92 PHYSICAL SCIENCE 



and the channels bridged over with the flowing 

 film of viscous metal. 



These experiments have an interest which 

 extends further than the immediate subject to 

 elucidate which they were undertaken — an experi- 

 ence not uncommon in physical research. The 

 existence of this viscous metallic film under 

 certain conditions suggests that, when minute 

 quantities of a solid alone exist — when there is in 

 effect inside the surface film no substance beyond 

 the range of molecular action — all crystalline 

 structure must disappear. The initial formation 

 of solid in the body of a saturated solution or of a 

 fused material will, on this view, be co-ordinated 

 exactly with the deposition of drops of water from 

 a mass of air saturated with aqueous vapour, and 

 the possibility of super-saturation will, in each 

 case, depend on the work required to form a 

 new surface of separation under the influence 

 of surface tension alone. It is only when the 

 individual solid structures attain a considerable 

 size that crystalline forms begin to appear. 



