ON RADIATION. 



465 



continued throughout the night, the greater cold of the surface is 

 found to be maintained until sunrise, and for some hours beyond it. 

 Had the air been pei - fectly still during the observations, the nocturnal 

 chilling of the surface would have been in this case greater : for you 

 can readily understand that even a light wind sweeping over the sur- 

 face, and mixing the chilled with the warmer air, must seriously in 

 terfere with its refrigeration. 



Hind Head, elevation 850 feet ; sky cloudless; hoar-frost; wind light, from 

 northeast. Course of temperature, March 4, 1883. 



Glacial wind from northeast. Stars very bright. 



Various circumstances may contribute to lessen, or even abolish, 

 the difference between the two thermometers. Haze, fog, cloud, rain, 

 snow, are all known to be influential. These are visible impediments 

 to the outflow of beat from tbe earth ; but we have now to deal with 

 the powerful obstacle to that outflow to which reference has been al- 

 ready made, and which is entirely invisible. The pure vapor of water, 

 for example, is a gas as invisible as the air itself. It is everywhere 

 diffused through the air ; but, unlike the oxygen and nitrogen of the 

 atmosphere, it is not constant in quantity. We have now to examine 

 whether meteorological observations do not clearly indicate its influ- 

 ence on terrestrial radiation : 



TOL. XXIII. 30 



