March 3, 1892] 



NATURE 



427 



Lapland and Greenland, and inquiry was raised as to the causes 

 of these facts. 



Mr. A. North read a paper on " The Truthful Treatment of 

 Brickwork." 



At the closing meeting of the Council, on February 14, the 

 following general officers were appointed : —Treasurer, Mr. H. 

 C. Russell, Sydney ; Secretary for Tasmania, Mr. A. Morton ; 

 for New Zealand, Prof. Packer, Prof. Thomas, and Mr. D. B. 

 Brandon ; for Victoria, Mr. A. H. S. Lucas ; for Queensland, 

 Mr. J. Shirley. 



THE DRAPER CATALOGUE OF STELLAR 

 SPECTRA. 



T^HE Observatory of Harvard College has played a promi- 



-'■ nent part in the development of astronomical photography. 

 It was here, on July 17, 1850, that Prof. Bond obtained the 

 first photographic image of a star, and from that time forward 

 much important work has been accomplished, culminating in 

 the Draper Catalogue of the photographic spectra of 10,347 

 .«tars. The progress of this latter branch of astronomical work 

 has been but slow, and it is a remarkable fact that its extra- 

 ordinary development during the last few years has followed 

 from the revival by Prof. Pickering of the method of observa- 

 tion first employed by Fraunhofer in 1824. Accounts of the 

 progress of the work have been published from time to time, and 

 have been noticed in our columns. A complete account of the 

 " Preparation and Discussion of the Draper Catalogue," which 

 has recently been issued, forms vol. xxvi., part i., of the 

 Annals of the Astronomical Observatory of Harvard College. 



The earlier attempts to photograph the s])ectra of the stars 

 were made with spectroscopes having slits, although, from the 

 time of Fraunhofer, it was recognized that a slit was not an 

 essential part of a stellar spectroscope. In 1863, Dr. Huggins 

 succeeded in photographing the .'spectrum of Sirius, but none of 

 the characteristic lines were visible. In 1872 Dr. Henry 

 Draper, to whose labours in the field of astronomical photo- 

 graphy the Draper Catalogue forms a fitting memorial, suc- 

 ceeded in obtaining a photograph showing four lines in the 

 spectrum of Vega. Dr. Huggins again took up the work, and 

 since 1879 has obtained a considerable number of photographs, 

 none of which, however, appear to show anything approaching 

 the amount of detail now obtainable. In all these attempts the 

 spectroscope was attached to the eye end of the telescope, so 

 that the image of the star was formed on the slit, a cylindrical 

 lens being interposed in order to give width to the spectrum. 



In the method which has been so pre-eminently successful, 

 the slit and collimator, which form an essential part of an 

 ordinary spectroscope, are dispensed with, the rays from a 

 star already possessing the necessary parallelism and its image 

 being almost a perfect slit without length. It is only necessary, 

 therefore, to fix a prism in front of the objective of a telescope, 

 and to introduce some means of widening the spectrum, to 

 obtain a complete stellar spectroscope. For eye observations 

 the necessary width is obtained by the use of a cylindrical lens 

 in conjjnction with the eye piece of the telescope. For photo- 

 graphic work, the prisms are so arranged that the spectrum lies 

 along a meridian, and it is then only necessary to allow the 

 driving clock to be slightly in error to obtain a widened 

 spectrum. The clock error must of course vary according to the 

 magnitude and declination of the star. 



The great advantage of the " slitless spectroscope " depends 

 upon the fact that every scrap of light passing through the object- 

 glass is utilized ; with the ordinary .'spectroscope it will seldom 

 happen that all the light passes through the slit, and it is further 

 reduced by absorption in the lenses and prisms of the spectro- 

 scope. F'urther, on account of the large focal length of the 

 telescopes employed, a high dispersion is obtained even with a 

 prism of small angle ; and a large number of spectra can be 

 photographed at a single exposure. Prof. Pickering has photo- 

 graphed the spectra of as many as 260 stars on the same plate, 

 and the labour involved in the construction of the Draper 

 Catalogue has thus been enormously reduced. Indeed, the 

 whole of the 10,347 spectra were photographed on 585 plates. 

 The improvement in photographic processes has undoubtedly 

 done much to facilitate the work, but it is lamentable that the 

 "wholesale" method was not applied twenty years ago, for 

 even with the less perfect processes then in vogue, our knowledge 

 would have been much advanced. 



An important feature of Prof. Pickering's work is the method 

 of enlargement of the negatives, which renders the fainter 

 lines clearly visible. "The negative is covered by a diaphragm, 

 having a slit in it which is made to coincide with the spectrum. 

 An image is then formed by an enlarging lens in the usual way. 

 A cylindrical lens is next interposed near the enlarging lens, 

 with its axis perpendicular to the lines in the spectrum. The 

 width of the latter may thus be increased indefinitely without 

 changing the length. In the case of faint stars very narrow 

 spectra only can be obtained. Their energy is so feeble that 

 they are capable of decomposing the silver particles only if 

 allowed to fall upon them for a long time. In the enlargement 

 the energy of the sun is substituted for that of the star, and thus 

 an indefinite number of silver particles may be decomposed." 

 (Introduction, p. xix.) The original negative may, perhaps, 

 be compared to a "relay " in electrical apparatus. 



The preparation of the Draper Catalogue involved five 

 different steps, which are thus stated on p. 74 : — 



I. Measurement of the spectra on each plate, including the 

 determination of their positions, intensities and the classes to 

 which they belong. 



II. Identification of each spectrum with that of a star in the 

 Durchmusterung or other catalogue. 



III. Reduction of the measures of brightness to the scale of 

 the Harvard Photometry. 



IV. Catalogues of plates. 



V. Preparation of the final catalogue, bringing forward the 

 places of all the stars to 1900, including various methods of 

 checking and correcting the results. 



That a catalogue of spectra may be of service to astronomers, 

 a sound system of classification is essential, and this, as far as 

 possible, should have some reference to chemical or physical 

 constitution. The notable classifications which were suggested 

 by eye observations were those of Secchi, Vogel, and Lockyer, 

 but it is not surprising to find that the greater detail shown on 

 the photographic plates requires modifications of these in order 

 that all the spectra may be included. A detailed but some- 

 what arbitrary classification has been adopted by Prof. Pickering, 

 the chief merit of which is that it readily lends itself to transla- 

 tion into other systems. Varieties of Secchi's first type are 

 indicated by the letters A, B, C, D, those of the second type by 

 the letters E to L, of the third type by M, and of the fourth 

 type by N ; bright line stars are referred to as O, planetary, 

 nebulas as P, and other spectra as Q. Of the varieties of the 

 first type, A includes all the stars with spectra similar to Sirius, 

 and B those with spectra of the Rigel type, in which, in addition 

 to lines of hydrogen, there is a small number of strong lines of 

 which the origins are at present unknown. 



Results of special interest, such as the discovery of bright 

 lines in the spectra of variable stars of long period, have already 

 been referred to in Nature, and we shall now confine ourselves 

 to the more general results. As some of the most interesting 

 spectra belong to stars of small magnitude, it is necessary to be 

 very guarded in making generalizations. Still, the fact that 

 Prof. Pickering's researches have extended in some cases to 

 stars of the ninth and tenth magnitude perhaps justifies the 

 assumption that all types of spectra are now included. We 

 cannot do better than let Prof. Pickering speak for himself. 



"The general conclusion derived from the study of these 

 spectra, is the marked similarity in constitution of the different 

 stars. A large part of them — those of the first type — have a 

 spectrum which at first sight seems to be continuous, except that 

 it is traversed by broad dark bands due to hydrogen. Closer 

 inspection shows that the K line is also present as a fine dark 

 line. If the dispersion is large and the definition good, many 

 more dark lines are visible, as stated above. These lines may 

 be divided into two classes — first, those which predominate in 

 many stars in the Milky Way, especially in the constellation of 

 Orion ; and, second, those present in the solar spectrum. 

 Nearly all the brighter stars may be arranged in a series, begin- 

 ning with those in Orion, in which the auxiliary lines are nearly 

 as intense as those due to hydrogen. Other stars may be found, 

 in which these lines successively become fainter and fainter, 

 until they have nearly disappeared. The more marked solar 

 lines then appear, become stronger and stronger, and the 

 hydrogen lines fainter, until they gradually merge into a 

 spectrum identical with that of the sun. At least, several 

 hundred lines appear to be identical, and no differences can be 

 detected. Continuing the sequence, the spectra pass gradually 

 into those of the third type. Certain bands become more 



NO. II 66, VOL. 45] 



