GLACIERS OF THE CANADIAN ROCKIES AND SELKIRKS. 127 



pletely liquified condition. That they may have been melted partly upon 

 one face and frozen upon another, or the water derived from one by melting 

 added to another granule, is in harmony with known properties of ice. Pulfrich 

 found that when an ice crystal was pressed against a wet surface of glass and 

 allowed to freeze the water between the ice and plate was incorporated into 

 the crystal, so as to make a homogeneous mass. 



Much interest is attached to the methods of granular development, since the 

 more modem theories of glacial movement are more or less dependent thereon. 

 It has been shown by Emden, Drygalski, Crammer and others that when n€v€ 

 granules are made into a water slush, such as might originate from excessive melt- 

 ing, or heavy rainfall, the granules grow in size and, under favorable conditions, 

 quite rapidly. In the nevd as well as in the body of the glacier, however, there 

 must be a maximum limit which the granules may attain by this means for the 

 ice will presently become too compact for more water to enter and no space will 

 be left for the growth of individual crystals.' That further growth of the granules 

 does not take place by the simple freezing together of neighboring granules, is 

 conckisively shown by the homogeneous structure of the mature granule. That 

 new granules cannot originate by the complete and simultaneous melting of a 

 number of adjacent smaller ones, is believed to have been just shown in the 

 preceding paragraph. Of the various theories of granular growth remaining 

 we may recognize three divisions, based upon evaporation, melting, and "dry 

 union." 



First. — It has been shown by Chamberlin and Salisbury that in dry granular 

 snow, kept continually below the freezing temperature, certain granules will 

 grow in size at the expense of the others, presumably by the giving off and con- 

 gealing of water vapor.^ In the porous snow of the ndve it seems probable that 

 the principle would be operative and that the granules would diminish in numbers 

 and increase in size, even when not immersed in water. For the body of the 

 glacier, with the granules in such intimate contact, the authors do not believe 

 that evaporation and condensation can take place to any appreciable extent. 



Second. — -Making use of the principle of Thompson that ice may be melted 

 by pressure, without any change in temperature, many investigators, as Mtigge, 

 Drygalski, Chamberlin, Crammer, etc., have accounted for the growth of the gran- 

 ules in the main body of the glacier by assuming a partial or complete melting 

 and refreezing. Those granules which owing to their location are subjected to 

 the greatest pressure, or internal friction, or those portions of granules similarly 

 affected will melt, thus redistributing the pressure and allowing the free molecules 

 to attach themselves to the most favorably located granules. The liquefaction of 

 the granules may be confined to their outer surfaces, or, as Drygalski believes, take 

 place locally in the bodies of the granules. Chamberlin believes that the granules 



' See Hagenbach-Bischoff's criticism of Forel's infiltration theory, "Weiteres Uber GletscViereis," Ver- 

 handlungen der Naturjorschenden Gesellschaft in Basel, viii, 1889, p. 822. 



^Geology, vol. i, Chamberlin and Salisbury, p. 296 (Chamberlin, Peet, and Perisho). "A Contribu- 

 tion to the Theory of Glacial Motion," Chamberlin, Decennial Publications of the Univ. of Chicago, vol. 

 IX, p. 194- 



