77 



densed by a double convex lens, was transmitted. At the moment 

 the beam crossed the transparent solid, the track of the beam became 

 instantly starred by little lustrous spots, like shining air-bubbles. 

 Round each of these a figure, shaped like a flower of six petals, was 

 formed. The petals were manifestly liquid water. When the beam 

 was permitted to traverse different portions of the ice in succession, 

 the sudden appearance of the stars, and the formation and growth of 

 the flowers around them, could be distinctly observed through an 

 ordinary pocket lens. 



To test whether the brilliant spots at the centres of the flowers 

 contained air or not, portions of ice containing them were gradually 

 melted in warm water. The moment a liquid connexion was esta- 

 blished between the cavities and the atmosphere, the bubbles col- 

 lapsed, and no trace of air rose to the surface of the water. The 

 formation of each liquid flower is therefore accompanied by the 

 formation of a vacuum at its centre. 



The perfect symmetry of these flowers at once enables us to infer 

 that ice is a uniaxal crystal, the line perpendicular to the planes in 

 which the flowers are produced being the optic axis. 



For a long time during the investigation it was found that the 

 flowers were formed in planes parallel to those of freezing ; but some 

 apparent exceptions to this rule were afterwards noticed, which are 

 described in the paper. 



In some masses of ice, apparently homogeneous, the flowers were 

 formed on the track of the beam, in planes which were in some cases 

 a quarter of an inch apart. This proves that the interior portions 

 of a mass of ice may be melted by radiant heat which has traversed 

 other portions of the mass without melting them. 



In a second section of the paper the author describes the gradual 

 liquefaction of masses of ice by the formation of drops of water within 

 them ; and he infers from his observations that the melting-point of 

 ice oscillates within small limits on each side of the ordinary standard. 

 Through weakness of crystalline texture, or some other cause, some 

 portions of a mass of ice melt at a temperature slightly under 32 

 Fahr., while others of stronger texture require a temperature slightly 

 over 32 to liquefy them. The consequence is, that such a mass, 

 raised to the temperature 32, will have some of its parts liquid and 

 some solid. 



