Davis.] 18 [Nov. 16, 
SO soon ti) (lirt'crent estimates of the value of various processes 
tliat, from this point omvanl, our views liave little similarity. 
We differ again as to the correctness of the inference that can 
he drawn from the lieavily drift-covered margin of the Malaspina 
glacier of Alaska, as bearing on the condition of the ice surface 
near the margin of our waning glacial ice-sheet. It seems to me 
that the supply of the Malas])ina glacier from among high moun- 
tains, whence it descends along deep narrow valleys, may largely 
explain the occurrence of a large amount of englacial drift in its 
middle layers ; and that the melting of the upper ])art of the ice 
will transform the englacial drift to superglacial drift near the ice 
margin, without calling on an oblique ascent of the drift from 
the bottom of the ice. The ice of the Greenland sheet is not so 
heaAnly drift-covered near its margin as that of the Alaskan 
sheet; and while I understand that Mr. Uphamexjdains this pecu- 
liarity by contending that the Greenland ice is in a state of increase, 
instead of decrease, it seems fair on the other hand to explain a 
considerable sliare of the difference between the two cases by the 
absence of high mountains on the Greenland plateau. 
Mr. Upham objects to the derivation of our druinlins from the 
half-wasted moraines of an earlier advance of the ice, because the 
(b-umlins do not contain stratified and washed gravels ; yet he 
advances the theory that the superglacial drift, leached out from 
the ice by surface melting, gathered from a large area of ice sur- 
face and concentrated on a smaller one by water action, shall be 
transformed into drundins of true till, by a re-advance of the ice. 
His objection appears to be as aj^plicable in the latter as in the 
former case. Moreover, as the whole mass of the accumulated and 
concentrated surface drift miist be worked over, part by part, by the 
ice before it can gain the structure of till, I do not see the useful- 
ness ot the concentration, except as affording a possible excessive 
su])])ly of ready material for ice dragging, of which more below. 
Me maintains further that the newer advance of the ice will climb 
over the wasting ice of the earlier advance, and that drumlins 
may therefore often have been formed with a layer of older ice 
beneath them, and a cover of heavier, thicker ice above them ; 
a<hluciiig in evidence of this,*lirst, the theoretical requirements of 
the process that he describes ; second, the occasional occurrence 
of cracks or crevices in the till of our drumlins, more or less 
lilleil in by sand, as if the cracks had been made when the under 
