TEMrEKATUKE OF TUE SUKFACE OF THE MOON. 37 



lunar measincinents. The following deflections were observed, the sensitiveness of the measuring 

 apparatus being the same as during the lunar observations of the previous evening: 



Teiiiperatiire Galvaiiome- 



of tcr 



Leslie cube. (k'lloctioii. 



°C. Div. 



95 408 



9-2 400 



b9 :i84 



86 370 



83 369 



HI 353 



73 '«t7 



From a smooth curve we adojit 435 as the presumable deflection, which would have been ob- 

 served under these conditions at 100° C. The screen itself acquired a minute amount of heat 

 during the experiment, but the correction for this is negligible. 



The bolometer strips attain thermal equlibrinm under ordinary circumstances within a traction 

 of a second, while on account of the slowness of change of the temperature of the case, we can 

 assume its radiation (0) to be constant during the experiment. The temperature of the bolometer 

 strips may always be taken to be proportional to the angular area of the part of the surface radiat- 

 ing to them, to its temperature, and to its emissive quality. 



Thus the aperture of the moon bolometer occupies 0.00565 of the sphere. The temiierature 

 of the room December 3, 1SS4, was 0°.0 0. If the aperture had pointed to a surface at the 

 absolute zero, having the same emissive jjower as its case, a fall of temperature of 0.00565 multi- 

 plied by 273° = 10.542 would have been experienced. We assume that, within the limits of this 

 experiment, the iSTewtonian law of radiation holds. If the pointing had been to a surface at 

 100° C. of the same emissive power, the temperature of the bolometer would have risen 0°.565. 

 Now, we have seen that a Leslie cube at 100° C. would have produced a deflection of 435 divis- 



Qo_565 

 ions on the galvanometer. One division, therefore, indicated ^W— =0°.0013 change of temper- 

 ature of the bolometer strips (the full sensitiveness of the galvanometer not being used). The 

 deflection produced by the full moon on the previous evening, if reduced to zenith, would have 

 been over 3.50 divisions, and the temperature of surrounding objects being 0° C. the effective radi- 

 ation of the moon, if we suppose its emissive power the same as that of the case, was such as would 



350 

 correspond to a temperature of jo^Xl00°i= + 80°.5 C, or 80o.5 C. above the temperature of sur- 

 rounding tei-restrial objects, which happened to be zero Centigrade. This on the absolute scale gives 

 800.5+2730=3530.5; and if one-fourth of the lunar radiation is reflected sun heat the true aver- 

 age temperature of the moon at the full is 80o.5 — j x353o.5 = — 7o.9 0. If we assume that one-half 



only is reflected sun heat, we have — 99o.3 C, if that one-sixth is reflected, + 21o.C C. 



A correction for atmospheric absorption which we have not applied would somewhat increase 

 these values, but it is evident that not only the experimental conditions here do not favor accu- 

 racy but that the results, such as they are, are subject to a wide latitude of interpretation. 



Class 3. — Trassmissiox of lunar heat bv tub earth's atmosphere. 



The remarks already made as to the difficulty of comparing observations at different altitudes 

 but at different phases, when the law of change of heat with the phase is so imperfectly known, 

 apply with peculiar force to this class of observations. Only a series of observations made exclu- 

 sively on the full moon in favorable circumstances (and therefore occupying many years) could 

 bring anything like satisfactory evidence. The reductions which we now give of the few values 

 we possess lead to conclusions to which we cannot attach great weight. 



The observations of December 2, 1884, are given below in tabular form, with the computa- 



