1921] on Cloudland Studies 429 



oxygen above the cell in the central vessel A. To maintain this 

 condition the flask employed required the addition of about a quarter 

 to one-third of a litre of liquid each day, preferably about an hour 

 before observing. 



Sky Observations. 



Observations were made through various sequences of weather 

 conditions, to note the resulting deduced zenith temperatures. 

 Exceptionally low values were obtained in June and November, 1920. 

 On the nights of June 13 and 14, the Z/E ratios at about midnight 

 or a little later were of the order of • 4, or even less. With the 

 earth temperature (in London) between 13° and 15° C, the equiva- 

 lent zenith temperatures were - 33' C. on the 13th, and - 47° C. 

 on the 14th, the corresponding values of the T z being therefore 



- 46" and - 62°. The first of these observations was associated 

 with exceptionally calm conditions, rapid temperature decrement, 

 and sharp ground frosts at night in the open outside London, while 

 the second values followed the clearing of the atmosphere by a 

 thunderstorm. Beyond these exceptionally low records, the Z/E 

 ratio only occasionally fell to 0*5, with a T z of the order of - 40° C. ; 

 but a good clear sky quite frequently gave a ratio round about 0*6. 

 From sunset to past midnight on Sundays low values were not 

 unusual, possibly associated with the clearing of the air over London 

 as the traffic subsided in the week-end lull. On several occasions a 

 clear atmosphere following heavy rain also gave low ratios. Readings 

 taken before sunset were as a rule lower than those just after, and 

 frequently quite as low as the midnight values. The increase in the 

 ratio at sunrise was very sharply marked, but only occasionally did 

 it rise above unity when the sky remained clear. Thus, on May 4, 

 1920, the Z/E ratio just before dawn (the observation was made 

 at Mill Hill, ten miles out of London) was 0*63, with the black 

 standard (in this case a strip of velvet) at +0*5° C, giving a T z of 



- 29F C. Air temperature just above grass was - 2*6° C, and in 

 the grass - 1 • 5° C. Half an hour later with sufficient dawn light to 

 read the scale (bird life just stirring), water was dripping from the 

 thermometer above the grass which read * 5° C, the other thermo- 

 meter being unchanged. The Z/E ratio was then 0*72, correspond- 

 ing to a T z of - 22° C. 



As already stated, during daylight the values of the Z/E ratio 

 were frequently below unity, apart from the more common low values 

 already noted that preceded sunset. A clear blue zenith was of 

 course associated with such readings, while cloudless haze or sunlit 

 cloud gave values equal to or greater than unity. The first occasion 

 on which a low daylight ratio was recorded was with a deep clear 

 blue zenith above brilliant white flecks of clouds extending about 60 c 



