212 Prof. Forbes on the Diminution of Temperature 



Saussure, in the part of his work already cited, finds, that, 

 whilst the minimum temperature occurred at 4 A. M. (in the 

 month of July) both at the Col du Geant and at Geneva, the 

 former station had acquired the mean temperature of the day 

 at 6 A. M., which at Geneva occurred three hours later ; and, 

 during the decline of temperature in the afternoon, the mean 

 recurred at the Col du Geant from half an hour to an hour 

 sooner than at Geneva. 



There are other causes besides those just mentioned which 

 contribute to distinguish the daily from the annual curve. Of 

 these the more important are the more gradual character of 

 the annual change of temperature, and the influence of humi- 

 dity. The former affects our experiments by preventing the 

 ascending and descending currents from being instantly esta- 

 blished, in the manner that the law of specific gravity would 

 assign ; and when radiation is least intense (as in winter), and 

 the moving power therefore small, this transfer is often im- 

 peded, and even the law of densities violated altogether. Thus 

 we know to be often the case in winter and in cold climates, 

 that the higher strata are the warmer. To place this in a 

 clear point of view, I shall add a table shewing the number of 

 times in each month that this has occurred, which is indicated 

 under the column headed " Number of times negative :" con- 

 sidering the differences of temperature simultaneously observed 

 at Bonally and Colinton as positive when the former (the 

 higher station) was colder than the latter ; and vice versa. I 

 have also added the extreme positive and negative values for 

 each month ; and though here, more than any where else, 

 errors of observation and registration are likely to have crept 

 in, yet we cannot but be struck with the number of times in 

 which the common law of density has been reversed, and the 

 great excess of warmth observable at the higher station on 

 some occasions, especially in autumn and winter. I would re- 

 peat, however, that the observation of these extremes is less 

 likely to be invariably correct than any other part of the tables. 

 Dividing the year into four seasons, the following summary, 

 which includes both morning and evening observations, is, from 

 the extent of the induction, entitled to considerable confidence. 



