1878.J last two total solar eclipses. 213 



In 1860 Mr Prazmowski made an important observation indicating 

 that a maximum of polarisation existed a few minutes away from 

 the sun ; and that from that point both away from the sun and 

 towards the sun the polarisation decreased. This result was 

 confirmed by Mr Janssen in 1871, and during the same eclipse 

 Mr Winter made a series of measurements which gave a few 

 minutes away from the sun a greater polarisation than close to the 

 edge of the sun. Other and partly contradictory observations 

 exist, but the great mass of them goes only to confirm the results 

 which I mentioned. The existence of polarisation at once suggests 

 the analogous case of our own atmosphere, and it is natural to 

 refer the two cases to a common source. The polarisation of our 

 atmosphere follows the law of polarisation due to the scattering of 

 fine particles. The mathematical theory of this law has been 

 given by Lord Rayleigh, and I have taken his formulae as starting 

 point. The case of our own atmosphere is much simpler than that 

 of the sun's surroundings, because the sun's rays are sensibly 

 parallel when they reach our atmosphere. 



I have calculated the amount of polarisation due to the 

 scattering of a particle near the sun. If we consider only a 

 luminous point the plane of polarisation will be that of a plane 

 passing through the luminous point, the scattering point, and the 

 observer's eye. The polarisation will be complete, if the scattered 

 ray is at right angles to the incident ray. It is easy to see that 

 the polarisation will not be complete in any direction when the 

 luminous point is replaced by a luminous sphere, and it is also easy 

 to see that as the point is removed from the sphere the polarisa- 

 tion ought to become more and more complete. The result how- 

 ever is somewhat startling, that close to the sun the light ought 

 not to be polarised at all. According to the formulae which I 

 have obtained the polarisation increases very rapidly as the point 

 is removed from the surface of the sphere ; the per centage of light 

 polarised in a ray which makes a right angle with the line joining 

 the centre of the sphere and the scattering point, vanishing on the 

 sphere, as I have said, is 52 per cent. 3 minutes of arc away from 

 the sun ; 67 per cent. 6 minutes away ; 86 per cent, when the 

 distance is that of the solar radius, and 97 per cent, when it is a 

 diameter and a half. If we look at the corona, however, each line 

 of sight takes in a number of particles; so that even if we look 

 at the edge of the sun the line of sight pa,sses through parts of 

 the corona which are removed from the sun and which therefore 

 partially polarise the light. We cannot therefore tell what the 

 polarisation ought to be along the different lines of sight, unless 

 Ave know in what way the scattering matter is distributed along 

 the line of sight ; but we can, I believe, prove one thing, that in 

 whatever way the amount of scattering matter varies within a 



