OF ARTS AND SCIENCES. Ill 



over an image of any given diameter. As the solar light entered 

 only through a diaphragm of known dimensions, it was easy to say 

 how much the sun's heat was weakened to balance that from the 

 metal. It must be borne in mind, however, that there was no account 

 taken of the loss of solar heat by reflection and absorption in the 

 lenses, by reflection from the mirror, and more than all by the fre- 

 quent clouds of smoke and steam, while the furnace heat suffered no 

 diminution whatever. Further, every other condition of the experi- 

 ment was designedly such as to weigh in favor of the furnace and 

 against the sun's heat. The value found for the latter, then, is a 

 minimum value. I should perhaps have remarked that experiments 

 had shown that the trifling heat from objects near the melted metal 

 might be neglected. That from the atmosphere about the sun was 

 also insignificant. Except, then, for the diminution of solar heat by 

 absorption, reflection, and so on, our method is equivalent to bringing a 

 specimen piece from the sun's surface (if I may so express myself) face to 

 face with one from the furnace, placing our thermopile mid-way be- 

 tween them, and determining how much we have to diminish the size of 

 the former to make its heat-radiation no more than equal the latter's. 

 The result of these experiments was that the minimum value we 

 can assign to the solar radiation is eighty-seven times that from an 

 equal area of the pouring metal. This, it will be remembered, is not 

 an actual but a minimum value. The true value may be indefinitely 

 greater. 



Part Second. 



Photometric Comparisons. 



Of the complex radiations from any source of high temperature, a 

 part is interpreted by the pile as heat, a part by the eye as light ; but 

 as the temperature is raised, it is now well known that the waves of 

 shorter length increase in amplitude much faster than the longer ones. 

 If the temperature of the sun, then, be much greater than that of the 

 furnace, we shall have a quite independent proof of the fact in a 

 photometric comparison, which, we can safely pronounce a priori, 

 will then give a very much greater ratio of sunlight to furnace-light 

 than that of sun-heat to furnace-heat. To make this comparison, a 

 photometer box, about 8 inches in square section and GG inches in 

 length, is placed so that its central axis lies as before in the path of 

 the reflected beam from the mirror to the furnace. Two similar 



