350 Lord Rayleigh on the Lower Limit 



to appear in all their characters — at first faintly, and, as it were, 

 sketched in by a dimness and dulness of the otherwise shining 

 and reflective surface of the wetted paper ; but this is speedily 

 exchanged for perfect whiteness, marking by a clear and 

 sharp outhne the lateral extent of the calorific rays, and by 

 due gradations of intensity in a longitudinal direction their 

 law or scale of distribution both within and without the lumi- 

 nous spectrum." 













1 



>> 



€ 



i 



/ 



fl 



a- 



fc 



s 





o 



O 



o 



/- ^ 





















Herschel's Thermal Spectrum. 



Sir John Herschel thus describes the results : — ^^ The most 

 singular and striking phenomenon exhibited in the thermic 

 spectrum thus visibly impressed, is its want of continuity. It 

 obviously consists of several distinct patches, of which «, /S are 

 the most continuous and intense, but are less strongly sepa- 

 rated, and of which, when the sun is very strong and clear, it 

 is even difficult to trace the separation. The spot y, on the 

 other hand, is round and well insulated ; it begins to appear 

 on the paper soon after the ovals a, (3 are fully formed, and 

 when P has assumed a sharply rounded outline. The first 

 symptom of its appearance is the dulling of the wet and sK- 

 ning surface of the paper, which is speedily followed by the 

 appearance of a small round w^hite speck ; this continues to 

 increase rapidly in size and whiteness, and at length assumes 

 a definite and perfectly circular outline, within which the 

 paper is entirely white. By degrees the oval /3 and this spot 

 join and run together, forming a white streak deeply indented 

 at the point of junction. It is not till this happens that similar 

 symptoms begin to betray the existence of a still more remote 

 spot, S. Indeed it generally requires another wash of alcohol 

 before this can be fully brought into evidence. It is, how- 

 ever, perfectly unequivocal, though very much feebler and 

 rather worse-defined than 7, with which also it is somewhat 

 better-connected than y with /3. Of the existence of a still 



