214 Proceedings of the Royal Society of Edinburgh. [Sess. 
St Helena. The following are the particulars, the means and departures 
being given in Centigrade : — 
Station or Area. 
Lat. S. 
Long. 
1902- 
1903. 
1903- 
1904. 
1908- 
1909. 
1910- 
1911. 
1911- 
1912. 
Mean. 
M‘Murdo Sound . 
77 44 
166° 34 E. 
-1°0 
4-0°6 
4-2°0 
o 
4-01 
o 
- 1-8 
o 
-5-6 
South Orkneys . 
60 44 
44 39 W. 
-0-9 
4-0-4 
4-0-7 
-0-2 
4-0-3 
-0-3 
New Zealand* . 
42 18 
173 25 E. 
-1-0 
4-1-3 
4-0-1 
4-0-8 
-1-2 
15-2 
Perth (West Australia) 
31 57 
115 45 E. 
-0-3 
4-0-2 
4-0-6 
-0-4 
-0-2 
22-7 
Rio Gallegos 
51 53 
68 59 W. 
-0-7 
-FO-3 
4-0-5 
-0-1 
-0-1 
13-2 
St Helena . 
15 57 
5 40 W. 
+ 0*8 
-1-1 
-0-6 
-0-2 
4-0-9 
17-5 
Sydney 
33 52 
152 11 E. 
4-0 5 
-0-7 
-0-2 
-0-3 
4-0-6 
21-8 
During the first three seasons the agreement is good for all the stations, 
but in the summer of 1910-11 (December 1 910 interpolated for M‘Murdo 
Sound) the signs differ at the South Orkneys, Perth, and Rio Gallegos, and 
a divergence also occurs in the summer of 1911-12 at the South Orkneys. 
There is little doubt, however, that cold summers at M‘Murdo Sound are 
associated with cold summers in New Zealand and the far South Atlantic, 
and with warmth in New South Wales and the region of the south-east 
trades, as shown by the St Helena record. Mild summers at M‘Murdo 
Sound show a similar relation with those in New Zealand and the far 
South Atlantic, and the reverse at Sydney and at St Helena. 
In autumn and winter the relation between the temperature at 
MAIurdo Sound and that at other regions is not on the whole very definite. 
When the data are obtained from a short term of years, it is obvious that 
an exceptional season at M‘Murdo Sound — such, for example, as the autumn 
of 1903 or the winter of 1912 — must strongly affect the sign of the de- 
partures from the short-period averages. In lower latitudes, where condi- 
tions are much more equable from one season to another, this disturbing 
factor does not exist. There is little doubt, however, that in autumn, as in 
summer, the temperature variations at New Zealand and Sydney are 
strongly influenced by Antarctic conditions. In winter no connection can 
be traced between the conditions at the places examined and at M‘Murdo 
Sound, if we except a well-marked positive correlation with Santiago de 
Chile and a negative correlation with Perth (West Australia) and the South 
Orkneys. In spring the Antarctic influence becomes more prominent, since 
a well-pronounced negative correlation is shown between M'Murdo Sound 
on the one hand, and Hobart (Tasmania), Adelaide, Punta Arenas, and 
Rio de Janeiro on the other. The following are the values : — 
* Mean of six stations, viz. Auckland, Wellington, Hokitika, Lincoln, Dunedin, and 
the Chatham Islands, 
