BALCH— THE COUDERSPORT ICE MINE. 



555 



ing down (Fig. 2). These were melting, as was also the ice floor, 

 the glaciere in fact being in a state of thaw, with the thermometer 

 several degrees above freezing point. 



Fig. I. Coudersport Ice Mine. Entrance, with ice above. 



The ice, it is said, begins to form about April and to be at its best 

 perhaps in June. After this it slowly diminishes and vanishes by 

 about October. The ice goes quickest in rainy weather and more 

 slowly in warm weather. Both these times of the appearance and 

 disappearance of the ice and these effects of wet or dry weather are 

 normal glaciere phenomena. 



The only theory about the formation and the disappearance of the 

 Coudersport ice which meets all the facts is the theory which applies 

 universally to all glacieres. Two things are necessary for the forma- 

 tion of ice : cold and water. In glacieres the cold of winter furnishes 

 the cold and the thaws of spring furnish the water. That the win- 

 ter's cold furnishes the cold is proved by the fact that every known 

 glaciere is in a place where there is snow and ice in the open in 

 winter. The winter air sinks from its weight into the glaciere and 



