672 TAC CILAMIBIGRGIN: 
ranges. Unfortunately there is not complete unanimity on this 
point... But, if ‘this be*true; as, the temperature: fell thevsolvent 
action of the carbonic acid of the land waters upon the limestone 
was increased. Over against this was a probable reduction of 
the action of organic acids. Probably the decomposition of the 
silicates went on at a lower rate, but, as it was much less than 
one fifth of the whole action, its reduced rate is not very 
material here. 
As the carbonic acid of the air was diminished, its action on 
the land surface declined—though not at a proportional rate — 
but long before it could offset the enlargement of the contact 
area aided by the sea action, glaciation would be far advanced, 
if the previous estimates hold good. 
Setting aside, as being measurably balanced, or as being of 
minor or uncertain value, all but the first three items, which are 
clearly factors of great potency, we find at first a strong disposi- 
tion toward the acceleration of the depleting process. 
But this, although a pronounced influence in the early stages 
of refrigeration, could not continue indefinitely, for the process — 
involved the conditions of its own arrest. 
The arrest of the depletion and the inauguration of the reaction.— 
With the development of glaciation, the agencies that tended to 
counteract atmospheric depletion received a powerful ally in the 
ice-sheet itself., Dhe spread oisithe ice over they sugiaceppie= 
vented further effective weathering of the area so covered and 
correspondingly arrested atmospheric depletion. The total area 
covered by glaciation at its maximum was probably not far from 
8 million square miles, or nearly 15 per cent. of the land surface. 
Besides this, the area outside of that actually covered by the 
ice-sheets was probably affected by prolonged freezing during 
the winter stages, and was perhaps to some extent permanently 
frozen beneath the surface, and this arrested solvent and chemi- 
cal activity. If the modest figure of 5 per cent. be assigned for 
this supplementary effect, 20 per cent. of the functional area 
wovld be withdrawn from action. Whether this numerical esti- 
mate be correct or not, it may be assumed that if a given amount 
