ANTARCTIC ATMOSPHERIC CIRCULATION 
would be the result (although glaciologists [119] esti- 
mate that less than a third of the present Ross Shelf Ice 
is derived from glacier outflow). 
Snow and ice flow away more rapidly from the sea- 
ward faces of the coastal mountains than from their 
backs, “so that after a time the ice-shed lies inland 
from the mountains,” a phenomenon observed in Ant- 
arctica and deduced for the Pleistocene ice sheets of 
Scandinavia and Canada. Behind this divide, any in- 
flowing air is descending, so in the interior “‘precipita- 
tionis small and...theicein that region must have largely 
accumulated at a time when the Ross Barrier did not 
exist.” 
Finally, Manley reasoned, glaciers from the coastal 
mountains so cool the surrounding ocean that the open 
water, moisture source for the precipitation, 1s found 
farther and farther away, and precipitation decreases. 
This involves a cycle perhaps several centuries long: 
heavy coastal precipitation, glacial outflow, sea cooling 
and freezing, reduced precipitation, starved glaciers, 
less sea ice and warmer seas, increased precipitation. 
Manley’s glaciation cycle is self-generating, as op- 
posed to others which require some external change, be 
it variation in solar energy, in mountain heights, or in 
submarine topography affecting ocean currents. Simp- 
son has maintained consistently [138] that “the glacial 
epochs were the consequence of an increase and not a 
decrease in the solar radiation,” and that ‘higher tem- 
peratures and greater general circulation of the at- 
mosphere are absolutely essential to produce increased 
outflow of ice in the Antarctic.” 
Kidson [74] questioned this precept, because “in 
temperate, and probably also antarctic, regions of the 
southern hemisphere, precipitation is heaviest in 
winter,” so that “a marked fall of temperature would 
be necessary to produce the conditions at the maxima 
of glaciation.” Flint [7], who has reviewed Pleistocene 
glaciation in all parts of the world, poited out this 
conflict. His excellent treatise [123] expresses the pre- 
vailing view of geologists that Pleistocene glaciation was 
due to a general lowering of the earth’s mean tempera- 
ture, presumably by reduction in solar radiation, so 
that a greater proportion of the earth’s precipitation 
fell as snow. 
Modern meteorological principles led Willett [139] to 
support Simpson: “The general circulation and world- 
wide precipitation are greatly intensified during an ice 
age, results which cannot be brought about by a general 
lowering of the earth’s mean temperature.” From a 
detailed study of the present temperature and precipita- 
tion regime along 49°N, Leighly [136] concluded that 
the North American glaciation was due to increased 
wintertime transport of warm moist air from the Gulf 
of Mexico to the St. Lawrence region, such as obtains 
with low-index conditions. 
The various hypotheses about Pleistocene atmos- 
pheric processes, summarized extensively by Willett 
[139] and more briefly by Landsberg [135], do not help 
materially to define Antarctica’s circulation; instead, 
they require its independent elucidation for their full 
development. In one way, however, they have removed 
935 
one of the unknown variables from the circulation 
equation: production of precipitation in the continental 
interior is not a requirement for an acceptable descrip- 
tion of Antarctica’s present circulation. 
CONCLUSION 
Antarctica’s atmospheric circulation has been de- 
seribed in terms of various versions of a polar anti- 
eyclone—glacial, laminar, cellular, peripheral, and ab- 
stract. Their review in the light of present knowledge of 
topography and precipitation distribution, of pressure 
waves and oceanic storms, of upper-air temperatures and 
surface temperature and pressure regimes, shows first 
that Antarctica’s circulation cannot be described ade- 
quately by its annual pattern alone. 
Variations in the extent of ice around the continental 
edge, effectively expanding the land area in winter and 
spring and decreasing it in summer and fall, have pro- 
found influence. More important are the changes from 
continuous summer sunshine to continuous winter dark- 
ness with outgoing radiation. Because the snow surface 
absorbs little solar energy but readily radiates away its 
own heat, winter cooling is primarily at the surface 
while summer heating occurs im all layers, beginning in 
the stratosphere. 
Throughout the year, the planetary pattern causes 
pressures to be lower around 60°S than at the Pole; the 
lower temperature of the continental snow surface as 
compared to the subantarctic waters intensifies the sur- 
face pressure gradient. But in the upper air, as already 
mentioned, the gradients apparently reverse with the 
seasons: temperature and pressure in winter decrease 
from middle latitudes to the Pole, while in summer 
their decrease stops at about Antarctica’s margin and 
over the continent they increase poleward. The insola- 
tion which warms the lower stratosphere over Little 
America to almost —30C in November and December 
should cause even higher temperatures farther pole- 
ward, so that on the average Antarctica is overlain by 
a deep warm semipermanent anticyclone in early 
summer. 
By late summer, insolation is no longer able to main- 
tain midsummer temperatures in the upper air; pos- 
sibly the ozone concentration decreases markedly by 
midsummer, stopping the warming. The upper layers 
then cool slowly but steadily so that by midwinter the 
lapse rate in the stratosphere approaches that of the 
troposphere and the tropopause disappears. Cooling 
lowers all the isobaric surfaces, creating over the sur- 
face anticyclone an upper-level cyclone which deepens 
as the cooling progresses; at the surface, pressure falls 
during fall and winter to a September minimum. Spring 
sunshine affects the stratosphere first, warming it rap- 
idly; the cyclonic layers expand and rise, and by sum- 
mer the cyclone disappears, with surface pressure rising 
rapidly. 
In summer and winter, the surface distribution may 
have the form of a large central anticyclone, or one 
with several pronounced lobes, or several cells around 
a weak center. During periods of high zonal index the 
polar anticyclone is apparently central, during low- 
