246 PROCEEDINGS OF THE AMERICAN ACADEMY 



effect is probably subjective, owing to tbe great difference of color, 

 and the Bunsen readings are the ones to be relied upon. This would 

 show that the Rumford must never be used to measure lights of 

 different colors, unless the constant error is allowed for. In this case, 

 it amounts to 50 per cent of the whole reading. 



Sunlight. 

 My observations on this source were somewhat interfered with hy 

 clouds ; although on the days available, it was generally clear in 

 the mornings, it nearly always clouded up in the afternoons, which 

 latter were the only times the observations could be made. The first 

 R, Y, and G, were observed at 1 p.m., altitude of sun, 57°, and the 

 rest between 3 and 4.30 p.m., altitude of sun, about 30°. 



Mean Scale Readings. 



R Y o v 



592 352 276 54 



454 325 186 50 



369 201 88 



523 349 221 64 



Brilliancies. 

 r Y o v 



45 100 250 2971 



The enormous value of the violet as compared with that of the 

 preceding lights is very striking. (See Fig. 1.) 



Limits. 

 The spectroscope was exposed to the full rays of the sun. The 

 second V, could not be determined on account of the large amount 

 of diffused light admitted. 



R. V Slit. 



728 395 .030 mm. 



742 1J _. .076 mm. 



14 ... Advanced. 



The total brilliancy of the sun, when at an altitude of 25°, I found 

 to be 64,700 c. p. at 1 meter's distance. Another time, when at 40°, 

 I found it 82,000. That is, it would be equal to about 350,000 full 

 moons. To understand this comparison better, we may add that if 

 the whole visible heavens were turned into one extensive full moon, 

 it would give rather less than one quarter of the lijjht of the sun. 

 The brilliancy has previously been stated at 600,000 full moons. 



