We present a measurement of the evolution of the stellar mass function (MF) of galaxies and the evolution of the total stellar mass density at 0 < z < 5, extending previous measurements to higher redshift and fainter magnitudes (and lower masses). We use deep multicolor data in the Fors Deep Field (FDF; I-selected reaching I AB ∼ 26.8) and the GOODS-S/CDFS region (K-selected reaching K AB ∼ 25.4) to estimate stellar masses based on fits to composite stellar population models for 5557 and 3367 sources, respectively. The MF of objects from the K-selected GOODS-S sample is very similar to that of the I-selected FDF down to the completeness limit of the GOODS-S sample. Near-IR selected surveys hence detect the more massive objects of the same principal population as do I-selected surveys. We find that the most massive galaxies harbor the oldest stellar populations at all redshifts. At low z, our MF follows the local MF very well, extending the local MF down by a decade to 10 8 M ⊙ . Furthermore, the faint end slope is consistent with the local value of α ∼ 1.1 at least up to z ∼ 1.5. Our MF also agrees very well with the MUNICS and K20 results at z 2. The MF seems to evolve in a regular way at least up to z ∼ 2 with the normalization decreasing by 50% to z = 1 and by 70% to z = 2. Objects with M > 10 10 M ⊙ which are the likely progenitors of todays L > L * galaxies are found in much smaller numbers above z ∼ 2. However, we note that massive galaxies with M > 10 11 M ⊙ are present even to the largest redshift we probe. Beyond z ∼ 2 the evolution of the mass function becomes more rapid. We find that the total stellar mass density at z = 1 is 50% of the local value. At z = 2, 25% of the local mass density is assembled, and at z = 3 and z = 5 we find that at least 15% and 5% of the mass in stars is in place, respectively. The number density of galaxies with M > 10 11 M ⊙ evolves very similarly to the evolution at lower masses. It decreases by 0.4 dex to z ∼ 1, by 0.6 dex to z ∼ 2, and by 1 dex to z ∼ 4.

Abstract. We use the very deep and homogeneous I-band selected dataset of the FORS Deep Field (FDF) to trace the evolution of the luminosity function over the redshift range 0.5 < z < 5.0. We show that the FDF I-band selection down to I AB = 26.8 misses of the order of 10% of the galaxies that would be detected in a K-band selected survey with magnitude limit K AB = 26.3 (like FIRES). Photometric redshifts for 5558 galaxies are estimated based on the photometry in 9 filters (U, B, Gunn g, R, I, SDSS z, J, K and a special filter centered at 834 nm). A comparison with 362 spectroscopic redshifts shows that the achieved accuracy of the photometric redshifts is ∆z/(z spec + 1) ≤ 0.03 with only ∼1% outliers. This allows us to derive luminosity functions with a reliability similar to spectroscopic surveys. In addition, the luminosity functions can be traced to objects of lower luminosity which generally are not accessible to spectroscopy. We investigate the evolution of the luminosity functions evaluated in the restframe UV (1500 Å and 2800 Å), u , B, and g bands. Comparison with results from the literature shows the reliability of the derived luminosity functions. Out to redshifts of z ∼ 2.5 the data are consistent with a slope of the luminosity function approximately constant with redshift, at a value of −1.07 ± 0.04 in the UV (1500 Å, 2800 Å) as well as u , and −1.25 ± 0.03 in the blue (g , B). We do not see evidence for a very steep slope (α ≤ −1.6) in the UV at z ∼ 3.0 and z ∼ 4.0 favoured by other authors. There may be a tendency for the faint-end slope to become shallower with increasing redshift but the effect is marginal. We find a brightening of M * and a decrease of φ * with redshift for all analyzed wavelengths. The effect is systematic and much stronger than what can be expected to be caused by cosmic variance seen in the FDF. The evolution of M * and φ * from z = 0 to z = 5 is well described by the simple approximations Mfor M * and φ * . The evolution is very pronounced at shorter wavelengths (a = −2.19, and b = −1.76 for 1500 Å rest wavelength) and decreases systematically with increasing wavelength, but is also clearly visible at the longest wavelength investigated here (a = −1.08, and b = −1.29 for g ). Finally we show a comparison with semi-analytical galaxy formation models.Key words. galaxies: luminosity function, mass function -galaxy: fundamental parameters -galaxies: high-redshiftgalaxies: distances and redshifts -galaxies: evolution

Abstract. Using the Very Large Telescope in Multi Object Spectroscopy mode, we have observed a sample of 113 field spiral galaxies in the FORS Deep Field (FDF) with redshifts in the range 0.1 < z < 1.0. The galaxies were selected based on apparent brightness (R < 23 m ) and encompass all late spectrophotometric types from Sa to Sdm/Im. Spatially resolved rotation curves have been extracted for 77 galaxies and fitted with synthetic velocity fields taking into account all observational effects from inclination and slit misalignment to seeing and slit width. We also compared different shapes for the intrinsic rotation curve. To obtain robust values of V max , our analysis is focused on galaxies with rotation curves that extend well into the region of constant rotation velocity at large radii. If the slope of the local Tully-Fisher relation (TFR) is held fixed, we find evidence for a mass-dependent luminosity evolution which is as large as up to ∆M B ≈ −2 m for the lowest-mass galaxies, but is small or even negligible for the highest-mass systems in our sample. In effect, the TFR slope is shallower at z ≈ 0.5 in comparison to the local sample. We argue for a mass-dependent evolution of the mass-to-light ratio. An additional population of blue, low-mass spirals does not seem a very appealing explanation. The flatter tilt we find for the distant TFR is in contradiction to the predictions of recent semi-analytic simulations.

We explore the buildup of stellar mass in galaxies over the wide redshift range by studying the 0.4 ! z ! 5.0 evolution of the specific star formation rate (SSFR), defined as the star formation rate per unit stellar mass, as a function of stellar mass and age. Our work is based on a combined sample of ∼9000 galaxies from the FORS Deep Field and the GOODS-S field, providing high statistical accuracy and relative insensitivity against cosmic variance. As at lower redshifts, we find that lower mass galaxies show higher SSFRs than higher mass galaxies, although highly obscured galaxies remain undetected in our sample. Furthermore, the highest mass galaxies contain the oldest stellar populations at all redshifts, in principle agreement with the existence of evolved, massive galaxies at . It is remarkable, however, that this trend continues to very high redshifts of . We also 1 ! z ! 3 z ∼ 4 show that with increasing redshift, the SSFR for massive galaxies increases by a factor of ∼10, reaching the era of their formation at and beyond. These findings can be interpreted as evidence for an early epoch of star z ∼ 2 formation in the most massive galaxies and for ongoing star formation activity in lower mass galaxies.

Abstract. We present a catalogue and atlas of low-resolution spectra of a well defined sample of 341 objects in the FORS Deep Field. All spectra were obtained with the FORS instruments at the ESO VLT with essentially the same spectroscopic set-up. The observed extragalactic objects cover the redshift range 0.1 to 5.0. 98 objects are starburst galaxies and QSOs at z > 2. Using this data set we investigated the evolution of the characteristic spectral properties of bright starburst galaxies and their mutual relations as a function of redshift. Significant evolutionary effects were found for redshifts 2 < z < 4. Most conspicuous are the increase of the average C IV absorption strength, of the dust reddening, and of the intrinsic UV luminosity, and the decrease of the average Lyα emission strength with decreasing redshift. In part the observed evolutionary effects can be attributed to an increase of the metallicity of the galaxies with cosmic age. Moreover, the increase of the total star-formation rates and the stronger obscuration of the starburst cores by dusty gas clouds suggest the occurrence of more massive starbursts at later cosmic epochs.

We present the B-band Tully-Fisher relation (TFR) of 60 late-type galaxies with redshifts 0.1 − 1. The galaxies were selected from the FORS Deep Field with a limiting magnitude of R = 23. Spatially resolved rotation curves were derived from spectra obtained with FORS2 at the VLT. High-mass galaxies with v max 150 km/s show little evolution, whereas the least massive systems in our sample are brighter by ∼ 1 − 2 mag compared to their local counterparts. For the entire distant sample, the TFR slope is flatter than for local field galaxies (−5.77 ± 0.45 versus −7.92 ± 0.18). Thus, we find evidence for evolution of the slope of the TFR with redshift on the 3 σ level. This is still true when we subdivide the sample into three redshift bins. We speculate that the flatter tilt of our sample is caused by the evolution of luminosities and an additional population of blue galaxies at z 0.2. The mass dependence of the TFR evolution also leads to variations for different galaxy types in magnitude-limited samples, suggesting that selection effects can account for the discrepant results of previous TFR studies on the luminosity evolution of late-type galaxies.

Abstract. The FORS Deep Field project is a multi-colour, multi-object spectroscopic investigation of a ∼7 × 7 region near the south galactic pole based mostly on observations carried out with the FORS instruments attached to the VLT telescopes. It includes the QSO Q 0103-260 (z = 3.36). The goal of this study is to improve our understanding of the formation and evolution of galaxies in the young Universe. In this paper the field selection, the photometric observations, and the data reduction are described. The source detection and photometry of objects in the FORS Deep Field is discussed in detail. A combined B and I selected UBgRI JK s photometric catalog of 8753 objects in the FDF is presented and its properties are briefly discussed. The formal 50% completeness limits for point sources, derived from the co-added images, are 25.64, 27.69, 26.86, 26.68, 26.37, 23.60 and 21.57 in U, B, g, R, I, J and Ks (Vega-system), respectively. A comparison of the number counts in the FORS Deep Field to those derived in other deep field surveys shows very good agreement.

We present the redshift evolution of the restframe galaxy luminosity function (LF) in the red r , i , and z bands, as derived from the FORS Deep Field (FDF), thus extending our earlier results to longer wavelengths. Using the deep and homogeneous I-band selected dataset of the FDF, we were able to follow the red LFs over the redshift range 0.5 < z < 3.5. The results are based on photometric redshifts for 5558 galaxies derived from the photometry in 9 filters and achieving an accuracy of ∆z/(z spec + 1) ≤ 0.03 with only ∼1% outliers. A comparison with results from the literature shows the reliability of the derived LFs. Because of the depth of the FDF, we can give relatively tight constraints on the faint-end slope α of the LF; the faint-end of the red LFs does not show a large redshift evolution and is compatible within 1σ to 2σ with a constant slope over the redshift range 0.5 < ∼ z < ∼ 2.0. Moreover, the slopes in r , i , and z are very similar to a best-fitting value of α = −1.33 ± 0.03 for the combined bands. There is a clear trend of α to steepen with increasing wavelength: α UV&u = −1.07 ± 0.04 → α g &B = −1.25 ± 0.03 → α r &i &z = −1.33 ± 0.03. We subdivided our galaxy sample into four SED types and determined the contribution of a typical SED type to the overall LF. We show that the wavelength dependence of the LF slope can be explained by the relative contribution of different SED-type LFs to the overall LF, as different SED types dominate the LF in the blue and red bands. Furthermore we also derived and analyzed the luminosity density evolution of the different SED types up to z ∼ 2. We investigated the evolution of M * and φ * by means of the redshift parametrization M * (z) = M * 0 + a ln (1 + z) and φ * (z) = φ * 0 (1 + z) b . Based on the FDF data, we found only a mild brightening of M * (a r ∼ −0.8, and a i ,z ∼ −0.4) and a decreasing φ * (b r ,i ,z ∼ −0.6) with increasing redshift. Therefore, from z ∼ 0.5 to z ∼ 3 the characteristic luminosity increases by ∼0.8, ∼0.4, and ∼0.4 mag in the r , i , and z bands, respectively. Simultaneously the characteristic density decreases by about 40% in all analyzed wavebands. A comparison of the LFs with semianalytical galaxy formation models by Kauffmann et al. (1999) shows a similar result to the blue bands: the semi-analytical models predict LFs that describe the data at low redshift very well, but show growing disagreement with increasing redshifts.Article published by EDP Sciences and available at http://www.edpsciences.org/aa or http://dx.