Ferrel — Measures of Intensity of Solar Radiation. 383 



From this we get, proceeding in the same manner as in the 

 preceding case, 



_ T 1+cr 1-frr /TV -1 T j, / e— 1\ *)',„. 



^IX-^-/.^-.^) =Ixjl + (- + r + - T -Jj((ll) 



the latter form of expression being sufficiently accurate where 

 A — t— t q is not very large. 



From this expression it is seen that the observed electric 

 intensities, indicated by the deflections of the galvanometer 

 needle, must be multiplied into a factor which is a function of 

 A and T , in order to reduce them to I , the intensities which 

 would have been observed if the apparatus had remained at 

 the same temperature r , or corresponding absolute tempera- 

 ture T . The two metals used in the thermoelectric circuit of 

 Crova's apparatus are unknown to the writer, but we will sup- 

 pose, for the sake of illustration, that they are iron and copper. 

 In this case we should have c= — '00366. The value of <?=4 of 

 Stefan's law has been shown by the writer, in the paper already 

 referred to, to be too large for ordinary air temperatures, and 

 he has given the following formula for computing its value 

 for any given temperature r : 



e= 3 + -00032 r + '00032— --M— (12) 



in which M is the modulus of common logarithms. For 

 r=15°, this gives e= 3 '53 very nearly, but the value varies 

 very little throughout the usual range of working temperatures. 

 With these values of e and e, and that of r in (8), we get from 

 (11), with T = 272-f 15 = 288°, 



I o = IX(] +('O0366 + 'O038 + '00S8)J) = Ix(l + '01526z7) (13) 



In this expression if we assume the temperature r in (6) 

 from which A is reckoned to be that of 10 o'clock. Then the 

 values of A for the hours preceding that hour would be nega- 

 tive, and after that hour positive, in all ordinary cases of fair 

 weather, and hence the observed values of I would be decreased 

 before 10 o'clock and increased after that hour, and it is readily 

 seen that this would tend to throw the maximum at least 

 toward midday, and to fill up the observed depression in the 

 afternoon. It does not appear adequate, however, to render 

 the forenoon and afternoon curves symmetrical, and at least a 

 considerable part of the effect which destroys this symmetry is 

 undoubtedly to be attributed to the different hygrometric 

 conditions of the air in the forenoon and afternoon. But both 

 effects are in the same direction and have the same epochs of 

 maxima and minima, so that there is no way of separating 

 them and determining how much belongs to each. But even 



