100 Prof. P. W. Wood on Optical Properties of 



application of the flame, and the clean spot cooled by the- 

 application of a small pellet of cotton attached to the end of 

 a glass rod and wet with liquid air, the deposit which forms 

 on the cold spot, when the rest of the surface is brushed over 

 with the flame, is homogeneous and quite opaque to visible 

 light if sufficient metal is vaporized. Thin deposits are 

 blue-grey by transmitted light, the colour being quite 

 different from that of the blue granular films. As soon as- 

 the bulb warms up to room temperature the opaque film 

 speedily disappears, the spot becoming more or less trans- 

 parent, and if viewed from the outside in a strong light 

 shows the peculiar pearly lustre characteristic of deposits of 

 moderately coarse granules. If the film is thin enough to 

 transmit much light it melts away in a few seconds, if thicker 

 it lasts for some minutes. The phenomenon of its disappear- 

 ance can be followed under the microscope. Minute holes^ 

 appear which are mere pin-points at first, but speedily 

 enlarge into forms resembling bird tracks. These continue 

 to grow into intricate patterns, the metal finally gathering 

 into small patches of irregular outline. A series of micro- 

 photographs of the breaking up of the film are reproduced 

 on Plate I. fig. 1. The first picture, "a," was taken as 

 quickly as possible after focussing the microscope, " b " was 

 taken about two minutes later, " c " at the end of five 

 minutes, " d " half an hour, " e " one hour, and " f ,} at the 

 end of two hours. This was a very thick film quite opaque 

 to visible light. Thinner films change with much greater 

 rapidity. 



The phenomenon is evidently related to surface tension, 

 the cohesion between the metal and the glass being in- 

 sufficient to prevent the film from drawing itself up into 

 droplets. 



The reflexion and transmission of the films was investigated 

 with a small quartz spectrograph. An iron-cadmium spark 

 was placed close to and a little to one side of the slit, and 

 the quartz bulb mounted directly in front of the slit at a 

 distance of 20 cm. A sodium film was formed on the wall 

 of the bulb furthest from the spectrograph, and the instru- 

 ment pointed in such a direction that the real image of the 

 spark formed by the concave sodium mirror was seen 

 exactly at the centre of the field, when the eye was placed at 

 the focus of the instrument. There is no difficulty in dis- 

 tinguishing this image from the one reflected from the 

 convex wall of the bulb, as it is highly coloured (purple if 

 the film is moderately thick). The iris diaphragm of the 

 spectrograph in front of the collimator lens was now 



