78 Transactions of the South African Philosophical Society. 
ing apparatus for me by means of which an automatic record could 
be written with a pen upon a Richard drum.- This very ingenious 
contrivance admits of a wide range of speed multiplication ; as used, 
and found most convenient, the pen rises once for each two and a 
half miles of wind, falling to zero and rising again, and so on. It has 
worked uniformly well, the only break in the records occurring for a 
few hours when a screw of the shaft connecting the cups with the 
dials had worked-loose. Until the repairs were completed the velo- 
cities were interpolated from as many interim readings as could be 
made. The charts can run for twenty-four hours, but for a portion 
of the time the pen would be running over the brass strap binding 
the chart to the drum. For some considerable time the charts were 
changed at the same hour as the Osler charts, the mileage to be 
added for the time during which the pen was on the brass strap being 
taken from the readings of the dials. This was found, however, to 
give not quite correct values, and latterly, in consequence, the charts 
are changed an hour or so earlier each day, ei@ht being used in a 
week. A small correction is sometimes necessary to the registered 
hourly velocities because the hour-lines are not quite truly centred. 
There seems to be very little friction in this instrument. | 
The barometric records are taken with a Beckley Photo-barograph 
(by Messrs. Negretti and Zambra), the ordinates being converted into 
true hourly air-pressures by comparison with three readings per diem 
of a large station standard mercurial barometer. 
The temperatures of the air and of evaporation are read directly 
from the hourly registrations of a set of Negretti and Zambra’s 
reversing thermometers, supplemented in a subordinate capacity by 
the indications of a Thermograph and a Hygrograph, both by MM. 
Richard, of Paris.* 
Winp DIRECTION. 
The first circumstance to be investigated is the relative frequency 
of the winds, referred to the sixteen principal directions, for each 
month and for each hour. For this purpose the mean direction of 
the wind during each whole hour was determined as nearly as 
possible, and entered in its proper column in the summary. Only 
those hours are excluded in which the vacillation of the vane made 
* For a fuller description of the various instruments see the introduction to 
the Kenilworth observations, published, by permission of the directors of the 
De Beers Consolidated Mines, Ltd., in the annual report of the Meteorological 
Commission of the Cape Colony for 1898. The site is not a very good one for 
temperature observations. 
