40 Prof. S. P. Langley on the 



it is generally found best not to note deflexions of less than 

 lmillim. What has just been said refers particularly to mea- 

 surements of the diffused heat from the moon's soil in the 

 invisible lunar spectrum and to the corresponding spectral 

 analysis of the reflected heat. When, however, we place the 

 bolometer directly in the lunar image formed by the 8 -inch 

 aperture, the deflexion throws the needle at once off the 

 scale, and is found on more careful measurement to corre- 

 spond under favourable circumstances to a potential deflexion 

 of about 1500 millim. divisions. Melloni, it will be remem- 

 bered, obtained four or five divisions on his galvanometer 

 with the thermopile and the metre polyzonal lens on Vesuvius, 

 and the immense difference just noted is some indication of 

 the advance of experimental physics in this matter since his 

 day. 



Theory of Observation ; with typical example, 

 showing Method. 



Every observation on the moon, whether on its total heat, 

 as observed directly in the lunar image, or on its diffused 

 heat in the spectrum, should consist in a comparison of its 

 radiations with those of the adjacent sky on either side of it. 

 If our thermometric apparatus had an absolute scale, and 

 there were no intervening atmosphere, it appears, in accord- 

 ance with what has already been said, that such apparatus, 

 when directed not to the moon but to u space " more or less 

 adjacent, should indicate the temperature of this space, which 

 is sensibly that of the absolute zero ; and then, when it is 

 turned upon the moon, supposing it to receive only the 

 emitted and not the reflected heat, it would give, also on the 

 absolute scale, the temperature of the lunar soil. In fact, 

 with such an absolute thermometer, the preliminary com- 

 parison with space would be unnecessary. In reality, we use 

 not an absolute apparatus with a natural scale, but a differen- 

 tial apparatus with an arbitrary scale ; and if we could work 

 without an intervening atmosphere, w T e should, even in this 

 case, require to let the bolometer radiate to space in order to 

 determine the point on our arbitrary scale which corresponds 

 to zero. We should then observe a second point correspond- 

 ing to the temperature of the lunar surface, and having 

 determined the value of the units of our arbitrary scale in 

 terms of the natural one, we should evidently have the 

 quantity sought. 



