Following Shapley's brave application of brightest red giants to the Galactic distance scale, continued applications were ironically being made at the same time as new observational data were being assimilated (sometimes by the very same people) which were showing that the method was fatally flawed. Flawed in the optical, that is. Late into the 1960's and 70's Arp in his review and even Harris (1974) in his doctoral thesis used the apparent B-band luminosities of the brightest (few) red giants as statistical indicators of distances to globular clusters. And this proceeded in the face of compelling evidence (and theory) that the giant branch morphology was a sensitive function of age and most especially a function of metallicity (see Figure 18 for examples). Bolometric corrections, driven primarily by atmospheric line-blanketing effects in the optical resulted in differences of up to 2 mag between the absolute magnitudes of the brightest red giant stars in metal-rich (line-blanketed) systems as compared to those in metal-poor globular clusters. It was not until the detailed and calibrated studies of the color-magnitude diagrams (CMDs) of Galactic globular clusters by Arp, Sandage, Wildey and others, using more modern photometric systems, that the absolute UBV magnitudes of the brightest giants in globular clusters were found to be less than ideal as distance indicators. The detailed morphology of the giant branch, (especially its height above the horizontal branch) and its terminal color were found to depend critically on the mean metallicity of the system.
Figure 18. A schematic representation of
the effect of metallicity on the morphology of the red giant branch when
observed at optical (UBV) wavelengths.
The application of the then new CCD detectors to unexplored wavelengths led Mould & Kristian (1986) to reconsider the tip of the red giant branch as a potential distance indicator. With admittedly sparse statistics, they noted in a series of papers that the dominant feature in the CMDs of Local Group galaxy halos was the presence of a giant branch population of stars, showing a wide range of colors, but generally terminating at high-luminosity at a well-defined and fairly constant magnitude. Of the four galaxies that they initially imaged (M31, M33, NGC 147 and NGC 205) only M31 had an independent and reasonably secure distance published (based on Cepheids), while the distance to M33 was still under considerable debate. At the conclusion of their 1986 paper, Mould & Kristian reasonably called for more work before the TRGB method could be considered a mature distance indicator. Their requests included: (1) The examination of a larger sample of Local Group galaxies, and (2) an absolute calibration through Galactic globular clusters based directly in the Cousins I-band system.
Figure 19. A sampling of early
color-magnitude diagrams for Local Group galaxies and their halos.
Answering the call for a direct calibration of the method using Galactic globular clusters Da Costa & Armandroff (1990) published a grid of standard globular cluster giant branches observed in the Kron-Cousins VI system. And they found ``acceptable agreement'' between the bolometric magnitudes of the brightest giants and the theoretical predictions for the luminosities of helium core flash. At the same time, considerably more observational work was being undertaken which supported the use of the TRGB as a distance indicator in the I band. Freedman (1988) made the first combined CCD-based determination of a TRGB distance and a Cepheid-based distance for IC 1613. While confusion over the disparate Cepheid distance moduli to M33 finally resolved itself with the application of multiwavelength CCD observations of the variable stars in that galaxy (Freedman, Wilson & Madore 1991), the result again confirmed the TRGB calibration. A comprehensive review and survey of published results pertaining to the comparison of tip distances in comparison to RR Lyrae and Cepheid distances was given by Lee, Freedman & Madore (1993), and the interested reader is referred to that paper for additional details and references.
Figure 20. Sensitivity of the TRGB to
age and metallicity effects (adapted from
Lee, Freedman & Madore 1993).
The upper panel shows fiducial giant-branch lines for a series of
Galactic globular clusters having a range of metallicities. Note how the
terminal I-band magnitude is a very weak function of chemical
composition. The lower panel shows the (negligible) effect of age (7-17
Gyr) on the RGB peak luminosity, based on Revised Yale isochrones
calculated for [Fe/H] = -1.3 dex.
The TRGB method, as applied by Lee, Freedman & Madore (1993) is based on a calibration established for the metallicity range of -2.2 < [Fe / H] < -0.7 dex, derived by using empirical loci of the red giant branches for four Galactic globular clusters given in the paper by Da Costa & Armandroff (1990). Da Costa & Armandroff estimated distances by adopting the metallicity-MV magnitude relation for RR Lyraes based on the theoretical horizontal branch models for YMS = 0.23 of Lee, Demarque & Zinn (1990). At present, the exact calibration at the higher end of metallicity range (i.e., metallicities exceeding the most metal-rich Galactic clusters) and its practical effects on TRGB distances to composite systems still remains uncertain.