Next Contents Previous

3. THE DATA: RECENT HST OBSERVATIONS

Thus the time is ripe for re-observing the the resolved stellar populations of all the galaxies in our LG. Much of our detailed knowledge of the SFHs of galaxies beyond 1 Gyr ago comes from the Milky Way and its nearby dSph satellites or from HST CMDs. To date, the limiting factors have been crowding and resolution limits for accurate stellar photometry from the ground. The large collecting area, the field field of view, and the extremely impressive image quality and stability of UT1 combined with the extremely good seeing attainable at Paranal makes it possible to obtain more accurate CMDs to fainter magnitudes, (ie. do better) than HST. These facilities provide unique opportunities to extend beyond our immediate vicinity and encompass the whole LG. To date HST has observed the resolved stellar populations in variety of nearby galaxies (e.g., dE: NGC 147, Han et al. 1997; Irr: LMC, Geha et al. 1998; Spiral: M 31, Holland et al. 1996; M 33, Sarajedini et al. 1998; BCD: VII Zw 403, Lynds et al. 1998; NGC 1569, Greggio et al. 1998; dI: Leo A, Tolstoy et al. 1998; dSph: Leo I, Gallart et al. 1999). Every galaxy which has been looked at carefully something new has been learnt.

An HST program was initiated by S killman (e.g. Skillman et al. 1998), observing a sample of four nearby dI galaxies (Sextans A, Pegasus, Leo A & GR 8), using four orbits of telescope time per galaxy, in three filters (effectively B, V and I). The results have been dramatic and illustrate the tremendous advances possible, even with short exposures, when crowding has been virtually illuminated. Here I am going to provide a summary of this program.

Figure 4

Figure 4. Here we show the results for the analysis of the HST / WFPC2 data of Leo A (Tolstoy et al. 1998). In a. is the V-I, I CMD, 1 orbit exposure time per filter. In b. is the B-V, V CMD, 2 orbits in B. In c. is the best match Monte-Carlo simulation model (in V-I, I) found for these data convolved with the theoretical measurement error distribution, and in d. is the SFH that created the model CMD which best matches these data. See Tolstoy et al. 1998 for more details.

Next Contents Previous