32 



MEMOIES OF THE NATIONAL ACADEMY OF SCIENCES. 



the light intensity ratio of moonlight and sunlight, may also be taken as the ratio of the ortlinates 

 of the lunar and solar energy curves. 



For the ordinates of the normal solar energy curve we may take the values given by the meau 

 of all noon observations made with the spectro-bolometer at Allegheny during the spring of 1881. 

 Then, multiplying each ordinate by the corresponding factor given in the last colurau of the pre- 

 ceding table, we obtain the ordinates of the lunar energy curve. The results are exhibited below 

 in tabular form : 



If we wish the lunar energy curve to represent the distributiou of the same amount of energy 

 as the solar within the limits of the visible spectrum (say between Qi^.i and t'*'.7), we must multiply 

 each of the lunar ordinates by the fraction fff, which is determined by plotting the curves 

 and measuring the areas within the required limits. We obtain by this operation the following 

 table, which is also graphically represented by the curves in Plate 6. 



An inspection of these curves shows at once the effects of the selective absorptiou undergone 

 by the solar rays at the moon's surface. The maximum ordinate of the lunar curve falls much 

 lower down in the spectrum, and there is a con-esponding reduction in the height of the curve 

 0%'er the violet end. The visible part of the normal siiectrum forms, however, so small a ])ortion 

 of its entire length, that it would be unsafe to judge from the nature of the lunar curve obtained 

 by optical means, as to its probable course at points very far below the limit of the visible red. 

 Nevertheless, the evidence of these i)hotometric measurements as to the selective reflection exer- 

 cised by the moon's surface is, as far as it goes, decisive, and it is shown to be in such a direction 

 as to cause a preponderance in the lunar spectrum of the rays of long wavelength, and hence to 

 tend to cause a smaller percentage of lunar rays to be transmitted by glass than of solar, and this 

 independently of any effect from heat reradiated by the lunar soil. There is, theu, no doubt that 

 the observed phenomenon of glass absorption already described is due in part to this cause, though 

 in how large part we do not now determine. 



ADOPTED HEAT-MEASUEES WITH BOLOMETER AND GALVANOMETER. 



The galvanometer is so important an accessory of the bolometer, that we will describe the 

 arrangement we have used to make our own most eflective. 



The galvanometer employed is a Thomson differential astatic galvanometer, having a resist- 

 ance of 20.35 ohms, and ori ginally made by Elliott Brothers, with a short suspending fiber, a damp- 



