Star clusters have hierarchical patterns in space and time, suggesting formation processes in the densest regions of a turbulent interstellar medium. Clusters also have hierarchical substructure when they are young, which makes them all look like the inner mixed parts of a pervasive stellar hierarchy. Young field stars share this distribution, presumably because some of them came from dissolved clusters and others formed in a dispersed fashion in the same gas. The fraction of star formation that ends up in clusters is apparently not constant, but may increase with interstellar pressure. Hierarchical structure explains why stars form in clusters and why many of these clusters are self-bound. It also explains the cluster mass function. Halo globular clusters share many properties of disk clusters, including what appears to be an upper cluster cutoff mass. However, halo globulars are self-enriched and often connected with dwarf galaxy streams. The mass function of halo globulars could have initially been like the power-law mass function of disk clusters, but the halo globulars have lost their low-mass members. The reasons for this loss are not understood. It could have happened slowly over time as a result of cluster evaporation, or it could have happened early after cluster formation as a result of gas loss. The latter model explains best the observation that the globular cluster mass function has no radial gradient in galaxies.

This paper provides an introduction to IAU Symposium 266 on star clusters as basic building blocks in space and time. We define clusters as bound systems and discriminate them from general stellar clusterings or groups and unbound associations. We give a few examples of young, embedded, compact clusters which may evolve into looser, open clusters after dynamical relaxation due to mass loss and secular relaxation processes. We ask how and where star clusters form (in normal and interacting galaxies) and provide statistics of open clusters in terms of cluster masses and ages in the solar neighborhood (where observational data are most complete). Finally, we list a number of basic questions for current and future star cluster research and discuss the prospects for cluster studies with the next generation infrared and submillimeter telescopes (Herschel; JWST, E–ELT; ALMA, NOEMA).

NGC 3603 is one of the most massive, compact young star clusters in the Milky Way. The cluster has an age of only about 1 Myr and is embedded in a giant molecular cloud with ongoing star formation. We have analyzed deep imaging data obtained with the Wide Field and Planetary Camera 2 aboard the Hubble Space Telescope. We have obtained two epochs separated by 10 years, from which we derived proper motions which we used to determine cluster membership. After the removal of field stars, the resulting color–magnitude diagram shows a main sequence in addition to another clear sequence of pre-main-sequence stars. The cluster shows pronounced mass segregation and appears to have a very short crossing timescale. Our photometric, astrometric and kinematic data help us to evaluate the dissolution timescale of NGC 3603 and whether the mass segregation is likely to be primordial or evolutionary.

I review the progress made in understanding the physics and modes of star cluster formation through the use of direct self-gravitating hydrodynamical simulations, including those that have recently been performed incorporating radiative transfer and magnetic fields.

The Tarantula Survey is an ambitious ESO Large Programme that has obtained multi-epoch spectroscopy of over 1000 massive stars in the 30 Doradus region in the Large Magellanic Cloud. Here, we introduce the scientific motivations of the survey and give an overview of the observational sample. Ultimately, quantitative analysis of every star, paying particular attention to the effects of rotational mixing and binarity, will be used to address fundamental questions in both stellar and cluster evolution.

Young, massive stars influence their ambient medium through winds and radiation. The outcome of this feedback depends on the number of massive stars in a star cluster and on the density of the ambient medium. This contribution is based on a comparison of observations to the results of numerical simulations. We discuss the gravitational fragmentation of feedback-driven shells expanding from young stellar clusters. The thin-shell approximation is compared to 3D hydrodynamical simulations with smoothed-particle hydrodynamics and adaptive-mesh refinement codes. We explore the influence of external pressure and propose a thick-shell dispersion relation, where the pressure of the external medium is included. The mass spectrum of the shell fragments is constructed and we speculate about the origin of the deficit of low-mass objects.

In spite of significant recent and ongoing research efforts, most of the early evolution and long-term fate of young massive star clusters remain clouded in uncertainties. Here, I discuss our understanding of the initial conditions of star cluster formation and the importance of initial substructure for the subsequent dynamical-evolution and mass-segregation timescales. I also assess our current understanding of the (initial) binary fraction in star clusters and the shape of the stellar initial mass function at the low-mass end in the low-metallicity environment of the Large Magellanic Cloud. Finally, I question the validity of our assumptions leading to dynamical cluster mass estimates. I conclude that it seems imperative that observers, modellers and theorists combine efforts and exchange ideas and data freely for the field to make a major leap forward.

The centers of most galaxies in the local Universe are occupied by compact, barely resolved sources. Based on their structural properties, position in the Fundamental Plane, and integrated spectra, these sources clearly have a stellar origin. They are therefore called ‘nuclear star clusters’ (NCs) or ‘stellar nuclei’. NCs are found in galaxies of all Hubble types, suggesting that their formation is intricately linked to galaxy evolution. Here, I review some recent studies of NCs, describe ideas for their formation and subsequent growth, and touch on their possible evolutionary connection with both supermassive black holes and globular clusters.

We present a short summary of several 2D hydrodynamic calculations that suggest that upon the collision of galaxies two physical mechanisms lead to the formation of proto-super star clusters. These are condensation, induced by radiative cooling, and implosion caused by the shocked intercloud medium. Even in the absence of gravity, these lead to storage and compression of the dense cloud component into massive and compact gravitationally unstable condensations. The resulting entities exhibit enhanced surface densities that are several hundred times higher than their initial values. These are here postulated as the cradles of very efficient and rapid star-formation episodes, able to withstand the negative feedback effects associated with star formation, while leading to the formation of massive and compact super star clusters.

Star clusters are often used as tracers of major star-formation events in external galaxies as they can be studied out to much greater distances than individual stars. It is vital to understand their evolution if they are used to derive, for example, the star-formation history of their host galaxy. More specifically, we want to know how cluster lifetimes depend on their environment and on structural properties such as mass and radius. This review presents a theoretical overview of the early evolution of star clusters and the consequent long-term survival chances. It is suggested that clusters forming with initial densities of ≳104 M⊙ pc−3 survive the gas expulsion, or ‘infant mortality,’ phase. At ~10Myr, they are bound and have densities of ~103±1 M⊙ pc−3. After this time, they are stable against expansion through stellar evolution, encounters with giant molecular clouds and will most likely survive for another Hubble time if they are located in a moderate tidal field. Clusters with lower initial densities (≲100 M⊙ pc−3) will disperse into the field within a few 10s of Myrs. Some discussion is given on how extragalactic star cluster populations, and especially their age distributions, can be used to gain insight into disruption.

We show that we can obtain a good fit to the present-day stellar-mass functions of a large sample of young and old Galactic clusters with a tapered Salpeter power-law distribution function with an exponential truncation of the form dN/dm ∝ mα [1 − exp(−m/mc)β]. The average value of the power-law index α is ~−2.2, very close to the Salpeter value of −2.3, while the characteristic mass, mc, is in the range 0.1–0.6M⊙ and does not seem to vary in any systematic way with the present cluster parameters such as metal abundance, total cluster mass or central concentration. However, the characteristic mass shows a remarkable correlation with the dynamical age of the cluster, namely mc/M⊙ ≃ 0.15 + 0.5 × t3/4dyn, where tdyn is the dynamical time, taken as the ratio of cluster age and dissolution time. The small scatter around this correlation is likely due to uncertainties on the estimated value of tdyn. We attribute the observed trend to the onset of mass segregation through two-body relaxation in a tidal environment, causing preferential loss of low-mass stars from the cluster and hence a drift of the characteristic mass towards higher values. If dynamical evolution is indeed at the origin of the observed trend, it seems plausible that globular clusters, now with mc ≃ 0.35M⊙, were born with a stellar mass function very similar to that measured today in the youngest Galactic clusters and with a value of mc around 0.15 M⊙. This is consistent with the absence of a turn-over in the mass function of the Galactic bulge down to the observational limit at ~0.2M⊙ and argues for the universality of the initial mass function of Population I and II stars.

Violent relaxation, the protocluster dynamical response to the expulsion of its residual star-forming gas, is a short albeit crucial episode in the evolution of star clusters and star cluster systems. Because it is heavily driven by cluster-formation and environmental conditions, it is a potentially highly rewarding phase in terms of probing star formation and galaxy evolution. In this contribution, I review how cluster-formation and environmental conditions affect the shape of the young cluster mass function and the relation between the present star-formation rate of galaxies and the mass of their young, most massive cluster.

We present the distribution of Galactic bulge globular clusters and a method based on simultaneous detection of field and cluster horizontal branches to derive the cluster distances. This method has the advantage of being independent of both reddening and the reddening law, RV = AV/E(B−V). The vast majority of clusters projected in the direction of the Galactic bulge are located on the near side of the Galactic Center. Deviations from the reddening law do not seem to be responsible for this peculiarity. We need to introduce a peculiar, steep dependence of the absolute horizontal-branch magnitude with metallicity in the metal-rich regime if we want to reproduce a symmetrical distribution. Instead, if the observed distribution is correct, we expect a rather large number of bulge globular clusters are still to be discovered.

I analyze the stochastic effects introduced by sampling the stellar initial mass function (SIMF) in the derivation of the individual masses and the cluster mass function (CMF) from broad-band visible/near-infrared unresolved photometry. The classical method of using unweighted U BV photometry to simultaneously establish ages and extinctions of stellar clusters is found to be unreliable for clusters older than ≈30 Myr, even for relatively large cluster masses. On the other hand, augmenting the filter set to include longer-wavelength filters and using weights for each filter increases the range of masses and ages that can be measured accurately with unresolved photometry. Nevertheless, a relatively large range of masses and ages is found to be dominated by SIMF sampling effects that render the observed masses useless, even when using U BV RI JHK photometry.

Almost 80 years have passed since Trumpler's analysis of the Galactic open cluster system laid one of the main foundations for understanding the nature and structure of the Milky Way. Since then, the open cluster system has been recognised as a key source of information for addressing a wide range of questions about the structure and evolution of our Galaxy. Over the last decade, surveys and individual observations from the ground and space have led to an explosion of astrometric, kinematic and multiwavelength photometric and spectroscopic open cluster data. In addition, a growing fraction of these data is often time-resolved. Together with increasing computing power and developments in classification techniques, the open cluster system reveals an increasingly clearer and more complete picture of our Galaxy. In this contribution, I review the observational properties of the Milky Way's open cluster system. I discuss what they can and cannot teach us now and in the near future about several topics such as the Galaxy's spiral structure and dynamics, chemical evolution, large-scale star formation, stellar populations and more.

The globular cluster system in M31 is an ideal laboratory for studying the formation and evolution of M31 as well as the globular clusters themselves. There have been numerous surveys and studies of the globular clusters in M31. However, only recently has the entire body of M31 been searched for globular clusters using wide-field CCD images by our group. A new era for the M31 globular cluster system has begun with the advent of wide-field CCD surveys of M31. We have discovered more than 100 new globular clusters in M31. Our catalog currently includes more than 500 globular clusters confirmed either based on spectra or HST images, many more than in the Milky Way. We present the structure, kinematics and chemical abundance of the M31 globular cluster system based on this large sample, and the implications for the formation and evolution of M31.

Starburst clusters in the Milky Way have the advantage that individual stars down to subsolar masses can be resolved. Thus far, field contamination along the line of sight towards the Galactic Centre and spiral arms was the limiting factor in deriving an unbiased census of the stellar population in Milky Way starbursts and, hence, the spatial extent and initial mass function in starburst clusters. As the next generation of telescopes with higher sensitivity and spatial resolution are being developed, these resolved clusters become increasingly important as templates for young, massive extragalactic systems, which will be resolved at the high-mass end of the stellar mass function. With the aim to obtain a uniform characterisation of starburst cluster properties in the Milky Way, we have initiated a proper-motion membership survey. This technique became feasible for clusters out to distances of 8 kpc with diffraction-limited imaging using adaptive optics from the ground and with Hubble Space Telescope from space.

Galactic globular cluster (GC) stars exhibit abundance patterns that are not shared by their field counterparts, namely the well-documented O–Na, C–N and Mg–Al anticorrelations. Recent observations provide compelling evidence that these abundance anomalies were already present in the intracluster gas from which the presently observed stars formed. The current explanation is that the gas was polluted very early in the history of the GC by material processed through H burning at high temperatures and then lost by stars more massive than the long-lived stars we still observe today. However the ‘polluters’ have not yet been unambiguously identified. Most studies have focused on asymptotic giant brach stars, but rotating massive stars present an interesting alternative. Here, we critically analyse the pros and cons of both potential stellar polluters. We discuss the constraints that the observational data provide on stellar nucleosynthesis and hydrodynamics, as well as on the formation and early evolution of very massive star clusters.

The primordial lithium abundance inferred from WMAP and standard Big Bang nucleosysnthesis is approximately three times higher than the plateau value measured in old metal-poor Population II stars, suggesting that these stars have undergone atmospheric Li depletion. To constrain the physics responsible for such depletion, we conducted a homogeneous analysis of a large sample of stars in the metal-poor globular cluster NGC 6397, covering all evolutionary phases from below the main-sequence turnoff to high up the red-giant branch (RGB). The dwarf, turnoff, and early subgiant stars form a thin abundance plateau, with a sharpe edge in the middle of the subgiant branch, where Li dilution caused by the inward extension of the convective envelope starts (the beginning of the so-called first dredge up). A second steep abundance drop is seen at the RGB bump, again highlighting the need for the onset of nonstandard mixing in this evolutionary phase. Moreover, by also measuring the sodium abundances of the targets, we have gained insight into the degree of pollution by early cluster self-enrichement, and may separate highly polluted, Li-poor and Na-rich stars from stars formed from pristine material. Our observational findings strictly limit both the extent of lithium surface depletion, which in turn constrains the efficiency of mixing below the outer convection zone, and the resulting spread in lithium abundance in metal-poor turn-off stars.

We review the properties of globular clusters (GCs) which make them useful for studying the Galactic halo, Galactic chemical evolution and the early stages of the formation of the Milky Way. We review the evidence that GCs have a chemical inventory similar to those of halo field stars. We discuss the abundance ratios for dSph galaxies and show that it is possible to have formed at least part of the Galactic-halo field stellar population by dissolving GCs and/or accreting dSph galaxies, but only if this occurred at an early stage in the formation of the Galaxy. We review the constraints on halo-formation timescales deduced from the low magnesium isotopic ratios in metal-poor halo field dwarfs, which indicate that asymptotic giant-branch (AGB) stars did not have time to contribute significantly, while M71 contains two populations, one without and also one with a substantial AGB contribution. We review the limited evidence for GCs with a second population showing additional contributions from Type II supernovae, currently confined to ω Cen, M54 and M22, all of which may have been the nuclei or central regions of accreted galaxies. We check our own data for additional similar GCs and find preliminary indications that NGC 2419, a massive GC far in the outer Galactic halo, may also belong to this group.