Current galaxy formation models assume that the large scale mass distribution in the early universe is driven by the gravity exerted by dark matter. The evolution of the dark matter distribution follows from conditions in the early Universe. Gas dynamics, shocks, and radiative heating and cooling all play fundamental roles in the emergence of the first stars and proto-galaxies. The first stars are thought to be massive (10 to 100 Solar masses), and hotter than their modern counterparts. Thus, the first stars are thought to create giant HII regions whose red-shifted hydrogen and helium emission lines should be readily observable by Darwin/TPF-I.

While NASA's JWST is expected to make the first detections of these objects, its angular resolution will be limited to the diffraction spot size of its 6.5 meter primary mirror, about 0.2". Darwin/TPF-I will resolve scales of order 10 to 100 pc at all redshifts. The IR response will enable the detection of rest-frame near-IR to visual wavelength emission at very high redshifts (z > 5). Thus, Darwin/TPF-I will provide the hundred-fold gain in resolution needed to resolve these primordial HII regions. Models suggest that the birth of the very first stars may inhibit further star formation until these primordial stars die, a few million years after their birth. High angular resolution follow-up of JWST-detected "First Light" objects by Darwin/TPF-I will test the current paradigm for the formation of the first stars. Are they truly isolated, single objects, or are they surrounded by young clusters of stars?

Soon after the formation of the very first stars, their supernovae will pollute the surrounding medium, causing the condensation of dust. Dust heated by starlight, and the HII regions surrounding the very first stars will be visible and resolvable by Darwin/TPF-I. Subsequent growth of primordial galaxies occurs by a combination of merging and in-fall of primordial gas.

The baryonic matter in young galaxies is expected to be dominated by gas. As the first generations of stars explode in supernova explosions, they will pollute their environments with metals. Dust and molecule formation will drive star formation to increasingly resemble star formation in the current epoch. Hot dust, giant HII regions, and warm molecular clouds are expected to emerge. Darwin/TPF-I will play a crucial role in mapping the distributions of stars, clusters, super-giants, post-main sequence stars, supernovae, and emerging black holes in the highest redshift galaxies being detected at sub-mm wavelengths with today's instruments (Smail et al. 1997, Hughes et al. 1998, Barger et al. 1998) or in the future by ALMA (Atacama Large Millimeter Array) and in the thermal IR by JWST.

Darwin/TPF-I I will be especially sensitive to forming super-star clusters, the suspected progenitors of today's globular cluster systems. While JWST may detect galaxies containing such clusters at high red-shifts, Darwin/TPF-I will be needed to determine their galactic locations, relationships to other galactic structures, and to characterize their global properties.

Current theory, modeling, and observations indicate that galaxies grow and evolve by merging. How do these processes impact global and local star formation and the formation, growth, and evolution of black holes? Darwin/TPF-I will obtain milli-arcsecond resolution observations that can be directly compared to models.

Scheduling flexibility will enable Darwin/TPF-I to respond to targets of opportunity and transient phenomena such as ultra-high redshift, possibly population III, supernovae, flaring activity in AGN or even currently unanticipated time-dependent phenomena. Darwin/TPF-I will provide milli-arsecond characterization of these phenomena and their immediate environments.

Galactic evolution will remain a central theme of astrophysics for decades to come. The investigation of large samples of distant galaxies will be crucial for such studies. The "Lyman break technique" has defined samples of more than 1000 galaxies between 2.5 < z < 5 (e.g. Steidel et al. 1999). Lyα and Hα emitting galaxies have also been found with deep imaging through narrow band filters (e.g. Venemans et al. 2002; Kurk et al. 2003) or by selection of very red J-K colors (Franx et al. 2003). Spectroscopic follow-up of SCUBA galaxies, radio galaxies (e.g. de Breuck et al. 2001) and X-ray emitters (e.g. Rosati et al. 2002) have yielded significant samples of z > 2 objects.

The Faint Infra Red Extra-galactic Survey (FIRES, Franx et al. 2000), a very deep infrared survey centered on the Hubble Deep Field South using the ISAAC instrument mounted on the VLT (Moorwood 1997) demonstrates that there will be plenty of targets to investigate with Darwin/TPF-I. With integration times of more than 33 hours for each of the infrared bands J, H and K, limiting AB magnitudes of 26.0, 24.9, and 24.5 respectively are reached (Labbé et al. 2003). . Recently, this field has been imaged at 3 to 8 µm with the Infrared Array Camera (IRAC) on the Spitzer Space Telescope with the aim of accurately determining stellar masses for distant red galaxies (Labbé et al. 2005). From these studies, we conclude that, for the brighter objects, it is possible to obtain good images with a signal-to-noise ratio of 50 within integration times of 25-50 hours using Darwin/TPF-I.

Darwin/TPF-I will resolve individual OB associations, massive star clusters, and their associated giant HII regions. By observing multiple fields, interferometric maps of entire galaxies can be obtained at selected redshifts. By carefully selecting targets of a specific type, the evolution of galaxy structures can be tracers as functions of redshift and environment. The evolution of metallicity with cosmic age (and redshift) can be mapped using the various molecular tracers, ices, PAH bands, and noble gas lines that fall into the pass-band of Darwin/TPF-I.