452 On the Existence of an Atmosphere swrounding the Sun. 



portional to the time, if we apply corrections to the angles p 

 differing little from half of those in the first column of our 

 Table I. ; that their evidence is on the whole in favour of an 

 atmosphere causing refractions of about half the amount there 

 tabulated, but is opposed to refractions much exceeding those in 

 amount. 



As a definitive result, the conclusion will probably be con- 

 sidered as entitled to little weight ; yet it is such as would agree 

 pretty well with the idea that the red protuberances are cloudy 

 masses of great tenuity held in suspension, as well as with the 

 absorptive action on light and heat, which has been previously 

 noticed by observers. 



It is at least satisfactory to find that the motions of the spots 

 offer no objections to the admissibility of an atmosphere, such 

 as common air of a uniform density increased from 8 to 10 

 times, and extending to a distance from the surface equal to 

 one- fourth the radius, but rather favour the assumption of an 

 equivalent atmosphere, while they suffice to assign quickly a 

 superior limit to its amount. 



The observations discussed, not having been corrected for 

 terrestrial refraction, it is desirable to point out that the angles 

 p are sensibly independent of correction from that source, being 

 derived from the expression 



^=sin ip + p), 



where r and R are respectively the linear distances of spot and 

 limb from centre of disc, which, as is well known, are by re- 

 fraction contracted in very nearly the same proportion. 



The principal source of error which may affect our result is 

 the omission of the effect of perspective on the apparent centre 

 of a spot's nucleus, according as we adopt the conclusions of 

 Dr. Wilson or Sir W. Herschel on that subject. But if we ex- 

 clude observations very near the limb, and take only spots which 

 exhibit penumbra on the side next the centre, our conclusions 

 will not be affected by Dr. Wilson's hypothesis ; excepting, it is 

 true, that the radius of the disc employed should be diminished 

 by the depth of the spot's cavity. On Sir W. Herschel's hypo- 

 thesis (the word is perhaps objectionable), it would similarly be 

 necessary to diminish the radius employed by the distance 

 between the surfaces of the photosphere and the cloudy stratum. 

 But if a spot be taken, as in the second series of our second 

 spot, which exhibits no penumbra, no such objection will lie. 

 The remaining difficulty is inherent, the little permanence of 

 such spots. 



