in the Magnetic Forces of the Earth. 219 



place, it appears from equation 4, p. 48, that the loss of tempera- 

 ture at the surface (7), is equal to the entire loss of heat (L) minus 

 the sum of all the losses of temperature of the layers below the 

 surface. It follows therefore, as L is always the same for the 

 same number of hours, that I cannot become less unless the sum 

 of V, I", /'", &c. becomes greater ; now it is a matter of observa- 



tion that when I is less, V, V\ V", &c., are also less; the sum of 

 l\ V\ &c, cannot therefore be greater in winter than in summer, 

 unless the number of layers which experience a change of tem- 

 perature during the night increases to a sufficient extent to com- 

 pensate for the diminished loss of temperature of those nearer the 

 surface. This cannot be the case, for, in the first place the loss 

 of temperature of the lower layers is always comparatively trifling, 

 and in the next place the depth to which any perceptible cooling 

 in consequence of surface radiation extends during the same pe- 

 riod of ten or twelve hours, must depend almost entirely upon the 

 length of the interval. Or, we may arrive at the same conclu- 



sion m another way. 



f 



heat gained during the day is lost during the night. The same 

 is true in July. Now the quantities of heat gained during the 

 day are quite different at these two epochs, the one of minimum 

 and the other of maximum temperature, and therefore L must be 

 variable, whereas it is constant. In the next place, although we 

 are in want of the systematic observations of the diurnal varia- 

 tions of I, l f % yt % &c., which would enable us to ascertain with 

 certainty whether the sum of V, l u , <fcc. varied materially during 

 the night or not, and therefore whether the variations of I are at- 

 tributable to an unequal upward flow of heat ; still, if we assume 

 that the established law connecting the total variations of tem- 

 perature of the different layers, viz., that they decrease according 

 to a geometrical progression for equal increments of depth, is very 

 nearly true for the variations that occur in any given time, and 

 take £ for the ratio of the progression, (derived from the data on 



5, answering to the depths of l m , 2™, 3 m , &c, taking va- 

 cation at surface =15°,) recollecting that the cooling progresses 

 at . the rate of about 1£ inches per hour, we shall find that the 

 sum of the losses of temperature of the layers below the surface 

 »P to the hour of 4 a. m. is materially less than twice the sum of 

 the losses up to the hour of 10 p. m.— midway between the hours 

 of maximum and minimum temperature. It can only be made 

 *qual to twice this sum by supposing that the fall of surface tem- 

 perature is increased very nearly in a two-fold proportion— in 

 other words, that it is proportional to the time, or uniform. This 

 result does not depend upon the ratio £ supposed to be used in 

 the calculation, but upon the fact that the losses of temperature 

 of the new layers below the depth of 8« n , or 9 in , after 10 p. m., 

 are very small. If it be erroneous, it can only be because the 

 law which I have assumed does not accord with fact. In view 



page 45 



