SOUTHERN AFRICA. 57 
this substance has certainly an influence upon the tempera- 
ture of the air, causing a considerable degree of cohi. A full 
hour after the sun had risen the thermometer stood, in the 
shade, at 26°, or six degrees below the freezing point. At Lit- 
tle Loory fonteyn, where the soil was hard, dry, and stoney, 
it was ten degrees above freezing ; and about the same time 
on the preceding morning, on the banks of the Traka, where 
there was also much nitre, the mercury was five degrees be- 
low the freezing point. The weather during the three days 
was perfectly clear, and the wind had not shifted a point. 
That the great changes in the temperature of the air upon 
the desert, whilst the weather apparently remains the . same, 
arise from some local rather than general cause, is pretty evi- 
dent from another circumstance : in travelling at night upon 
the Karroo, if the wind should happen to blow upon the side, 
it is very common to pass through alternate currents of hot 
and cold air, whose difference of temperature is most sensi- 
bly felt. AVhether the cooler columns of the atmosphere may 
have been owing to the subjacent beds of nitre, Avhich fre- 
quently occur on the Karroo plains, or to some remoter cause, 
I have not grounds sufficiently strong to determine ; but a 
variety of circumstances seem to favor the former suppo- 
sition. 
In looking through the exhalations of these beds of nitre, 
a meteorological phenomenon, of a different nature, was also 
here accidentally observed. In marking, about sunrise, the 
bearing by a compass of a cone-shaped hill that was con- 
siderably elevated above the horizon, a peasant well ac- 
quainted with the country observed that it must either be a 
