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



The minus sign in the preceding table indicates that the plain sta- 

 tion to which it refers had a lower minimum temperature than the hill 

 station with which it is compared and the plus sign that it had a higher 

 temperature. 



The preceding table shews over what an extensive area in Northern 

 and Central India it is possible for the minimum temperature to be consi- 

 derably (from 1° to 12°) below that of the hill stations in Upper India. 



Table I. shews that the inversion of temperature obtained on at least 

 eleven nights during the month. The following examples from previous 

 years, which examination shews to be fairly average cases, will indicate to 

 what extent the temperature variations of January 1889 were abnormal. 

 In January 1888 the night temperature of Mussooree ranged from 5 - 6° 

 above that of Roorkee to 2 1'8° below it (giving a total range of 27'4). The 

 average difference of temperature was 8T° for the month, which is almost 

 identical with the normal average (8 - 5°). The minimum temperature at 

 Mussooree was in excess of that of Roorkee on only three nights of the 

 month. In January 1886 the night or minimum temperature at Simla 

 ranged from 28° above that at Ludhiana to 23'5° below (giving a total 

 range of 26'3°) and was above that at Ludhiana on three nights only 

 during the month. The difference between the minimum temperatures 

 at these two stations averaged 10°. It is not necessary to multiply cases, 

 as all that have been examined give similar evidence. Hence it appears 

 that in ordinary seasons the minimum temperature may be on two or 

 three nights in January in slight excess at the hill stations of Upper 

 India as compared with the adjacent plain stations of the Punjab and 

 North-Western Provinces. These figures hence establish that, although 

 inversion of the normal vertical temperature relations is not infrequent 

 in the month of January in Upper India, it was of abnormal frequency 

 in January 1889. It was undoubtedly related to or connected with the 

 holding off of the winter rains in that month. Anticyclonic conditions 

 prevailed in Upper India with unusual persistency, and it was not until 

 the end of the month that general rain accompanying a depression and 

 cold weather storm occurred in the plains and heavy general snow in the 

 hills. Hence the high temperature was undoubtedly associated with 

 anticvclonic conditions of pressure, as has been found to be the case in 

 Europe and the United States during similar vertical temperature re- 

 lations, and also with the protracted delay in the depression of the snow 

 line in the hills during winter produced by general snowfall. 



The preceding paragraphs have stated fully one important feature 

 of the anomalons temperature conditions of the month of January 1889. 

 Before proceeding to discuss the causes of these features, it is desirable 

 to trace the varying temperature relations between the plains and the 

 hills in Upper India more exactly. There are three prominent types of 



