1014 
elevation may not have exceeded 500 ft. At such times 
the mean temperature of the world would be expected 
to be about 9F higher than now, merely because of 
the lower mean height. It will be seen later that the 
changes of land and sea distribution associated with 
mountain building and erosion bring other factors into 
operation which magnify the effect of elevation. 
The Role of Polar Icecaps in Accentuating Climatic 
Changes. The geographic cycle affects the distribution 
of heat over the globe by means of ocean currents. 
Before discussing this factor we must consider the effect 
of floating ice. If the surface of the ocean were above 
the freezing point of sea water, there would be no ice. 
But if a general fall of temperature brought a small 
area near the North Pole below freezing poimt, a nucleus 
of ice would be formed, and this would act as an addi- 
tional source of cooling, reducing the temperature still 
further. Brooks [6] showed that once the nascent ice 
sheet exceeded a diameter of about 250 miles, this 
additional cooling would exceed the rise of temperature 
due to decreasing latitude, and the floating icecap would 
grow until it nearly filled the arctic basin. The freezing 
point of sea water is 28F. Three independent calcula- 
tions have shown [4] that if the arctic ice could all 
be cleared away, the mean winter temperature of the 
sea surface near the pole would be not far from 24F. 
This is the “nonglacial” temperature. A permanent rise 
of the nonglacial temperature by 5F would eventually 
sweep away the floating icecap and convert the Arctic 
Ocean into an open sea, with only some thin and scat- 
tered floating ice forming in winter and melting in 
summer. 
The present low winter temperature of the Arctic, 
estimated as —40F near the pole in January, is due 
almost entirely to the existence of the icecap itself. 
A small access of heat only sufficient in itself to raise 
the temperature by 5F would result im an actual rise 
of nearly 70F, that is, in a complete change of climate 
which would permit cool temperate vegetation to extend 
to high latitudes. 
There would also be a great, change in the atmospheric 
circulation. The area of relatively high pressure in 
high latitudes is a surface effect, due to the weight of 
cold air in the lowest few thousand feet; at a height 
of 10,000 ft pressure falls continuously from low lati- 
tudes to the neighbourhood of the pole. The polar 
front is the boundary between the surface easterly 
winds associated with this area of high pressure and 
the westerly winds of middle latitudes. It is a reason- 
able inference that the removal of the surface cooling 
due to the floating ice would remove this layer of cold 
air, so that westerly or southwesterly winds would 
extend to much higher latitudes than at present. This 
in turn would accentuate the warming of the Arctic. 
Depressions would still occur, but in the absence of 
marked differences of temperature they would probably 
be weak; the necessary balance between easterly and 
westerly winds would be maintained chiefly by an 
extension of the subtropical anticyclones into higher 
latitudes. 
The differences between completely ice-free and com- 
CLIMATOLOGY 
pletely glaciated polar regions add up to the difference 
between the warm periods constitutmg most of geo- 
logical time, which we may term “nonglacial,”’ and 
the ‘glacial’? periods such as the present, which cul- 
minated in ice ages. From this discussion it follows that 
the basic difference between a nonglacial and a glacial 
period is comparatively small—of the order of 10F in 
the temperature of high latitudes. Such a change could 
be brought about by a general rise of temperature over 
the whole earth, but it could also result from a greater 
transfer of heat from low to high latitudes by ocean 
currents. 
Ocean Currents. At a time of maximum mountain 
building, the continents are most extensive and irregu- 
lar in shape. The free circulation of the oceans between 
low and high latitudes is restricted and relatively little 
heat is carried to polar regions. At the end of quiescent 
periods the continents are small and ocean currents have 
free access to polar regions along several broad channels. 
The present situation in the Northern Hemisphere is 
rather unfavourable in this respect, access to the Arctic 
being possible only by the gap between Greenland and 
Europe, and part of this gap is occupied by the cold 
East Greenland Current, which owes its existence to 
arctic ice. There is no doubt that the Gulf Stream 
raises the temperature of the Arctic Ocean by several 
degrees, but this is not sufficient to keep the Arctic 
free of ice. In warm periods such as the Jurassic or 
early Tertiary, the Bering Strait was more open and 
there was a third channel, the Volga Sea, from the 
Indian Ocean across western Siberia. Brooks [4], on 
the basis of calculations by Kerner [18], estimated 
that the additional supply of heat was sufficient to 
raise the “nonglacial”’ temperature by 10—-13F, more 
than enough to keep the Arctic free of ice. That, given 
the geography of these periods, such currents would 
exist has been shown experimentally by Lasareff [20] 
by means of models. f 
The Late Palaeozoic Ice Age. The greatest problem 
of geological climates is presented by the Late Pal- 
aeozoic—Upper Carboniferous (Pennsylvanian) and 
Lower Permian—in which an apparently highly favour- 
able climate in the Northern Hemisphere, permittmg 
the rich vegetation of the coal measures, coincided 
with or only preceded by a short time very extensive 
ice sheets in low latitudes of both hemispheres. The 
continental drift theory gives a plausible explanation 
of this distribution but as we have seen, it suffers 
from other objections. There remains the possibility 
that it was due to a peculiar distribution of sea, land, 
and mountains. 
In the Late Palaeozoic a large continent, termed 
Gondwanaland, extended across South America, Africa, 
India, and Australia. As a result of a period of mountain 
building, this continent was most extensive and lofty 
in the Pennsylvanian and Lower Permian. To the 
north (Fig. 3) the Tethys Sea, a precursor of the 
Mediterranean, extended east-southeast to open into 
the equatorial Pacific, and was connected by the Volga 
Sea and the Atlantic with the Arctic. To the south 
the Southern Ocean had no direct connection with the 
