In the absence of spectroscopy, metallicities for dEs have been derived from their CMDs by comparing the colors of the red giant branches (RGB) with those of globular clusters. This is not entirely satisfactory because of the age-metallicity degeneracy in RGB colors, and because element abundance ratios, which also affect RGB colors, may not be the same in the dEs as in the globulars. A few efforts have been able to estimate metallicities from low-resolution spectra obtained with 4-5 meter telescopes; these studies are summarized in Mateo (1998). The photometric and spectroscopic studies show the dEs to have quite low <[Fe/H]>, ranging from about -1.3 in Fornax to about -2.2 in Ursa Minor. Most of the dEs show evidence for a significant spread in [Fe/H], ([Fe/H]) = 0.2-0.5 dex, based on the width in color of the RGB (disregarding possible age spread contributing to this).
With several 8-10 meter telescopes now available, medium-resolution (R 5,000) spectroscopy of RGs in the Milky Way satellite dEs can be almost routinely done, while high-resolution spectroscopy (R > 15,000) is possible for the brightest giants. These facilities offer exciting new possibilities for understanding the evolution of the Local Group dEs. One example of what can be done is the VLT study of Tolstoy et al. (2001), who obtained Ca II triplet measurements for 37 red giants in Sculptor, 32 RGs in Fornax, and 23 RGs in the dI NGC 6822. Having measured metallicities for individual stars and existing CMDs for these galaxies, it was possible for Tolstoy et al. to assign ages to each star directly by comparison with isochrones of the proper metallicity, and subsequently to derive the time evolution of both the star formation rate and the metallicity. This is shown for Sculptor in Figure 7 taken from Tolstoy et al. The results are consistent with an initial burst of star formation between 11 and 15 Gyr ago with a metallicity [Fe/H] -1.8, a sharp subsequent decline in the SFR, followed by a slow decline in star formation until it stops approximately 5 Gyr ago. The metallicity evolution is very modest, with the youngest stars having a mean [Fe/H] of only -1.4. Their data for Fornax, on the other hand, show a low star formation rate over the time period 10-15 Gyr ago, then a sharply increased rate over the next few Gyr, and a higher mean [Fe/H] of -0.7 for the youngest stars.
Figure 7. The star-formation and metallicity evolution history for Sculptor derived by Tolstoy et al. (2001) from Ca II triplet measurements and photometry of red giants. The upper panel shows a schematic plot of how the star formation rate may have varied over time. The lower panel shows the corresponding variation in metallicity over the same time period (dashed line). Overploted on the lower panel are the Ca II triplet measurements for individual Sculptor giants, with ages determined using isochrones.
The main source of uncertainty here is that the Ca/Fe ratio in these stars may not reflect that of the calibrator globular clusters stars - and the Ca/Fe ratio may actually vary with time within each galaxy! Nevertheless, with more data like this for the Local Group dwarfs it should be possible to derive very accurate evolution histories for these galaxies.