Temperature of the Moon. 39 



and on the further fact, established by Mr. F. W. Very and 

 the writer, that two distinct heat-maxima are observable in 

 the lunar spectrum, one corresponding to the radiation 

 reflected from the soil, the other to that emitted by it. It at 

 first seemed, in accordance with what has just been said, that 

 the accurate determination of the wave-length of this latter 

 maximum would give a correspondingly accurate determination 

 of the temperature of the sunlit surface of the moon ; and, 

 accordingly, to this object the main portion of the observa- 

 tions were given. We may anticipate what follows by here 

 saying that by this method, a perfectly correct one in theory, 

 the writer believes that the temperature of the lunar soil 

 could be determined with great exactness, were it not for the 

 intervention of the earth's atmosphere, which exercises, in 

 this part of the spectrum as in every other, a highly selective 

 absorption, indicated here, however, not by fine lines like the 

 Fraunhofer lines of the solar spectrum, but by enormously 

 wide " cold bands," which vary in size and even in position 

 from night to night *, rendering the exact position of this 

 maximum in a corresponding degree indeterminate. 



Another chapter is occupied by an example of a single 

 night's work in detail, with a statement of some of the pre- 

 cautions and corrections employed in practice. It may be 

 observed here in general as to the apparatus, that while the 

 rock-salt train, as already mentioned, is of such perfection 

 as to show the Fraunhofer lines very completely in the lunar 

 luminous spectrum, the accompanying bolometer and galva- 

 nometer enable us to measure cold bands in the non-luminous 

 lunar or air spectrum, whose heat is otherwise so inappre- 

 ciably small that it corresponds to a radiation of toooVooo 

 of a calorie per second, measured by the generation of a 

 current of 0*000,000, 001 ampere. This is the amount of 

 heat and current implied in moving the galvanometer image 

 over 1 millim. of the scale. In fact, the image is quite 

 steady enough under favourable conditions to admit of the 

 observation of less heat than this, giving deflexions of frac- 

 tional portions of a millimetre ; but owing to fluctuations in 

 the absorption of our atmosphere which transmits this radia- 

 tion, rather than to any limitations of the instrument itself, 



* That an absorption-band may vary in magnitude will excite no sur- 

 prise, but that it should vary sensibly in position may appear to some in 

 contradiction with our knowledge of the fixity of lines in the upper 

 spectrum. An explanation of the anomaly will be found in this Journal 

 for Dec. 1888. 



