118 W. J. McGee—Theory of Glacial Climate. 
factors progressively lose in relative efficiency, the fifth increases, 
the share of heat derived from warmer latitudes diminishes, and 
the periodicity of solar accession becomes more equable, whereby 
loss through radiation is accelerated. From this complex of 
diverse and antagonistic elements only the most vague estimates 
of the relative rates of addition and loss in higher and lower 
latitudes could be directly deduced without exhaustive analysis 
and computation ; but it is certain that the annual addition to 
the ice-sheet could never exceed the precipitation, while it is 
obvious that the annual loss must fail of the addition; whence 
the foregoing value, if doubled or tripled, and certainly if quad- 
rupled, would be ample for the whole glaciated area of the 
northern hemisphere. 
n 
the initial development of an ice-sheet, would then reach maxi- 
mum efficiency. The enormous dissipation of heat by icy sur- 
faces is seldom adequately appreciated: after a light snow-fall 
equal to but a fraction of an inch of ice, in the upper Mississippi 
valley, the temperature falls from freezing-point to zero, and the 
snow is not even softened by a day’s uninterrupted sunshine 
demonstrably sufficient to melt an inch and three-quarters of 
ice; the névé-fields of the Savoyan Alps receive enough solar 
energy in a year to melt 54 feet of ice, yet the actual superficial 
liquefaction must be trivial; an earlier paragraph indicates that 
less than a fortieth of the theoretical melting actually occurs in 
the frigid zones; the solar accession in the frigid zone in sum- 
mer is considerably greater than at the equator, as Meech has 
calculated,” yet the liquefaction annually effected there would 
be effected in a week were the available energy utilized in such 
work; it appears susceptible of mathematic proof that if the 
water of the earth were converted into a mantle of ice uniformly 
% “Relative Intensity of the Light and Heat of the Sun,” Smiths, Contrib. 
Knowl., 85, 1855 (=vol. ix, 1856), 18, pl. L 
