266 



R. M. Deeleij — The Viscosity of Ice. 



fed by snow falling evenly over its surface, llecent exploration has 

 shown that Kelvin's conclusions are not very wide of the mark. He 

 does not, however, give any detailed figures, and we are therefore in 

 the dark concerning the figure ho took for the viscosity of glacier ice. 

 1 have recalculated the mean hourly extensions given by Main, as he 

 had done this somewhat roughly, and in some instances had misplaced 

 the decimal points. 



The samples of ice tested were cylindrical, having been formed by 

 freezing water in a metallic mould. In experiment IS'o. 1, Table I, 

 the water was not freed from air, and the test bar contained minute 

 bubbles radiating in horizontal straight lines from the axis of the 

 cylinder. In experiments Nos. 2 and 3 the water was first frozen, 

 then boiled and refrozen in the mould, and only a core of minute 

 bubbles up the centre of the cylinder remained. 



The ice was secured in conical holders, and the extension was 

 measured on lengths of from 216 to 244 mm. To prevent rapid 

 cliauges of temperature the tests were made in two wooden boxes. 

 That the temperatures read are correct is probable, for the thermometer 

 was tested at Kew. 



