An Introduction to the Study of South African Bainfall. 9 



2. Lightning unaccompanied by audible Thunder . 



3. Thunder, Thunderstorms, and Lightning, all together. 



The numbers are got from eye-observation. They cannot pre- 

 tend to any completeness, particularly of lightning during the night. 

 They show all that was seen or heard of this kind of electrical 

 manifestation, but it is certain that many hours of it must have 

 escaped notice. Still, they probably give a fairly good rough idea 

 •of the diurnal variation, at any rate between VI. and XXIV. A 

 feature of special interest in the Table is the curious slackening in 

 the frequency of thunderstorms between XVI. and XVII. Since 

 thunderstorms evidently tend to a maximum about the time of 

 sunset, we have at once a reason why the greatest rate of rainfall 

 •should come about this time. 



Table 20 gives the most important harmonic constants in the 

 diurnal curve of rainfall at Kenilworth, counting from Oh. 30m. a.m. 

 The resemblance of the annual and semi-annual waves of both 

 -quantity and frequency to the same waves of barometric pressure 

 inverted is very strikingly exhibited in the epochs. Attention 

 should also be directed to the small fluctuation in the epoch of 

 the second term of rain -frequency at all seasons, its departure from 

 -the mean never much exceeding half an hour. It is important to 

 compare this fact with the resemblances between the corresponding 

 terms of barometric pressure and vapour tension mentioned in 

 previous papers of this series. * Seeing that there are not any 

 mountains on the tableland to turn the horizontal air currents 

 upwards, and that the air when it rises must do so of itself, the 

 importance of this result in its bearings upon the theory of the 

 semi-diurnal oscillation of the barometer will be obvious. 



The amplitudes are of some interest. In the hours for the year 

 the fourth harmonic term is practically as great as the third. But 

 in the case of quantity the fourth is very much smaller than the 

 third. Therefore the rainfall of six hours' period is very light. The 

 first term is the greatest at all seasons, but the relative magnitude 

 of the other three varies through the year. In the spring the 

 second term almost vanishes, while the fourth term is large. In the 

 summer the second term is large. 



The harmonic constants of thunderstorms bear a strong family 

 likeness to those of rain, particularly the summer frequency of the 

 latter to the mean of the former. It is evident that a large propor- 

 tion of the summer rain falls in thunderstorms, that of other classes 

 •being largely represented by the fourth term. 



The diurnal variation of thunderstorms is probably a general 



* See "An Elementary Synopsis, &c," Trans. S. A. Phil. Soc, vol. xiv., p. 185. 



