686 ME. HOPKINS ON THE THEOET OE THE MOTION OE OLACIEES. 
10. Interior Temperature of Glaciers . — There are two obvious causes by which the 
temperature in the interior of a glacier may be affected— (1) conduction of heat from 
the superficies of the mass according to the ordinary laws of conduction through solid 
bodies, and (2) infiltration of water from the upper surface. We may consider sepa- 
rately the operation of these causes, with the view of determining whether the lower sur- 
face of the mass is permanently at the temperature of freezing, as it must be to render 
the preceding experiments strictly applicable to account for the continuous and unaccele- 
rated motion of the glacier. To investigate the effect of the first cause, let us conceive 
the whole surface of the earth covered with a superficial crust of ice. The temperature 
of this crust will be subject to periodical annual variations to a certain depth, which 
will depend on the annual variations of the superficial layer of the icy crust, and the 
conductivity of the ice. It may be considered, for our immediate purpose, sensibly 
the same as the temperature of the glacier itself, at all points where the glacier and 
our imaginary crust of ice coincide. Now the surperficial changes of temperature in the 
crust of ice subject to conditions similar to those of a glacier, would be much less than 
those which the actual rocky crust of the earth experiences; for the temperature 
of the ice could never rise above 32° Fahr. in the hottest summer, nor in the coldest 
winter could it probably fall many degrees below that temperature, on account of the 
covering of snow like that which on all glaciers protects their outer surface against the 
effect of low winter temperature. Again, the depth through which these oscillations 
of temperature would be perceptible, would, cceteris paribus, be comparatively small if 
the conductive power of the mass should be so. I am not aware of any experimental 
determination of the value of this power for ice ; but that substance is known to be a 
very bad conductor, and probably worse than the average of the rocks which form 
the outer crust of the earth. For both these reasons, then, the depth of oscillatory 
annual temperature would be much smaller than it is found to be within the actual 
crust of the globe, under the same external climatal conditions. Now in the eai'th’s 
crust, and in our own latitudes, this depth may be approximately estimated at 70 
or 80 feet, according to the nature of the upper strata. I should therefore deem it 
probable that the variations of external temperature in a crust of ice like that above 
supposed, or therefore in an ordinary glacier, would not exceed at most perhaps some 
30 feet. 
Again, let us consider the probable mean annual superficial temperature of our hypo- 
thetical icy crust, or of a glacier of ordinary dimensions. The actual temperature could 
never rise, as above remarked, above 32° Fahr., and would never sink many degrees 
below that temperature. M. Agassiz has left us the only reliable observations on this 
subject^. He buried a self-registering thermometer in the glacier of the Aar, at the 
depth of 2T metres. It was taken up two years afterwards, and was foimd to have 
registered a minimum temperature of — 2°T (C.) =28°’22 (Fahr.). Thus the mean 
temperature for the winter was probably not less than 30° (Fahr.), and that for the 
* Systeme Grlaciaire, p. 425. 
