86 Transactions of the South African Philosophical Society. 
vane from its normal position. The inference seems to be that over 
and above instrumental faults which may contribute something to 
the result, storms, and other disturbances occurring here, are, on 
the whole, little else than exaggerations of normal conditions. The 
curves illustrating columns 7-10 are shown in Fig. 5. 
The fact-will now be readily appreciated that because the winds 
which prevail during the middle of the day come from between 
north-east and north-west, and also because the velocities are 
greater when the directions are normal, these, if only through the 
dust they raise, will attract the most attention from the ‘‘ man in 
the street,’ who as a rule only notices what is forced upon him. 
The published accounts owe their inspiration in the main to the 
observations taken for the Meteorological Commission of the Cape 
Colony by amateur observers. Now the hour of observation is 8 or 
9 a.m., and a glance at Table 7 will be sufficient to prove that either 
of those hours will return a great preponderance of winds lying 
between north and north-east. Thus we see how it comes about 
that both scientific and unscientific observation have independently 
evolved the same totally incorrect conclusion.* 
The number of miles of wind recorded from each divedtions irre- 
spective of the duration, for each hour during two years is given in 
Table 18. It will be seen that the highest values occur about 
midday—as might be expected from Table 12—and that on the whole 
the columns containing the highest numbers also contain the lowest. 
The components of Table 14 are calculated from Table 13 exactly as 
those of Table 9 from Table 7. Fig. 6 is a graphical representation 
of the results of Table 14. So far as its critical points are concerned 
it is substantially a reproduction of Fig. 3, each component possessing 
only one pair of simple maximum and minimum points, temperature 
affinities being equally apparent in the east component curve, and 
barometric affinities, so it seems, equally lacking in both curves. 
By dividing the directions into quadrants we get the following 
total mileages in two years :— 
Quad. 1. hile ese pees Cemecnee eee 32,674 miles 
UR PA eS obo o5csachcaosoon: 25,122 __,, 
PE ee eee n dnbachosuecboe 31,325 ,, 
WM ETP ees cogtoanatsor 29,960) 3) 
Motaleccecc: FO GIE 
* Dr. Buchan gives the following yearly totals for Kimberley derived from 
observations made at 8 a.m. and 8 p.m.: N., 87; N.E., 72; E., 57; S.E., 15; 
S,, 51; S.W., 24;. W., 28; N.W., 20. . 
The inference would be a prevalence of northerly winds. The observations are 
evidently quite honest, but they are insufficient. 
