the Hintereis Glacier. 157 



will be s" tan 0, and the ratio of the former to the latter 

 is 4 : tan 0, depending upon the angle only ; so that 

 the total resisting area is independent of the size of the 

 projections. 



According to this hypothesis the resistance of a rock 

 surface to the slip of a mass of ice resting upon it will be 

 proportional to the sum of the cross-sections of the projections 

 on the surface, provided that the pressure holding the surfaces 

 together is sufficiently great to prevent the ice from rising- 

 bodily up their inclined surfaces. 



If we regard the rate of slip as proportional to the force 

 parallel to the surface, and as being independent of the size 

 of the projections, it is possible to calculate the angle of the 

 sides of the conical protuberances which would allow a 

 glacier to slip at known speeds. 



Tarr and Rich found that the velocity with which the wire 

 cut through the ice was 0*0107 cm. per sec. under a load of 

 70*32 kg. per cm.-. These experiments were made with 

 metallic wires. Experiments have, therefore, been made 

 with glass, which has a coefficient of conduction somewhat 

 similar to rock. 



Fig. 5 shows the apparatus employed for the experiments. 

 a is a bar of wood extending from side to side of a well 

 insulated box. Below this bar, and resting upon laths, was 

 placed a slab of ice about 1*5 in. thick. Upon the bar a was 

 placed the piece of ice b to be experimented upon, and across 

 it was fixed the glass rod which had a diam. of 0*7 cm. 

 From the glass rod hung a sling of stout wire coated with 

 sealing-wax, the wires passing through holes melted in the 

 ice and then through the bottom of the insulated box. 

 Covering the bar b was a hollow cover of ice, enclosing 

 the ice bar b and the w r ooden supporting bar a, and resting 

 upon the ice slab below. The box was fastened to a wall at 

 a convenient height, a large piece of ice placed on the top, 

 and the whole covered with flannel. To the sling carrying 

 the weight was secured a lever multiplying 8*3 times. The 

 readings of this lever registered the result of the melting of 

 the lower side of the ice bar as well as the fall of the glass 

 rod, but the error introduced by melting in contact with the 

 wood could not have amounted to more than a Aery small 

 percentage of the total. 



After being fixed the apparatus was allowed to stand for 

 about three hours with the load on the glass, after which 

 readings were taken. When, after the experiments, the box 

 was opened only a very small amount of melting was round 

 to have taken place in the ice cavity surrounding the ice bar 



