858 PROCEEDINGS OF THE AMERICAN ACADEMY 



applied. The rough galvanometer deflections for different wave-lengths 

 are, where the slit used is so narrow as to give all the principal 

 Frauenhofer lines sharply on the screen, as follows. (In correspond- 

 ing curves, wave-lengths are abscissae ; reduced mean galvanometer 

 deflections, ordinates.) 



mm. 

 A = .00035 .0001 .0005 .0006 .0007 .0003 .0009 .0010 .0011 



div. 

 Defl. 12 55 207 24G 198 129 80 58 41 



The corrections for the underlying second and third spectrgi not 

 being fully applied, it can only be said that these values are trustworthy 

 (as first^ approximations) as far as from .00035 mm. to .0007 mm. 

 The values below .0007 which give the sum of first, second, and third 

 spectra are, perhaps, too large. They are hitherto unpublished, and 

 they at least, though as yet approximative, show that the heat maxi- 

 mum in a normal spectrum is not in the ultra-red, but is at least a^ far 

 up the spectrum as the orange near D ; and this result may be relied 

 on, any smaller values below A = .0007, as well as all favorable 

 atmospheric circumstances (high sun, blue sky, etc.), rather tending 

 to move it toward the violet. 



These measures show a certain approximation of the '' heat " curve 

 to the " light " curve, though these are commonly drawn with their 

 maxima in entirely different parts of the spectrum. Viewing the actual 

 distribution of energy here, and then comparing it with the so-called 

 " heat," " liglit " and " actinic " curves of the text-books, we have evi- 

 dences, I think, of interesting results already reached by this instru- 

 ment, and which have been possible only by its use. Among these, 

 we observe that (contrary to the statements of our text-books, and 

 contrary, as I think, to most present scientific opinion) the great pro- 

 portion of all solar heat received at the earth's surface does not appar- 

 ently lie in the non-luminous parts of the S2)ectrum. Not only is the 

 heat maximum in the luminous part, but the total sum of non-luminous 

 heat (as far at least as our measures extend) is relatively small, the 

 joint effects of ultra red and ultra violet radiations (so far as measured) 

 not making up the sum of those in the visible portion. This is a result 

 to me unexpected, but which I thuik may be relied on ; and it is, if true, 

 sufficiently important to make me hope that the society will feel that the 

 instrument whose construction they have promoted is already of utility. 



The above illustrations form i)arts of other researches, and are bor- 

 rowed for this paper merely to show the action of the instrument, and 

 to enable each reader to judge for himself of its value as a measurer of 

 radiations. 



I 



