 

 spark. If the films are thick enough to be opaque to all 

 visible light, they persist at room temperature long enough 

 to make two or three exposures. 



On Plate I. fig. 2 we have some of the records made 

 with potassium. The reflecting power of this metal is very 

 high from the red down to wave-length 3400, below which 

 point it falls rapidly. 



The transmission in the ultra-violet is most remarkable. 

 This film \\ as so dense that a tungsten lamp filament was 

 invisible through it, yet it transmits the entire ultra-violet 

 region below \==3600. This transparency is associated with 

 its low reflecting power in this region. 



Sodium behaves in a different manner (Plate I. fig. 3). 

 While it has a high degree of transparency for the region 

 below 3600, in which respect it resembles potassium, it has 

 -also a high reflecting power throughout the entire range 

 •of the spectrum. 



Rough estimates of the reflecting power based on com- 

 parisons made with the reflecting power of quartz indicate 

 that in the visible region the reflecting power is of the order 

 of 80 to 1)5 per cent., while in the ultra-violet between 

 \ = 2100 and X = 3000 it is of the order of 50 to 60 per cent, 

 Its ultra-violet transparency is not accompanied with low 

 reflecting power, as is the case with potassium. 



So far as I am aware the transparency of these two metals 

 *o the entire ultra-violet region is unique. Silver films, as 



