ANTARCTIC ATMOSPHERIC CIRCULATION 
meridional temperature gradient is strongest in winter 
south of 55° or 60°S, and also north of 30°S, and weakest 
in winter between 30° and 55°S. 
Atmospheric pressure around Antarctica is excep- 
tionally low; readings as high as 1000 mb are unusual 
and at Little America on 24 September 1940 sea-level 
pressure was only 932.5 mb. Meinardus [14] distin- 
guished two types of annual pressure variation: a deep 
minimum in the Ross Sea area in late winter or early 
spring (July to October) with a very steep rise to a 
sharp maximum in midsummer (December or January), 
and a double period with minima around the equinoxes 
and maxima near the solstices in the Palmer Peninsula 
area. Whether this difference is regional, latitudinal, or 
MAY|JUN 
Fre. 3—Mean surface air temperature for 10-day periods 
at Little America, based on three years of data. 
due to the differing years of observation is not certain; 
the more recent data from the Palmer Peninsula [36, 
37, 40] show no pronounced pattern. In the Ross Sea, 
the lowest monthly values at Little America have al- 
ways come in September, while in McMurdo Sound each 
year has had its minimum in a different month. 
Antarctica is the only part of the world where the 
atmosphere appears to have less oxygen than ‘‘normal.” 
Everywhere else oxygen represents 20.95 per cent of the 
dry constituents of surface air, but the only analyses 
made of antarctic air [58] gave an average of only 20.56 
per cent in summer, and 20.64 in winter except for one 
day (22 July 1940) which averaged 20.50. Although two 
leading authorities on atmospheric composition [132] do 
not regard this deficiency “‘as well established, since no 
check analyses with normal air were carried out,” 
these are the only analyses to date and their implica- 
tions must be considered until they are proven definitely 
erroneous. 
‘Atmospheric oxygen is being removed constantly by 
rock weathering, while burial of organic material results 
in a net gain. Rock weathering in Antarctica, however, 
seems insufficient to maintain the observed deficiency, 
even if all the presently unknown continental areas 
contained bare rocks undergoing weathering. A possible 
link between antarctic air and that of the middle- 
latitude stratosphere, which was considered in 1936 
to have an oxygen content of about 20.50 per cent, has 
been destroyed by recent indications [129] ‘that the 
chemical composition of stratospheric air at 70 km is 
practically the same as that of the troposphere.” 
933 
PRECIPITATION 
Precipitation in Antarctica cannot be measured as it 
falls, but its total accumulation over a year can be 
determined by measuring the daily changes in height of 
the snow surface, using a graduated stick imbedded in 
the snow. At Little America during 1940 [47], such a 
program, involving four pairs of sticks several hundred 
feet distant in four directions from the camp, showed a 
total accretion during 348 days of 170 cm due to new 
snow, hoarfrost, or drift snow, and total decrease of 85 
em due to ablation, deflation, or settling. The net in- 
crease of 85 cm had an average density of 0.35 and there- 
fore represented a net accumulation of 380 ecm or 12 
inchesof water during 1114 months, mest of it in March, 
June, and July. 
In the western Ross Sea the 1902-1909 average 
(computed from the accumulation over a depot on the 
Ross Shelf Ice) was 19 em (water equivalent); other 
estimates for various years were: Cape Evans 41 cm, 
Cape Royds 23 em, Cape Adare 36 cm. Annual pre- 
cipitation at the Gauss was estimated at 80 cm, at 
Charcot’s two Palmer Peninsula bases 34 and 38 cm, 
and at the Deutschland only 11 cm. 
At the South Pole, Meinardus [14] estimated annual 
precipitation to be 2 cm, ablation 1 cm, for a net accu- 
mulation of 1 em; independently, Kidson [74] found it 
“dificult to imagine that the precipitation at the Pole 
is greater than 2 inches per year,” or 5 cm. For all 
Antarctica, Meinardus estimated precipitation 7 cm, 
ablation 3 cm, net accumulation 4 em; Kidson deduced 
total continental precipitation at less than 9 cm. 
Kidson’s estimate assumed that precipitation of all 
the moisture in the lowermost 2 km of an air mass 
initially saturated at 32F would give 0.5 cm, and that 
this process would require twenty days. An entirely 
different approach was used by Meinardus [107]: “If 
the interior of Antarctica is covered with snow and ice 
and there is a discharge of the same to the marginal 
oceans, it follows that over Antarctica as a whole the 
precipitation exceeds the evaporation.”’ Outflow of the 
inland ice, 250 m high and 17,000 km in perimeter, at 
150 m per year, totals 640 km® of ice or 550 km* of 
water, requiring an annual net precipitation of about 4 
em over Antarctica to maintain balance [108]. 
Although Meinardus considered his figures conserva- 
tive, each of the three factors in this computation seems 
excessive in the light of recent exploration. Further- 
more, there is considerable doubt that the ice which 
does flow out around Antarctica originates very far 
inland. 
Height. Highty years ago, an ice thickness of 24 
miles at the South Pole was estimated by Croll [117], 
assuming that the ice cap extended to 70°S and had a 
one-degree slope, as required “‘to produce the necessary 
outflow of ice.” However, “to avoid all objections on 
the score of over-estimating the thickness of the cap,” 
he used a thickness of only 6 miles in his computations, 
explaining that “the greater thickness of an ice-cap at 
its centre is a physical necessity not depending on the 
rate of snowfall,” which anyhow would not be negligible 
