r 35^ 1 



The u/>J>er term of congelation is that above which no viftbic _ 

 vapour afcends, and its temperature is conjlantly at leaft 32*^ of 

 Fahr. Bouguer has ventured to lay down its height over the 

 fnowy mountains of Peru, under the equator, but it is highly 

 probable that over the fea, and efpecially over land much heated, 

 its height is flill greater within the fame latitudes, fince its height 

 is proportioned to the intenfity of the heat below. And hence, 

 between fun-.fet and fun-rife, if the heat below be much diminifli- 

 ed, it muft defcend in proportion to the diminution. Bouguer 

 has determined its height only in a fingle inftance, but its impor- 

 tance has induced me to invcftigate the variations of this height in 

 all latitudes above latitude 5'='. The principles which I employed 

 in this inveftigation may be feen in the EJlimate of the temperature 

 of different latitudes, publilhed in 1787, p. 8. Since that time I 

 have correded the heights appropriated in that effay to latitude 55°, 

 apd every other 5*^ higher, in the ncnhern hemifphere, by taking 

 the inverfe ratio of the height in latitude 45^, divided by the 

 fine of 30°, namely, 2. The intervals I have filled by taking 

 arithmetical means. Without the tropics, it is evident that not 

 only night and day, but alfo winter and fummer, muft vary this 

 height, and, in the higher latitudes, very confiderably ; hence I 

 have confined this determination to the fummer months, namely, 

 May, June, July, and Aiiguft, wifhout the tropics. Daring the 

 \V.inter months, Odober, November, January, and February, this 



line 



