Shaler. J 
164 
[March 16. 
then the mass will creep faster, and the stream poured into the 
glacier will he deeper and of more rapid flow. If the supply he 
scant, in the course of years there will he a diminution in the 
supply working downwards into the glacier. Now these varia- 
tions will he differently timed in each glacial system. In regions 
of small and deep rivers lying on steep slopes, the effect of a 
series of extraordinary seasons will he more immediately mani- 
fested than in those fed hy hroad, thin fields of snow lying upon 
low angles of slope. It is easy to see that years or decades 
may separate the periods when accident, accumulations or depri- 
vations of snow will make themselves felt in the variations of 
supply of ice to the lower glacier. When the ice enters upon its 
lower path helow the line of perpetual snow yet other perturbing 
influences assail it. If the stream he of great length relatively 
to its depth, then rainfall and the sun’s heat in warm summers will 
have more effect upon it than upon those glaciers that present 
opposite conditions. 
The accidental occurrence of a single warm rain in any valley, 
may so far affect the depth of the ice over the glaciers that lie 
within it, that their terminal points may he made to retreat. Yet as 
this rain in the region of perpetual snow is added to the thickness 
of the neve, it }M’ovides materials for a future extension of the 
ice. A careful study of the relations of deposition and melting to 
the rate of movement of the ice, and the position of the terminal 
points of the glacier, would probably show many causes of varia- 
tion, which would still further help us to account for the varia- 
tions of advance and recession which at first sight appear to he so 
puzzling. But I have given reason enough to make it clear that the 
position of the terminus of any glacier depends upon a very com- 
plicated question, and cannot he readily and directly referred to 
the immediate changes of climate in the given region. 
There can he no doubt that glaciers are, though in a very ob- 
scure way, singularly delicate measures of rainfall. These neve 
regions retain their snows for ages, and only yield them slowly to 
the glaciers beneath. If we try to get metric results from their 
conditions, we must not expect to find them manifesting the 
changes in the lower levels of their streams. If we are to use 
them as indices of change, we must go nearer to the snow-fields 
