xxv.] A PREDICTION. 359 



The test was first applied during the eclipse which occurred 

 in 1882 in Egypt, and afterwards, in 1886, in the island of 

 Grenada in the West Indies. Nor was this all ; the lines to be 

 first seen as short and brilliant, and the lines to be seen later 

 as long and dim, were predicted a year before. 



In May, 1881, in an address delivered to the Astronomical 

 Society I pointed out the importance of observing the eclipse 

 of 1882. The following part of the address is all I need 

 give here : 



" There is an eclipse of the sun next year, lasting only, I am 

 sorry to say, a minute and a very few seconds ; but there is to be 

 another the year after, lasting nearly six minutes, but it happens to 

 be in a part of the world where it is always afternoon. In the 

 observations of the future we must pay attention to these lines 

 which have been picked out by Nature herself in the spots and 

 prominences. If I observed either of these eclipses, I should be 

 content to fix my instrument on these three iron lines between 

 4,900 and 5,000 ten millionths of a millimetre, because, of these 

 three lines which are in the Fraunhofer spectrum, two have always 

 been seen in spots without the third, and the third has always been 

 seen in the prominences without the other two. If, then, the 

 spectrum of the flames represents the lowest part of the atmosphere, 

 and the spectrum of the spots represents the atmosphere above the 

 flames and below the corona, then we ought to see these lines 

 different in the corona, and in the corona we ought to see the lines 

 which are dropped in these two regions. Of the twelve lines 

 between 4,900 and 4,957, only one is picked out by Thalen for 

 intensification, and that particular line is the line seen alone in the 

 region of the prominences. There are eleven lines which are 

 absolutely untouched by Thalen, showing that absorption must be 

 proceeding somewhere, and it is most interesting to determine 

 where it is going on. 



"In the Indian eclipse, in 1871, I saw lines reversed before 

 totality. I saw, as it were, hundreds of lines ; but if I had con- 

 fined my attention to these three lines I should have got a better 

 idea of what the magnificent flashing out of those lines meant. It 

 has been called the reversing layer ; but I do not now believe it is 



