932 
bend upward, rather than continue downward as in the 
published figures. 
These temperatures, warmer in interior Antarctica 
than at the periphery, coupled with slightly higher 
surface pressure, would cause pressures near the Pole 
to be higher at all levels than around the coasts. Conse- 
quently, the isobaric surfaces should be lowest at 60° to 
80°S and then rise slightly toward the interior, rather 
than continue to slant downward (as in Fig. 4 of Loewe 
and Radok, and up to 300 mb in Flohn’s Fig. 8). More 
precisely, the trough in the isobaric surfaces in early 
summer probably inclines poleward with height up to 
the tropopause, so that the 300 mb surface is lowest 
(5 km) at about the coastline and the 200 mb surface 
lowest (11 km) some distance inland; Flohn’s diagram 
suggests such a pattern, with a trough around 65°S at 
150 mb and higher. Thus upper-level west winds at 
. coastal stations may not indicate an upper-level polar 
cyclone, but the southern edge of the westerly flow. 
Implications of these suggestions are discussed in the 
final section. 
The Loewe-Radok cross section for 150°E shows the 
tropical tropopause extending to 40°S and the polar 
tropopause beginning under it at 30°S, where it is at 
12.6 km in summer, 8.5 km in winter. At 55°S the height 
is the same, 10 km, throughout the year, and farther 
south the height decreases in both seasons, but is 
slightly greater in winter. The twenty-five soundings 
from the Commandant Charcot [71] around 65°S in the 
coastal icepack, some 400 miles south of Macquarie 
Island, during February 1949 found the tropopause at 
6 to 9 km, —50 to —57C, slightly lower but in gen- 
eral agreement with this diagram. Flohn’s summer 
cross section for the South Atlantic, with Little America 
data added, shows only one continuous tropopause. 
The Loewe-Radok cross section shows tropopause 
height as nearly constant between southern Australia 
and Macquarie Island (86° to 55°S), but 1000 miles to 
the east, in New Zealand, a winter ‘‘reversal of the 
slope of the tropopause” has been noted [57], the aver- 
age height during July 1945 increasing® from 9.7 km at 
Auckland (37°S) to 10.1 at Hokitika (43°S) and 10.3 
at Taeri (46°S). Harlier, Kidson [11] assumed that the 
“annual variation of temperature in the stratosphere 
can be little less than that at the surface in subantarctic 
latitudes and. ..the base of the stratosphere must vary 
little,” because ocean temperatures between 45°S and 
60°S change little through the year. 
Other Aspects. Water temperatures at Campbell Is- 
land (52°32’S, 169°08’E), about 400 miles south of 
New Zealand, vary [56] only 6F through the year, from 
42.8F in July to 49.1F in February (1942-47). Air 
5. This anomaly is shown also on an excellent winter cross 
section along 170°H from Little America to the equator, pub- 
lished since preparation of this article. Otherwise it and the 
companion summer diagram are quite similar to those of 
Loewe and Radok for 20° farther west, but do not attempt 
extrapolation poleward from Little America. (See Hutchings, 
J. W., ‘‘A Meridional Cross-Section for an Oceanic Region.” 
J. Meteor., 7: 94-100 (1950).) 
POLAR METEOROLOGY 
temperatures there are 2F warmer in winter, 7F warmer 
in summer than at Macquarie Island (54°30’S, 
158°57’E), 500 miles to the west, which in turn is 
noticeably warmer [73] than Heard Island (52°01’S, 
73°08’E), 3400 miles farther west. Marion Island 
(46°51’S, 37°52’), 1600 miles farther west-northwest 
of Heard, appears from preliminary reports [92] to be 
substantially warmer than Heard, but colder than 
Macquarie. 
Month by month during 1948, Macquarie was 10F or 
more warmer than Heard [73], with the difference 
gradually diminishing with height so that at the tropo- 
pause, around 300 mb at both stations, the modal 
temperature at Macquarie was —50C, Heard —53C. 
Despite local topographic effects (both stations are a 
few feet above sea level on the low, narrow necks of 
peninsulas extending northward from their main rocky 
islands), winds at Macquarie were northerly, at Heard 
southerly, showing that the wind regime is far from 
zonal, 
Why should the maximum temperature on clear days 
during the antarctic night occur, almost without ex- 
ception, in the hours after midnight? Simpson pondered 
this problem longer than any other of the many he 
investigated in compiling the meteorological reports 
[112] of the last Scott expedition, without developing a 
satisfactory hypothesis. “In McMurdo Sound there is 
an excess of temperature at 4 a.m. compared with 
2 a.m. and 6 a.m. [only bihourly readings were made] 
in every month from April to September, and the same 
effect is possibly present in October, November, and 
March,. . .is clearly seen in the temperature observations 
at the Gauss. . .[and] the Snow Hill observations also 
show the same effect in July and August.” In all cases 
it was characteristic of cloudless or partly cloudy 
weather; on cloudy days “during the three winter 
months when there is no direct solar radiation the day 1s 
warmer than the night.” 
Similar relations were found by Rouch on the Charcot 
expedition and later in the data of the two Byrd expedi- 
tions [65], where the average daily variation of 2C on 33 
clear days “is a simple curve with maximum around 
1 or 2 a.m. and minimum after noon...I cannot give a 
satisfactory explanation.” Nor is van Everdingen’s ad- 
vective explanation [24] very plausible. 
Less mysterious, but still not satisfactorily explained, 
is the annual march of temperature at Little America 
[47], which shows a single maximum in early January, 
but three distinct and progressively colder minima: in 
early May, in July, and in early September. Evidenced 
in monthly means, this is shown most clearly by 10-day 
means (Fig. 3). Temperature averages at other antarctic 
stations do not show such triple minima, although indi- 
vidual years at many stations do show double minima 
as found, although not adequately explained, at arctic 
coastal stations. 
Other aspects of Antarctica’s temperature regime 
offer no peculiarities. The interdiurnal variation and 
the mean daily range [16] change little through the year 
between 50° and 60°S. They are greatest in winter 
south of 60°, greatest in summer north of 50°. The 
Ve eS 
a i 
