NATURAL PHILOSOPHY. 155 



ATMOSPHERIC REFRACTION. 



Notwithstanding the great importance of solar eclipses in astronomical 

 calculations, their value has been hitherto much diminished by a certain 

 want of agreement of the phenomena observed with the calculations of the 

 most competent astronomers. The moment when the eclipse becomes total, 

 as well as the places over which the shadow passes, and the duration of ob- 

 scurity, all commonly differ, in a most provoking manner, from what theory 

 would seem to indicate. On this subject, M. Liais has written a letter to the 

 Astronomer Royal, of Great Britain, in which he points out a source of error 

 which had hitherto escaped the researches of the most distinguished samns. 

 The law by which a ray of light, passing obliquely from a rare into a denser 

 medium, is deflected from its path so as to enter the dense body less obliquely 

 than it could have done by pursuing a straight course, is well known to hold 

 good with respect to atmospheric strata of different density. This refraction 

 causes the heavenly bodies to appear higher up in the sky than they really 

 are; and the denser the atmosphere, and the nearer the luminary is to the 

 horizon, the more Avill this effect be apparent. This refraction M. Liais calls 

 the regular refraction; but, besides this, there exists an abnormal refraction, 

 which takes place only on occasions of eclipses of the sun. It will be readily 

 understood that the sun's rays being cut off from a portion of the earth by 

 the interposition of the moon, the temperature decreases, and the strata of 

 the atmosphere becomes denser over the place where the moon's shadow 

 falls; thus a cone of comparatively dense air, surrounded by that which is 

 expanded by the sun's heat, is created, which will cause a variation in the 

 refraction of the solar rays. The tendency of this refraction will evidently 

 be to diminish the extent of ground covered as well by the umbra as the 

 penumbra, and to make the eclipse at any given point to commence later 

 and end sooner, in other words, to be shorter, than previous calculation 

 would indicate, if this abnormal refraction had not been taken into account. 

 The amount of these refractions, depending as they do on the height of the 

 sun and on variations in the atmospheric density, from a variety of causes, 

 can never be calculated beforehand, but the necessary data can easily be 

 obtained as the moment of eclipse approaches, by which to make the neces- 

 sary corrections. Besides alterations in the apparent position and duration 

 of the eclipse, these refractions produce several remarkable phenomena, 

 which are only to be observed during a total solar eclipse; of these the pe- 

 culiar blood-red color of the moon may be mentioned, as well as the appar- 

 ent projection of the red flames, of which we know so little, upon the 

 moon's disk during the eclipse of 1842. Again, slight irregularities in the 

 refraction of different portions of the sun's edge may tend towards the pro- 

 duction of what are known as Baily's beads, which have been frequently 

 observed in cases where it is difficult to suppose the existence of lunar 

 mountains to hare caused them. M. Liais proposes, with a view to correct- 

 ing eiTors in the determination of longitudes by eclipses, that the different 

 phases of the phenomenon, as well before as after the moment of total ob- 

 scurity, be photographed, and the least distance of the centres at the place 

 of observation calculated from the variation of the angle of position of the 

 cusps. The intersection of the line between the centres to be determined bv 

 calcul vtion, together with the latitude of the place of observation, will give 

 the longitude independently of these abnormal refractions. 



