140 Baron Humboldt on the Mean Temperature 



not remain a sufficient time in the countries whose climate 

 they desire to know, and they collect for the annual means 

 only observations which others have made, and most frequent- 

 ly at hours and with instruments which are far from giving 

 correct results. Owing to the constancy of the atmospheric 

 phenomena under the zone nearest to the equator, a short 

 space of time is without doubt sufficient to give approximate- 

 ly the mean temperature at different heights above the level of 

 the sea. 



I have always pursued this class of researches ; but the on- 

 ly precise result which I have been able to obtain, and which is 

 deduced from observations made twice a-day, is that of Cu- 

 mana. (Compare with respect to the degree of confidence 

 which the mean temperatures merit, Relat Hist. torn. i. p. 41 1, 

 547, 631-637, 584 ; torn. ii. p. 73, 418, 463 ; torn. iii. p. 

 314-320, 371-382.) 



The true numerical elements of climatology can only be fix- 

 ed by skilful persons established for a great number of years 

 in different parts of the earth ; and, in this respect, the intel- 

 lectual generation which is preparing itself in the free part of 

 equatorial America, from the coast to 2000 toises of altitude 

 on the back and on the declivity of the Cordilleras, between 

 the parallels of the Isle of Chiloe and San Francisco in New 

 California, will have the happiest influence on the physical 

 sciences. 



In comparing what has been known for forty years on the 

 mean temperature of the equatorial regions with what we now 

 know, we must be astonished at the slow progress of positive 

 climatology. I do not know, at the present day, more than one 

 mean temperature observed with any appearance of precision, 

 between 3° north and 3° south lat, and it is that of St Louis 

 de Maranham in Brazil, 2°.29' S. lat., which Colonel Antonio 

 Pereira found from observations made in 1821, three times 

 a-day, (at 8 h a. m., 4 h p. m., and ll h p. m.) to be 81°.32. 

 (Annaes das Sciences das Artes e das Letras, 1822, torn. xvi. 

 Plate II. p. 55-80.) This is still 0° 54 less than the mean 

 temperature of Cumana.* Below 10J° of lat. we know only 

 the mean temperature of 



* See a preceding notice on this subject in p. 117 of this Number* 



