ON THE SPECTRA OF STARS OF SECCHI'S FOURTH

TYPE. I.

By GEORGE E. HALE and FERDINAND ELLERMAN.

THE rapid progress of stellar spectroscopy during recent years, which has been due almost entirely to the development of photographic methods, has naturally followed three principal lines:

1. The determination of the general characteristics of stellar spectra, permitting a classification of stars on the basis of spectral types.

2. The measurement of the wave-lengths of dark and bright lines, for the purpose of identifying the substances present in stellar atmospheres.

3. The measurement of the displacement of stellar lines with reference to the lines of an artificial comparison spectrum, giving a means of determining the velocity of stars in the line of sight.

The measurement of the radial motion of stars, first attempted by Huggins with the inadequate instruments of a third of a century ago, has already reached an advanced stage of development. This is largely due to the work of Vogel, who first employed photographic methods in this field. The photographs taken at Potsdam for the purpose of Vogel's investigations have served, in the hands of Scheiner, for the accurate determination of the wave-lengths of numerous stellar lines. With the exception of this research, however, but little systematic work has been done in this direction. An extensive field of investigation, hitherto almost unexplored, here lies open to students of astrophysics. Fortunately many of the brighter stars are within the reach of moderate apertures, and may be investigated with but small

I Professor Keeler is engaged in an important investigation of the spectra of stars of Secchi's third type.

additions to the instrumental equipment of the average observatory.

The study of the general characteristics of stellar spectra has been greatly facilitated by the use of the objective prism, which has been applied with such striking success by Pickering at the Cambridge and Arequipa stations of the Harvard College Observatory. The recent work of Vogel and Wilsing with the small spectrograph of the Potsdam Observatory, and the photographs of stellar spectra made in England and at the Cape by Huggins and Lockyer and McClean afford additional material for an extensive study of stellar development.

An examination of these results will show, however, that our knowledge of the spectra of stars of Secchi's fourth type (Vogel's IIIb) has advanced but little since the publication in 1884 of Dunér's memoir "Sur les Étoiles à Spectres de la Troisème Classe." Two causes sufficiently explain this fact. In the first place none of the stars of the fourth type are brighter than the 5.5 magnitude; and further, the spectra are so faint in the more refrangible region that but little is recorded on ordinary plates used with an objective prism. Thus the great store of negatives belonging to the Harvard College Observatory, so rich in other respects, contains few data available for the study of the spectra of these stars.

In planning the work of the Yerkes Observatory it was felt. that the 40-inch refractor should preferably be employed in fields of investigation where its great light-gathering power would be likely to prove of most service. The objective of this instrument, which is corrected for visual observations, is especially adapted for the examination of the less refrangible regions of stellar spectra. The present investigation, which relates particularly to the visible spectra of faint red stars, was accordingly undertaken in November 1897.

RESULTS OBTAINED BY PREVIOUS OBSERVERS.

In his early survey of stellar spectra Secchi divided the stars into three classes, of which the third was given up to red stars.

Most of the stars of this character examined were naturally of the true third type; but one, Lalande 12561, is of the fourth, though it was classed by Secchi in the first part of his memoir, Sugli spettri prismatici delle stelle fisse (1867) with a Herculis, in the following words (Catalogo, p. 14.)

In conclusion, this is of the type of a Herculis, but with the dark zones lacking, while its own zones are so broad that some of them embrace two of those of a Herculis.

In the second part of the memoir it appears that the distinctive characteristics of fourth type spectra were recognized in the course of a survey of some twenty red stars from Schjellerup's catalogue. In describing the spectrum of 152 Schjellerup as characteristic of the class Secchi remarks (p. 9):

This type is composed of but three principal zones; a bright one in the green, a fainter one in the blue, and a pretty bright one in the red. This latter zone is frequently subdivided into other lesser zones.

This type differs essentially from the third, not only by the division of the zones, which have twice the breadth, but also because the zones have the greater luminous intensity on the opposite side; i. e., in the fourth type the light increases in intensity from the red toward the violet, while in the third type the reverse is true. Thus if the third type were represented by a system of columns, the fourth type would be represented by cavities, supposing the illuminating light to be directed from the same side.

These stars also contain bright lines like those of the metals, and it is singular that these occur at the brightest extremity of the colored zones.

Few objects of this class were known to Secchi, but many were discovered in the subsequent observations of Vogel, D'Arrest, Pickering, and Dunér. The memoir published by Dunér in 1884, "Sur les Étoiles à Spectres de la Troisème Classe,"together with Vogel's observations made with the 27-inch refractor of the Vienna Observatory,2 afford the best data for a study of the spectra. A complete list of all fourth type stars hitherto discovered has recently been compiled by Espin, who has himself made many additions to the number.

'Svenska Vetenskaps-Akademiens Handlingar, Vol. XXI, No. 2.

Publicationen der Astrophysikalisches Observatorium zu Potsdam, Vol. IV, Part 1. 3 M. N., 58, 443, June 1898.

Dunér's important observations, which will frequently be referred to in the present series of papers, were made with several direct vision spectroscopes of different dispersive powers attached to the 10-inch refractor of the Lund Observatory. In spite of the small aperture of his telescope, the low dispersion which was necessarily employed, and the serious difficulties that are almost invariably encountered in visual observations of faint spectra, Dunér's results are of the highest value, and have been confirmed in almost every particular by our photographs. For purposes of comparison Dunér's drawings of fourth type spectra are reproduced from his memoir (Plate II).

While it is undoubtedly true, as Dunér has pointed out in his memoir, that many of Secchi's results are unreliable, it should nevertheless be remembered that the earlier observer was a pioneer in this field, and that his instruments were hardly adequate for such difficult observations. He deserves credit for having discovered this type of spectrum, and for giving evidence of the presence of carbon in the atmospheres of fourth type stars. His most conclusive measures, as quoted by Dunér (p. 122) are as follows:

As Dunér remarks, Secchi's many inaccuracies cast doubt upon these as well as other results, but they are interesting as the earliest recorded evidence of the carbon origin of the bands.

In his observations at Bothkamp and in his later work with the 27-inch Vienna refractor Vogel measured the spectra of stars Nos. 51, 78, 152 and 273 in Schjellerup's catalogue, and that of DM. 34° 4500.1 His results are given in the following table, which is taken from Vol. IV of the Potsdam Publications:

I

VOGEL'S drawings of the spectra of 152 Schjellerup and DM. 34° 4500 are reproduced in Plate II from the memoir cited above.