TRANSACTIONS OF SECTION E. 461 
otherwise occur with increased intensity of solar radiation due to a nearer 
approach to the equator. 
A closer examination showed an increase of temperature along the coast, from 
north to south along the west coast, from west to east along the south coast, 
and from south to north along the east coast, due chiefly to the modifying influence 
of the cold Benguela current in the west and the warm Mozambique current in 
the east. 
The extremes of mean temperature were Disa Head (2,500 ft.), part of Table 
Mountain in the Cape peninsula, with 54°7 F., and Tuli in Rhodesia in the 
Shashi Valley (1,750 ft.) with an annual temperature of 72°-4 F. 
The average temperature of ninety-seven stations scattered over South Africa 
was 62°'8 F., or nearly the same as Sydney, N.S.W. The mean temperature curve 
was at its maximum in February, fell rapidly till June, continued to fall slightly 
in July, then rose, with a peculiar flattening of the curve in September, to the 
maximum in February. The continued fall in the July mean temperature was 
closely associated with a peculiar and, apparently, regular cold spell about the 
middle of the month; the minimum for the year occurred on July 16 over the 
Cape peninsula, and on July 17 at Kimberley. The flattening in September was 
associated with an increase in the cloud-curve, which was ‘coincident with the 
change of the prevalent wind-direction from north-west in August to south in 
September. 
Rainfall.—This curve showed two maxima—one in November and one in March 
—the minimum occurring in July. A comparison of the thunderstorm-distribution 
curve showed that the two maxima in the rain-curve were not coincident 
with the two maxima in the thunderstorm-curve, the maximum in the latter 
oceurring in February (month of maximum mean temperature) and falling till 
June, then rising again to the maximum in February, with a dip down in Novem- 
ber. South Africa might be divided into three rainfall areas according to its 
seasonal distribution: (1) winter-rainfall area in the west, (2) constant rains (small 
area) in the south, and (8) summer rains in the east. In Cape Town and the west 
generally the rain fell chiefly with north-west winds, with south-west along the 
south coast, and with south-west and some north-east along the east coast. here 
was little evidence in support of the ‘ south-east rain’ theory, which would apparently 
have to be abandoned so far as the coastal areas were concerned. 
Berg Winds.—These were foehn-like winds experienced pgactically all along 
the coasts blowing from off the plateaux at right angles to.the coast-line, being 
easterly in west, northerly along south coast, and north-westerly in the east. 
Those at Port Nolloth caused the winter temperature there to be higher than that 
at Ookiep (the reverse holding during the rest of the year), and actually delayed the 
occurrence of the minimum mean temperature till August. These were closely 
connected with the occurrence of secondaries, especially during autumn and the 
early spring. 
Storms.—The storms visiting South Africa seemed to he closely connected with 
moving anticyclones and assumed apparently and principally the forms of inverted 
V-depressions, as in Australia. 
Sunshine.—The largest proportion of sunshine occurred in the Cape peninsula 
in summer and was lowest in winter ; whereas at Kimberley, which was typical of 
the greater part of the central plateau regions of Cape Colony, the largest propor- 
tion of sunshine occurred in the winter months, when the days were mostly bright 
and cloudless, although the night temperature was frequently low (below freezing- 
point) and severe. This fact had an important bearing on the suitability of the 
Karroo for phthisical patients and other invalids, 
