xxxii 



in January, and lowest in June ; which are respectively the 

 coolest and hottest months in the year. As at all other 

 • tropical stations, the diurnal variations of the Barometer are 

 here found to be so regular in their recurrence, that there 

 are very few days in the year on which the atmospheric tides 

 are not plainly discernible. The highest readings- occur 

 about 9^ a. m. The mercury then commences falling, and 

 arrives at its lowest minimum a little before 4 p. m. The 

 secondary maximum and minimum occur during the night, 

 at 10 p. m. and 3^ a. m. respectively, but are far less marked 

 than those observed during the day. The greatest daily 

 range occurs in March, and the least in December. The 

 mean temperature, although somewhat higher at Madras 

 than at many other parts of the Presidency, is subject to 

 such small changes, that the severe extremes which else- 

 where prove so prejudicial to health are never experienced. 



The Dew Point, or Temperature at which rain would 

 begin to fall if the air were suddenly cooled down thereto, 

 does not rise and fall conformably to the temperature ; for 

 though its greatest depression below the Dry Bulb (13° *2), 

 occurs during the hottest month, June, the least difference 

 (6° '6) falls in October, when the temperature nearly agrees 

 with the average of the whole year. The per-centage of 

 Humidity changes but slightly ; the average amount of 

 aqueous vapour in the atmosphere being exactly three- 

 fourths of the whole quantity sustainable thereby. The 

 average Rain-fall is deduced from observations extending 

 over a period of fifty-two years, viz., from 1804 to 1807, and 

 again from 1813 to 1860. The dryest year on record was 

 1832, when only 18i inches of rain were collected: the 

 greatest quantity fell in 1827, when nearly 89 inches were 

 recorded. In October, 1857, not less than 37f inches fell ; 

 while no shower occurred throughout the first five months of 

 1860. By summing up separate portions of this column, it 

 will be found, that under ordinary circumstances only about 

 a tenth part of the year's rain falls during the first five 

 months, but a full half at the setting in of the N. E. Mon- 

 soon, in October and November. It is also worthy of re- 

 mark, that, whereas in Europe a rain gauge placed on the 

 ground invariably collects more than one elevated some 



