Diurnal Meteorological Conditions at Kimberley. 141 



It will be noticed here that the two most important terms in the 

 cloud series are the first and third, and that the second has an 

 amplitude scarcely more than one-half that of the third. Also that 

 the first and third differ little in phase-time from the first and third 

 in the temperature series. Now the Trevandrum temperature con- 

 stants in Yj and V3 are not very materially different, excepting that 

 their maxima come somewhat earlier, from those of Kenilworth ; 

 nor, as it happens, from those of Allahabad and Cordoba. And it is, 

 as we have seen, the same two angular magnitudes, and only these 

 two, that vary so conspicuously in the transition from a clear to a 

 cloudy sky. 



To test this matter further, the cloudy days at Kenilworth have 

 been divided into two sets, containing respectively — 



1. All days on which the percentage of cloud averaged 50 per cent., 

 or more, and — 



2. All days on which the percentage of cloud was greater than 

 20 per cent, and less than 50 per cent. 



The respective monthly distribution will be found in Table 11. 

 There were 412 days in the four years upon which the sky was more 

 than one-half covered with cloud, and the average percentage of cloud 

 for these days was 72 per cent. There were 472 cloudy days in the 

 same period in which the sky was less than one-half covered with 

 cloud, the average percentage in this case being 29 per cent. The 

 first set of days may be styled " very cloudy," and the second 

 " moderately cloudy." The monthly distribution will be found in 

 Tables 12 and 13. As before, the last columns, giving the hourly 

 temperatures for the year, are the means of the monthly columns. 

 It would appear, according to these results, that the mean tempera- 

 ture of a typical very cloudy day is somewhat greater than the mean, 

 but that the air finds itself fully a degree cooler at the end of the day 

 than it was at the beginning. On a moderately cloudy day, however, 

 the temperature tends to a secular increase. It is nearly half a 

 degree warmer at the end of the day ; and also the mean temperature 

 is greater than the mean. And since the maximum on a moderately 

 cloudy day rises almost to that of a clear day, it seems ta follow that 

 a clouded sky, so long as it is not more than half obscured, may to a 

 certain limited extent conserve the solar heat, admitting it with a 

 little more freedom than it permits its escape. The actual range, 

 under such conditions, is about seven degrees greater than under 

 very cloudy skies, this total being created more by the excess of 

 maximum than the decline of minimum. 



The harmonic constants are given in Table 14, and for purposes 

 of comparison, others appearing in isolation elsewhere in this paper. 



