88 Prof. Liveing, On the probable presence in the Sun 



a distance from the sun's surface, while much more volatile 

 metals, such as sodium, are not regularly present at a quarter 

 of that height. 



We have been expecting from Mr Baly for a long time past 

 a list of wave-lengths of the gases neon, krypton and xenon 

 determined with great exactness by the use of a grating. The 

 announcement that these measurements were in progress was 

 made more than two years ago, and when the list appears we 

 shall be in a better position to judge of the nearness of the 

 coincidences. Meantime I have remeasured some of the rays of 

 the most volatile gases, still, however, with only a prismatic 

 spectroscope, so as to get, approximately, the fifth figure of the 

 wave-length, and the result has been in many cases to bring my 

 figures into closer agreement with those of Humphreys. The 

 exactness of Humphreys' wave-lengths is quite as important 

 a factor in the settlement of the question as that of mine. He 

 obtained six photographs on celluloid films with a large concave 

 grating of 30 feet focal length, used without any slit. The first 

 had an exposure of 2 seconds immediately after second contact. 

 Two seconds sufficed for a change of film, and then the second 

 film was exposed for 5 seconds, and the third for 15 seconds, the 

 fourth had a lengthened exposure of 2 minutes during the middle 

 of totality, the fifth 15 seconds, and the sixth 8 seconds, ending 

 3 seconds before third contact. The dispersion of the spectrum 

 was so great that one millimeter on the film corresponded to 

 3'66 Angstrom units, and the range of the spectrum depicted 

 extended from X3118 to A, 5204 The latter limit was probably 

 determined by the want of sensitiveness of the films to lower 

 green rays, for it would have been of great interest if the range 

 had included the most characteristic rays of coronium, krypton, 

 and neon. At the upper end the Fraiinhofer line S is just ex- 

 cluded. This too is to be regretted, as S is a triple line of iron 

 very unlikely to be reproduced by any mixture of elements 

 excluding iron. The dispersion is ample for very exact deter- 

 minations of wave-length if there were no difficulties in the way 

 of exact measurement. Humphreys has met these difficulties as 

 best he could, and I do not propose to discuss them now, but 

 I may perhaps explain one or two of them. The station where the 

 photographs were taken was near the northern limit of totality, 

 so that the duration of totality was less than half what it was 

 in the centre of the shadow. From these durations of totality 

 I have calculated that the points of contact of the limbs of sun and 

 moon were each nearly 63° from the diameter of the sun parallel 

 to the line of the moon's apparent motion ; and the instrument 

 was so arranged that the ruling of the grating was parallel to 

 that line, and in consequence the dispersion at right angles to that 



