318 Transactions of the South African Philosophical Society. 



at Habana, together with similar decreases at Potsdam from 9 p.m. 

 to midnight, are : — 



Day. 



Habana. 



Potsdam. 



• -3 



27% 



+ 2-6% 



4-6 



26 



-0-6 



7-9 



15 



+ 2-6 



10-12 



14 



+ 2-2 



13-15 



13 



-2-4 



0-18 



6 



+ 0-3 



19-21 



25 



+0-3 



22-24 



29 



-1-8 



25-27 



26 



+ 3-7 



28-30 



29 



H-5-4 



According to this the Habana results show no cloud-dispersing 

 power in the moon ; on the contrary, they show the effect of lunar 

 illumination. The Potsdam results are so irregular that they prove 

 (or disprove) anything, except that when the sky is dark, about #, 

 the decrease of cloud seems to be greatest. 



The distribution of thunderstorms throughout the lunar month at 

 Kenilworth, for the 87 lunations 1900-1906, is given in the last 

 column of the large Table. The numbers fall upon a somewhat 

 irregular curve, showing nevertheless a well-marked minimum at 

 new moon, and maxima in the first and fourth quadrants. There is 

 also some tendency to a maximum at full moon. This does not 

 agree with European experience. And contrary to what might have 

 been anticipated, the thunderstorm numbers do not agree very well 

 in phase with those of cloud. Trihemera are : — 



Day. 



•-3 



No. of Thunderstorms 

 34 



4-6 



48 



7-9 



33 



10-12 



42 



13-15 



36 



0-18 

 19-21 



39 



38 



22-24 



45 



25-27 

 28-30 



46 

 26 



The inference that may be fairly drawn from the results tabulated 

 in this paper seem to me to be that if they do not go far enough to 



