The MK spectrum-luminosity class diagram (S-LC diagram) between types 09 and G0 is a two-dimensional array containing approximately 100–150 information elements; each information element is a single cell denoted by a spectral type and a luminosity class. These individual, non-dimensional cells (i.e. AO III) can be considered as carriers of observational information toward the HR diagram. Some categories of peculiar spectra (for example, spectra resembling that of the “silicon star” θ Aur) can be attached to certain cells, and can be considered as local perturbations upon the underlying continuity of the spectrum-luminosity class sequences. By such a procedure, the S-LC diagram cells can be considered open-ended, in that progressive additions to the cells can be made indefinitely into the future. Such a procedure can incorporate successfully an unlimited variety of “normal” and “peculiar” categories of spectra - both those already known and those still to be discovered. A morphological system that is unitary thus comes into existence - a system of considerable astrophysical interest. The translation of the detailed language of this system into that of the classical HR diagram is a task for the future.

The MK system of spectral classification does not include luminosity classes for types earlier than 09. Subsequent work with higher dispersions has allowed a luminosity classification for early O stars to be proposed. One of the principal criteria at types 09-B0 is the selective, negative luminosity effect in the Hell 4686 Å absorption line; a similar effect was found in the NIII 4634-4640-4642 Å absorption lines at 09-B0. These particular lines weaken markedly with increasing luminosity, while other HeII lines maintain a constant ratio to HeI (by definition), and other NIII lines strengthen considerably. These anomalous lines are the same ones which selectively appear in emission in Of stars, and the derived luminosity classification for earlier types is based upon the hypothesis that the negative absorption effect at 09-B0 is due to emission filling by the same mechanism producing the Of phenomenon, and that hence the latter may also be identified as a luminosity effect (Walborn 1971). One observes strong 4686 Å absorption and often weak 4640 Å emission [denoted ((f))] on the main sequence to the earliest types. As luminosity increases at a given spectral type, the HeII absorption weakens while the NIII emission increases [denoted (f)] and finally the most luminous stars have both 4686 and 4640 Å strongly in emission [denoted f]. A sequence of spectra all with types very near 07, but showing the marked changes with luminosity class in 4686 and 4640 Å, is illustrated by Walborn (1973). These effects show good agreement with other spectroscopic luminosity criteria such as SiIV absorption (Conti and Alschuler 1971) and Balmer-line strength (Crawford 1975, Morrison 1975), which are however less sensitive at the earliest types.

Even if we confine ourselves to the classical HR diagram, postponing the transformation of spectral types and visual absolute magnitudes into basic physical variables, there are several ways to choose the data. If, for example, we want the density of points to represent the spatial density of stars, then we must concentrate on reducing sampling errors, usually at the cost of accuracy in some parts of the diagram. The Hess diagram is very useful, but an equally useful alternative is an HR diagram which includes only those stars whose positions are determined as accurately as possible. The objective is to classify the fine structure in the pattern in the luminosity-spectrum domain, and thus to define more precisely the borders of zones which are either occupied or avoided by stars of a given population. It is this pattern which is compared with theoretical evolutionary tracks and it is such a diagram based on the revised MK classification of the recent Atlas of Keenan and McNeil (1976) that I shall present here.

In the lower main sequence, the coolest halo population stars do not appear to be particularly subluminous in either the standard HR diagram or the (MI, R-I) plane. Earlier type M subdwarf's (e.g., Kapteyn's Star) that lie one or more magnitudes below the main sequence exhibit enhanced hydride absorption bands relative to M dwarfs of the same temperature. From visual region (4500–7000 Å) low dispersion image tube spectrograms we find that the cool halo M dwarfs exhibit enhanced absorption in the CaOH 5530 Å band and Na I resonance lines. Easily noticeable changes in the strength of the CaH “B” band relative to that of the TiO bands was found to be more difficult to determine in halo dwarfs later than spectral type M4. Several M dwarfs with old disk type motions have spectral peculiarities similar to those of the halo dwarfs. In both cases, Balmer emission is not present in most stars, in the rest it is very weak. Therefore, the anomalous strength of the CaOH 5530 Å band and Na resonance lines appear to be a more conveniently utilized qualitative criterion in spectroscopically identifying very cool M dwarfs which are subluminous with respect to most main sequence stars of the same temperature class.

Spectroscopic parallaxes are an important tool for determining distances of individual stars in the Galaxy. Unfortunately, for many astrophysical applications this method does not produce sufficiently accurate results for distant stars. Therefore, we started an attempt to improve the accuracy of spectroscopic parallaxes by evaluating spectroscopic temperature and luminosity criteria such as those of the MK classification spectrograms which were analyzed automatically by means of a suitable computer program. If necessary, the grain noise is reduced by Fourier filtering techniques. Then, the computer derives an approximate spectral type and the wave length scale from selected prominent spectral features. As a next step the computer searches the spectrum for adsorption features and produces an internal catalog of detectable absorption lines. Then, depending on the approximate spectral type, line ratios, line widths, and line depths of selected lines are calculated and compared to the corresponding values of standard stars. Since the investigation is still in progress, no final result on the achievable accuracy can be given here. However, preliminary results indicate that spectroscopic parallaxes derived by the computer method are at least as accurate as those derived from the conventional visual classification technique. A more detailed account of this investigation will be published elsewhere.

The Large Magellanic Cloud has been surveyed on ultra-low dispersion (10,000 Å mm−1) plates taken with the 102 centimeter reflector at Kavalur. Many new red and blue stars were found. The LMC Bar was found to have a strong component of red stars and a very weak component of faint 0B stars. In general the regions other than the bar contain very luminous 0B stars, faint 0B stars and a few very red stars. The 30 Doradus complex is rich in luminous red and blue stars.

The objective prism (or grating) is the most efficient stellar spectrograph since it suffers no light losses at a slit and - compared to a slit spectrograph - only small losses in the camera optics. However, objective prism spectra have two principal disadvantages:

1. (1) the spectral resolution (and consequently the sharpness of spectral features) depends on the seeing, which may vary from plate to plate.

2. (2) quantitative measurements are difficult because of the difficulty in obtaining an exact photographic calibration; no comparison spectrum can be placed near the stellar spectra.

The discovery of different ratios of carbon (C-type) to oxygen stars (M type) in the direction of the central regions of our own and the nearest galaxies, the Large and Small Magellanic Clouds, is the result of surveys made by Blanco, Blanco and Hoag (unpublished) who observed Baade's Window near NGC 6522 in the direction of our Galaxy's nuclear bulge and by Blanco, Blanco and McCarthy (in press) who observed selected regions in the center of the Large and Small Magellanic Clouds. We have observed each of these central regions with a combination of a grating and prism set in the converging beam of the prime focus of the CTIO 4 meter telescope (Hoag 1976). We used Eastman IV-N plates hypersensitized with silver nitrate solution; the spectral plates were exposed through a Wratten 89 filter for 60 min. We limited our spectral survey to M type stars later then M5 and to the carbon stars, using the criteria of Nassau and colleagues (1949). Thus far no S type stars have been discerned in central regions of the LMC or SMC at the dispersion used (2300 Å/mm).

This paper reviews the main sequences defined by members of the Hyades cluster and by the field stars in the solar neighborhood. For this purpose, the discussion is limited primarily to the stars of the lower portions of the main sequence, especially those of spectral classes K and early M. There are two reasons for emphasis on the faint red dwarf stars. First, the value of a parallax depends on its size or, more accurately, on the error in parallax divided by the parallax itself. Large parallaxes of high precision occur in large numbers only for stars inhabiting the lower main sequence. Furthermore, brighter stars of earlier spectral classes are more likely to be influenced by evolutionary effects which may differ between the Hyades and field stars, and which are difficult to calibrate.

Our algorithm for stellar luminosity calibrations (based on the principle of maximum likelihood) allows the calibration of relations of the type:

Where n is the size of the sample at hand,

• Mi, are the individual absolute magnitudes,

• Cij are observational quantities (j = 1, …, N), and

• qj are the coefficients to be determined.

If we put N = 1 and CiN = 1, we have the mean absolute magnitude of the sample. As additional output, the algorithm provides us also with the dispersion in magnitude of the sample σM, the mean solar motion (U, V, W) and the corresponding velocity ellipsoid (σu, σV, σw).

A statistical parallax algorithm (Clube and Dave, 1978a,b), using the technique of maximum likelihood, has been applied to a set of 103 ‘ab’ - RR Lyrae stars in the solar neighborhood (r 2<kpc), using observational data from the Royal Greenwich Observatory Bulletins. A second set of 130 ‘ab’ - RR Lyrae stars has been kindly supplied to us by Dr. A. Heck (Université de Liège) to permit a comparison between our analyses. The purpose of this investigation was:

(a) to investigate the variations of kinematical parameters and absolute luminosities of these stars as functions of Preston's index ΔS and of log (Period).

(b) to identify those RR Lyrae stars in the solar neighborhood which most closely resemble those in the galactic halo, those near the galactic center, and those in the Magellanic Clouds.

Twenty years ago, Wilson and Bappu (1957) published their observational correlation of MV and the logarithm of the full width at half maximum of the CaII K-line central emission for G, K and M stars. The technique was stated to have ± 0m.5 accuracy, while a tabulation of MV(K) by Wilson (1976) suggests the error may only be ± 0m.3. This accuracy makes the approach valuable for late-type supergiants since other methods suffer from comparable errors. However, for F through M supergiants (Ia, 0), circumstellar absorption obscures the chromospheric K-line core emission and excludes such objects from the Wilson-Bappu correlation. I report here on a new class of emission lines in late-type giant and supergiant spectra that exhibit MV correlated widths, yet are detectable among the brightest stars.

Since regular HR diagrams require apparent magnitudes, colors or spectra, and parallaxes, and such complete data are available for relatively few stars, there may be some advantage in making up diagrams which utilize proper motions instead of parallaxes, and are thus statistically similar to an HR diagram. The reduced proper motion, first used by Hertzsprung, is defined as H = m + 5 + 5log μ, but may also be written as H = M + 5log T, where T is the tangential velocity, and is expressed in astronomical units per year. A diagram plotting H against color will thus contain the considerable dispersion in tangential velocity which is a serious disadvantage. However, this is outweighed by two practical advantages. First the one and the same person who does the proper motion survey can, and does also determine the other two quantities needed. Second, when using data obtained from such a proper motion survey one deals, statistically, with all the stars within a given distance and the results, therefore, are much more representative of the real situation in space than many HR diagrams which often contain an unrealistic preponderance of giants.

Scales of fundamental bolometric connections (B.C.) and effective temperatures (Teff) as a function of spectral type or color are necessary for the comparison of observations and theory in the HR diagram.

Observational data for stars in the solar neighborhood received before January 1969 were included in the Catalogue of Nearby Stars (Gliese, 1969) which contains more than 1800 objects, single stars and components of binaries, nearer than 22 pc. Since then more than 8 years have gone by and many new data have been published, especially new trigonometric parallaxes as well as photometric distance determinations. Today we know more than 2000 stars with parallaxes measured to be greater than 0m.044. But many of these distances are still very uncertain and for a considerable number of objects, especially the faint ones, spectral classificiation is lacking or the photometry is still not satisfactory.

Since the absolute magnitude of the Sun is known accurately and the solar spectrum defines spectral type G2 V, it would appear that the position of the Sun in the HR diagram is well established. However, the color of the Sun and its metallicity relative to other stars remains controversial. It is known from theory that stellar groups of differing composition will have different main sequences.

The observed difference in the strength of the metallic lines in Coma stars relative to Hyades stars is caused by a systematic difference in [Fe/H] of less than 0.2 dex or 50%. Photographs and quantitative measures of the systematic differences between the Hyades and Coma spectra are presented. Visual comparison and quantitative analysis of the solar spectrum relative to Hyades and Coma spectra reveal that the Sun is metal rich similar to the Hyades stars rather than of normal disk metallicity as represented by Coma stars. Evidence is presented showing that the color and spectral line strengths of the Sun are more similar to stars classified G3 and G4 in the literature than G2. If the Hyades modulus is y = 3m.25, the (B-V) color of the Sun must be 0m.65 or redder for the Sun to be on or above the Hyades main sequence.

In this paper we present some HR diagrams constructed using data from the Michigan Spectral Catalogues. Houk (1975) has been systematically reclassifying the Henry Draper stars on the MK system, from the south pole northward. Objective-prism plates, with a reciprocal dispersion of 108 Å/mm, have been taken with the Michigan Curtis Schmidt telescope at Cerro Tololo Inter-American Observatory in Chile. The spectra are classified visually from the plates, and the results are put onto IBM cards and magnetic tape from which the catalogues are produced.