The Winds of Kimberley. 17 
thing less than 4,000 feet above sea-level. It lies on one of the 
gentle slopes characteristic of the undulating veld of Griqualand 
West, and is rather lower than the crest upon which the greater part 
of Kimberley is built. Though thickly wooded, the trees do not yet 
offer any impediment of importance to the free circulation of the air 
at a moderate height. 
Observations of wind directions are derived from the automatic 
records of an Osler anemometer mounted with its vane about 36 feet 
above the ground. It has had a fair exposure hitherto, albeit this 
advantage is not likely to last very much longer unless it can be 
raised considerably higher. The friction of the various parts is con- 
siderable, though not greater, apparently, than that of other instru- 
ments of the same class. It is indeed less than that of a small, and 
very much lighter, vane mounted close by. Of the wind-pressure 
records from the same instrument little need be said, more especially 
since they are not used here. The motion of the vane is communi- 
cated by means of a rack and pinion to an aniline pencil beneath, 
which writes the directions continuously upon a moving band of 
paper controlled by clockwork. In the lighter winds of less than, 
say, five miles per hour the vane remains fairly steady approximately 
parallel to the direction, the pencil record being then a clear, if 
crooked, line ; but with higher velocities the vane may swing rapidly 
and continuously as much as 40° on either side of its mean position, 
making the pencil-record a broad shaded band. In the latter case 
the medial line of this band is taken as the true direction of the wind. 
The charts are changed every twenty-four hours, as punctually as 
possible at 10 p.m. As originally constructed the instrument only 
allowed a transverse motion of the pencil equal to a range of 675° 
for the vane, and in consequence the pencil was sometimes pushed 
out of gear, portions of the record being thereby lost. To guard 
against this, charts and pencil-rack were both altered to admit a 
range of 1260° in the vane, with satisfactory results. In general the 
portions of the record lost as above were interpolated in either of two 
ways, according as the vane carried the pencil gradually and slowly, 
or by one swing across the chart out of gear. In the former case it 
was assumed that the changes were also gradual for the hours of lost 
record; in the latter case that the vane had swung at once to the 
point at which it stood when the accident was discovered. 
Observations of wind velocity are derived from the records of a 
Robinson anemometer of the standard Kew pattern, mounted on a 
post at a height of 40 feet. At first the velocity was only registered 
on dials, no hourly record being taken. In March, 1897, however, 
Mr. Henderson, of the De Beers Crushing Mill, constructed a count- 
