STScI Logo

The Evolution Of Low Luminosity Galaxies And Faint Blue Galaxies
The Star-Forming Histories of the Nearst Dwarf Galaxies

Tammy Smecker-Hane  45 min  Review of nearby dSphs and the LMC 
Taft Armandroff     10 min  Andromeda I dE - HST photometry of giant/horizontal
                                branches which illustrate the classic second 
                                parameter problem                           
Mario Mateo         10 min  Leo II dSph - HST deep CMD and the age 
Discussion          25 min

We will open the meeting by reviewing what is currently known about the star formation and chemical evolution histories of the dwarf galaxies in the Local Group. For if we are to claim to understand the physics of galaxy evolution, we must first be able to successfully model these, the simplest of galaxies. Because of their close proximity, we can determine ages and chemical compositions on a star-by-star basis and accurately determine their evolution from t=0 to the present day (with varying degrees of time resolution). The density of stars in regions of the color-magnitude diagram (CMD), particularly at main-sequence turnoff, gives us the star formation history. We can hope to disentangle the degeneracy of age-metallicity in these galaxies (as opposed to more distant ones for which only integrated light is available) by determining metallicity from the color of the giant branch, or by explicitly determining the metallicity distribution from spectra of individual stars. Using these methods, we are finding that, contrary to past prejudices, even the dwarf spheroidals in the Local Group have had complex evolutionary histories.

The evolution of the Carina, Fornax, Leo I and Leo II dwarf spheroidals and the Large Magellanic Cloud, which span a wide range in galaxy luminosity from M_v=-9 to -18, will be discussed in detail. (I have chosen these because CMDs reaching the critical main-sequence turnoff region have been obtained for each.) In these galaxies, episodic "bursts" of star formation appear to be common. I use quotes around bursts to highlight the fact that we do not yet know how bursty these were. Theoreticians have argued that low luminosity dwarfs form at early times on a cloud cooling timescale, typically 10^7 yrs. However, we will argue that evidence points to these bursts happening on timescales of greater than, or of order, 10^8 yrs with duty cycles of ~5 Gyr. The duration of the bursts and the duty cycle have obvious implications for the contribution of intrinsically low luminosity dwarfs to the faint blue galaxy counts.

My review talk will consist of:

  1) Carina dSph (M_V=-9)
     The Carina dSph has had bursts of star formation 3, 5, 6 to 8 and 
     ~15 Gyr ago with little or no star formation between these episodes. 
     Roughly 70% of the stars are of intermediate-age and 15% are
     ~15 Gyr. The colors of stars on the giant branch imply the mean 
     metallicity is [Fe/H]=-1.8 and the dispersion is low (~< 0.2 dex).
     The lack of chemical enrichment and the high M/L ratio (32 Mo/Lo),
     suggests that galactic winds ejected gas through out the evolution 
     of the galaxy and lead to an inefficient conversion of gas to stars.

     references: Mighell (1990) A&AS 
                 Mighell & Butcher (1992) A&A 
                 Smecker-Hane, et al. (1994, 1996) - AJ and in progress
                 Mateo et al (1993) AJ
      Carina illustrates some puzzling problems that are ubiquitous to
      dwarf galaxies:

              - The energy input from the ~10^3 supernovae that
                exploded during the first episode of SF is much
                larger than the binding energy of the galaxy (even with 
                90% of the mass in a dark matter halo) -- naively, this 
                galaxy should have ejected its ISM during the first SF 

              - But it did not. Multiple episodes of SF occurred
                so either ejection of gas was inefficient, or new 
                gas was accreted. The observed metallicity argue for
                pristine gas, rather than gas that had been ejected 
                during the first burst of SF.  However, accretion 
                would be difficult because the total mass is very low 
                (10^7 Mo), the cross section is small (r_t=660 pc) and 
                the kinematic speed in the outer halo is ~220 km/s, 
                hence for gas to successfully merge with the galaxy 
                it would have had to be moving in nearly the same orbit.

                I will briefly discuss my objections to the idea
                put forth by Lin and Murray (1994) that Carina is the 
                remains of a piece ripped off a larger-LMC galaxy, in which
                a second episode of SF was triggered by subsequent
                accretion of gaseous tidal debris.
              - I will argue that the best explanation for such an episodic 
                SF rate is that cooling and SF operates slowly (on ~ 10^8 yr)
                and ejection of gas is inefficient. Supernovae drive winds 
                which  rapidly accelerate the low density 
                ISM. However, because of the small physical 
                size and diffuseness of the galaxy, the wind rapidly 
                breaks through holes in the ISM -- leaking out metals, heat,
                energy and momentum -- but only slowly disrupt the densest 
                molecular gas clouds. I argue that SF in the first "burst"
                occurs slowly over ~10^8 yr, some gas remains in the galaxy 
                after SF is quenched, and the gas stays below the threshhold 
                for SF for a few Gyr. Exactly what causes the few Gyr
                timescale between bursts remains to be answered -- maybe
                the remaining gas is optically thin to the UV background
                and remains photoionized for long times, or maybe 
                tidal torques or increased ambient pressure can repeatedly
                nudge it over the threshold for SF (the suspected orbital 
                peroid around the Galaxy is ~ few Gyr). 

    2) Leo I dSph (M_V=-10) 

       A CMD of Leo I obtained with HST compliments the ground-based CMD and
       shows that Leo I also has experienced two distinct epochs of SF. 
       Roughly 90% of the stars have an age of ~3 Gyr, and ~10% have ages 
       of ~> 12 Gyr.

       references: Lee et al (1993), AJ
                   Mateo et al (1994) BAAS

    3) Leo II dSph (M_V=-12)

       The CMD of Leo II obtained with HST shows a period of extended SF 
       from 7 to 14 Gyr ago. However, because of its larger distance,
       subsequently larger photometric errors, and the narrowing of the 
       isochrones for older ages, it is unclear whether how constant
       the SFR was during this interval.

       reference: Mighell et al (1996), ApJ

    4) Fornax dSph (M_V=-14) 

       Fornax dSph has a higher luminosity, mass, surface density and a 
       retinue of 5 globular clusters. The latest CMDs of Fornax show that it
       has had an extended SF history. Stars began to form early, ~15 Gyr ago.
       There are hints that the SFR rose between 5 and 8 Gyr ago, and a 
       trickle of stars continued to form until very recently. Indeed, there 
       is a small population of very young (few 10^8 yr old) stars. Unlike 
       lower mass dSphs, Fornax stars show a wide range of metal abundances 
       from -1.7 < [Fe/H] < -0.7. Its lower M/L (~10 Mo/Lo) suggests a 
       more efficient conversion of gas to stars.

       references: Buonanno et al (1985), A&A
                   Beauchamp et al (1995), AJ
                   Smecker-Hane et al (1996), in preparation
                   Stetson et al (1996), in preparation
                   Mateo et al (1991), AJ

    5) Large Magellanic Cloud (M_V=-18)

       For many years, the uneven distribution of LMC star cluster ages 
       (0.1 Gyr, ~3 to 4 Gyr, and ~12 to 15 Gyr) hinted at a non-constant
       SFR. Ground-based CMDs of LMC field stars have been analyzed. 
       Because of the limitations of the photometry (driven by crowding), 
       an unknown mix of metallicities, and possible inadequacies of 
       stellar evoln models, the Padova group have advocated "region-fitting"
       to quantify the LMC's SFR. They count stars in key regions of the 
       CMD that are age sensitive to test simple models of the SFR. Assuming
       a constant SFR + one burst, they determine that the best fit model 
       is one in which a burst that increased the SFR by a factor of 10 
       occurred ~ 2 to 4 Gyr ago (with mean age possibly varying with position 
       across the LMC) and having a duration of  ~10^9 yr or more. Although 
       the data hint that this simple model (a single burst) may be

       In support of these results, a new CMD of field stars obtained with 
       HST shows structure in the main-sequence turnoff region - evidence of 
       an episodic SFR. A burst with an age of ~2 Gyr, duration 10^8 yr, 
       superimposed on a nearly constant SFR in the last 1 to 3 Gyr is derived.
       Of order 25% of the stars may have formed in the 2 Gyr burst with 80%
       of the stars forming from 1 to 3 Gyr ago. This was preceded by low SFR
       in the preceeding few Gyr. The mean age of the older disk population is
       ~8 Gyr and with a very low fraction of disk stars being ~15 Gyr old. 

       Complimenting this, new results on the age and chemical abundances of 
       planetary nebula in the LMC show that a rapid increase in the chemical
       abundances (factor of 2) occurred 2 Gyr ago. 

       But these data/analysis do not preclude an more complex evolution at 
       ages ~> 4 Gyr. In fact, they hint that the single burst picture is not
       complete. It is imperative that we use HST to overcome crowing 
       problems at faint mags, in combination with ground based photometry 
       and spectroscopic metallicity determinations of bright evolved stars,
       if we are to resolve the evolutionary history of one of our nearest, 
       most luminous, neighbors.

       references: Bertelli et al (1992), ApJ
                   Westerlund, Linde & Lynga (1995), A&A  
                   Vallenari et al (1996a, 1996b), A&As, in press 
                   Gallagher et al (1996), ApJ, in press
                   Smecker-Hane et al (1996), in progress
                   Dopita et al (1996), in progress