20 MEMOIRS OF THE NATIONAL ACADEMY OF SCIEXCES. 



of a screen, therefore, the aj^parent transmission of lunar heat by glass is larger than it otherwise 

 would be. * 



These and other preliminary observations, then, are not used in the final results, but it has 

 been thought worth while to refer to them to indicate some of the subtle causes of error which 

 beset the commencement of such a research. 



SUBSEQUENT OBSERVATIONS ON THE TRANSMISSIBILITY OF GLASS FOR LUNAR RATS. 



In October, 1884, these observations were again taken up and the transmissibility of the same 

 pieces of glass redetermined. 



The method of procedure was as follows : The apparatus being in adjustment and the galvano- 

 meter needle in a position of equilibrium, the lunar image was thrown on the bolometer by turning 

 one of the concave-mirror screws, and the deflection of the galvanometer noted. Then the lunar 

 image was thrown oft' the bolometer, and the new position of equilibrium noted, to which the gal- 

 vanometer needle returned. This was generally slightly different from the original position. 

 From these readings was obtained the effect of the uninterrupted lunar beam. A piece of glass 

 was then interposed immediately in front of the bolometer case, and alter the galvanometer needle 

 had taken a\t a new jiosition of equilibrium, caused by the alteration of conuitions in regard to 

 its thermal exchanges with the glass itself and with outside objects, the same operation was re- 

 peated, and the effect of the lunar ray obtained after it had suffered reflection at the surfaces and 

 absorption in the substance of the glass. The following resnl's, which are the means of repeated 

 observations, were obtained bj' the writer under favorable conditions, except that exposures were 

 made by touching the adjusting screw of the concave mirror, instead of tliat of the siderostat 

 mirror. The image of the moon is thus replaced by that of the neighboring sky, but since the 

 concave mirror is within the building, an alteration of the thermal conditions may be produced by 

 an increased reflection of heat from the walls of the ai)artment in the mirror. 



* To put this into syniljolical form, let C = the amonnt of heat received from the walls of the bolouieler case, « = 

 the heat received frniii tlie sereeu, g = the heat received from the substance of the glass, S = the heat received from 

 the sky, »ii = the portion of lunar heat transmitted by glass, and m-z = the portion of lunar heat absorbed by glass. 



First Ktrji of uhxervulion. — Tlie bolometer is exposed to the radiation from the screen, which is interposed betwc en 

 it and the sky. The heat received is then C -\- s. 



Second step of ohsen'alion. — The screen is then withdrawn and the bolometer exposed to the moon. The heat re- 

 ceived is C -|- Mil + i«2 + S. 



The deilectiou of the galvanometer which is produced is due to the difference between these two amounts of heat 

 or to nil + 1112 + <S' — ». 



'Third itep of oliserrution. — The screen is again interposed, and the plate of glass, completely cutting oft' the radia- 

 tion from the screen, is ])laced in front of the bolometer, which then receives the amount of heat C -\- g. 



Fourth itip of ohserfntiDii. — The screen is now withdrawn and the bolometer again directeil toward the moon. The 

 he.at from the sky, S, which consists entirely of radiations of long-wave length, is also completely cut otf by the glass, 

 and the amount of heat received by the bolometer is C -f »»n -f- g, so that the resulting deflection of the galvanometer 

 is proportional to (C -j- m.; + ff) — (C -|- g) = iii^. 



The ratio of the two deflections obtained as above is the apparent transmissibility of glass for lunar heat, and it 

 is therefore 



= (, 



nil +"11! + *' — s 



If observations had been made without the use of a screen, by moving the siderostat mirror, the transmissibility 

 would have been 



JBi -I- nki 



= f2 



The quantity of heat, <S, depends upon the apparent temperature of the sky: that is, npon the temperature of the 

 external air, and as this, except iu an unusual combination of circumstances, is lower than the temperature inside 

 the building, S — 8 is negative, and consequently /i ]> t^. 



