8 4 



NATURE 



[Nov. 27, 1879 



contrived for lunar photography and to avoid the moon's 

 motion in declination, the instrument was not suitable 

 for the spectroscopic work contemplated. A reflector of 

 28 inches aperture was therefore commenced in 1866, and 

 in 1871 it was ready for use. 



On May 29, 1872, my first photograph of the spectrum 

 of a star was taken, the spectrum of Vega being photo- 

 graphed by the aid of a quartz prism. At this time I 

 did not happen to know that Dr. Huggins, who is so 

 distinguished for his thorough and accurate researches 

 on the visible portion of the spectra of the heavenly 

 bodies, had already made some attempts in this direc- 

 tion, as is shown by the following paragraph from the 

 Transactions of the Royal Society for 1864: — "On 

 the 27th of February, 1863, and on the 3rd of March of 

 the same year, when the spectrum of Sirius was caused 

 to fall upon a sensitive collodion surface, an intense 

 spectrum of the more refrangible part was obtained. 

 From want of accurate adjustment of the focus, or from 

 the motion of the star not being exactly compensated by 

 the clock movement, or from atmospheric tremors, the 

 spectrum, though tolerably defined at the edges, presented 

 no indications of lines. Our other investigations have 

 hitherto prevented us from continuing these experiments 

 farther, but we have not abandoned our intention of 

 pursuing them." 



During August, 1872, I took several photographs of the 

 spectrum of Vega, and these showed four strong lines at 

 the more refrangible end of the spectrum, the least 

 refrangible being near G. On pursuing the subject and 

 seeking to ascertain what substances gave rise to these 

 lines, it became obvious that a photographic study of this 

 part of the spectrum for the metals and non-metals was 

 necessary to interpret the results. This, of course, opened 

 out a large field for experiment, requiring many years for 

 its study, and hence, as several physicists were engaging 

 in the study of the spectra of the metals, I concluded to 

 discontinue the experiments commenced in 1870 on the 

 spectra of the metals and to confine the investigation 

 mainly to the non-metals. The initial step was, however, 

 to obtain a fine photograph of the normal solar spectrum, 

 so that the wave-lengths of the lines up to O [wave-length 

 3440] might be determined with precision. 



In the spring of 1873 I published a paper on the dif- 

 fraction spectrum of the sun, illustrated by a photograph 

 embracing the region from wave-length 4350, near G, to 

 3440, near O. and in the fall of the same year took photo- 

 graphs of the spectra of several non-metals, notably 

 nitrogen, carbon, and oxygen. The experiments were 

 interrupted, in the spring of 1874, by going to Washing- 

 ton to superintend the photographic preparations for 

 the United States observations on the transit of Venus. 



Since that time my experiments have been divided into 

 two parts, an astronomical portion occupying principally 

 the summer season, and a laboratory portion during the 

 rest of the year. The former consisted of photographs 

 and observations on the spectra of the stars, planets, and 

 sun ; the latter of photographic work on the spectra of 

 the elements, and particularly the non-metals, and has led 

 to the discovery of oxygen in the sun. 



In 1876 Dr. Huggins published a note in the Pro- 

 ceedings of the Royal Society, accompanied by a wood- 

 cut of the spectrum of Vega, with a comparison solar 

 spectrum. Seven lines were observed in the spectrum of 

 Vega. In the summer and autumn of 1876 I made 

 several photographs of the spectra of Vega, a Aquite, 

 and Venus, and sent a note concerning them to the 

 American Journal. 



Since that time Dr. Huggins has pursued the subject 

 actively in spite of the London atmosphere, and has 

 attained very fine results, which I had the pleasure of 

 seeing at his observatory last spring. These he is preparing 

 to publish shortly. In my observatory photographs have 

 been taken of the spectrum of Vega, Arcturus, Capella, 



a Aquilae, Jupiter, Mars, Venus, the moon, &c. Recently 

 the plan has been to have a comparison solar spectrum 

 on every plate, derived either from the diffused light of 

 our atmosphere or from the moon or from Jupiter. In 

 this way no difficulty in determining the wave-lengths of 

 the lines is encountered, and the changes produced by our 

 atmosphere are eliminated. The telescope and spectro- 

 scope are now in good working order, but to secure the 

 requisite degree of precision of movement it has been 

 necessary to make seven different driving-clocks before a 

 satisfactory one was attained. 



It has been remarked that on account of the faintness 

 of the light of stellar spectra, prolonged exposures of the 

 sensitive plate are required. In former times, when the 

 dry processes of photography were much less sensitive 

 than the best wet plates, the exposure was limited by the 

 length of time the plate could be left in the camera with- 

 out being stained by drying. But now, since the gelatino- 

 bromide process has been introduced, this obstacle has 

 been removed and a sensitive plate is sometimes exposed 

 two hours to the spectrum of a star and then almost an 

 hour to Jupiter for the comparison spectrum. The best, 

 and most sensitive, gelatine plates I have used are those 

 made by Wrattcn and Wainwright, of London ; Dr. 

 Huggins was good enough to call my attention to 

 them. 



It is not worth while to describe the various forms of 

 spectroscope that have been employed in the last ten 

 years ; quartz, Iceland spar, hollow prisms and flint glass 

 have been the materials, and they have been sometimes 

 direct vision and sometimes on the usual angular plan. 

 Gratings on glass and speculum metal given to me by 

 Mr. Rutherford have been tried. The length of spectro- 

 scope has been sometimes 28 feet and sometimes not 

 as many inches. 



The especial spectroscope for stellar work that is now 

 on the telescope is intended to satisfy the following con- 

 ditions : (1) to get the greatest practicable dispersion with 

 the least width of spectrum that will permit the lines to be 

 seen ; (2) to use the entire beam of light collected by the 

 2S-inch reflector or 12-inch achromatic without loss by 

 diaphragms ; (3) to permit the slit to be easily seen so 

 that the star may be adjusted on it ; (4) to avoid flexure 

 or other causes that might change the position of the 

 spectrum on the sensitive plate in pointing the telescope 

 first on one and then on another object ; (5) to admit of 

 observing the spectrum on the sensitive plate at any time 

 during an exposure without risk of shifting or disarrange- 

 ment. The dispersion is produced by two heavy flint 

 prisms which are devoid of yellow colour ; the telescopes 

 are about 6 inches in focal length, and the slit has a 

 movable plate in front of it, enabling the operator to 

 uncover either the upper or the lower portion at will. 



During the past summer this spectroscope has been 

 used with the Clark refractor of 12 inches aperture, partly 

 because the 28-inch reflector has been kept unsilvered 

 since it was used in taking photographs of the transit of 

 Mercury, on account of its employment in certain experi- 

 ments on the sun. Moreover, there is an advantage 

 possessed by the refractor for this work which does not 

 appear at first sight. Naturally one supposes that a 

 reflector which brings all the rays from the star, no 

 matter what their refrangibility, to a focus in one plane, 

 would be best, because when the slit is put in that plane 

 it is equally illuminated by rays of all refrangibilities, and 

 the spectrum will be parallel-sided in its whole length. 

 On the other hand a refractor is not achromatic, for the 

 violet end of the spectrum comes to a focus either inside 

 or outside of the plane of the rays in the middle of the 

 spectrum, and in observing the spectrum it is not parallel- 

 sided. This peculiarity was used by Mr. Rutherford to 

 enable him to correct a telescope lens for the ultra-violet 

 rays. It is easy, therefore, with a refractor, so to adjust 

 the position of the slit that you may have a spectrum 



