BRITISH ANTARCTIC EXPEDITION—SHACKLETON. 367 
region was dragged over powerfully toward the southeast. On such occasions 
one usually saw evidence of two high-level currents, the one coming from a 
northerly direction, its under limit being about 15,000 feet above sea level, and 
the other, or middle current, from a southerly quarter, usually blowing toward 
the east-northeast, having its upper limit at 15,000 feet normally, while its lower 
limit was between 6,000 and 7,000 feet above sea level. While these two ecur- 
rents were blowing strongly, there would frequently be a surface current blow- 
ing gently from the north. This would bring up very dense masses of cumulus 
cloud from off Ross Sea. The cumulus would drift up to the 6,000 or 7,000 feet 
level on the northwest slopes of Erebus, and then the tops of the cumulus would 
be cut off by the lower edge of the northward-flowing middle current. Wisps 
of fleecy cloud would be swept along to the east-northeast, torn from the tops of 
these cumulus clouds by the middle current. Our observations showed that 
during blizzards the whole atmosphere from sea level up to at least 11,000 feet 
moves near Cape Royds from southeast to northwest, and the speed of movement 
is from 40 up to over 60 miles an hour. After and during the blizzard the 
middle air current, normally blowing from the west-southwest, is temporarily 
abolished, being absorbed by the immense outrushing air stream of the south- 
east blizzard. During a blizzard the air was generally so thick with snow that 
we were unable to see the top of Hrebus. At the end of a blizzard the air 
current over Erebus became suddenly reversed, the steam cloud swinging round 
from the south to the north. After a time, following on the conclusion of a 
blizzard, a high-level current was seen to be floating the cirrus clouds from the : 
southeast toward the northwest, and the steam of Erebus would stream out 
toward the northwest. We could not account for this high-level southeasterly 
current. It looked like a reversal of the usual upper wind, and it appears to be 
a fact new to meteorological science. 
In this article I can only indicate the scientific results of the ex- 
pedition, as apart from the new geographical knowledge secured. 
We were able to throw some additional light on the problem presented 
by the Great Ice Barrier. The disappearance of Balloon Bight shows 
clearly that the recession noted since the days of Sir James Ross con- 
tinues, and suggests that very large portions of the Barrier edge may 
oceasionally “calve off.” The trend of the mountains discovered 
on the southern journey indicates that the Barrier is bounded by 
mountains which run eastward along the eighty-sixth parallel, about 
300 miles from the sea edge. The great glacier up which we marched 
to the polar plateau shows that the Barrier is fed to some extent 
from the highlands of the interior. It would seem, however, that in 
the main the Barrier is formed of superimposed layers of snow, and 
some interesting observations were secured in this connection. We 
formed the opinion that at Cape Royds the annual snowfall is equal 
to about 9.5 inches of rain. The southern depot party, in January, 
1909, found depot A, left by Captain Scott in 1902 on the Barrier 
off Minna Bluff. A careful examination showed that the depot had 
been moving bodily to the east-northeast at the rate of a little over 
500 yards a year, while there had been an accumulation of about 138 
inches of hard snow above the depot during each year. A deter- 
mination of the density of the snow showed that the snowfall on 
