of the newly discovered gases of the Earth's Atmosphere. 



89 



line, and the definition affected as little as possible by the moon's 

 motion relative to the sun. During the exposures the tangents 

 at the centres of the small arcs photographed would at first be 

 inclined at about 27° to the direction of the moon's path and 

 would gradually swing round until at mid totality they would 

 be parallel to it, and then on until they were again inclined 

 at about 27° to that direction, but on the opposite side. This 

 motion tends to blur the definition of the arcs, but the effect 

 would be very small on the 1st, 2nd, and 6th films, and not 

 great on the 3rd and oth, but it would be very great on the 

 4th film for rays emitted only from parts of the chromosphere 

 at much less distance than seven seconds of arc from the sun's 

 limb. Rays emitted from the chromosphere at heights exceeding 

 7 - 3" would give arcs crossing the direction of dispersion at right 

 angles, and would have well defined edges on the inside, un- 

 affected by the moon's motion. There are however very few of 

 such rays besides H and K and the strong hydrogen and helium 

 rays. 



The wave-lengths had to be determined by measuring the 

 distances between the arcs on the films in the direction of dis- 

 persion and interpolating between such of the arcs as could be 

 identified as due to rays of which the wave-lengths were well 

 known. Rays answering to this test would be those above 

 mentioned, namely, H and K and the strong rays of hydrogen 

 and helium, and one or two very characteristic groups such 

 as the b group of magnesium and, perhaps, the iron triplet 

 4046 — 4071. Indeed on looking through the list I do not see 

 any ray other than these which I could identify with certainty 

 without assuming it to be a reversal of a Fralinhofer line. 

 Humphreys does not say which rays he used as fiducial rays 

 between which he could interpolate. Those which I have just 

 named seem to me too few, and too unequally distributed 

 through the spectrum, to give very trustworthy wave-lengths 

 in all parts of the spectrum. No scale of reference, made with 

 artificial light and the same grating used without a slit, could 

 well be made, or had in fact been attempted. 



In his list Humphreys gives, against each chromospheric ray, 

 the Fraiinhofer line in Rowland's list which has the wave-length 

 nearest to that found for the chromospheric ray. As there are 

 13,595 lines in Rowland's list in the range of the spectrum 

 embraced by Humphreys' 339 lines it is not surprising that 

 a Fraiinhofer line is to be found for every one of Humphreys' 

 lines, so close that the difference of wave-length is in most cases 

 less than a tenth of a unit and rarely exceeds two-tenths In 

 discussing his results he evidently assumes the rays to be identical, 

 and ascribes them to the metal (if any) indicated by Rowland. 



