On Refraction. 451 



" It results from this table," he says, " that the effect of 

 moisture iu the air on the refraction is very small ; the ex- 

 cess of the refractive power of the aqueous vapour on that 

 of the air being compensated in a great measure by its less 

 density. We may nevertheless attend to it by means of the 

 preceding table, in cases of extreme humidity. Observa- 

 tions of the hygrometer will point- out the ratio of the quan- 

 tity of vapour spread in a given volume of air to the quan- 

 tity which would produce extreme humidity in this volume* 

 The increase of refraction which corresponds with extreme 

 humidity must then be multiplied by this ratio. " He con- 

 cludes the subject with the following remark : 



iC If we would take into account the figure of the earth 

 in the theory of refraction, it is to be observed that at the 

 point where the observer is situated, we may always con- 

 ceive an osculatory circle to the surface of the earth, whose 

 plane passes through the star: now the figure of the atmo- 

 spheric layers is very nearly the same as that of the earth ; 

 the circles, concentric to the circle in question, are there- 

 fore oscillators likewise of these different figures; and we 

 may determine the refraction of the star by supposing 

 the earth to be spheric, and of a radius equal to this oscula- 

 tory circle. Thus wee see, 1st, That the refraction always 

 takes place in the vertical plane: 2dly, That it is not the 

 same on all sides of the horizon, since the oseulatory cir- 

 cles are not the same in every direction; but we may rest 

 assured, that the error is insensible, when the star is a little 

 elevated. At the horizon, however, differences of some se- 

 conds may occur." Thus terminates one of the most 

 masterly chapters on this subject ever written ; after which 

 he proceeds to treat of the terrestrial refraction. 



Upon these theorems * found by La Place, reduced to 

 rather a more convenient form, and with coefficients differ- 

 ing a little from his,Delambre has computed a set of Tables 

 bv means of which the refraction may be found with great 

 facility. They were first published by Puissant f in a work 

 closely connected with this subject, and are well arranged 

 for use. The first of them gives the refraction for every 

 degree of apparent zenith distance down to SO', and for 

 every 30' from thence to 00\ They are adapted for 0*71 

 metre -of height of the barometer, and 35° of the centigrade 

 thermometer]:. Besides the refraction and its difference for 



* Paje 27 1 , and page 264 of the Mecan. Cel. vol. iv. 

 f Traite de Geodesic, fit the end, 4to, 180J. J Or, £7 953 English 



inches, and 0j° of Fahrenheit's thermometer. 



2 F 2 each 



