336 M. Humboldt m the Difference of the 



tion to the mean temperature of the year, yet the distribution of 

 heat in the different seasons is strikingly different, although 

 the mean annual temperature be one and the same, — a circum- 

 stance which has a very great influence on the growth of plants 

 and on the health of man. I have endeavoured to determine 

 the law of this distribution, according to different situations 

 and heights. But comparative results in numbers ought to con- 

 tain the mean temperature of every month, derived from the 

 two extremes of every day, supposing an arithmetical series to 

 be formed. This method was first adopted by Reaumur in 

 1735 : he compared the produce of two harvests, not (like Her- 

 schel) with the numbers and size of the spots in the sun, but 

 with the quantity of heat which the corn received in the time of 

 vegetation. Many labours have of late been directed towards as- 

 certaining the hour, the mean temperature of which expresses 

 also that of the whole year. I here only mention the observa- 

 tions carried on in Scotland at Leith Fort. The night watch 

 of a military post has been employed for establishing observa- 

 tions of the thermometer during two years, from hour to hour ; 

 and from the mass of these observations, which ought to be re- 

 peated in other latitudes, it has been calculated, that, in the la- 

 titude of Edinburgh, a single daily observation at 9 o'clock 13 

 minutes in the morning, and in the evening at 8 o'clock 29 mi- 

 nutes, would be sufficient to fix the average heat of the year *. 

 Of the months, it is April and October which give this important 

 result (a fact, first discovered by Leopold and Von Buch, which 

 is connected with remarkable modifications of the upper currents 

 of the atmosphere), except when, as in the island of Grand Ca- 

 nary, local causes carry the maximum of heat to a later period, 

 and place it in October. 



If I frequently allude to the great increase of meteorological 

 observations within the last twenty years, I by no means wish 

 to express an opinion that the perfection of climatology is parti- 

 cularly founded on such an increase. Here, as in all collec- 

 tions of knowledge derived from experiments, which are too 

 soon denominated sciences, every thing depends on " an accu- 



• A result, which does not differ from the true by one-half degree of 

 Reaumur's thermometer, is also obtained by the mean of two hours of 

 the same denomination. — Results of the Thermometrical Observations made at 

 Leith Fort every hour of the day and night during the years 1824 and 1825, p. 19. 



