It has become clear in recent years that widespread and complex abundance variations exist among the stars of several globular clusters. ω Cen is the prime example, containing CH stars, weak-G-band stars, CN strong stars, TiO stars, and (not least) RR Lyrae stars possessing an anomalous range in calcium line strengths. (Harding 1962; Freeman and Rodgers 1975; Norris and Bessell 1975, 1977; Dickens and Bell 1976; Bessell and Norris 1976; Lloyd Evans 1977) The extent to which these phenomena result from primordial abundance variations or from the mixing to the surface of material processed by nuclear reactions is currently the subject of some debate. The present paper presents a short account of investigations of the giant and horizontal branches of 47 Tuc and of several giants in ω Cen. The results were obtained in collaboration with M. S. Bessell and K. C. Freeman, and will be fully reported in papers being prepared for submission to Astrophysical Journal. Most emphasis will be placed on 47 Tuc, where we have sought to address four problems. (i) How widespread are the CN anomalies reported by McClure and Osborn (1974), Bell, Dickens, and Gustafsson (1975), and Hesser, Hartwick, and McClure (19 77) - both throughout the HR diagram and in relative frequency? (ii) Is there any correlation between the relative frequency of the CN anomaly with radial position in the cluster, which might correlate with the color gradients observed by Gascoigne and Burr (1956) and Chun (1976)? (iii) Is there any evidence for G-band variations similar to those found generally in the metal poor ([Fe/H]< −1) globular clusters? (cf. Zinn 1973, 1977; Mallia 1975; Norris and Zinn 1977) (iv) Are the CN variations accompanied by variations in the abundance of the heavy elements?

The giant branch of ω Cen is much narrower in the I, (V-I) diagram than in the V, (B-V) diagram but at least 24 stars which lie on the red side of the giant branch are radial velocity members (Lloyd Evans 1977), (Lloyd Evans, to be published). Spectra of these stars show, with only one exception, stronger metal lines than stars of similar (V-I) color on the main giant branch (comparable to stars in 47 Tuc), strong CN bands where the stellar temperature allows and strong Ba II 4554 Å. The Ba II star RGO 371 (Dickens and Bell 1976) is a typical member of the group. The Cepheid V29 also shows strong CN bands and strong Ba II. A start has been made on a spectroscopic search for stars on the main giant branch showing these anomalies and the sample of stars with VI photometry is being enlarged. An explanation of the properties of these stars would appear to require both a large primordial metal content and mixing of processed material to the surface.

A review of the available Strömgren four-color data concerning horizontal-branch (HB) stars is presented. Several observers have studied seven globular clusters (with [Fe/H] values from −2.2 to −1.3) and more than 100 HB stars in fields at high galactic latitudes over the last ten years at Kitt Peak, Cerro Tololo, Mt. Wilson and Steward Observatories. The predictions of model atmospheres (Kurucz, 1976) allow one to calculate atmospheric parameters such as θeff, log g, and MV (Philip, Miller and Relyea, 1976). These parameters then can be compared with the predictions of evolutionary models (Sweigart and Gross, 1976).

A comparison of the observed data with evolutionary tracks indicates that within the rms error of observation and the computational error of the models the data and the tracks agree quite well. This matches the case for Population I stars, where a similar analysis shows a good match between observational data and evolutionary tracks (Philip, et al. 1977).

On the basis of available theoretical evaluations of the H-burning evolutionary phases of Population II stars, it is at present quite simple to predict the expected location of Zero-Age Horizontal-Branch (ZAHB) stars for assumed values of the age and original chemical composition. A result of such a computation is disclosed in Fig. 1, for the labelled values of the assumed evolutionary parameters.

During the last decade theoretical and observational research on globular cluster stars has fairly conclusively demonstrated that the mass of horizontal-branch (HB) stars is substantially smaller (by ~ 0.2 M⊙) than the mass of their main-sequence progenitors (cf. Rood 1973, Fusi-Pecci and Renzini 1975, hereafter MPS V and FPR respectively, and the literature quoted therein). Also well established is the fact that HB stars in a single cluster do not follow a unique evolutionary path but that some dispersion in at least one HB parameter (e.g. total mass, core mass, or composition) is required. Further, a growing body of astrophysical evidence shows that the HB morphology is poorly correlated with the abundance of low-ionization potential elements (ALIPE) such as Mg, Ca, Si, and Fe. This has raised the well known problem of the “second parameter” in the sense that something else besides ALIPE has to vary from cluster to cluster, the most popular candidates being age, t, helium, Y, and CNO abundance, ZCNO. Therefore, the main problems which research on globular cluster stars is presently facing are i) a quantitative knowledge of the mass loss process, ii) the identification of the origin of the HB dispersion in a single cluster, and iii) the identification of the second parameter(s). The solution of the last point has profound implications for the theory of the early chemical and dynamical evolution the Galaxy.

At the Leiden Southern Station a large program on galactic pulsating variables has been carried out with the Walraven five-channel photometer. For 170 Cepheids and 100 RR Lyrae stars south of δ =+15° there are now available 30000 VBLUW measurements (Walraven et al., 1964; Pel, 1976; Lub, 1977b). These data are used to derive physical properties of the variables. A general discussion of the VBLUW system and its calibration for A-G stars by means of the theoretical fluxes of Kurucz (1975) has been given by Lub and Pel (1977). Following this calibration, reddening and blanketing corrections, as well as effective temperatures, gravities, and their cyclical variations were determined, and luminosity and radius curves were derived (Cepheids: Pel, 1977; RR Lyrae stars: Lub, 1977a). We will present here some interesting results that have been obtained from these data. A more detailed discussion is in preparation.

Mapping of the Cepheid region of the Hertzsprung-Russell diagram has been difficult to do from a theoretical viewpoint because the masses and compositions of the Cepheids have always been very uncertain. It now appears that the various mass anomalies have been solved by distance scale changes and by the realization of very helium rich convection zones. This helium enrichment is caused by a Cepheid wind which blows away more hydrogen than helium, just as in the solar wind. Now with the evolutionary theory masses, radii, luminosities, and our new composition structures, the predicted blue edges of the instability strip and periods throughout the strip agree very well with observations.

Balona (1977) has described a modification of the Wesselink technique of radius determination which takes into account the effects of observational errors using the Principle of Maximum Likelihood. This leads to a tighter period - radius relationship for Cepheids than obtained by the classical method and suggests that some of the problems associated with radii estimation of RR Lyraes may be due to the incorrect statistical treatment of the data. With this in mind, we have applied the Maximum Likelihood technique to well-observed RR Lyraes. Except for two stars, the data used were taken from the literature.

Baade (1944) based his concept of stellar populations in galaxies on the HR diagrams that he inferred from the magnitude at which their brightest stars could be resolved. His type I population had bright blue supergiants like those in the disk of the Milky Way, while the brightest stars in type II were the red giants found in globular clusters. He postulated that the Hubble sequence of galaxy types from irregulars to ellipticals contained increasing proportions of Population II relative to Population I, and that similar differences characterized nuclear bulges of spirals relative to their disks. A very important revision of this picture came with the discovery by Morgan and Mayall (1957; Morgan, 1956, 1959) that the integrated blue light of the nuclear bulges of M31 and the Galaxy is dominated by strong-lined CN giants, not by the weak-lined type found in globular clusters. On the basis of integrated spectra of galaxies, Morgan developed a revised population scheme, in which the extreme types are a young-star rich population, like Baade's extreme Population I, and a young-star deficient population, analogous to Population II but generally metal-rich. Different proportions of these two types are still thought to represent the main differences among stellar populations in different regions of galaxies.

Our knowledge of the spectra of stars in other galaxies is essentially limited to those which are extremely luminous. In the LMC, slit spectra of ~ 200 stars have been obtained (Feast, Thackeray and Wesselink 1960, Ardeberg et al. 1972) down to Using an objective prism technique, Brunet et al. (1975) have been able to detect 272 OB stars with in the LMC. At the fainter limit this survey should include relatively normal 0 and early B dwarfs and giants. The principal aim of the present investigation was to obtain spectra of these stars for comparison with galactic stars and to apply the techniques of MK classification and measurement of Hγ equivalent widths to determine their luminosities and hence a distance modulus to the LMC.

HR diagrams for the stellar populations in other galaxies play a fundamental role in our understanding of the progress of stellar evolution and the effects of possible variations in chemical composition. It is important to compare what little information we have about stars in other galaxies with the same types of stars in our own Milky Way. Basically we are asking - are they the same, and how universal are the processes we observe in our Galaxy?

The Ursa Minor dwarf galaxy (RA = 15h08m.2, Dec = +67°18′ (1950)) was one of four dwarf ellipticals discovered by Wilson (1955) in a search of the Palomar Sky survey plates. The general properties of these galaxies resemble those of galactic globular clusters, linear size and density excepted (see Hodge, 1971, for a review). The characteristics of Ursa Minor are very similar to those of the Draco dwarf, for which a color magnitude diagram was published by Baade and Swope (1961). Van Agt (1967) published a study of the variable stars in Ursa Minor based on Baade's 200-inch plates.

A quarter of a century ago Keenan and Keller (1953) showed that the majority of high-velocity stars near the Sun outline a Hertzsprung-Russell diagram similar to that of old Population I. This result, which did not appear to fit into Baade's (1944) two-population model of the Galaxy was ignored (except by Roman 1965) for the next two decades. Striking confirmation of the results of Keenan and Keller was, however, obtained by Hartwick and Hesser (1972). Their work appears to show that high-velocity field stars with an ultraviolet excess (which measures Fe/H) of δ(U-B) ≃ +0m.11 lie on a red giant branch that is more than a magnitude fainter than the giant branch of the strong-lined globular cluster 47 Tuc for which δ(U-B) ≃ +0m.10. Furthermore Demarque and McClure (1977) show that the red giants in the old metal poor [δ(U-B) ≃ +0m.11] open cluster NGC 2420 are significantly fainter than are those in 47 Tuc. Calculations by these authors show that the observed differences between the giants in 47 Tuc and in NGC 2420 can be explained if either (1) 47 Tuc is richer in helium than NGC 2420 by ΔY ≃ 0.1 or (2) if 47 Tuc has a ten times lower value of Z(CNO) than does NGC 2420.

In an attempt to determine the Hyades distance (Golay, 1973), it was assumed that stars of the same “photometric 0m.01 box” (see Golay, 1977a) have the same visual absolute magnitude. The large amount of photometric data in the UBV B1B2V1G photometric system allows a discussion on this hypothesis (Golay, 1977b). We have 60 “photometric 0m.01 boxes”, each containing a central star of known trigonometric parallax and at least one Praesepe star. We select the 16 boxes (Table I) containing single stars or binaries with an estimated mass ratio, a relative probable error < 30% for the parallaxes and a standard deviation for colors <0m.007. The UBV B1B2V1G colors, the indices (B-V), (B2-V1) and the magnitude mV are taken from the Second Catalogue (Rufener, 1976) and the internal catalogue of the Geneva Observatory. The color index (B-V) is taken from Johnson (1952, 1957), Johnson and Knuckles (1955), the trigonometric parallax from Jenkins (1952, 1963) and Gliese (1969) and the spectral type for Hyades stars from Morgan and Hiltner (1965). The listings of all 0m.01 photometric star boxes in the UBV B1B2V1G system are given by Golay (1977c). The parallax obtained for Praesepe is π(0″.001) = 6.175 ± p.e. 0.1, i.e. a distance modulus (m-M) = 6m.05 and a distance of 162 parsecs. Golay (1977c) published the differences of the distance moduli for pairs of clusters having stars in the same box. The distances of these clusters is given in Table III, assuming a distance of 162 pc for Praesepe. The accuracy of this method is independent of both the distance magnitude and the chemical composition of the stars of a cluster since the stars have to be in the same box as a star with a known trigonometric parallax. The main sequence of Praesepe and a sample of Hyades stars, in the same photometric box with a Praesepe star is given in Table II. The depth effect in Praesepe being very small, the main sequence is very thin and the main sequence fitting procedure is better starting from Praesepe than from the Hyades.

When the system of correlations between stellar parameters that we know as the HR diagram was discovered, the physical basis for them was unknown and remained essentially unknown during four decades of eager application.

The detection of Am and Ap stars is possible in various photometric systems. Those most suitable for this purpose are certainly the uvbyβ (Strömgren 1967, Cameron 1967), the Barbier Morguleff system completed by Gerbaldi (1972, 1977) and the Geneva system. It seems important to me that the Am and Ap stars should be detected with a system which also allows one to classify normal stars, and to obtain quantitative information about metallicity or other peculiarity. The Geneva system satisfies both these criteria and numerous Am (about 400) and Ap (about 250) stars are in a catalogue prepared on magnetic tape by Rufener, which follows the one published in 1976 (Rufener 1976).

We have observed over a hundred A-type stars in the ∝ (16) ∧ (9)-photometric system defined by Mendoza (1976a, b; 1977). These stars also have UBVRI measurements. Sixty five stars have been observed in the thirteen color-photometric system (Mendoza 1971, and Johnson and Mitchell 1975).

The Vilnius photometric system permits one to determine photometrically temperatures and log g's or spectral types and MV's of stars in spite of the presence of interstellar reddening. The position of any normal star in the HR diagram can be determined by its direct calibration in five reddening-free quantities Q. The position of the star in the HR diagram corresponds to the point at which five Q-isolines intersect. This is the most direct way to find the position of the star in the HR diagram from photometry.

Half a centry ago Henry Norris Russell and Heinrich Vogt independently made a conjecture concerning the structure of spherical stars which are in hydrostatic and thermal equilibrium (Russell, 1927; Vogt, 1926). This conjecture has later come to be known as the Vogt-Russell theorem and is usually formulated as follows: The structure of a star is uniquely determined by the mass and the composition. In other words, the statement claims the existence and uniqueness of a stellar equilibrium configuration for given parameters mass and composition, and you may find what is called a mathematical proof in many textbooks on stellar structure.

In a paper by Perrin et al. (1977), we have constructed an empirical HR diagram for 138 nearby F, G and K stars, for which we had: i) an effective temperature and a metal content derived from a detailed analysis; ii) a reliable bolometric magnitude obtained from an absolute magnitude My, based on a large parallax and a rather small bolometric correction.