806 THE JOURNAL OF GEOLOGY. 



formed not so very long ago. But the gravel is evidently of glacial 

 origin, being of angular and subangular pebbles, of great variety 

 of material. The conclusion seems inevitable, therefore, that 

 these deltas were formed during the retreat of the ice-sheet. 



4. Stagnant, melting ice. — In the retreat of the ice-sheet there 

 were parts at least which became too thin to move. As Professor 

 Davis has said : 



During this time it must have melted irregularly, presenting a very uneven, 

 ragged front, from which residual blocks may have been frequently isolated ; 

 and it must have endured longest in the valleys, where it was thickest, not 

 only by reason of its greater depth, but also because its surface there, where 

 motion had been fastest and longest maintained, must have been higher than 

 on the hills — this being homologous with the variation in the thickness of a 

 Swiss valley glacier from middle to sides." 1 



It seems to me that we must consider the change to have 

 been gradual from a moving glacier to a stagnant one, and that 

 there may have been times of renewed activity with a forward 

 motion, even in the period of decline. Such forward motion may 

 have had some influence in shifting the course of esker rivers 

 and so have determined where the next sand-plain was to be built. 

 So far as I know, this point has not been worked out in the field. 



Crevasses are formed as the ice moves, and change their posi- 

 tion according to the tensions in the mass of the glacier. When 

 the tension from motion has ceased, and the ice has become a 

 diminishing, drift-covered mass, the condition represented in 

 Fig. 2, we should not expect to find any crevasses remaining. 

 They would either have been closed by the forward motion of 

 the ice, or would have lost their distinctive character by the 

 excessive melting of their sides, while the water would have 

 washed detritus into them covering the underlying ice, and pre- 

 venting it from melting as fast as that on either side. Such 

 protection of the ice by detritus must have had great influence in 

 determining the surface forms of the stagnant ice-sheet, as is 

 shown in Professor Russell's account of the sand cones and the 

 deposits in glacial lakelets. 



I Bull. Geol. Soc. of Am., Vol. I., p. 196. 



