260 Transactions of the South African Philosophical Society. 
The middle stations have also a minimum in March and also a 
second, much less pronounced, in the spring. The separate stations 
exhibit considerable variation among themselves, Worcester having 
the greatest: range in the summer, while the others, on the whole, 
have theirs in the winter. The coast stations have a decided 
minimum range in February, and a smaller and, in some cases, abor- 
tive minimum in October. They are all alike in having the greatest 
range in winter, Simonstown excepted. At this station there is a 
third minimum, which also happens to be the greatest, and runs con- 
currently with that at Worcester, in the winter. Now Worcester 
and Simonstown are both in the Western Province, and _ their 
minimum winter range is no doubt a consequence of the greater 
rainfall, cloudiness, and humidity there at that season. The average 
annual ranges of monthly mean ranges of temperature are— 
Abert Niel be walC ESUERAOINS) Angelus cassonbescsdonssaoscoo oc 7°:0 
Middle shy. We pg ARNEL SARIS RBI et OE eat 
Coast pea ae NRL 8 due SRA eS DS) 
The results of the four Tables (33-36) seem to indicate that in the 
same way as the power of the sun to heat the air is limited by radia- 
tion, so is that same radiation limited (at night mainly) by the air’s 
hygrometric condition. The curves of Fig. 3 suggest the conclusion 
perhaps better than the tabular numbers. 
The influence of local surroundings is very clearly shown when we 
compare the temperatures of the different places with what some 
theory should indicate. The coast stations do not, for example, 
follow any simple cosine law in either their maxima, minima, or 
means of temperature. Neither do the various stations, when allow- 
ances have been made for latitude, conform to any law of decrease 
with altitude. Supposing the average in free air to hold good over 
South Africa, Prince Albert should have, approximately, a mean 
temperature 64° less than that of Hast London, instead of 44°; Wor- 
cester nearly 2° less than Port Elizabeth, instead of 1°; Bloemfontein 
about 13° less than Durban, instead of 92°; and the Table-Land 
stations generally quite 5° less than that of the middle stations. The 
common reduction of temperature to sea-level, made with more or 
less (usually less) success in other countries, is therefore quite illusory 
here. We may conclude that the effect of the table-land is to reduce 
the fall by about one-third of its amount in free air, bringing it to, 
say, 1° in 440 feet. This is rather less than the rate (1° in 410 feet) 
connecting the temperatures of Mount Abu and Karachi, India, both 
near 25° N. Lat., at the respective altitudes of 3,945 feet and 49 feet, 
