38 THE ROYAL SOCIETY OF CANADA 



whole floor of the chambers had been flooded to a depth of about four 

 feet. The ice which formed on the surface was quite clear enabling 

 observations to be made of the condition of the under side of the ice. 

 Direct observations were commenced on Feb. 17th of the system of 

 cracks which had developed over the surface. In order to avoid 

 any serious effect of ice expansion, a long channel had been cut through 

 the ice. Over this channel a cover of ice blocks had been placed, which 

 subsequently became cemented by freezing to the general surface. 



In Fig. 3 we reproduce a sketch to show the position of the 

 cracks with respect to the sides of the chamber and the pillars. When 

 the temperature rose above the freezing point these cracks closed and 

 disappeared, except where the upper edges had become chipped or 

 rounded. A succeeding hard frost opened a new system of cracks 

 similar to them. On Feb. 17th, and at the times when the tempera- 

 ture was well below the freezing point, the cracks started and opened 

 as a rule from the pillars, and intersected, like the meshes of a net. 

 Where blocks of ice had been removed from the channel and placed 

 on the surface and became frozen to it, the cracks were observed to 

 run directly through them without being deflected. This showed 

 us that the tensile force of the ice which was exerted by the contrac- 

 tion of the general surface, acted in such a way as to split a thick 

 block of ice, not contributing to the contraction. 



All these cracks were observed to be wedge shaped, ending be- 

 fore the under surface of the ice was reached. This showed us that 

 there was unequal contraction from the surface downwards. The 

 breadth of the cracks seems to bear a close relation to the tem- 

 perature. 



Table II. 



Size of Cracks in Relation to Air Temperature. 



In examples measured there was only one crack in the 19 ft. 

 between the pillars. 



Opening of crack. 

 Feb. 17th Temperature 8°F. -130 inch. 



" 26th " 21°F. -079 " 



Mar. 12th " 22 °F. -063 " 



We show in Fig. 4 a curve of the linear contraction of the ice 

 from Andrews' measurements. The range of breadth of some of the 

 cracks measured at various temperatures is indicated also. This shows 

 us conclusively that the opening and closing of the cracks is affected 

 by the temperature changes. 



