6 John Eliot— On the occasional Inversion of the Temperature [No. 1, 



tween mountain stations and the neighbouring valley or other low-lyino- 

 stations. They are hence assumed to be phenomena restricted to 

 hills and the neighbouring confined valleys and hence of limited extent. 

 The explanation generally given, whilst making the inversion a pheno- 

 menon of terrestrial radiation, attaches much weight to the flow of cool 

 air down the mountain sides into the valleys, and hence suggests that it 

 is peculiar to mountain districts. 



The present paper will, I believe, prove that inversion may occur 

 over very large plain areas, and that it has, in some cases at least, little 

 or nothing whatever to do with air motion between hills and valleys. 

 It will also shew that the vertical temperature relations during the cold 

 weather in Northern India are much more variable and complicated 

 than they have been hitherto supposed to be, and that the descensional 

 motion which accompanies cooling of the air during the night in fine clear 

 weather is almost entirely one of slow compression, and is not the 

 opposite of the ascensional and convective movement which takes place 

 largely during the day, or, in Professor Ferrel's suggestive words, " the 

 effect of the heating of the earth's surface is not confined to the lower 

 strata merely, as that of the cooling of the surface is, but as soon as the 

 first stratum in contact with the earth is heated, the effect is carried to 

 these above." The principle is, I believe, of great importance generally, 

 and more especially in India, in connection with the production of the dry 

 winds of the Gangetic plain during the hot weather months of , March, 

 April, and May. 



The paper consists of three parts ; — 1st, a statement of the normal 

 meteorological temperature conditions of the plain and hill districts of 

 Upper India in the month of January and of certain meteorological 

 conditions and actions upon which temperature mainly depends ; 2nd, 

 a statement of the more striking abnormal temperature relations of the 

 month of January 1889 and of the cold weather period generally in 

 Upper India ; and 3rd, a discussion of the causes which produce these 

 unusual temperature conditions and variations. 



It may be premised that one or two of the actual observations 

 quoted for the month of January 1889 appear to me to be somewhat 

 doubtful. I have, however, thought it best to include them, as it is on the 

 whole more probable that they are exaggerated examples of the peculiar 

 temperature relations about to be discussed than that they represent 

 instrumental or observational errors. 



The following table gives the average maximum temperatures of 

 the month of January of certain selected pairs of stations in Upper 

 India, each pair consisting of a hill station and the nearest plain 

 station at which there is an observatory : — 



