110 Prof. R. W. Wood on Optical Properties of 



accuracy the dependence of reflecting power on wave-length. 

 The other curves are more or less qualitative in character, as 

 a result of the use of ray filters having rather broad bands 

 of transmission. 



Further increments in thickness cause the absorption band 

 to move up into the red, giving a blue reflexion ; but now 

 we begin to have a diffusion of white light, and if the 

 thickness is still further increased, the amount of white 

 diffuse reflexion increases, the film resembling a matt surface 

 of silver, such as is obtained by electrolytic deposition. AVe 

 now have optical evidence of the crystalline structure of the 

 film, a circular " rainbow " halo appearing around the 

 regularly reflected image of the source of light. 



Films showing this effect are best made by gradually 

 driving the deposit around to the further side of the bulb by 

 .playing the small flame continually against the edge of the 

 deposit, keeping the front of the bulb clean by occasional 

 brushing with the flame, and examining the nature of the 

 reflexion every few seconds. A deposit made in this way 

 will be thickest along the edge, and if properly made will 

 show a rainbow of great intensity mixed with very little white 

 light. 



The rainbow is best seen with the light incident normally 

 on the centre of the metallic patch, which can be accom- 

 plished by standing with the back to the light. It will be 

 found that the rainbow is verv sensitive to changes in the 

 angle of incidence, and can be made to disappear by a slight 

 rotation of the bulb. 



From considerations of the angle of incidence at which 

 the rays meet the oblique surface along the curve of the 

 coloured halo, it seems probable that we are dealing with 

 minute octahedral crystals of sodium. 



Moderately thick deposits show also a colour by trans- 

 mitted light which is very sensitive to small changes in the 

 angle of incidence. The colour sequence for transmission, 

 as the thickness increases, is blue, blue-green, green, and pale 

 red. This latter colour is very sensitive to changes of the 

 incidence angle, a rotation of the bulb of a few degrees 

 causing its disappearance. One feels as if the structure 

 concerned must be of the nature of vertical spicules or 

 needles somewhat as shown at the right-hand part of fig. 1, 

 or it may even be possible that a dendritic type of crystalli- 

 zation occurs, as is frequently the case with metals. 



I have never succeeded in producing these colours except 

 with the alkali inetals, though an analogous phenomenon can 



