Mr. J. J. Waterston on Solar Radiation. 499 



circumference of the bulb touched the brass, and its upper side 

 was enclosed with cork, while the lower was exposed to the air 

 within the tube, but was untouched by the rays of the sun that 

 passed through. The internal diameter of the tube was 0*9 inch, 

 and length 6 inches. The bulb of the solar thermometer was 

 0*42 inch in diameter, spherical in shape, and fixed in the centre, 

 as shown in the figure. Its shadow was an easy guide in moving 

 the tube in altitude and azimuth to keep pace with the sun. It 

 is difficult, if not impossible, to demonstrate that the thermo- 

 meter Y shows the exact temperature of the inner surface of the 

 tube. It was subjected to three tests. 



1. The instrument being out of the sun's rays, and X and Y 

 showing the same temperature, it was removed to a place where 

 the atmospheric temperature was 10 degrees lower. Both thermo- 

 meters descended, and showed a difference equal to about one- 

 tenth the amount they had to fall to arrive at the atmospheric 

 temperature, Y being so much in advance of X. 



2. A bat's-wing flame of gas was brought within 3 inches 

 fronting the middle of the tube ; both X and Y rose together, 

 keeping pace exactly. 



3. While taking observations, the heat absorbed by X from 

 the sun, and again emitted from it and transferred to the tube, 

 gradually raised its temperature until a maximum was obtained. 

 Now, comparing X and Y while both are rising, and after having 

 obtained their maximum, a difference of o, 3 was remarked ; and 

 this difference, no doubt, affected isolated observations when this 

 maximum was not attained in consequence of interruption by 

 clouds passing, when it was usual to heat the solar thermometer 

 artificially to near the stationary point, in order to save time, 

 the great inconvenience of the apparatus in this climate being 

 the slowness with which r obtained its final value. An arrange- 

 ment with a differential air-thermometer would, no doubt, be 

 preferable in this respect, but the absolute value of the degrees 

 indicated does not seem capable of being exactly determined*. 



The thermometers were carefully graduated and compared by 

 myself, and the divisions between two fixed points, 60° and 100° 

 (which included all the observed temperatures), were drawn 

 as nearly equal as a good ivory scale and magnifying lens would 

 admit. The length of a degree on the scale of Y was about 

 •05 inch, and upon X *0G7 inch : with practised eye it was easy 

 to read off the temperature to g^-th of a degree with lens; 

 but such accuracy was unattainable for other reasons, and chiefly 



* Another mode, which is perhaps the least liable to uncertainty, is to 

 have the surface against which the sun-thermometer radiates maintained at 

 212° by surrounding it with the steam from boiling water : only one thermo- 

 meter would be required in this arrangement. 



2L2 



