428 R. D. SALISBURY 



three times as great as that of I. By the time the same ice has 

 spread over 3, the basal layer of drift derived from I will have 

 been reduced to one fifth its original thickness, since the area of 

 3 is five times the area of 1. Still supposing none, of it to have 

 been deposited when 4 has been overspread, the thickness of 

 the drift derived from I when the ice has covered 4, will be but 

 one seventh of that which it possessed over 1. Even supposing 

 all the drift from I to be carried to 4 over the intervening areas,, 

 it is thus seen that at such a point as c, the amount of distant 

 material should be relatively small. If the ice keeps an equal 

 amount of drift in and beneath itself by gathering enough new 

 material to counterbalance the thinning of that previously held,, 

 the till at c should be made up as follows : 



From formation 1 - - - - - - -fV 



From formation 2 ------ -fV 



From formation 3 - - - - - - -fV 



From formation 4 ------ tV 



Thus the spread of the ice will tend constantly to decrease the 

 proportion of basal drift derived from any formation, once that 

 formation is passed. 



Since some of the drift derived from I would doubtless be 

 lodged as the ice advanced over formation 2, the average thick- 

 ness of that carried from 1 to 3 will be correspondingly 

 diminished beyond the figure given (i of that over 1). Since 

 more of the material from 1 will in all probability be deposited 

 on 3, as the ice advances to 4, the figures given for the thick- 

 ness on 4 of drift derived from I (\ of that on 1), will need to 

 be still further reduced, and that by a larger amount than that 

 to which the \ was subject. In view of the continual lodgment 

 of drift, it will be seen that the series of fractions given above,, 

 namely Jg, T 3 6 , f^, and ^ will need to be changed by some unde- 

 termined amount, but so that the extremes well be farther apart,, 

 and the differences between successive members greater. If -| of 

 the drift carried from I, were deposited on 2, 3, and 4, and \ of 

 that from 2 on 3 and 4, and \ of that from 3 on 4, the last mem- 

 ber of the preceding series of fractions would be increased at the 



