1 Introduction

Starburst galaxies churn gas into stellar light at tremendous rates, so fast that the time to exhaust this material is short compared to the age of the universe (e.g., Mirabel 1992). In recent years they have also become notorious for their propensity to form swarms of star clusters, including massive, compact super star clusters (SSCs), which have been suggested by many authors to be the progenitors of globular clusters (e.g., Hunter et al. 1994; Whitmore & Schweizer 1995; Ashman & Zepf 2001).

The starburst phenomenon occurs rarely in the local universe, but increases in frequency at larger lookback times (e.g., Huchra 1977; Liu et al. 1998; Glazebrook et al. 1999 and references therein), which may be connected to the observed excess of distant "faint blue galaxies'' (Ellis 1997). Studies of moderate redshift ($z \sim 1$) galaxy samples reveal substantial populations of luminous galaxies with high star formation rates (SFRs) (Cowie et al. 1995; Omont et al. 2001). These include the blue "compact narrow emission line galaxies'' (CNELGs; Koo et al. 1994; Koo et al. 1995; Guzmán et al. 1996, 1997, 1998; Phillips et al. 1997).

Similar objects have been categorized into various groups with various names, such as the "luminous blue compact galaxies'', or LBCGs (our preferred term for such objects; Jangren et al. 2001). Galaxies of this class have small linear sizes (<20 kpc), luminosities near or above L^*, high surface brightnesses, strong emission lines, and blue colors, all features that are indicative of enhanced SFRs.

The evolutionary fate of moderate redshift LBCGs is a subject of some debate, and depends strongly on the current and future SFRs in comparison with the extent of their gas reservoirs. One scenario holds that the most compact of the LBCGs will evolve into spheroidal stellar systems covering a range in mass (e.g., Babul & Rees 1992; Guzmán et al. 1998). The similarity of this subclass of LBCGs to luminous, young, star-forming H II galaxies has led to a classification of some systems as "H II-like'' (Phillips et al. 1997; Guzmán et al. 1997).

Passive evolutionary models predict that after 4-6 Gyr, the luminosities and surface brightnesses of these H II-like LBCGs could resemble local low-mass ellipticals, like NGC 205 (Guzmán et al. 1998). The viability of this scenario depends on the star formation in LBCGs having a duration of less than 1 Gyr, and in that time they must lose almost all their gas (Pisano et al. 2001), which is difficult to accomplish in any but the least massive dwarfs (Mac Low & Ferrara 1999; Ferrara & Tolstoy 2000).

Alternatively, LBCGs could be analogs to intense starbursts in nearby disk galaxies, especially when observed in near face-on orientations that favor the escape of blue/ultraviolet light from the starburst region (Gallagher et al. 1989; Gallagher et al. 2000a; Barton & van Zee 2001). The evolutionary scenario for disk-like LBCGs does not necessarily lead to their descendants being spheroidal galaxies. In nearby examples of this starburst mode, relatively minor interactions between galaxies can yield major starbursts in which the disk is only moderately perturbed (e.g., M 82). If the post-burst systems have sufficient gas supplies to allow star formation to continue, albeit at lower rates than in the current bursts, then they can appear at the present epoch as comparatively normal star-forming disk galaxies (Phillips et al. 1997; Hammer et al. 2000).

Related to these evolutionary questions is the issue of bulge formation. Theoretical work has shown that gas-rich (>$10\%$ in H I) disks subject to gravitational instability may form large clumps of gas, up to $10^{9}~M_{\odot}$, which rotate in the plane of the disk (Elmegreen et al. 1993; Noguchi 1999). Such clumps experience strong star formation, resulting in a morphologically peculiar galaxy, and could appear like the "clumpy'' galaxy investigated here.

The clumps suffer dynamical friction from the surrounding visible and dark matter, leading them to spiral inwards and accumulate in the central region, potentially forming a bulge. In this scenario the disk forms first, hosts the initial round of star formation, and produces a bulge. The remaining gas could then fuel a more sedate course of evolution. However, a key factor in this model is whether the clumps can retain their identities over the $\sim$100 Myr time scales required for dynamical friction to act (Noguchi 1999).

To clarify the general issues of how starbursts connect to star formation processes and galaxy evolution, we concentrate on understanding nearby starbursts whose internal properties are observable. With this goal in mind, the Wide Field Planetary Camera 2 (WFPC2) Investigation Definition Team GTO program included exploration of small scale structures in blue starburst galaxies with M_B < -18 and high optical surface brightness, many of which are members of the LBCG class (Gallagher et al. 2000b). NGC 7673 (a.k.a. IV Zw 149, Markarian 325, UGC 12607, and UCM 2325+2318), a "disturbed spiral'' or "clumpy irregular'' galaxy was chosen for a multi-wavelength imaging investigation extending from the mid-ultraviolet to the near infrared. NGC 7673 is a member of the LBCG class, and a luminous infrared source (e.g., Gallagher et al. 1989; Hunter et al. 1989; Sanders & Mirabel 1996). As such, it presents a unique opportunity to understand the inner workings of a high SFR object and its potential evolutionary connections to distant and present-day galaxy populations.

This paper presents an analysis of star clusters in the clumpy starburst regions of NGC 7673 based on archival images obtained with WFPC2 on the Hubble Space Telescope. We assume H₀ = 70 km s^-1 Mpc^-1; the recession velocity of 3408 km s^-1 from the NASA/IPAC Extragalactic Database for NGC 7673 then implies a distance of 49 Mpc, and a projected scale of 250 pc per arcsec. The distance modulus for this galaxy is 33.45 and one WFC pixel covers a projected scale of approximately 25 pc.

The next section summarizes properties of this unusual galaxy. Our observations and photometry of star clusters are detailed in Sect. 3, and Sect. 4 provides information on our choice of models for the spectral evolution of the clusters. In Sect. 5 we discuss the morphology of NGC 7673 as seen with WFPC2, Sect. 6 discusses the star cluster colors and magnitudes in terms of their ages, and Sect. 7 briefly characterizes the star-forming clumps. Our results are discussed Sect. 8, and in Sect. 9 we present our conclusions.