LIMIT OF THE ATMOSPHERE. 49 



Arago has ingeniously shown, on optica* grounds,* that the 

 variable stars which always exhibit white light without any 

 change of colour in their periodical phases, might afford a 

 means of determining the superior limit of the density to be 

 assumed for cosmical ether, if we suppose it to be equal to 

 gaseous terrestrial fluids in its power of refraction. 



The question of the existence of an ethereal fluid filling 

 the regions of space is closely connected with one warmly 

 agitated by Wollaston, 41 in reference to the definite limit 

 of the atmosphere, a limit which must necessarily exist at 

 the elevation where the specific elasticity of the air is equi- 

 poised by the force of gravity. Faraday's ingenious experi- 

 ments on the limits of an atmosphere of mercury (that is, 

 the elevation at which mercurial vapours precipitated or 

 gold-leaf cease perceptibly to rise in an air-filled space; 

 have given considerable weight to the assumption of a 

 definite surface of the atmosphere " similar to the surface 

 of the sea.'' Can any gaseous particles belonging to the 

 region of space blend with our atmosphere and produce 

 meteorological changes ? Newton ** inclined to the idea that 



43 ' En assimilant la matiere ires rare qui remplit les espaces 

 celestes quant d ses proprietcs refringentes aux gas terrestres, la 

 densite de cette matiere ne saurail depasser une certaine limite 

 dont les observations des etoiles chanyeantes, p. e. celles d' Algol 

 ou de /3 de Persee, peuvent assigner la valeur." Arago in the 

 Annuaire pour 1842, pp. 336-345. "On comparing the 

 extremely rare 1 matter occupying the regions of space with 

 terrestrial gases, in respect to its refractive properties, we 

 shall find that the density of this matter cannot exceed a 

 definite limit, whose value may be obtained from observations 

 of variable stars, as, for instance, Algol or /3 Persei." 



41 See Wollaston, Philos. Transact, for 1822, p. 89: Sir 

 John Herschel, op. cit. 34, 36. 



42 Newton, Pn'n>\ Mat/tern., t. iii. (1760) p. 671 " 



VOL. 111. * 



