[ 5- ] 



fcarcely exceeds 5000 feet, as we learn from the tcftimony of 

 aeronauts and the height of clouds ; and indeed the winds that 

 prevail on the furface of the earth, and which are the primary 

 agents of a change of temperature, feldom reach higher, and in the 

 more northern regions not fo high. Thus, on the ift of Decem- 

 ber, 1783, at Paris, while a S. wind prevailed below, a N. wind 

 prevailed at the height of 1280 feet*; and the fame oppofition 

 was found in the currents of air at the fame height at Pifancon 

 in January, 1784. At Ponoi the clouds are frequently feen 

 unmoved during the moft violent ftormSf ; yet on the 21ft of 

 December, 1779, the thermometer in the open air being at 49°, 

 and the wind S. S. E. the barometer flood at 28,91 inches; but 

 the next day, the wind turning to N. N. W. the thermometer 

 fell to 30°, and the barometer rofe to 29,89 inches. Here the 

 difference of temperature is 19°, and the variation of the baro- 

 mfetrical height nearly -rs- of an inch. Let us now examine how 

 this fad can be explained on the fuppofition that the mafs of the 

 lower atmofphere is encreafed in proportion to the condenfation of 

 its volume. \ 



In the firft place, we may aflume'that heat in its progrefs 

 upwards decreafes nearly in an arithmetical progreffion. The 

 barometrical method of meafuring heights is in a great meafure 

 founded on this fuppofition ; and as the errors of this method do 

 not eSccecd 2 or 3 feet, and feldom i foot in 1000, it may be 

 looked upon as fufficiently exad, and confcquently fo may the 



• Mem. Par. 1782, p. 650. 



f II. Faujas Balons, p. 274. X\''I. N. A£l. Petrop, p. 68. 



H 2 fuppofition. 



