ON LIGHT. 337 



which, therefore, do not interfere with or enfeeble one 

 another in any part of tlieir previous course. The image 

 of a star on the retina is formed by the union in a focal 

 point of the wjiole bundle or/<i7/<f/7 of parallel rays con- 

 tained within a cylindrical space or column, having the 

 circular opening of the pupil for a base or section, con- 

 tinued through the whole atmosphere, however far it may 

 extend. Now the air, though a very fcel)ly refracting 

 medium, has still a certain amount of refractive power, 

 and that a variable one, depending on its density, tem- 

 perature, and moisture ; and corresponding to the 

 degree of this power is the velocity with which it is tra- 

 versed by the luminous undulations. Now ; however 

 the density, temperature, and moisture of the lower and 

 upper regions of the air may differ ; if throughout the 

 whole extent of this column it were perfectly uniform in 

 these respects, at eveiy point of each cross section of it 

 (however it might differ in different sections) all the rays 

 traversing its length from the star to the eye would have 

 their undulations equally retarded by the aerial medium : 

 and therefore all the rays belonging to any one wave 

 setting out at the same instant of time from the star 

 would reach tlie focal point on the retina at the same 

 moment ; such being the condition which determines the 

 focal point of a lens. But if the air in one side of the 

 column should for any considerable distance along it be 

 slightly different in these respects from that in the other, 

 the undulations transmitted along that side would be 

 differently retarded from those along the other, and would 

 not arrive on the retina at the same instant. The one 



Y 



