Ferret — Measures of Intensity of Solar Radiation. 381 



pression E and R are functions of the temperature ; and such 

 that, for not very great ranges of temperature, we can put 



E=C(1 + cr)d\ R=:R (l+rr). (2) 



In the first of these the temperature r is to be taken for the 

 mean of the range o, the difference of temperature between 

 the two ends of the thermopile, and C and c are constants. 

 In the latter R G is the resistance at the temperature r=0, and 

 r is a constant. With these values of E and R, we get 



0(1 + CT) 



R o (l+rr) w 



Hence I is not proportional to d, except for a constant tem- 

 perature, and for different temperatures there are different 

 relations between I and d. 



Since d is the change of temperature of the plate from being 

 exposed to the sun's rays, and consequently the difference of 

 temperature between the plate and the surroundings, its value, 

 in a state of equilibrium or constancy, must be such, that the 

 plate loses heat just as fast as it receives it from the sun. Now 

 the loss of heat may be by radiation, conduction and convec- 

 tion, but is mostly by the former. The conduction of heat 

 through the air becomes considerable, in comparison with the 

 radiation of it, mostly in the case of curved surfaces only with 

 small radii of curvature, as those of small thermometer bulbs 

 and small wires, and is inconsiderable generally in the case of 

 flat surfaces. But of course much depends upon the nearness 

 of the surface to the surroundings receiving the heat ; for the 

 rate of radiation is independent of this. From the construc- 

 tion of the instrument there cannot be any sensible effect from 

 convection currents, except perhaps from such as arise from 

 the outside disturbances of the atmosphere in windy weather. 



Considering here the effect of radiation only, we would 

 have by the Newtonian law of radiation, the rate of losing 

 heat by the plate, and so the intensity of solar radiation, pro- 

 portional to d for all temperatures and all ranges of £, but this 

 is not the case for any other law. By the law of Dulong & 

 Petit, putting i for the intensity of solar radiation, we have 



i = Bf/{fi 6 — 1 ) = B//o — 1)6 (4) 



in which ^=1*0077 and B is a constant depending upon extent 

 of surface and radiativity, and r is reckoned in degrees of the 

 Centigrade scale. The value of d from this expression in (4) 

 gives 



0(1 -her) i 



R o(l+-) • B(^-l)/' 



