(A, 4861) to K (A. 3934), and for the quartz 3 inches from A 4000 to A 3300. In order to obtain a greater length of spectrum in focus, the plate holders were made, at Mr. Davidson's suggestion, to carry two plates 3| inches wide, inclined at a suitable angle, instead of one 6J inches wide.
Both spectroscopes were mounted horizontally, and supplied with light by a heliostat furnished with a 12-inch flat mirror. The. heliostat was adjiisted in the usual manner by means of the attached theo- dolite.
Programme of Exposures. The two spectroscopes were adjusted to be as nearly as possible on the Sun's limb simultaneously, and the programme of exposures was the same for both. The cameras were provided with rack movements so that a number of exposures could be made on the same plate. The programme arranged to take
(i.) A number of short exposures from 10 s before to 6 s after the beginning of totality for the " flash" spectrum.
The plate carriers were then changed, and
(ii.) An exposure for the spectrum of the corona made from 15 m or 20 s after the beginning of totality, and continued for 5i- minutes the exposure ending at 5 m 50 s from the beginning of totality.
The plate carriers were again changed, and
(iii.) From 6 ra 10 H from the beginning or 10 s from the end of totality, exposures were to be made for the " flash " at third contact.
Between (i) and (ii) the image of the Sun was to be brought by the slow motion of the heliostat from being nearly tangential to the slit to a position in which the horizontal diameter was divided by the slit in a ratio of 1 : 3. Again, between (ii) and (iii) it was to be brought nearly tangential.
A hitch occurred between (ii) and (iii), and the "flash" at second contact was not obtained.
The plates used for the " flash " at the beginning of totality were Edwards' Isochromatic and Ilford Empress for the flint spectroscope, and Ilford Empress for the quartz.
For the corona, " Snap-shot," Isochromatic, and Kocket plates were used for the flint spectroscope, and Imperial Special Rapid for the quartz.
The Corona Spectrum. The continuous spectrum of the corona is shown from D 3 (A 5876) to A 3350. The green line at 5304 is strongly shown. The lines at 4568, 4233 and 3987 are shown. No new lines are shown. In the ultra-violet the lines at 3801, 3456 and 3381 are strongly shown, and possibly some fainter lines. The wave-lengths have not yet been determined accurately. No absorption lines are shown.
The Chromospheric Spectrum. This is obtained on the different plates extending from D 3 (A 5876) to a little beyond 3300. The photographs are not on the whole so good as those obtained at Ovar, but will serve
