i 



290 Mr. S. P. Langley on the Amount 



fhe atmosphere to see what the radiation really is, all of 

 them, however skilful, must depend on inference to determine 

 what it would be, if they could thus observe it. It is certain 

 that nearly all have used a formula of which it seems capable 

 of absolute demonstration that it is not only erroneous, but 

 that its error always lies in one direction, so as to invariably 

 make the calculated absorption too small ; and it may be 

 further shown, with an evidence which seems little less than 

 demonstrative, that the numerical value of the error is very 

 large in relation to the quantities involved. 



I have been led, however, not by theoretical considerations 

 alone, but by experimental investigation (during the course 

 of which I have observed both near the sea-level and at great 

 altitudes), to the conclusion that the laws under which solar 

 and stellar light and heat are absorbed by the atmosphere 

 are so complex that their complete comprehension is still 

 beyond our power, but that we may at least now improve 

 our present ways of studying them ; for investigators have, 

 however skilful in their observations, commonly ignored the 

 complexity of the problem of their reduction, and, assuming 

 that it is as simple as we could wish it to be, have proceeded 

 to compute the result by such a formula as would be most 

 convenient for us, if nature would follow. Thus, owing to a 

 natural tendency to accept as sufficient any scientific dogma 

 which has respectable sponsors and which saves labour, the 

 simple rule, established over a century ago by Bonguer, and 

 consecrated by the use of Herschel and Pouillet, to whom it 

 embodied all the knowledge of their time, is commonly used 

 by us to-day ; though we may easily be convinced that it has 

 ceased to express the facts known to us. 



To justify this language, let us consider what the problem 

 appears to be at first glance, and what the first suggestion is 

 for solving it. If a beam of sunlight enters through a crevice 

 in a dark room, the light is partly interrupted by the particles 

 of dust or mist in the air, the apartment is visibly illuminated 

 by the light laterally reflected or diffused from them, and the 

 direct beam, having lost something by this process, is not so 

 bright after it has crossed the room as before. In common 

 language, the direct light, to an observer in the path of the 

 beam, has been partly " absorbed, " and the problem is, to 

 determine in what degree. If a certain portion of the light 

 (suppose one fifth) was thus scattered, the beam after it 

 crossed the room would be but four fifths as bright as when 

 it entered it ; and, if we were to trace the now diminished 

 beam through a second apartment altogether like the other, 

 it seems at first reasonable to suppose that the same proper- 



