112 PROCEEDINGS OF THE AMERICAN ACADEMY 



telescopes of 1.66 inch aperture and 20.01 inches focus, having their 

 objectives outside the extremities of the box and their optical axes in 

 the path of the beam, project, by their eye-pieces, images of the sun 

 and of the pouring metal on the two sides of a Bunsen disk, whose 

 normal position is in the centre of the box. Both images are viewed 

 simultaneously by mirrors attached to the disk, which is movable 

 along a graduated scale. (I here omit certain small corrections 

 applied in practice, and describe the use of the instrument in brief 

 terms.) We do not now need to consider the relative angular areas 

 of the sun and furnace, for so long as both are of appreciable size the 

 images of both falling on the screen, when nearly midway between 

 the two telescopes, will be sensibly proportioned in brightness to the 

 absolute intensities of the sun-light and furnace-light. We do in 

 fact, however, at the outset find the sunlight so immensely brighter 

 that no direct comparison is possible. We then diminish the aperture 

 of the solar telescope (which we will call A), till it has a small known 

 ratio to that of the furnace telescope (which we will call B). In 

 practice Bwas always left with the full aperture of 1.66 inch diameter, 



while that of A was 0.192. Were the original sources of equal inten- 



(0 192 \ - 

 TftMi) — 0.013 -f- or a 



little over one one-hundredth of the other. But it was surprising to 

 see that the image from A was even now incomparably stronger than 

 that formed either by the flame from the blast at its brightest, or by 

 the pouring metal. Under these circumstances, the Bunsen disk was 

 moved from its central position toward B, thus approaching the apex 

 of one light cone and withdrawing from the other, so as to diminish 

 the sunlight still further in an exactly determinable ratio. The low- 

 est value obtained in a series of accordant measures gave intensity of 

 sunlight over (5300) five thousand and three hundred times that from 

 the metal ; and this value is, I think, considerably below the truth. 



It results from these experiments : (1) That direct observation dis- 

 proves the statement that the sun's effective temperature does not 

 exceed 1500° C. It is demonstrably over 1800° C, and may for any 

 thing here shown to the contrary be indefinitely greater. 



(2) The solar heat-radiation, so far from being comparable to fur- 

 nace heat, is at a minimum something like 100 times that from melted 

 platinum, area for area, and probably much greater. 



(3) The solar-light radiation (which offers a more trustworthy 

 indication of the total difference between the sum of all degrees of 

 radiant energy than the heat) is over 5300 times that from a tem- 

 perature above that of melted platinum. 



