314 Transactions of the South African Philosophical Society. 



Meissner attributes the belief in an exactly opposite state of things, 

 as also Ernst Schmid did half a century ago/" to a mere psychological 

 impression. My own idea, however, is that, even if there be no 

 lunar influence whatever upon the clouds, we should expect to get 

 Meissner's result or something like it. For there must often be 

 clouds in the sky at night which moonlight could make visible, but 

 starlight alone cannot. At Kimberley it often happens that cirrus 

 and thin cirro-stratus appear to dissolve at sunset, whereas the 

 rising moon has shown them again quite plainly. Moreover, 

 summer sheet-lightning, on the horizon, will often show clouds 

 overhead in a sky that otherwise will seem perfectly clear. For this 

 reason it is perhaps not altogether improbable that Meissner's cloud 

 sequence is in reality due to a variation of illumination rather than 

 to a true variation of cloud. This view receives some confirmation 

 from the fact that various observers have found thunder-clouds at 

 any rate to be more frequent at new moon than at full moon. 

 Koeppen, for example, finds that the frequency of thunderstorms at 

 new moon to that at full moon is in the ratio of 29 to 21. t From 

 which it seems to follow that the lower and heavier clouds, at least, 

 are more numerous at new moon. 



My own impression has been that a sufficiently long series of 

 observations would show — providing we allow for the variation of 

 illumination at night throughout the lunar month — practically no 

 influence on the part of the moon upon the clouds. Short series 

 can of course show anything, as we see from the thousand and one 

 meteorological cycles that are deduced from the sunspot cycle. | 

 With the idea of substantiating or disproving this impression I have 

 worked up my cloud observations § for the 87 lunations extending 



* E. E. Schmid, Lehrhuch der Meteorologie, 1860, p. 681: " Wenn aber Baco 

 annimmt, der Himmel sei bei Vollmond besonders heiter, und dabei im Winter die 

 Kalte besonders streng, so ist diese Meinung wohl nur darauf begriindet, dass der 

 Vollmond in heiteren Nachten unsere Aufmerksamkeit besonders auf sich zieht, 

 und dass kalte Winternachte besonders hell sind." 



t See J. Hann, Lehrhuch der Meteorologie, 1901, p. 662, Art. " Mondperiode der 

 Gewitter ; " also W. M. Davis, Elementary Meteorology, 1894, Art. 327, "Weather 

 Cycles." Arago found a slight preponderance of rain at new moon as compared 

 with full. Seeing that a good proportion of rain in France falls in thunderstorms, 

 Arago's result is in conformity with that of Koppen. 



I See, e.g., J. A. Westwood Oliver's interesting little brochure, Sunspottery, 

 1883, p. 36 et seg. The late E. A. Proctor had a decidedly clever chapter, entitled 

 *' Sunspot, Storm, and Famine," in Pleasant Ways in Science (reprinted 1893). 

 He showed in a very amusing way that there is a sunspot period in the Oxford 

 and Cambridge Boat Race. 



§ Made at Kenilworth (Kimberley), South Africa. For a brief general discussion 

 of the Kenilworth observations of cloud, see J. R. Sutton, " The Climate of 

 Kimberley," in the Report of the Eighth International Geograpliic Congress, 1905. 



