The spectroscopic observations hereafter discussed were made with instruments belonging to the Royal Society, and in accordance with certain suggestions which a Committee were good enough to make in connexion with my letter to Sir Edward Sabine, President, dated 13th February, 1866. In view of my residence at a considerable height, and the exceedingly clear atmosphere prevailing at some periods of the year, it was suggested that the locality was peculiarly favourable for comparing the solar spectrum when the sun was high with the corresponding spectrum at sunset; any differences between these aspects which might appear were to be noted on Kirchhoff’s well-known maps. Accordingly I set to work with the spectroscope first supplied to me (hereafter distinguished by the prefix
old
), and during the autumns of 1868 and 1869 I mapped the differences in question from the extreme red to D: these results appeared in the ‘Proceedings of the Royal Society,’ No. 123, 1870, the Map being marked vol. xix. pl. 1; it is unnecessary, therefore, to dwell on this portion of my labours, excepting to add that the definitions and general procedure there adopted have been retained in the remarks which follow. 2. The observations hereafter noticed were always taken in the
autumn
, when, the rainy season having passed away, the atmosphere on these mountains is exceedingly clear, so that the sun, the object of inquiry, is bright even to his setting, and a spectrum may therefore be then obtained through a long stretch of terrestrial atmosphere corresponding to the height of the station of observation; on the other hand, with the sun about the meridian, the height of station places the observer
above
a relative amount of atmosphere, so that the spectrum obtainable at this time and about sunset are highly eligible for the comparison in view. Accordingly the two spectra are given in the accompanying map (Plate 25); and for easy comparison they are placed in juxtaposition. By “sun high” is to be understood any position for the sun within a couple of hours of the meridian; by “sun low” that the sun was within 3 or 4 diameters of his
setting
and yet quite bright. Indeed it is only when
very
near sunset that the marked alterations in the lines appear; so that the spectrum required is not only
rarely
obtainable, but it hardly lasts beyond 10 minutes of an evening. In this short period (when, moreover, the observer is fatigued with previous watching) changes from the
sun-high
spectrum must first be detected; then their position must be identified, and, failing this, found by measurement; next, the appearance should be drawn, and finally the drawing should be compared with the original: under these conditions a week may be easily absorbed by a single group. It is also to be borne in mind that no human eye will endure, without at least temporary injury, protracted watching of the bright solar spectrum for more than four or five weeks at a time; indeed, though I habitually used both eyes as a relief to one another, they both invariably suffered, and continued to do so for several weeks after every autumn. The following facts may be here mentioned:—
V. Account of spectroscopic observation of the Eclipse of the Sun, August 18, 1868, in a letter addressed to the President of the Royal Society. By Captain C. T. Haig, R. E. Communicated by the President