258 Observations on Humboldt's Works 



easily have, for such a comparison, places in the Old Continent, 

 whose mean temperature, or geographical latitude, agrees ex- 

 actly with the given places of the New Continent, the author 

 has, in each of these comparisons, made use of two places in 

 the Old Continent, from which the geographical latitude, or the 

 mean temperature of the place is determined, that is to be 

 compared with one in North America. He must, then, in order 

 to institute a comparison with Natchez, first fix upon a place in 

 the Old Continent, which lies under 31° 28' of latitude, and 

 the mean temperature of which should be known ; to obtain 

 this, he compares Funchal 32° 37', and Orotava 28° 25', and 

 calculates from the difference between the mean temperatures 

 of both places, the probable mean temperature of 31° 28' in the 

 Old Continent ; this is 20° 5', which number, compared with 

 18° 2', gives the difference = 2° 3'. Secondly, he must have 

 a place in the Old Continent, whose mean temperature should 

 be 18° 2' ; he obtains this place in the same manner through a 

 comparison between Rome 15° 8' (41° 33' of latitude), and 

 Algiers 21° 1' (36° 43' of latitude). I must, however, declare 

 that this last comparison, as well as some of the following ones, 

 do not afford the given difference ; which I can only explain 

 through the author's having chosen round numbers. 



But, if we compare these two methods, we shall perceive 

 that the author, according to the first, gives 7° as the difterence 

 of latitude between two places of the same mean temperature, 

 both in the first and third parallels. According to the second 

 method, however, the difference of mean temperature in the 

 first parallel is only 2° 3', in the third 5° 3'. It is besides 

 evident, that if the difference of temperature between two places 

 near the North Pole, is greater than between two others more 

 to the southward, the difference of latitude between two places 

 of the same temperature cannot, in both cases, be the same. 

 This surprising difference, in the author's results, proceeds from 

 his having employed, in the first method, places of a tempe- 

 rature comparatively too low, and in the second, too high. For 

 instance, in the first comparison, he takes Orotava, which city, 

 though 8° more southward, has a lower mean temperature than 



