158 Transactions of tJie South African Philosophical Society. 



when it is clear ; or, in other words, is westerly for at least an hour 

 longer. 



The north component is positive for between six and seven hours 

 under a clear sky, whereas it is negative for about the same space 

 under clouds. The maximum excursion of this component, posi- 

 tively or negatively, falls about half an hour later when the sky is 

 clear. 



Upon a clear day the minima of E, in the order of their magni- 

 tudes fall at about X., XVII. , and III. ; upon a cloudy day they 

 fall at about 6.15 p.m., and 7.30 a.m. 



Taking the four years as they stand, there was practically as 

 much wind with an east component upon a clear as upon a cloudy 

 day ; but the south components changed into half as many again 

 north. There were, however, more than three days cloudy to every 

 two days clear ; hence if we allow for these we get the following daily 

 average unbalanced hours of component direction : — 



N. 



Clear days — 3*6 



Cloudy days + 3-5 



All days + 0-7 



If we compare the angle at any hour upon a clear day with what 

 it is upon a cloudy day, we shall see that the effect of a clear sky is 

 to accelerate the velocity of the resultant, whereas a cloudy sky 

 retards it. E begins the clear day 25° behind the mean, and ends 

 about 8° behind. E begins the cloudy day 18° in front of the mean, 

 and ends about 10° in front. This statement of the acceleration and 

 retardation of velocity will require some sort of qualification, of 

 course, because in general the greater number of cloudy days come 

 in summer, and the clear days in winter ; and there is, as has been 

 proved before, a prevailing wind-direction for each of these seasons. 



Eeason has been given for supposing that apart from the 

 seasonal changes of prevailing direction, there is not, strictly 

 speaking, a true prevailing direction for the year. It is seldom, 

 indeed, that any two consecutive years have a like resultant. For 

 the four years under consideration, out of a gross total of 35,040 

 hours of wind (of which 142 were variable hours, and therefore not 

 taken into account) the unbalanced components were — 



From the north 1,059 hours, 



From the east 4,725 hours ; 



E. 



E. 



-f 4-2 



5-5 



-f 2-6 



4-4 



-f 3-2 



3-3 



