TEMPKIJATLTEE OF TDK SURFACE OF TOE MOON. l9 



a pii't-e of plnte glass ("A") 0.9 niilliiiieters tliick, ajiparOntly of Eiiglisli make; tlie tliird ("B") a 

 piece of Ameiicaii plate glass, sliglitl.v greoiiisli in hue, C.(! inillinieteis tbick, and tlie fourtli a large 

 pane of American window glass of good (piality. 



On July n, 1883, October -1, 18S4, and November 20, 1884, the diathermancy of these specimens 

 of glass was determined for solar rays. The values obtained were : 



Tor cent. 



For tlie Hilger glass 86 



For Glass A S(i 



For Glass B 77 



For the large pane 70 



the zenith distance of the sun being about .50°, or in the last case over 00°. 



We have not as yet reduced these observations taken at somewhat different altitudes of the 

 sun and moon to one common altitude, because it is sntliciently obvious from a comparison of the 

 above figures without such a reduction with those indicating the absorption of the lunar heat by 

 the same siiecimens of glass, that these absori)tions are strikingly different in the two cases, and 

 far more so than any difference in the altitudes of the bodies under consideration can account for. 

 "We will iiass, then, to our ]ireliminary obser\ations on this unequal absoriition before introducing 

 small corrections which would be sn])erfluous at this stage of the inquiry. 



Preliminary measurements were made in pursuance of this system on the night of June 21, 

 1883, when all atmospheric conditions appeared to be favorable. A screen was, however, employed 

 ou this night to cut off the lunar rays, and the exposures made by withdrawing it (a method not 

 capable of giving exact results). 



If the glass be interposed while the lunar rays are falling ou the bolometer, radiations from or 

 to its substance will in general be confused with the effect to be studied, and further the radiations 

 of low refrangiljility from surrounding objects, such as those from the substance of the condensing 

 mirror, will be cut off by it. It is always a cundition of good observation, tlien, that the glass be 

 placed in front of the bolometer and the instrument allowed to register its separate effect before 

 the lunar rays are allowed to fall upon it; and this method has always been used. 



It was Ibund from the series of observations that the i)ercentage of the total lunar heat trans- 

 mitted by glass A was 70 per cent, and from a second series 54 percent., giving a mean for glass 

 A of 02 per cent. 



For the ])ercentage transmitted by glass 1> were obtained the values GO i)er ceut.,.4Gper cent., 

 and 58 per cent., giving a mean of 55 per cent. 



The discrepancies in these preliminary results were partly due to the above mentioned 

 erroneous method of exposure by withdrawal of a screen, but also to the difficulty of telling when 

 the luiuir image was exactly coincident with the bolometer face, since, when this was not the case, 

 part of the heat was wasted and the deflection obtained too small, and the eye could not be 

 safely brought into a position where the strijis could be seen, since then the radiation from the ob- 

 server's face gave a large detlection. This difficulty was subsequently overcome by placing a large 

 sheet of glass behind the mirror, through which the observer could regard the bolometer strips 

 without producing any disturbance of the galvanometer, the radiations from his face being com- 

 pletely intercepted by the glass. The conclusion, apparently resulting from the preliminary ex- 

 periments where a screen was used, is that the specimens of glass appeared to transmit the greater 

 part of the moon's rays; but that the ai)parent transmission of the glass should be greater w hen ex 

 posure is made by the withdrawal of a screen may be inferred from the following considerations. 

 The screen is in general warmer than the external air, and if withdrawn while the bolometer was 

 directed to the sky near the moon, a negative detlection of the galvanometer would be produced. 

 When the screen is withdrawn while the bolometer is directed upon the moon, the heating effect of 

 the latter is partly counteracted by the cooling effect of the sky or air between us and the moon, 

 and the detiectiou obtained is smaller than that which would have been produced if the bolometer 

 had been continuously exposed to the radiation irom the sky. When, however, the glass is inter- 

 posed, it forms a barrier to the interchange between outside objects of low temperature and the 

 bolometer, and nearly the same detlection is obtained whether a screen is used or not. By the use 



