When work commenced on the Bonn Atlas for Objective Prism Spectra some years ago it was our intention to supply a tool for astronomers working with such widely different dispersions and resolutions as are used in the field of objective prism spectroscopy. This was to be accomplished through an atlas which contained and compared spectra of considerable difference in dispersion.

The Bonn Schmidt telescope which is used to obtain the observational material is equipped with three prisms giving linear reciprocal dispersions of 240, 645 and 1280 Å mm−1 at Hγ, thus covering a good part of the range generally used.

The first part of the atlas, containing spectra of the largest dispersion only, was published as soon as it was finished, while work on the second part with the two lower dispersions was still in progress. Thus, one of the main intended features of the complete publication, a comparison of criteria displayed at different dispersions, with a special interest in the appearance and disappearance of certain criteria as one goes from higher to lower dispersions, was not yet possible.

Now, work on the second part of the atlas has progressed far enough for sample pages of the lower dispersion plates to be distributed for inspection and discussion. Plate I gives examples of spectra that will be used on the S-plates (sequences of different spectral types for a given luminosity class) showing both spectra of 645 Å mm−1 (right hand side) and 1280 Å mm−1 (left hand side).

While the first part of the atlas tried to point out as much detail in the spectra as possible to make the atlas applicable for work with even higher dispersions, the second part concentrates only on those features which are important in the mediumto low-resolution range.

Originally, it was intended to use greater enlargements on the L-plates (sequences of different luminosity classes for a given spectral type). Yet, general agreement upon inspection of the sample plates, which are not shown here, was that the smaller enlargements are better and thus they will be used in the final version of the atlas.

A detailed explanation of the different spectral features useful in classification from medium- to low-dispersion plates will be possible only after completion of all observational work. So far it is indicated, as was to be expected from the work of other authors, that some of the most useful luminosity criteria overlap seriously with population criteria, e.g. the CN bands.

The discussion at the symposium revealed an interest in the sampling of peculiar spectra. Following this suggestion work has began on preparing for a third part of the atlas which is planned to contain spectra of about 70 peculiar stars taken with all three above-mentioned dispersions.

The present state of a spectral survey of the southern Milky Way for stars earlier than A5 is described.

As a result of the spectral survey at the Stockholm Observatory finding lists are prepared for early and late type stars in the Southern Milky Way. In order to make the lists more useful we present the principles of the stellar classification.

The spectral classification experiments with the 70-cm meniscus telescope and 8° objective prism of Abastumani Observatory (dispersion 166 Å mm-1) have shown that if we apply the Kodak IIa-O plates at the 0.4 mm widening of spectra and the 20 minutes exposure time, we can successfully classify with relatively high accuracy in two dimensional MK system the stars down to the 11th magnitude, with simultaneous discrimination of peculiars.

We have undertaken such determinations first of all in Kapteyn Selected Areas aiming to fill the gaps in existing data for these standard areas.

Finally we could mention the high quality of the meniscus telescope optics together with good seeing in Abastumani as favourable factors making the Abastumani spectra of 10–11th magn. stars comparable with the slit spectra of the same dispersion.

The Abastumani spectra allow to apply the Morgan and Kitt-Peak criteria of MK classification only with a slight modification.

We have two-dimensionally classified in MK system about ten thousand stars down to the 11th magnitude in KA Nos. 2–43 (4.°5 × 4°5 each). Among them 137 peculiar stars have been revealed. They are mainly Ap, Am and BaII (7 stars) and the stars with composed spectra (5). 97 stars have been first identified as peculiars.

Surveys of M and C stars down to infrared magnitudes 12.m0 in two galactic regions in the direction Cygnus and anticentre are described. Infrared magnitudes are derived by a new method of calibration based on the use of a combination of crossed calcite plates and Polaroid HN32. Spectral classification is based on spectrophotometric criteria which measure the relative strengths of the molecular bands visible in the infrared in the low dispersion objective prism spectra. Infrared absolute magnitudes and interstellar extinction in the infrared are obtained from the present material and published data, and the space densities of M stars are determined.

Full accounts have been published in Publ. Edinburgh Roy. Obs.7, 1 (1970) and 7, 73 (1971).

Well-widened objective prism spectra of carbon stars south of −25° declination have been examined to determine their suitability for classification. These spectra, taken by Henize in the early 1950's from South Africa as part of the Michigan-Mt. Wilson Hα survey of the southern sky, have a reciprocal dispersion at Na I 5890 Å of 300 Å mm−1 and are in good focus in the range between 5100 and 6600 Å.

Compared with spectra of northern carbon stars taken by the writer at the Lindheimer Astronomical Research Center with the one meter reflecting telescope and Cassegrain spectrograph at a grating reciprocal dispersion of 263 Å mm-1, and compared with spectra taken by Sanford of carbon stars overlapping the Henize sample, the objective prism spectra have been found of sufficient quality to support their classification into five groups following the scheme as proposed by Keenan and Morgan in 1941 and as refined by Gordon in 1967.

The first group exhibit weak C2 and CN features and Hα in absorption; the second, features attributable to a low C12/C13 ratio; the third, strong C2 and CN features and only moderately strong Na i in absorption; the fourth, very strong Na i in absorption; and the fifth, Hα in emission. Examples are given of each group, stars not fitting well into any of these groups are discussed, and the significance of each group is mentioned in light of work by Bouigue, Gordon, Peery, and Richer.

Methodological problems and hardware/software requirements of efficient automatic classification of objective prism stellar spectra are described on the basis of experience from ‘microphotometer-computer’ classification. Significant advances in hardware (two-dimensional scanners, on-line computers) as well as in software (image processing techniques) have now brought high-speed, exhaustive, automatic classification within reach. Some astronomical implications of automatic spectral surveys are discussed.

Three-dimensional classification of late-type stars is obtained through computer-processed digitized microphotometer data. Plate transmissions are converted to relative intensity through spot-sensitometer calibration data. A pseudo-continuum consisting of several straight-line segments is formed by connecting high points in the spectrum. Absorption-line strengths and line ratios are then measured. The temperature-and-luminosity-sensitive ratios are relatively insensitive to seeing effects. For plates with a dispersion of 108 Å mm-1, preliminary results indicate an accuracy in derived absolute magnitude comparable to, and possibly better than, that of MK classification.

The CN anomaly serves as the abundance index. The CN index was defined in a manner to give the largest figure of merit (the ratio of total range of the index to the mean error). The adopted method is the same as that of Yoss and Lutz (1971) and has a figure of merit over twice that of the index similar to that of Griffin and Redman (1960).

For field stars, plate X and Y positions are converted to equatorial coordinates and printed out in order of increasing right ascension, making quick and easy identification of the Henry Draper numbers.

A technique for narrow-band photometry in the 3600–8400 Å spectral region from objective prism spectra is briefly described and applied to the Hyades.

Different methods, used for measuring the strength of hydrogen lines in objective-prism spectra, are discussed and the parameters are compared with the photoelectric β-index. It is found that for A5V–G1V stars β can be given with a mean error of ±0.m020 (one spectrum).

This photometric system which was described for the first time in 1963 by Golay has been in use at the Geneva Observatory since 1959. We are involved in a long term observational programme having as objective the study of cluster stars, peculiar stars and interstellar extinction. This research aims at the determination of physical parameters such as effective temperature, gravity, abundance, or those of spectral type, absolute magnitude, [Fe/H]. To permit such an analysis the photometric effects of rotation, of multiplicity and of interstellar extinction were given particular attention. Attempting to attain the above mentioned purposes, we have had to give great care to the precision of the measurements during their acquisition as also during their reduction. Finally, in view of allowing a valid comparison between observations and the theory of stellar atmospheres, particular care had to be given to the maintenance of the pass-bands and to the knowledge of their form.

We intend here to give the main bibliographical references relative to the articles which present the general properties of the system, its applications and the catalogues.

A classification method has been developed for early–type stars observed on the Dominion Astrophysical Observatory photometric system. Two reddening-independent parameters, Q(35) and Q(38), are used. Q(35) is a measure of the Balmer discontinuity, while Q(38) is a measure of the strength of the upper members of the Balmer series. A preliminary calibration of Q(35) and Q(38) in terms of spectral types and luminosity classes is given, and applications to several groups of stars are shown.

An intermediate bandpass filter system is described that is useful for classification of G and K stars in terms of effective temperature (spectral type), surface gravity (luminosity class), heavy element abundance, and interstellar reddening. The system measures the strength of the cyanogen band in the region shortward of λ4216, the break in the continuum at the G band, and a colour index in the λ4500–λ4900 region of the spectrum. The surface gravity parameter is independent of the cyanogen index, and the cyanogen band strength, after a surface gravity correction, is used as a measure of heavy element abundance.

Application of the system to the determination of the atmospheric parameters of open cluster giants, as well as of field stars, is discussed. The system has been used to isolate stellar populations among field stars without any prior knowledge of their distances. From observations on this system, the only open cluster found so far to have giants with stronger cyanogen band absorption than the Hyades giants is the intermediate age southern cluster NGC 2477.

The David Dunlap Observatory (DDO) intermediate-band photometric system described in the previous paper has been used to observe a number of red giant branch, horizontal branch, and asymptotic branch members in each of the five globular clusters M3, M5, M10, M13, and M92. A calibration of the DDO system is described by which it was possible to determine the effective temperatures, surface gravities, [Fe/H] values, and masses of the observed stars. The mean [Fe/H] values for the clusters were found to be −1.01 for M3, −0.68 for M5, −1.44 for M10, −1.69 for M13, and −1.96 for M92. Evidence was found that the masses of the horizontal branch and asymptotic branch stars are systematically smaller than those of red giant branch members. Two stars were discovered to have CN bands that are anomalously strong for Population II objects. The observational results have been compared with the theoretical predictions of two detailed Population II evolutionary tracks and in general the agreement is good.

The programs of three-dimensional classification of stars using eight-color intermediate band photometric system are described. They include (1) investigation of interstellar reddening and absorption in the direction of objects significant from point of view of stellar evolution, (2) investigation of spatial distribution of stars of different spectral classes, luminosities and chemical composition, (3) the detection of unique objects.

Several photometric techniques for classification of B-type stars exist. This note describes another one, now nearly ready for publication.

U, B, V colors of CPD stars brighter than 12.m0 in the central region of NGC 2362. Color excess. Discussion of the distance modulus of the cluster.

The results of over 200 stars measured with interference filters in Hα, Hβ and Hγ are given here. 113 Be stars from the northern and southern hemispheres were included. The comparison of 60 of them, in common with the observations made by Mendoza about 15 years ago, shows that nearly half of them have differences larger than 0.m030 in Hβ (Crawford system). On the other hand the comparison with the observations made recently by Crawford indicate variations in a short time for some active Be stars.

At the National Astronomical Observatory in San Pedro Mártir, B. C. (Mexico), we have performed photoelectric photometry for 31 stars in a narrow-band system. The system allows the measurements of total absorptions of neutral oxygen at 7774 Å through three interference filters (20–25 Å half-width), one for the OI lines and two for the continuum.

The preliminary results are very encouraging: Supergiant stars, ranging in spectrum from the later subdivisions of type B to early G, can be clearly separated from other luminosity classes. The total absorptions depend strongly on the stellar luminosity and, to a lesser degree on the effective temperature (spectral type) of the star. The system may be improved by using narrower filters and detectors with a higher quantum efficiency, already in existence.