304 CLAUSIUS ON THE CONSTITUENTS OF THE ATMOSPHERE 



a quite definite assumption, and indeed prove the same to be 

 almost the only one possible. These conclusions constitute the 

 object of the following considerations. 



In the first place, it is clear that the different strata of air 

 cannot, by reason of their different densities, produce this re- 

 flexion. As the limiting surfaces of these strata, in so far as 

 it is permitted to conceive of them separately, are -nearly pa- 

 rallel to the earth's surface and to each other, these surfaces 

 could only reflect the sun-light in particular directions, which 

 indeed would never strike the earth's surface at all, so that the 

 firmament must appear dark throughout. From the actual 

 circumstance that at every point of the earth's surface light is 

 received from all directions, we must rather conclude that re- 

 flecting surfaces of all possible directions exist in the atmosphere. 

 Hence through the atmosphere we must imagine innnumerable 

 masses to be distributed, the densities of which differ from that 

 of the surrounding matter, and at the limiting surfaces of which 

 the reflexion takes place. 



Leaving the question whether these masses are foreign ones 

 which float in the atmosphere, or the constituents of the atmo- 

 sphere itself, for the present undecided, one question must be 

 stated here and answered ; and that is, whether these masses 

 are transparent or opake. With regard to the latter, it might 

 perhaps be affirmed as possible, that small non-transparent 

 bodies which strongly reflect the light at their surfaces may 

 float as dust in the atmosphere, and although imperceptible 

 singly, on account of their minuteness, may still by the number 

 modify the transparency of the atmosphere and effect a sensible 

 reflexion. Against such an assumption, however, so much 

 might be urged, and it is even at first sight so improbable and 

 inexplicable how" such masses, which in optical respects at least 

 behave like polished metal, could be so uniformly distributed 

 throughout all regions of the atmosphere, that the supposition 

 may be permitted to fall to the ground. It may perhaps be 

 asserted more definitely, that oi foreign bodies, only water as the 

 precipitate of vapour can exist in sufficient quantity in the at- 

 mosphere to produce so considerable a reflexion. 



Assuming, then, the masses which cause the reflexion of light 

 in the atmosphere to be transparent, an important starting- 



