and near the times of beginning and ending of totality, in accordance with the method first used successfully by Mr. SHACKLETON at the eclipse of 1896. The spectrum of the reversing layer, i. e. the "flash spectrum flash spectrum" near the beginning and end, was especially desired. The optical parts of the instrument consisted of two objective-prisms of moderately dense flint glass, refracting angle 60°, refracting edges 2% inches long, with faces 334 and 3% inches long, respectively, placed immediately in front of a triple lens, aperture 2% inches, focal length 60 inches, corrected for Hy central. The plate-holder, holding seven Seed No. 27 plates, each 134 X 10 inches, was carried in a long slide. Brass racks fastened to the back of the plate-holder, and pinions supported by the slide and working in the racks, were the simple means of giving motion to the plate-holder. As soon as an exposure had been made on one narrow plate, one rotation of the pinions (by means of a small crank) brought the next narrow plate into position. The exposing shutter was a simple flap of zinc directly in front of the sensitive plate. The mounting of the instrument was of sugar-pine wood, and it was supported upon timbers set in stone and cement. The spectrograph received its light from the 15-inch plane-mirror of a cœlostat kindly loaned to the expedition by the Yerkes Observatory. The various parts of the instrument were adjusted to each other, and the instrument as a whole was adjusted to bring the Hy region approximately to the center of the plate and the length of spectrum parallel to the edge of the plate several days before the eclipse. It was brought into final position a few minutes before totality. In accordance with Mr. CAMPBELL'S instructions, it had been intended to make the first exposure several seconds before totality, to record the Fraunhofer spectrum, and to begin the second exposure at three seconds before totality, to record the flash spectrum. The coming of totality 17 seconds before it had been expected interfered with carrying out this programme, but fortunately Dr. HARTMANN, who was watching the progress of the eclipse with a small hand spectroscope, recognized its earlier arrival, and made his first exposure three seconds before totality, with estimated duration 08.4. Development of the plate showed that the photosphere was still visible on one section of the Sun's limb, but that the reversing layer was isolated from the photosphere on an adjoining section. The result is a magnificent photograph, showing the continuous and dark-line spectrum for one section and the bright-line spectrum (many hundreds of bright crescents) for the other section. The plate is in splendid focus from about A 4800 to A 3700. The Hẞ bright line near one end of the plate is out of focus, unavoidably, just enough to make it clearly doubled. The seven exposures made by Dr. HARTMANN were: Plate No. 2 shows the high-level bright lines. Plate No. 5 is overexposed for the continuous spectrum, but shows the flash spectrum well on one side of the Fraunhofer spectrum. A study of this very successful series of photographs, involving much labor, should supply extensive and accurate information as to the structure of the reversing layer. OBJECTIVE-PRISM SPECTROGRAPH WITH MOVING PLATE. In charge of Professor ARRHENIUS. Photographs of the reversing-layer spectrum, taken in the usual manner with objective-prism instruments, such as that described in the preceding paragraphs, are integrated effects. Changes taking place during the exposure are not differentiated, and changes taking place between exposures are entirely lost. A continuous record of the changing spectrum is a great desideratum. A simple addition to the objective-prism spectrograph enables this to be obtained for a short length of solar limb. The usual reversing-layer spectrum consists of a series of crescents, each crescent an image of the uneclipsed portion of the Sun. A slit running centrálly through the spectrum, placed all but in contact with the photographic plate, permits a short central section of each crescent to fall upon the plate. If the plate is given a slow continuous motion by suitable mechanical means, a fresh part of the plate will be brought under the slit, and the changing spectrum will be recorded continuously. This method was devised and used by Mr. CAMPBELL in India in 1898 and in Georgia in 1900. At the latter eclipse an exposure beginning about 10 seconds before the end of totality, and continuing until 12 seconds after totality was over, recorded more than nine hundred bright lines and their succeeding dark lines. The instrument used by Professor ARRHENIUS was equipped in the manner just described. The two objective prisms were of moderately dense flint glass, Jena No. 0.102, refracting angle 63° 27′.5, length of refracting edge 2 inches, lengths of faces 3.47 and 3.95 inches, respectively. The triple lens had aperture 2% inches and focal length 60 inches. The slit in front of the plate was 0.05 inch wide and 91⁄2 inches long. The plateholder was moved by a hydraulic piston actuated by a weight, modeled after the piston of the Potsdam spectroheliograph, The controlling valve was regulated to a speed of about 0.06 inch per second. It had been intended to secure a moving photograph extending from 12 seconds before to 12 seconds after the instant of totality, and a similar photograph from 12 seconds before to 12 seconds after the end of totality. The first of these was prevented by the arrival of totality 17 seconds earlier than was expected, but the second exposure, extending from 3m 338 to 3m 57s after totality began, was entirely successful. The changing spectrum is shown admirably through all the phases from high-level strong bright lines, through the bright-line stage of the reversing layer, and into the ordinary dark-line stage. There are six or eight hundred lines recorded in the region à 3800 - 5200, with Hy central on the plate. During totality the slit in front of the plate was rotated to one side of the field of view, and the full prismatic image fell on the plate, from Im to 3m after the beginning of totality. The coronal rings at λ 4231 and λ 3987 are strongly recorded. For this exposure the plate was fixed in position. This spectrograph also received its light from the Yerkes Observatory colostat mirror. ULTRA-VIOLET SPECTROGRAPH, WITH MOVING PLATE. In charge of Mr. CAMPBELL, assisted by Dr. KOBB. In order to extend further into the violet the study of the reversing-layer spectrum with a continuously moving plate, a spectrograph was constructed whose optical parts should be efficient for ultra-violet light, as follows: Two objective prisms of Jena ultra-violet glass No. 3199, refracting angles 60°, refracting edges 60mm, and lengths of faces 100mm and 110mm, respectively. A special lens of ultra-violet glasses, aperture 66mm and focal length 2000mm. A coelostat mirror of Schroeder's metal No. 1, diameter 110mm and thickness 15mm, attached to the lower end of the Lick Observatory colostat's polar axis. The foregoing parts were ordered from CARL ZEISS in October, 1904, delivery promised in February, 1905. As they did not reach Mt. Hamilton until the first week in June, after the necessarily unfinished wooden mounting was packed for shipment, the instrument was completed in Spain with time that could ill be spared for the purpose. The plate-holder was moved by a weighted piston, as in the case of the preceding spectrograph operated by Professor ARRHENIUS. This ultra-violet spectrograph is shown in position at the observing station in the accompanying illustration. The Lick Observatory cœlostat, seen in the same photograph, carries a 12-inch flat mirror by PETITDIDIER, as well as the Schroeder'smetal mirror. The exposures planned for this spectrograph were identical with those for Professor ARRHENIUS's instrument. Those actually made, on account of the earlier arrival of totality, were identical with his for the exposure with fixed plate, from Im to 3m, and for the exposure with moving plate, 3m 338 to 3m 57s. The thin clouds interfered with the passing of ultraviolet light; and while the record extends to perhaps à 3200, the spectrum is weak in intensity, only the stronger ultra-violet lines being shown. The dispersion of the ultra-violet glass is very low; the linear dispersion of A 3200 and λ 5200 is but 150mm for the two 60° prisms and focal length of camera 2000mm The exposure with fixed plate recorded two strong coronal |