i;6 LIGHT SCIENCE FOR LEISURE HOURS. 



light). In the eclipse of August 18, 1868 ,the shape of 

 the corona resembled what was seen during the eclipse 

 of July 29, 1878, nearly enough to correspond to the 

 difference of time. 



In fine, so far as the pictures and other evidence 

 in my possession enable me to judge, 1 all that is known 

 about that part of the meteor stream region beyond 

 the sun as seen during eclipses in July and August, 

 and therefore lying between the earth and sun in 

 January and February, corresponds precisely with what 

 we should expect if the cold spell of February is due 

 to the earth's passage through the shadows of meteor 

 streams at that time. It so happens that we have 

 no good pictures of total eclipses in October (corre- 

 sponding to the April cold spell) or in November 

 (corresponding to the cold days in May). I have very 

 little doubt that when such eclipses come to be observed, 

 similar appearances will present themselves. 



It will form an interesting subject of study during 

 future solar eclipses to determine how far the observed 

 meteoric structure of the corona corresponds with the 

 variations of the earth's mean temperature during the 

 year in other words, to determine the relative density 

 of the meteor shadows in different months and in dif- 

 ferent days. One cannot but regret the more seriously, 

 that owing to unwillingness to recognise the clearest 

 mathematical evidence as to the real nature of the corona, 



1 For illustrations of the corona, as seen on these occasions, the 

 reader is referred to my treatise on the sun. This particular point is 

 fully discussed, with abundant illustrations, in the pages of Know- 

 ledge. 



