Star Clusters in Evolving Galaxies

Florent Renaud Department of Astronomy and Theoretical Physics, Lund Observatory, Box 43, SE-221 00 Lund, Sweden Department of Physics, University of Surrey, Guildford, GU2 7XH, UK

Abstract Their ubiquity and extreme densities make star clusters probes of prime importance of galaxy evolution. Old globular clusters keep imprints of the physical conditions of their assembly in the early Universe, and younger stellar objects, observationally resolved, tell us about the mechanisms at stake in their formation. Yet, we still do not understand 5 the diversity involved: why is star cluster formation limited to 10 M objects in the Milky Way, while some dwarf galaxies like NGC 1705 are able to produce clusters 10 times more massive? Why do dwarfs generally host a higher specific frequency of clusters than larger galaxies? How to connect the present-day, often resolved, stellar systems to the formation of globular clusters at high redshift? And how do these links depend on the galactic and cosmological environments of these clusters? In this review, I present recent advances on star cluster formation and evolution, in galactic and cosmological context. The emphasis is put on the theory, formation scenarios and the effects of the environment on the evolution of the global properties of clusters. A few open questions are identified.

I Introduction 2 5.5 Galactic centre ...... 13

1 Outline, framework and objectives 2 6 Young massive clusters in interacting galax- ies 14 2 A multi-scale and multi-physics topic 3 6.1 Nuclear starburst ...... 15 3 What is a star cluster? 3 6.2 Cloud-cloud collisions, shocks . . . . . 15 3.1 Bound and part of a galaxy? ...... 3 6.3 Tidal and turbulent compression . . . . 15 3.2 Small and compact? ...... 5 3.3 Dark matter-free? ...... 5 7 In gas-rich, clumpy galaxies 16 3.4 Relaxed? ...... 5 3.5 Hosting a simple population? ...... 6 8 Globular clusters: formation at high redshift 17 3.6 Is there a definition? ...... 6 8.1 Uncertainties on the ages, and bi- modalities ...... 17 4 Cluster mass function 6 8.2 Formation in dark matter halos . . . . . 18 8.3 Formation in wet mergers ...... 18 8.4 Multiphase collapse ...... 18 II Formation 8 8.5 In situ formation and accretion . . . . . 19 5 Resolved star formation, hints from local 9 Nuclear clusters and their connection with arXiv:1801.04278v1 [astro-ph.GA] 12 Jan 2018 galaxies 8 ultra-compact dwarf galaxies 19 5.1 Molecular clouds, star formation and 9.1 Two and a half formation scenarios . . . 20 feedback ...... 8 5.2 The role of the environment ...... 10 9.2 Ultra compact dwarf galaxies ...... 20 5.3 Arms and inter-arm regions ...... 12 5.4 Cloud-cloud collisions, tips of the bar . 13 III Evolution 21

Email address: [email protected] (Florent Renaud) 10 Collisional systems 21

Preprint submitted to New Astronomy Reviews January 16, 2018 11 Modelling dense stellar systems 22 fade. There is now a strong need to merge the two 11.1 A numerical challenge ...... 22 and build a more holistic picture of star clusters within 11.2 Accounting for the environment . . . . 24 their galactic and cosmological environments. How- ever, bridging the gap between previously distinct fields 12 Dynamical friction 24 of research is a demanding task. With this review, I aim to draw a picture of our current knowledge and under- 13 Adiabatic, secular tides 25 standing of this topic, by focussing on the multi-scale 13.1 Tidal radii and Jacobi surface ...... 25 and multi-physics aspects of the question and the cou- 13.2 Potential escapers and kinematics in the pling between the two sides of this coin. This fields outer regions ...... 26 being vast, technical aspects and details on the results 13.3 Stellar streams ...... 28 presented here are often skipped, and I strongly encour- 13.4 Evolution in a fixed, isolated galaxy . . 29 age the reader to consult the studies mentioned and the 13.4.1 Constant tides: circular orbits . 30 references therein. This contribution could not be more 13.4.2 Time-varying tides: the exam- than a mere starting point to explore a field too rich and ple of eccentric orbits ...... 31 diverse to make an exhaustive review of it. 13.4.3 In non-axisymmetric, barred The focus being on the link between the clusters and potentials ...... 32 their galactic and cosmological context, it does not ad- dress the topics of internal cluster physics and their stel- 14 Tidal shocks 32 lar populations in details, nor the state of the art on the 14.1 Disc crossing ...... 32 observational side of the subject. I encourage the reader 14.2 Passages of spiral arms ...... 33 to explore complementary reviews on these topics, in 14.3 High speed encounters with clouds . . . 33 particular in

15 Evolution in cosmological context 35 • Naab and Ostriker (2017): galaxy formation 15.1 Adiabatic growth ...... 35 15.2 Accretion of satellite galaxies ...... 35 • Hennebelle and Falgarone (2012): turbulence of 15.3 Interacting galaxies and major mergers . 36 the interstellar medium 15.4 Full cosmological context ...... 36 • Krumholz (2014): star formation • IV Open questions 37 Brodie and Strader (2006): formation of extra- galactic clusters, mainly from an observational per- 16 Are YMCs young globular clusters? 37 spective • Charbonnel (2016): multiple stellar populations 17 Zone of avoidance 39 • Portegies Zwart et al. (2010) and Adamo and Bas- 18 Maximum cluster mass and star formation tian (2015): young massive clusters rate 39 • Heggie and Hut (2003) and Vesperini (2010): in- 19 The halo mass - cluster mass relation 40 ternal dynamics of clusters

20 The role of globular clusters in reionization 42 • Aarseth (2003): N-body simulations of clusters.

21 And many more ... 43 The interested reader could also compare our current understanding of this field with the challenges identified more than a decade ago by Davies et al. (2006), and Part I even confront the timeline they proposed for progress in each domain to what actually happened... Introduction In this introduction, I set the stage for this review, list a few possible definitions of star clusters, and highlight 1. Outline, framework and objectives connections with dwarf galaxies. The second part ad- dresses our theoretical understanding of star cluster for- Over the last years, the boundary between the fields mation in different galactic environments, in the Milky of galaxy formation and stellar populations started to Way, in interactions and mergers, in the local Universe, 2 at high redshift etc. The main focus of this part is the to propose holistic theories. Yet, significant progresses physical phenomena responsible for the accumulation on individual concepts have been made in the last few of star forming dense gas, and the first Myr of the re- decades, and coupling several aspects together is now sulting stars clusters, until they become gas-free. I also routinely achieved in theories, simulations and interpre- propose short incursions in the neighbour topics of nu- tation of observations. By focussing on the interactions clear clusters and ultra-compact dwarf galaxies. The between clusters and their galactic context, I will high- third part concerns the evolution of gas-free clusters, fo- light here some of these progresses and pin down a few cussing on the effects of the galactic and cosmological possible directions for future research. environment but leaving aside most of the cluster inter- nal physics. I address here the processes modifying the orbits of clusters, like dynamical friction, and the tidal 3. What is a star cluster? effects (adiabatic and impulsive) that alter their proper- ties. Finally, I identify in the last part a few open ques- This section gives the general tone of this review: tions on these topics, the state of the art and why they asking a simple question and showing that the complex- still keep us puzzled. ity of the problem still causes a (frustrating but exciting) lack of definite answer. 2. A multi-scale and multi-physics topic Until about a decade ago, star clusters in general and globulars in particular were considered as simple Many introductions of papers and proposals about (and sometimes bland) objects, made of gravitationally star clusters would tell you they are the building blocks bound stars sharing a common origin. However, the im- of galaxies, that most if not all stars form in clusters provements in observational resources and techniques and that globular clusters have witnessed all steps in the allowing us to deeper probe these objects and study the evolution of their host. Yet, the formation and evolu- composition of their stars, and this in an always widen- tion of star clusters remains a poorly known topic. The ing range of environments, have revealed a much more reason for this situation probably lies in the multi-scale complex picture. Up to the point that, today, there is and multi-physics nature of the field. Star clusters form no clear definition of what a star cluster is. More than in dense gas clouds, of which assembly is triggered, just semantics (or our urge to sort the Universe), the set and regulated by galactic-scale hydrodynamics, it- underlying question is to establish whether star clusters self influenced by the inter-galactic and cosmological and galaxies follow comparable physical formation pro- environment. Once formed, stars in clusters alter the cesses, and to which extend parallels can be drawn be- morphology, energy and chemistry of their nurseries tween theses classes of objects. through stellar feedback. Details on the injection of feedback are set by stellar evolution and depends on the clustering of star formation, the presence of binary and 3.1. Bound and part of a galaxy? multiple stars, the metallicity etc. These effects propa- gate to larger scales, up to e.g. driving galactic outflows As it is often the case in astrophysics, the community and enriching the inter-galactic medium. The details de- attempted to define this concept in contrast with other pend on the porosity of the interstellar medium (ISM), objects. For instance, star clusters differ from stellar as- the cooling rate, the turbulence and other processes sociations and moving groups by being gravitationally which are themselves largely set at galactic scales. In bound (see Zuckerman and Song, 2004, and references parallel, the internal physics of clusters (regulated by therein). Gieles and Portegies Zwart (2011) further pro- stellar evolution and star-star interactions) rules the evo- posed that objects with crossing times longer than the lution of these stellar systems and plays an important age of their stars would likely be unbound and should role in the evolution of their mass and size, on top of thus be classified as associations. One can also attempt external factors like tides. to discriminate clusters and galaxies using a hierarchi- In the end, the formation and evolution of star clus- cal argument: one contains the other. In that case, some ters is the results of a complex interplay of dynamical dwarf galaxies like Aquarius and Tucana, which do not and hydrodynamical (and probably magnetic) processes host any cluster (Forbes, 2005), would not qualify as from intergalactic scales (∼ 10 Mpc, ∼ 10 Gyr), down galaxies. Furthermore, some of the so-called ultra-faint to stellar scales (∼ 10−3 pc, ∼ 1 day), all this happening dwarfs count fewer stars than massive globular clusters for most of the life of the Universe. This wide diver- and can even be fainter than a single star (Belokurov sity of scales and physical processes makes it difficult et al., 2007). 3 4

3 ) ] c p [

h r ( g

o 2 l

1

TDGs NCs 0 YMCs

0 5 10 15 20 25 MV

Figure 1: V-band absolute magnitude versus half-light radius of dynamically hot stellar objects. Contours indicate the number density distribution of the ∼ 13000 globular candidates in galaxies of the Virgo cluster from Jordan´ et al. (2009), that would overload the figure if plotted individually. We give rough indications of the loci of the different classes of objects: globular clusters (GCs), extended clusters (ECs, also known as faint fuzzy clusters), ultra-faint objects (UFOs), dwarf spheroidal galaxies (dSphs), dwarf elliptical galaxies (dEs), tidal dwarf galaxies (TDGs) nuclear clusters (NCs), ultra-compact dwarf galaxies (UCDs), Young massive clusters (YMCs), compact ellipticals (cEs), ellipticals (and lenticulars) and giant elliptical galaxies (Es), as well as a few examples that illustrate the difficulties of drawing boundaries between these classes. We distinguish some objects that are undoubtedly classified, because of their position in their host galaxy (TDGs, NCs) or the age of their stars (YMCs). Data points from Portegies Zwart et al. (2010), the SAGES database (Brodie et al. 2011), McConnachie (2012), Duc et al. (2014) Norris et al. (2014), Georgiev and Boker¨ (2014), Homma et al. (2016), Contenta et al. (2017) and from Karina Voggel (private communication, as an update of Voggel et al. 2016) and references therein.

4 3.2. Small and compact? the core remnant of a tidally stripped dwarf (Freeman, After the hierarchical proposition above, it is natural 1993; Bekki and Freeman, 2003). Ibata et al. (2013a) to try to discriminate star clusters and galaxies in term estimated that NGC 2419 could have a dark halo twice of size and luminosity. Fig. 1 compiles a large vari- more massive than its stellar component. This would ety of stellar systems in the magnitude - size plane. A then explain the flattening of the velocity dispersion few years ago, the dense clusters seemed to be separated profiles observed in the outskirts of some clusters (see from the galaxies, both on the bright side (GCs and NCs e.g. Scarpa et al., 2007, and Section 13.2). However, vs. cEs) and the faint one (GCs vs. dSphs). These gaps such clusters would have a different formation scenario between the classes disappeared, thanks to deeper ob- than less massive clusters, but possibly more similar to servations in a variety of environments. The “chrono- that of UCDs (see e.g. Mieske et al., 2012; Pfeffer and logical evolution” of this figure can be seen by looking Baumgardt, 2013; Renaud et al., 2015a, and the oppo- at previous versions (each from their own perspective) site conclusion in Caso et al. 2014). Here again, the in e.g. Mackey and van den Bergh (2005), Belokurov classification is not obvious. et al. (2007), Misgeld and Hilker (2011), Brodie et al. Furthermore, a high mass-to-light ratio can be inter- (2011), Hwang et al. (2011), Forbes et al. (2013), Nor- preted differently. For instance, the high ratio observed ris et al. (2014), Voggel et al. (2016) and Contenta et al. in UCDs (Has¸egan et al., 2005; Dabringhausen et al., (2017). However, there is still a clear minimum in the 2008; Mieske et al., 2008) can be explained by invoking size distribution of objects around 100 pc. Gilmore et al. an initial stellar mass function (IMF) with an excess of (2007) proposed this to be a transient zone, populated massive stars. In old UCDs (Evstigneeva et al., 2007), by objects out of equilibrium and being close to disso- the massive stars have transformed into the remnants of lution, but recently found UCDs challenge this interpre- their evolution, like black holes and neutron stars, thus tation at the bright end of the distribution (e.g. Voggel increasing the mass-to-light ratio (Dabringhausen et al., et al., 2016). 2009, 2012). This interpretation would imply that the Filling these gaps has made challenging (and even star formation in UCDs differs from that of solar neigh- impractical) to define boundaries between these classes bourhood, where the IMF appears to be invariant (Bas- in term of luminosity (or mass) and size (or density). tian et al. 2010, see also Leigh et al. 2012 on a dynami- For instance, in Fig. 1 the loci of nuclear clusters (which cal reconstruction of the IMF in globular clusters). It is can otherwise be unambiguously defined based on their still not clear whether the same could apply to massive central position in their host galaxies) overlap with that globular clusters. of ultra-compact dwarf galaxies, and the brightest end of the globulars. However, different formation scenar- 3.4. Relaxed? ios have been proposed for these classes of objects (see In their attempt to define galaxies, Forbes and Kroupa Part II). (2011) proposed to compare the relaxation time of ob- jects to the age of the Universe. The relaxation time 3.3. Dark matter-free? quantifies how long a stellar system takes to erase signs Recently discovered extended clusters and faint of a perturbation, through star-star encounters and ex- dSphs around nearby galaxies appear to make a contin- changes of energy toward equipartition (see Section 10 uous connection between the bulks of their respective for more details). For old systems (i.e. with an average 5 – 8 families. In this regime, the difference is often made stellar mass of 0.5 M ) in the mass range of 10 M , a on the mass-to-light ratio. When high, it is assumed relaxation time shorter than a Hubble time corresponds to depict the presence of dark matter that would only to a half-light radii of 1 – 6 pc. Most of the globular be found in galaxies. Following such a definition, tidal and nuclear clusters do indeed yield relaxation times dwarf galaxies, that originate from galactic disc mate- shorter than a Hubble time (which qualifies them as col- rial being tidally stripped during an interaction (Duc and lisional systems), while it is not the case for the more Mirabel, 1999) and that hence are free of dark matter, diffuse, collision-less, UCDs and cEs. The overlap of should be classified as clusters. Furthermore, Baum- these classes in Fig. 1 implies that adopting such a defi- gardt and Mieske (2008) argued that globular clusters, nition would require to update the classification of some if formed in dark matter-dominated environments and of these objects, but the majority would keep their his- evolving in weak tides, would retain a dark halo around torical type. If convenient from a semantic point of them (but not in their central parts, see also Bekki and view, this definition however sets a somewhat arbitrary Yong 2012). This could be the case of massive glob- boundary, and hence could not be used to distinguish ular clusters like Omega Cen which is suspected to be (or not) different physical processes at stake during the 5 4 formation and/or evolution of these objects. This un- All clusters more massive than 10 M seem to yield satisfactory situation pushed the community to propose a Na-O anti-correlation (Carretta et al., 2010), while it alternative definitions, based on a common composition is absent in less massive objects like open clusters (e.g. or a common origin of stars. Bragaglia et al., 2014), yet with the exception of NGC 6791 (Geisler et al., 2012). Therefore, the presence of 3.5. Hosting a simple population? multiple stellar populations has been proposed to define For long, globular clusters have been thought to con- the class of globular clusters. This would exclude a few sist of stars with a common origin, i.e. that formed at the objects like Palomar 3, 12, 14, Terzan 7, 8 or Rup 106 same place, same time and thus from the same material (Cohen et al., 2004; Sbordone et al., 2004; Koch et al., (e.g. Ashman and Zepf, 1998). On the contrary galax- 2009; C¸alıs¸kan et al., 2012; Villanova et al., 2013; Car- ies yield more complex star formation histories and thus retta et al., 2014), notably smaller than the average glob- encompass a variety of stellar populations. The com- ular in the Milky Way. plexity of the stellar populations would then set a clear Star clusters would then be almost coeval collec- boundary between the two classes. The situation is how- tions of stars, with comparable but not exactly iden- ever more complicated. tical chemistries and ages. However, massive objects Anti-correlations between the abundances of several like NCs and UCDs (Section 9) potentially assembled chemical elements (Na-O, C-N, but also Mg-Al to some by merging several clusters, either shortly after the for- extent, Pancino et al. 2017) seem to exist in all glob- mation in a common gaseous structure, or later else- ular clusters (see e.g. Cohen, 1978; Kraft et al., 1997; where in their galaxy, and could thus fulfill the above- Gratton et al., 2001; Carretta et al., 2009, and refer- mentioned criteria to define clusters. ences therein). Such variations, from stars to stars, results from precise reaction rates depending on the 3.6. Is there a definition? abundance of the catalysts species during the CNO The situation depicted above leads to the absence of a and NeNa cycles of hydrogen burning at high tempera- clear, unambiguous and practical definition of star clus- ture (Denisenkov and Denisenkova, 1990; Kraft, 1994). ters (and thus also of galaxies, see Forbes and Kroupa Therefore, theses abundances and the anti-correlations 2011). Even defining star cluster members as being have been set at the earliest stages of the clusters, by gravitationally bound can be a dangerous shortcut. Stars the now long-gone massive (and thus hot) stars, and with sufficient energy to potentially escape their cluster are therefore expected to leave a unique fingerprint of can be trapped for a significant amount of time and ap- each cluster. Anti-correlations are indeed used to esti- pear as cluster members (see Section 13.2). In this re- mate cluster memberships in the always deeper surveys view, I cowardly adopt flexible definitions by focusing where contamination of background stars must be cor- on the most commonly agreed star clusters and men- rected for (Mesz´ aros´ et al., 2015). However, these varia- tioning, here and there, a few more debated cases. tions imply a sequential star formation, with a first gen- eration of stars producing the chemical elements (with the above mentioned spreads and others) used to form 4. Cluster mass function a second population, and even more. This translates into multiple main sequences and subgiant branches vis- Probing the exact nature of the stellar systems formed ible in colour-magnitude diagrams (Piotto et al., 2007; in a diversity of galactic and cosmological environments Milone et al., 2008). is made difficult by limitations on both observational To date, the field of multiple stellar populations has and numerical resolutions. For instance, depending on raised more questions than it has provided answers. the physical conditions in their gas nurseries, star clus- While several formation scenarios have been proposed, ters could either form monotonically, or through the our lack of detailed understanding of stellar evolution, merger of sub-systems, parts of larger molecular com- star cluster formation and feedback has prevented us to plexes. The latter was suggested by Bonnell et al. reach a comprehensive theory that would account for (2003) and Fellhauer and Kroupa (2005) as a formation all the strong observational constraints (abundance ra- channel of the most massive objects (see also Fujii and tios and spreads, relative importance of the populations Portegies Zwart 2015, Smilgys and Bonnell 2017, and etc.). Discussing these scenarios and their respective is- an observational confirmation in 30 Doradus by Sabbi sues is out of the scope of the present contribution and et al. 2012). Simulations resolving these sub-structures the reader is referred to extensive reviews by Gratton account for both formation channels being active simul- et al. (2012) and Charbonnel (2016). taneously in different locations of interacting galaxies 6 (e.g. Renaud et al., 2015a). The young age of the YMCs mass function follows a Schechter function, i.e. a in the Antennae galaxies and comparable mergers im- power-law with an exponential cut-off at the high mass poses that the merging process should happen rapidly end: ! ( 10 Myr). dN M . ∝ M−β exp − . (1) At the heart of these questions lies the initial clus- dM Mc ter mass function (ICMF) and its evolution. The nature (shape and bounds) of the ICMF can be used as a hint This is comparable to the shape of the mass function of on the formation mechanism of clusters, but an impor- galaxies, derived from the Press and Schechter (1974) tant step must be taken to infer this from observational formalism. This function is parametrised by the slope data, since observations can only probe the present-day of the power-law β ≈ 2±0.3 (Zhang and Fall, 1999; Bik cluster mass function (CMF), i.e. an evolved version et al., 2003; Hunter et al., 2003; Lada and Lada, 2003; of the ICMF. Even young clusters experience physical Kennicutt and Evans, 2012), likely set by the organisa- mechanisms altering their masses. It is thus far from tion of the ISM regulated by turbulence (e.g. Fujii and obvious to infer similarities and/or differences on the Portegies Zwart, 2015) and the efficiency of cluster for- formation process(es) of e.g. old globular clusters and mation, and the characteristic mass Mc of the exponen- YMCs by simply comparing their observed present-day tial cut-off of which physical origin is still debated (see mass functions, as described in the rest of this review. Section 18). However, due to low number statistics, the The present-day mass function of globular clusters in very presence of a truncation has not been unambigu- the Milky Way is described by a log-normal distribu- ously demonstrated (Adamo et al. 2017, see also several 5 tion, peaking at ≈ 2 × 10 M (Whitmore et al. 1999, examples in Portegies Zwart et al. 2010, their figure 10, see also Gnedin et al. 2014, i.e. at the magnitude MV ≈ Chandar et al. 2017). Demonstrating the presence of −7.5 in the corresponding cluster luminosity function, a physical truncation at high mass will possibly allow Harris 1991). Uncertainties arise when converting the to tell apart the formation mechanisms of clusters from observed luminosity function into the mass function, those of dwarf galaxies. i.e. when adopting a mass-to-light ratio derived from The lower limit in this function (in nearby galaxies) 2 single-age and mono-abundance models. These models goes down to ∼ 10 M , where the definition (and pos- assume a stellar initial mass function and often a simple sibly formation mechanisms) of clusters conflicts with stellar population, i.e. coeval stars in each cluster, de- that of associations (e.g. Piskunov et al., 2006; de Grijs spite a growing body of evidence for abundance spreads and Anders, 2006; Baumgardt et al., 2013; Fouesneau and multiple stellar populations, and a still open debate et al., 2014). Such a functional form provides a good on the single-age assumption (see e.g. Mackey et al., description of young clusters across a number of galax- 2008; Dib et al., 2013; Mucciarelli et al., 2014; Piatti ies, including mergers (see Portegies Zwart et al., 2010), and Cole, 2017). Additionally, for lower mass systems, with Mc (or the maximum mass detected, in case of the stochastic presence of massive stars may cause sys- non-truncated mass functions) being of the order of 5 tematic errors and biases in estimating the properties of ∼ 10 M for spiral galaxies (e.g. Larsen, 2009; Chan- the clusters, and care must be used (see a discussion dar et al., 2010), but significantly larger for starbursting 6 and method in Fouesneau and Lanc¸on, 2010). Other ef- galaxies (& 10 M , e.g. Bastian et al., 2008; Whitmore fects, like those originating from the sample size, are et al., 2010; Linden et al., 2017). discussed in Adamo and Bastian (2015). Simulations showed that stellar systems as massive Mandushev et al. (1991) showed that the mass-to- and extended as UCDs can form during galaxy interac- light ratio of clusters increases with mass. As a con- tions (Renaud et al., 2015a). These could, for instance, 7 sequence, and as predicted by models and simulations match the properties of W3 (8 × 10 M , 17 pc, Fell- (e.g. Ostriker et al., 1972; Baumgardt and Makino, hauer and Kroupa 2005; Maraston et al. 2004) detected 2003; Gieles, 2009) and detailed in Part III, it is very in the merger NGC 7252. In this case, these objects likely the result of several evolutionary processes most merely lie at the end of the mass and size distributions efficient at damaging the low-mass clusters. As a con- of star clusters, as noted observationally (recall Fig. 1, sequence, the low-mass end of the CMF is very dif- and see also Mieske et al. 2012), and share formation ferent from that of the ICMF, while the high-mass end mechanisms with young massive clusters. The galactic (made of more robust clusters) has only be mildly af- interactions would then represent the necessary trigger fected (Boutloukos and Lamers, 2003; Gieles, 2009). to reach the required extreme physical conditions, but When considering young systems however (i.e. get- other formation channels of UCDs have been proposed ting closer to birth epoch and thus to the ICMF), the (see Section 9.2). Here again, the absence of a clear, 7 physically motivated definition of star cluster introduces radiate energy and produce chemical elements. In this potential biases, in particular in the high-mass end of the part, I focus on this first phase of cluster evolution, from CMF. The same question arises about the (dark matter the stage of increasing the gas density, to that when gas free) tidal dwarf galaxies and lower mass clusters de- and stars becomes decoupled. tected as beads on a string along tidal tails of interacting One key question in this field is the formation of the systems (e.g. de Grijs et al., 2003; Mullan et al., 2011; globular clusters in the early Universe. However, the Knierman et al., 2003). Among other scenarios, the for- lack of direct observational constraints makes this a dif- mation of such objects would be favoured by the accu- ficult topic. Our current approach is thus to study it in- mulation of gas near the tip of the tails, in which gravita- directly, by exploring cluster formation in the local Uni- tional collapse would then proceed in a similar manner verse, where is can be resolved, and extrapolating (and as for massive, ordinary clusters (Duc et al., 2004). often speculating) to higher redshift. The situation is more complex in dwarf galaxies, where (keeping in mind low-number statistics effects) 5. Resolved star formation, hints from local galaxies only a couple of massive clusters dominate the (I)CMF (e.g. Anders et al., 2004; Larsen et al., 2012; Pasquali 5.1. Molecular clouds, star formation and feedback et al., 2011). This unexpected specific frequency (i.e. Each of these topics easily deserves its own review. the number of massive cluster per galactic stellar mass, A lot of progresses have been made over the last years, Georgiev et al. 2010; Harris et al. 2013; Larsen et al. thanks to increasingly powerful resources like the Her- 2012, 2014) is yet to be explained, possibly through shel Space Telescope, ALMA, MUSE, NOEMA, and formation mechanisms leading to a non-Schechter, top- supercomputers capable of handling the large amount heavy ICMF in such environments. of data on these interconnected subjects and running al- Making a connection between the Schechter-like ways improving models. I present here only a tiny frac- ICMF and the peaked present-day CMF requires a de- tion of our current understanding of these aspects and scription of the mass-loss and dissolution processes ex- their interplay. perienced by clusters along their entire evolutions (i.e. The first step of star formation consists in gathering possibly over ∼ 10 – 12 Gyr), and the dependence of gas at high enough densities to start the collapse pro- these mechanism on mass. Jordan´ et al. (2007) pro- cess that would eventually lead to the formation of pre- posed an evolved version of the Schechter ICMF (equa- stellar cores, and then proto-stars. Leaving aside the tion (1)), replacing the initial mass M with a M + ∆M questions of chemical composition and internal proper- term, but keeping the same functional form and param- ties of clusters, I skip here considerations on the initial eters. This thus assumes that the mass-loss ∆M is inde- mass function, binary stars and many others, to focus on pendent of the mass, contrarily to predictions by simu- the assembly of the star forming sites themselves. lations of a number of disruption mechanisms and series Observational surveys in the Milky Way show that of their coupling (see Part III). the dense, cold gas (e.g. traced by CO) is organised in ∼ 6 The mere exploration of the properties of the ICMF clumps with typical masses and radii of 10 M and ∼ − and its evolution reveals the need for an holistic descrip- 10 30 pc (Solomon et al., 1987; Dame et al., 2001). tion of the physics governing both the formation and the These giant molecular clouds (GMCs), embedded in en- evolution of star clusters, and equality importantly, how velops of atomic gas (with comparable mass, see e.g. Blitz et al. 1990), host dense enough gas to poten- these mechanism depends on the environment. Our cur- −3 rent understanding of these questions is presented in the tially form star clusters (& 100 – 1000 cm ). Measur- rest of this contribution. ing observationally the inner properties of these clouds (and a fortiori the sites of individual star formation) is limited to the Milky Way and nearby galaxies (LMC, SMC, M 31, M 33, etc., see e.g. Rosolowsky 2007; Part II Rosolowsky et al. 2007, Miville-Deschenesˆ et al. 2017 with estimates of cloud properties over the entire Milky Formation Way disc, and Hughes et al. 2013 in the more distant M 51), all representing relative quiescent environments, −1 The first effect of the cosmological and galactic envi- with low star formation rates (∼ 1 M yr ). Compa- ronment on the history of star clusters is the assembly rable studies in more active galaxies, including mergers of their formation sites. When dense enough, gas ig- and starbursts like M 82, will only become possible in a nites thermo-nuclear reactions that actually make stars few years. 8 All stars seem to form in clusters (Lada and Lada, (2014) noted that rotation, specially in the outer parts, 2003), and even though all these clusters are not neces- of old globulars could results from the interplay of inter- sarily bound and could be dynamically destroyed within nal dynamics (relaxation) and external effects (tides, see a few Myr (Bressert et al., 2010), the important point Part III), whereas Gavagnin et al. (2016) invoke mergers here is that star nurseries host the formation of more of clusters to induce rotation. Rotation of clusters, po- than one star, in turbulent molecular clouds (Hennebelle tentially set at birth, is thought to accelerate the pace of and Falgarone, 2012). It must be noted that no star their evolution (Einsel and Spurzem, 1999; Hong et al., formation mechanism actually requires the gas to be 2013) and modify their velocity dispersion profiles, thus molecular to form stars. Yet, after the cooling from free- potentially biasing the interpretation of observational 6 free emission in charged plasma (& 10 K, i.e. impor- data (Varri and Bertin, 2012; Bianchini et al., 2013). tant to cool gas in galactic halos), the collisional recom- Once formed, stars alter their surrounding through bination (i.e. the inverse of the ionization process) and proto-stellar outflows, photo-ionisation, radiation, the atomic de-excitation (∼ 104 – 5 K, Baugh 2006), the winds, the emission of cosmic rays, supernovae blasts scarcity of free particles in atomic gas to exchange en- and chemical enrichment (see, among many others, ergy with (and thus radiate energy away and cool down) Kessel-Deynet and Burkert 2003; Joung and Mac Low makes atomic cooling rather inefficient below ∼ 104 K.1 2006; Nakamura and Li 2007; Krumholz and Thompson Molecular gas (and dust) however yields more numer- 2012; Dale et al. 2012; Offner et al. 2009; Grisdale et al. ous and lower energy levels that allow efficient cool- 2017 and a review in Dale 2015). Because of the wealth ing below 104 K (e.g. the CO rotational line emission), of mechanisms and the non-linearity of their coupling thus reducing the internal pressure of clouds and allow- (e.g. Hopkins et al., 2012; Agertz and Kravtsov, 2015; ing self-gravity to increase the density enough to trigger Gavagnin et al., 2017), the exact role of stellar feed- fragmentation and collapse into cores (see e.g. Glover back is not fully understood yet. We know it can in- and Clark, 2016). clude stopping nearby star formation, destroying clouds, Star formation then only proceeds in the densest cores ionising the ISM, injecting turbulence, launching galac- (McKee and Ostriker, 2007; Andre´ et al., 2014), mean- tic outflows and even quenching star formation activity ing that a large fraction of the gaseous surrounding at galactic scale (Murray et al., 2010; Krumholz, 2014; medium is not converted into stars. This is quantified Hopkins et al., 2014; Dale, 2015; Grisdale et al., 2017; by the star formation efficiency (SFE), i.e. the mass ra- Semenov et al., 2017), but can also have a positive effect tio of stars formed and the initial gas mass. This ratio by compressing the gas (e.g. in shock fronts) and trig- obviously depends on the scale over which it is mea- gering sub-sequent star formation (Koenig et al., 2012; sured but for typical molecular clouds (∼ 10 – 100 pc) Shima et al., 2017). Furthermore, in dense environ- up to galactic scales, values ranging from 1 to 30 per- ments (e.g. gas-rich galaxies, mergers), feedback might cent are commonly accepted (Krumholz and Tan, 2007; not be powerful enough to clear all the gaseous left- Faucher-Giguere` et al., 2013). overs from the star forming sites (Rahner et al., 2017, Metallicity plays an important role in shaping the re- 2018; Howard et al., 2017, Guillard et al., in prepa- sulting star clusters, including the speeds of the stellar ration), possibly allowing for the retention and/or the winds (e.g. Goldman et al., 2017) and the escape ve- re-accretion of gas and, the formation of several popu- locity for ejecta (Georgiev et al., 2009), which possibly lations of stars (Pflamm-Altenburg and Kroupa, 2009). influences self-enrichment and the properties (or the ex- Such mechanism could, at least qualitatively, partici- istence) of multiple stellar populations. pate in the formation of multiple stellar populations de- The internal kinematics of the cloud may imprint ro- tected in globular clusters (thus formed at high redshift tation on the cluster, by conservation of the angular mo- in gas-rich, dense regions) and also in some young clus- mentum during collapse (Lee and Hennebelle, 2016), ters (Vinko´ et al., 2009; De Marchi et al., 2011; Beccari and/or through large scale tidal torques when massive et al., 2017). The other types of feedback, active be- clusters assemble hierarchically (Mapelli, 2017). Such fore the supernovae (i.e. before 10 Myr on average), can a rotation has been detected in young (Henault-Brunet´ however form hot, ionized cavities around the young et al., 2012), intermediate (Mackey et al., 2013) and stars (e.g. HII regions), in which the supernova blasts old clusters (e.g. Davies et al., 2011; Bianchini et al., expand more efficiently, thus increasing their net im- 2013; Lardo et al., 2015). However, Vesperini et al. pact on the ISM (e.g. Hopkins et al., 2012). Finally, the accretion onto black holes also represents a channel 1Note however the non-negligible cooling effect provided by the to remove the gas left-overs (Krause et al., 2012; Leigh fine-structure energy levels of e.g. carbon and oxygen. et al., 2013). 9 Star forming regions which can be observationally re- 2007; Bolatto et al., 2008; Heyer et al., 2009; Gratier solved are restricted to the quiescent environment of the et al., 2012; Donovan Meyer et al., 2013; Hughes et al., Milky Way and nearby galaxies where the gas density 2013; Faesi et al., 2016; Freeman et al., 2017). of clouds and their surroundings is low enough to allow On the other hand, although technical limitations feedback to efficiently remove gas from the star forming have restricted for long the theoretical studies of star 3 sites and preventing further accretion (but see e.g. Rah- formation to isolated, low mass clouds (. 10 M ), ner et al. 2017). When the gas is removed, the gravita- modern simulations now emphasize the importance of tional potential of the region rises rapidly, such that the galactic (hydro-)dynamics on the properties of clouds stars left alone, and which were in a relative dynami- and thus on the process of star (cluster) formation (Fu- cal equilibrium before, yield an excess of kinetic energy jimoto et al., 2014; Smilgys and Bonnell, 2017). For with respect to the new potential energy. Depending on instance, Rey-Raposo et al. (2015) showed that clouds the SFE (i.e. the relative importance of gas left-overs in experiencing shear or compression would host star for- the net gravitational potential), the re-adjustment of the mation at different rates than equivalent clouds in isola- stars to the new potential can lead to the dissolution of tion (with the same virial parameter). the cluster, a phenomenon known as the infant mortal- Observationally, Miville-Deschenesˆ et al. (2017) ity (Goodwin, 1997; Boily and Kroupa, 2003a,b; Smith measured a large range of masses and sizes of clouds 7 et al., 2013). Even if the cluster is self-bound enough to (between 10 and 10 M and 0.5 and 200 pc) over the survive, it likely loses a significant fraction of its mass Milky Way disc. The outer Galaxy (& 15 kpc) also (Bastian and Goodwin, 2006), such that after the aver- counts clouds significantly smaller and lighter than the age timescale for supernova feedback (≈ 10 Myr), only rest of the Galaxy (e.g. Izumi et al., 2014; Sun et al., ∼ 10% of all stars remain in bound stellar structures 2017), indicating an influence of the galactic environ- (Lada and Lada, 2003). Therefore, the SFE is suppos- ment on the properties of clouds. edly higher in regions producing bound clusters (Par- In this context, M 51 represents a prime target for mentier and Fritze, 2009). observations, being seen face-on and close enough to In such a picture, both the geometry and the timescale resolve individual clouds, showing a grand-design spi- of star formation (and feedback) are critical in setting ral pattern and thus offering a strong contrast in envi- the properties of the stellar objects formed, in term of ronmental physical conditions (see Fig. 2, Koda et al., binarity (Goodwin and Kroupa, 2005; Raghavan et al., 2009). Furthermore, it hosts a large number of clus- 2010; Kuruwita et al., 2017) and even the IMF (Chabrier ters (e.g. Bastian et al., 2005; Scheepmaker et al., 2009; et al., 2014), but also in term of the sub-structures en- Chandar et al., 2011). Schinnerer et al. (2013, see also compassed in a given cloud. Once the first star or stel- Meidt et al. 2013) particularly highlighted the role the lar aggregate formed, its feedback affects the surround- galactic potential and gas streaming motions in shaping ing media which can alter, quench or trigger the for- the molecular clouds of M 31. The effect of the spiral mation of other stars. The assembly of a cluster is arms and the differences with the inter-arm volumes is thus highly dependent on these aspects (Smilgys and discussed in the next section. Bonnell, 2017). Yet, the structuring of the ISM in a Simulations of isolated galaxies further emphasized hierarchical manner (Elmegreen and Falgarone, 1996; the role of the kpc-scale effects, through bar and spi- Burkhart et al., 2013) being first set by large-scale ral patterns and the global dynamics of the galaxy (e.g. mechanisms (& 100 pc, like flows, shear, tides and tur- Dobbs et al., 2012; Benincasa et al., 2013; Renaud et al., bulence), the details of star and star cluster (or stellar 2013; Fujimoto et al., 2014; Tasker et al., 2015; Ward association) formation are likely strongly dependent on et al., 2016; Nguyen et al., 2017). In addition to af- the galactic environment (e.g. Rey-Raposo et al., 2017). fecting the large scales flows of gas, the gathering of cloud material (e.g. Dobbs et al., 2011) or their destruc- 5.2. The role of the environment tion by shear (e.g. Weidner et al., 2010; Emsellem et al., In isolated galaxies, the exact role of the kpc-scale 2015, but see also Dib et al. 2012), galactic dynamics environment on star forming clouds and thus star clus- also influence the collision rate between clouds (see e.g. ter formation remains unclear. On the one hand, num- Renaud et al., 2015b; Fujimoto et al., 2016, and Sec- ber of observations reported that the average properties tion 5.4). and the scaling relations (e.g. mass-size, size-velocity When ignoring the environment, the break point dispersion, see Larson 1981; Solomon et al. 1987) of in the balance between internal pressure and self- molecular clouds are remarkably similar in a diversity of gravitation can be expressed by considering a density galaxies, ranging from dwarfs to spirals (Rosolowsky, perturbation propagating in a homogeneous, isolated 10 Figure 2: CO(1-0) line emission across the disc of M 51, seen by the PAWS survey (Schinnerer et al., 2013, their figure 1, c AAS, reproduced with permission). The morphology of the dense gas clouds varies significantly in between and along the spiral arms. Note for example the elongated feathers at the top-middle zone, in contrast with the regularly spaced beads on a string in more central areas. Not only the morphologies of the GMCs vary, but also their densities and their positions with respect to the spirals, which likely translates into different dynamical evolutions.

11 medium. This leads to the Jeans (1902) formalism, tually forming them and triggering star formation. Yet, providing the maximum scale-length and mass stable results from simulations emphasize that the gathering against collapse. Such a formalism neglects the large of gas in spirals tends to favour the formation of GMCs scale galactic effects, in particular, shear, tides and the (Dobbs et al., 2011), while the transition from the com- external pressure which significantly modify the stabil- pressive, pressurized and deep potential well of spiral ity criteria (Dobbs et al., 2011; Field et al., 2011). These to the inter-arm medium leads to an enhancement of points seem to (at least qualitatively) explain the ob- the destruction of clouds (Dobbs and Bonnell, 2006; served spreads and deviations from the classical scal- Roman-Duval et al., 2010). Koda et al. (2009) noted ing relations of molecular clouds in regions where en- that the giant molecular associations assembled in spi- vironmental effects play major roles (e.g. Leroy et al., rals would then split into smaller clouds when leaving 2015; Grisdale et al., 2018). To reconcile this diversity the arm (see also Hirota et al. 2011 on the mass of star of clouds with our search of a universal star formation forming clouds downstream the arms, and Schinnerer law, Meidt et al. (2013) suggested that accounting for et al. 2017 for a cautionary note on not resolving the in- the dynamical pressure on star forming regions would ner structures of giant molecular associations). Meidt unify the otherwise diverse star forming relations in dif- et al. (2013) and Colombo et al. (2014) showed that ferent environments. shear and shocks in the arms would yield a stabilizing Although we still ignore how and to what extend this effect on the gas structures, preventing star formation. diversity of star forming regions translates into a diver- sity of star clusters, it is worth noticing that Bastian et al. (2012, see also Konstantopoulos et al. 2013) de- tected variations a factor ∼ 3 of the truncation mass of The velocity pattern of spiral structures also triggers the ICMF (recall Section 4) with galactocentric radius the development of instabilities that give rise to a di- within a given galaxy (M 83 and NGC 4041). Varia- versity of features, like beads on a string, feathers and tions from galaxies to galaxies can be even stronger and spurs. Not only the spiral structure but also its kinemat- are discussed in Section 18. ics set the relative roles of gravity, shear and instabili- ties. For instance, in spirals with low pitch angles, the 5.3. Arms and inter-arm regions steep velocity gradient across the arm due to the differ- Observations resolving the inner structures of galax- ential rotation of the galactic disc could favour Kelvin- ies (spiral arms, bars) reveal the importance of the en- Helmholtz instabilities, in the form of spurs (see Fig. 2, vironment in setting the star formation activity. For in- Wada and Koda 2004; Kim and Ostriker 2006; Shetty stance, Schinnerer et al. (2013) showed the connection and Ostriker 2006; Dobbs and Bonnell 2006; Renaud between CO emission tracing dense molecular gas and et al. 2013, but also Kim et al. 2014; Sormani et al. 2015 the gravitational features of M 51. By concentrating for different interpretations). Spurs (and feathers when gas, spirals arms yield an excess of molecular cloud more elongated and populating the inter-arm regions) formation with respect to the surroundings, as com- are offset with respect to the bulk of the spiral and thus monly observed (Heyer and Terebey, 1998; Hou and outside of the local minimum of the gravitational po- Han, 2014) and modelled (Dobbs and Bonnell, 2006). tential of the spiral. This possibly protect them against This however does not imply that either the efficiency of dissolution as they do not experience rapid changes of cloud formation and that of star (cluster) are increased their external pressure and gravity, contrarily to clouds in arms. Foyle et al. (2010) reported indeed a very weak formed in the spiral and then leaving it. Conversely, dependence of the ratio between molecular and atomic beads on a string are found along the spiral (Elmegreen hydrogen when comparing arm and inter-arm regions in and Elmegreen, 1983; Foyle et al., 2013), and could spiral galaxies, concluding that the efficiency of cloud form there in the relative absence of Kelvin-Helmholtz formation is independent of spiral structures (see also instabilities, i.e. in arms with a higher pitch angle (Re- Eden et al., 2012, 2013). Moore et al. (2012) estimated naud et al., 2013). A given galaxy (observed, Schin- that 70% of the increase of the star formation rate (SFR) nerer et al. 2013 or simulated, Renaud et al. 2013) can density is arms is a mere reflection of the higher num- simultaneously host both types of clouds. The effects ber density of star forming regions found there, and that of this diversity on the formation (and early survival) of only the remaining 30% result from a more efficient for- clusters remain to be quantified, but it is likely that the mation (see also Schinnerer et al., 2017). variation of external pressure, gravitational energy, tidal From this, it follows that spirals are more important field and intrinsic rotation (as induced in spurs) would at re-organising the ISM and gathering clouds than ac- play a role in shaping the resulting stellar objects. 12 5.4. Cloud-cloud collisions, tips of the bar 5.5. Galactic centre The innermost ∼ 100 pc of the Milky Way host a Convergent flows in some environments like spiral significant fraction of the total molecular gas content arms and the tips of bars favour the accumulation of of the Galaxy (∼ 10%, see e.g. Ferriere` et al. 2007; gas structures (e.g. through orbital crowding, see Ken- Molinari et al. 2011). Classical stability criteria predict ney and Lord 1991), and thus cloud-cloud collisions. In that gas at such a high density should form a significant such events, shocks between the two components at dif- amount of stars. However, observations report a very ferent velocities boosts the local density, in addition to weak star formation activity despite the large quantity the natural gathering of gas mass. The resulting effect of (theoretically) dense enough molecular gas (e.g. Oka is a clear increase of the Mach number at cloud scale, et al., 2005; Ginsburg et al., 2015). Identifying young possibly converting previously sub- or transonic clouds stellar objects in this region of complex extinction re- into a supersonic medium (Renaud et al., 2015b). The quires spectroscopic confirmations, such that the exact collision thus triggers star formation (Loren, 1976; Tan, SFR remains debated. Yet, in the central molecular zone 2000; Tasker and Tan, 2009; Inoue and Fukui, 2013), (i.e. the inner ∼ 300 pc of the Milky Way contain- 7 – 8 possibly in the form of massive stars (Anathpindika, ing about 10 M of molecular gas), Morris (1989, 2010; Motte et al., 2014; Takahira et al., 2014). Fu- see also Longmore et al. 2013) estimated the SFRs of rukawa et al. (2009) and Fukui et al. (2014) further re- clouds to be about one order of magnitude lower than ported that YMC form in the volumes of interaction dur- what expected from the high gas surface density and the ing cloud-cloud collisions, probably because of the in- classical star formation laws established empirically in crease of external pressure (Elmegreen, 2008). One ex- other environments (e.g. Yusef-Zadeh et al., 2009; Lada ample of this can be found at the near tip of the Milky et al., 2012; Krumholz et al., 2012; Kauffmann et al., Way bar: the molecular complex W43 is thought to be 2017; Barnes et al., 2017). Kruijssen et al. (2014) thus the remnant of a recent (≈ 20 – 30 Myr) cloud merger interpreted this as a higher critical density for the on- 7 and hosts large quantities of gas (∼ 140 pc, ∼ 10 M ) set of star formation in the central molecular zone than leading to an active episode of star formation (Nguyen- in other regions of the galaxy. Emsellem et al. (2015) Luong et al., 2017). Due to the complex interactions showed that the physical origin of this peculiarity is set between gas structures, such giant molecular associa- by the galactic rotation curve being steep in this region, tions yield a number of sub-structures, and therefore inducing strong differential rotation (i.e. shear) on ex- star cluster formation does not proceed in a bulk, mono- tended structures like clouds. This shear would there- lithic manner, but rather hierarchically by merging sub- fore overcome self-gravity and smooth-out overdensi- clusters. This mode of assembly has been proposed ties, preventing gas structures to collapse and form stars, to explain the formation of massive stellar systems, in- despite their high densities. The tidal field also plays a cluding UCDs in other contexts (Fellhauer and Kroupa, destructive role in the central molecular zone, but likely 2005; Banerjee and Kroupa, 2015). As a zone of conver- of lower amplitude than the shear. gence, tips of the bar(s) would thus be a prime location The very centre of the Galaxy offers both the deep for the formation of massive stellar systems in isolated potential well and the symmetry to shield clouds and galaxies. clusters against destruction (with however the notable In this picture, the gas circulates along the bar on perturbation from the super massive black hole, Zhong x1 orbits (and is also fuelled inwards along the spiral et al. 2014; Arca-Sedda and Capuzzo-Dolcetta 2017). arms connected at the extremities of the bar). Thus, the Star formation could thus happen in the very centre crowding and collisions happening at the tips of the bar and participate in building the nuclear cluster (see Sec- and leading to cluster formation implies that the newly tion 9), but the above-mentioned destructive effects formed cluster roughly follows a comparable x1 orbit, would take over already within offsets of only a few and thus should be preferentially found on the leading ∼ 10 – 100 pc from this spot. In this context, it is some- sides of the bar (Renaud et al., 2015b). what surprising to find young massive clusters in the The relatively high degree of symmetry of the bar im- central molecular zone, like the Arches and the Quintu- 4 plies that comparable physical conditions should exist at plet. For instance, Arches (2 × 10 M , Clarkson et al. the other extremity. The presence of a YMC candidate 2012) hosts very young stars (≈ 2.5 Myr, Figer et al. has indeed been reported at the far tip of the Milky Way 2002; Najarro et al. 2004) which thus indicates an in situ bar by Davies et al. (2012), further confirming the role formation, and rules out formation in a more favourable of the kpc-scale (hydro-)dynamics in triggering the for- environment, followed by migration. Arches hosts a mation of massive clusters. top-heavy IMF (Espinoza et al., 2009; Habibi et al., 13 2013), which calls for specific, yet unknown, physical now would actually resemble present-day globulars (see conditions for star formation. The origins of these pe- Section 16). In any case, YMCs and in particular those culiarities and the reason for the very existence of these in interacting galaxies represent a prime laboratory to clusters in such a hostile environment remain unknown, explore the physics of massive cluster formation, which and possibly linked. is closely related to that of the bursts of star formation. In a comparable fashion to the orbital crowding seen Elmegreen (2008) proposed that star forming clouds at the tips of the bar (i.e. at the apocentres of x1 orbits), in mergers experience an excess of external pressure, it is possible that the intersection of the x1 and x2 or- which would keep confined otherwise unbound gas bits (i.e. the elongated orbits along the major and minor structures (see also Kruijssen, 2012; Maji et al., 2017). axes of the bar, respectively) could favour the accumu- Such a pressure originates from a number of causes, in- lation of gas (possibly via cloud-cloud collisions) at the cluding convergent flows (including cloud-cloud colli- apocentres of the x2 orbit, outside the central molec- sions) and tidal compression, as discussed below. ular zone (Stolte et al., 2008). Clouds could in prin- Note that mergers also play the important role of re- ciple form there, collapse, form massive clusters that distributing the material of their progenitors. If this can would then follow their x2 orbit bringing them back to- lead to star formation quenching (Martig et al., 2009), ward the galactic centre. Yet, the measured velocity of it can also eject pre-existing clusters into the halo, i.e. Arches seems incompatible with this scenario (Stolte in a low-density, weak tidal environment where the dis- et al., 2008). ruption mechanisms are less efficient than in the much The fact that the Milky Way is not currently hosting denser (and likely more central) regions of their forma- more than a handful of massive clouds and YMCs is tion. This would then considerably increase the survival a strong indication that the physical conditions lead- rate of clusters that would otherwise get dissolved by ing to the formation of even more massive clusters shocks and tides (Kravtsov and Gnedin, 2005, see also 4 – 5 (& 10 M ) must be drastically different than those Part III). found in our Galaxy. Clues can be obtained by explor- In interacting galaxies, starbursts are not restricted to ing more extreme, but less resolved, environments. the merger phase, i.e. the coalescence of the progen- itors, but can take place earlier in the interaction (e.g. 6. Young massive clusters in interacting galaxies Gao et al., 2003). The enhancement of the star forma- tion activity (and thus likely that of YMC formation) is The old age of the Milky Way globular clusters even detected in pairs of galaxies with a large separa- has long suggested that massive stellar systems could tion (∼ 10 – 100 kpc, Ellison et al., 2008; Scudder et al., only form in the early Universe, and that the required 2012; Patton et al., 2013). At the largest distances, this physical conditions are not matched anymore. Yet, probably results from a triggering during a closer pas- some galaxies in the local Universe, including the Mag- sage before the galaxy actually separate. ellanic Clouds (e.g. Glatt et al., 2008), host young Resolving the location(s) of star formation in inter- 4 – 5 (. 0.1 – 1 Gyr) and massive (& 10 M ) clusters acting systems is usually limited to dedicated observa- (YMCs). Interacting galaxies and mergers are partic- tional programs focussing on nearby pairs, potentially ularly good candidates to find such objects, because harbouring a large fraction of the diversity of the phys- of their extreme physical conditions being supposedly ical conditions leading to YMC formation (see Moreno closer to those at high redshift than those in quies- et al., 2015, and references therein), but a number of cent discs. Predicted by Schweizer (1987) and Ash- large surveys also address the question in a more sys- man and Zepf (1992), starbursting mergers of gas-rich tematic way (e.g. S4G, Sheth et al. 2010, SAMI, Croom galaxies do trigger the formation of (de facto, young) et al. 2012, CALIFA, Sanchez´ et al. 2014 and MaNGA, massive clusters. This has been confirmed in a number Bundy et al. 2015). of interacting systems (e.g. van den Bergh, 1971; Arp It happens that the closest ongoing major merger to and Sandage, 1985; Lutz, 1991; Holtzman et al., 1992; us encompasses a large variety of these environments: Whitmore et al., 1993; Holtzman et al., 1996; Zepf et al., the Antennae galaxies (NGC 4038/39), where the pres- 1999; Whitmore, 2003; Larsen, 2006; Trancho et al., ence of young massive clusters have been reported in 2012; Roche et al., 2015). Despite many similarities be- several areas of the system (Whitmore and Schweizer, tween globular clusters and YMCs (see e.g. Brodie and 1995; Mengel et al., 2008; Bastian et al., 2009; Herrera Strader 2006 and Section 8), it is yet not clear whether et al., 2011). This system is found to be representative YMCs are what globular clusters were ∼ 10 Gyr ago, for clusters found in larger sample of luminous infrared or whether present-day YMCs observed in 10 Gyr from galaxies (Linden et al., 2017). 14 6.1. Nuclear starburst (1992) suggested that, during penetrating interactions, The formation activity in interacting galaxies, espe- the shock between marginally stable clouds from one cially at a late stage of the merger is mostly found in the galaxy and the extended gas reservoir from the other, or central region of the remnant (Schweizer, 1982; Soifer even cloud-cloud collisions (recall Section 5.2) would et al., 1984). The reason for this is gas inflows from the trigger the necessary instabilities leading to the collapse strong gravitational torques one galaxy induces on the of the clouds and the formation of star clusters. This is other (Keel et al., 1985; Hernquist, 1989; Barnes and notably visible in the overlap region of the Antennae, Hernquist, 1991, 1996). Inside the co-rotation radius where the two galactic discs get imbricated. Number of (i.e. where the rotation velocity of one galaxy equals YMCs have been detected in this region, and their age ≈ the orbital velocity of its companion), the torques on ( 3 – 10 Myr, e.g. Bastian et al. 2009) further confirm- the disc material of a galaxy from its companion are ing the role of the recent galactic collision in triggering opposed to the rotation motion, which thus induce in- their formation (see also Herrera et al., 2011, 2012). In flows of material toward the galactic centre (Bournaud, such a region, the shocks increase the pressure on star 2010; Maji et al., 2017). The gas density rapidly in- forming clouds by several orders of magnitude, favour- creases, leading to the formation of a star cluster (see ing the formation of massive star clusters (Elmegreen Section 9). Such nuclear inflows also fuel the central and Efremov, 1997; Ashman and Zepf, 2001; Maji et al., super-massive black hole (Cox et al., 2006) and could 2017). However, overlap regions usually are rather ob- trigger a violent active galaxy nucleus (AGN) activity, scured by nature, and observational data (particularly possibly populating the radio-loud class of AGNs (Chi- kinematics) on the young clusters and the surrounding aberge et al., 2015). Then, feedback from the AGN reg- ISM are difficult to interpret. For instance, it is not clear ulates gas flows in the central region and slows down whether these objects keep signatures of their distinct the growth of the black hole itself, the host galaxy and progenitor clouds (in kinematics and/or chemical con- the central cluster (Di Matteo et al., 2005; Hopkins and tent) or not. Shocks and collisions thus play an impor- Quataert, 2010). tant in penetrating interactions, which preferentially oc- curs at the beginning of the final coalescence, and dur- The nuclear activity being triggered by gravitational ing fly-bys with small pericentre distances2. torques, it becomes important mainly during close en- counters, in prograde discs (i.e. where the disc angular 6.3. Tidal and turbulent compression momentum is roughly aligned with the orbital angular momentum of the companion galaxy, see Duc and Re- Finally, enhanced and off-nuclear star cluster for- naud 2013). During the interaction, the orbital pericen- mation is also detected in other regions of interacting tre distance of the galaxy pair decreases with time, due galaxies, usually before the coalescence. The spatial to dynamical friction, which implies that nuclear bursts distribution of such regions is rather complex, but some are predominantly triggered at the last passage before of them are usually found opposite side of the galaxy coalescence. This enhanced activity can however be with respect to the overlaps. This is the case for in- maintained for ∼ 10 – 100 Myr after coalescence (e.g. stance of the so-called Northern arc of the Antennae Di Matteo et al., 2007). (e.g. Bastian et al., 2009), or the so-called “Feature i” Note also that fly-by interactions can trigger the for- zone of NGC 2207 (Elmegreen et al., 2000; Kaufman mation of galactic bars (Noguchi, 1988) which, by cre- et al., 2012; Mineo et al., 2014; Elmegreen et al., 2016, ating their own torques, could play an important role 2017). These regions host dense and massive molecular in nuclear inflows. Therefore, non-antisymmetric struc- clouds and complexes, high SFRs, and YMCs compara- tures formed by a passage could have a comparable ef- ble to those found in overlaps. The absence of inflows fect as a close encounter on star cluster formation in the (Section 6.1) and shocks (Section 6.2) in such regions nuclear region. calls for another triggering mechanism for star forma- tion. 6.2. Cloud-cloud collisions, shocks The symmetry of the starbursting regions with respect to the galactic centre and along the axis connecting the The enhancement of star (cluster) formation is not confined to the galactic centres. Off-nuclear activity has been reported, also in the form of YMCs (Wang 2Note however that, in order to allow the galaxies to separate after et al., 2004; Hancock et al., 2009; Chien and Barnes, a fly-by, dynamical friction must be weak with respect to their relative pre-interaction velocities. Small pericentre passages complying with 2010; Smith et al., 2014; Elmegreen et al., 2016), espe- this requirement thus imply high velocities, which translates into a cially before the coalescence stage. Jog and Solomon short interaction phase and a short-lived star formation trigger. 15 two galaxies originates from the differential aspect of et al., 2016, in a different context). It is however chal- the tidal forces (Duc and Renaud 2013, the same way lenging to propose an observational diagnostic to iden- lunar tides on Earth have the same effect on the side tify compressive tides and turbulence, such that this the- facing the Moon and the opposite one). The combina- ory has not yet been tested observationally. However, no tion of the gravitational potentials of the two galaxies alternative exists to explain extended, off-nuclear, off- (including the contribution of the dark halos, i.e. even overlap starbursts. without overlap of visible matter3) induces the forma- tion of cores in the net potential (Renaud et al., 2008). 7. In gas-rich, clumpy galaxies In cores, i.e. the convex regions of the potential, the tidal forces are compressive in all directions (Valluri, Along their evolution, at least some disc galaxies 1993; Dekel et al., 2003; Renaud et al., 2008). This sit- experience a phase during which their morphologies uation is found in all interacting systems, with an inten- are dominated by giant clumps (Cowie et al., 1996; sity and duration that depend on the galaxies and their Elmegreen et al., 2007; Genzel et al., 2008; Tacconi orbits (Renaud et al., 2009). These fully compressive et al., 2013). Typically at z = 1 – 4, once the disc tides then induce an enhancement of the total turbu- is formed, a galaxy with a large gas fraction (& 30% lence of the ISM (as observed, Irwin 1994; Elmegreen of the baryonic mass) can host violent disc instabilities and Elmegreen 1995; Bournaud et al. 2011; Ueda et al. which, because of a high velocity dispersion compared 2012), but also a change of its nature by making it to the disc circular velocity, favour Jeans instabilities in compression-dominated (Renaud et al., 2014a). a handful of kpc-scale gas clumps (Agertz et al., 2009; This forcing drives turbulent motions out of the clas- Dekel et al., 2009; Bournaud and Elmegreen, 2009; sical equipartition found (on average) in isolated galax- Forster¨ Schreiber et al., 2009). These clumps are typ- 8 – 9 ies (Kritsuk et al., 2007; Federrath et al., 2010, see Hen- ically made of 10 M of gas and stars (Elmegreen nebelle and Falgarone 2012 for a review). The com- and Elmegreen, 2005) and are dense enough to host star pressive mode (curl-free) of turbulence takes over the formation (Zanella et al., 2015), possibly in the form of classical mixing, solenoidal one (divergence-free), and massive clusters. Note that the same process also oc- generates denser and more numerous gas over-densities curs in local dwarf irregulars and blue compact dwarfs than usual. Once it has cascaded down to parsec scale, which also yield a small ratio of circular velocity to dis- this effect enhances the SFR (Renaud et al., 2014a), in persion velocity (Elmegreen, 2015). Such galaxies thus particular in the form of massive clusters (Renaud et al., host a higher fraction of their star formation in the form 2015a; Li et al., 2017). Note that tidal and turbulent of massive, bound clusters. compressions also occur in the nuclei, the overlap re- Observations suggest that such clumps could encom- gions and, to a lower extend, along the tidal tails where pass several sub-clumps (with stellar masses of the or- 6 – 8 TDG-like objects form (Ploeckinger, 2015), and thus der of 10 M , Dessauges-Zavadsky et al. 2017), al- participate, to various degrees, to the starburst activity though it remains difficult to establish whether such in all concerned regions. small clumps share a common envelop in a giant clump Classical stability criteria like the Toomre Q parame- or not. Details on the cooling, fragmentation and galac- ter (Toomre 1964, used in the context of axisymmetric, tic tides set whether these small clumps are gravitation- mono-component discs, and Lin and Shu 1966, Bertin ally bound to each other or rather independent struc- and Romeo 1988, Elmegreen 1995, Rafikov 2001, and tures. In the former case, each clump could form a rela- Romeo and Falstad 2013 for multi-component discs) tively small star cluster that would contribute to a much and the Jeans formalism (Jeans 1902, for propagation larger object by merging with its neighbours, follow- of a perturbation in a homogeneous medium) have been ing the hierarchical scenario of Bonnell et al. (2003). adjusted to account for the additional effect of tides In this process, because all sub-clumps might not be (Jog, 2013, 2014), confirming that compressive tides dense enough to form stars (or because all the sub- favour the instabilities of gas structures (see also Inoue structures from the same giant clump might not form their stars simultaneously), young clusters could inter- 3The absence of stabursting regions between the nuclei and the act with nearby dense gas structures. Such impulsive zones of compressive tides, e.g. between the nucleus of NGC 4038 tidal perturbations could be sufficient to destroy the and the Northern-arc in the Antennae, rules out shocks with diffuse least massive clusters and damage the others (see Sec- HI reservoir that would extend the effect of shocks outside the overlap tion 14), as proposed by Elmegreen (2010). This topic of the dense discs. Only a balance between the (usually) destructive tides from the host galaxy, and the contribution of the other can form is sensitive to the details of the cooling, fragmentation the observed star forming regions and their peculiar morphologies. and turbulence, and is still open. 16 Because of their high gas densities, the giant clumps host galaxies if they form in situ (Strader et al., 2005), manage to survive the effects of stellar feedback (Bour- this point could be important to assess the link between naud et al., 2015)4 and can, through clump-clump in- globular clusters and the reionization of the Universe, teractions and dynamical friction, merge in the galac- which is actually a twofold question. On the one hand, tic centre and participate in the assembly of the bulge. could globular clusters have played a role in reioniza- Therefore, star clusters formed in the massive clumps tion? And on the other hand, how has reionization af- would not be detected as such at present-day, but instead fected the formation of globulars? On top of the uncer- would be dissolved (by tides) in the bulge. tainties on the intensity of the processes involved them- For such star clusters to survive over a long time- selves (see Section 20), the first topic to address remains scale, one must invoke a modification of their orbit the actual concomitance of the formation of globulars (Elmegreen, 2010). Galaxy interactions provide such (i.e. the presence of massive stars emitting a strong ion- an effect and could account for several aspects. First, isation flux) and the reionization. by scattering the stellar material above and below the Determining the absolute ages of clusters reveals it- original disc, the distribution of star clusters would re- self to be more complicated than estimating relative dif- semble the flatten spheroid, observed in the Milky Way ferences between cluster populations, yet done with sig- (e.g. Portegies Zwart et al., 2010). Second, by ejecting nificant uncertainties of ∼ 1 – 2 Gyr (corresponding to the clusters off the disc into a low density environment, variations in redshift between roughly z ≈ 4 and 10 at it would significantly reduce the tidal harassment they this epoch, Strader et al. 2005; Forbes et al. 2015). Yet, experience. However, while long-term, secular tidal ef- the features detected in age-metallicity distributions of fects would be largely decreased (see Section 13), in- globular clusters can be used to constrain their forma- clined orbits would necessarily lead to disc-crossings tion scenarios, and the number of formation channels and the associated tidal shocks that could severely dam- involved, as illustrated in the following Sections (e.g. age the clusters, and even destroy the least massive ones Muratov and Gnedin, 2010; Renaud et al., 2017). (see Section 14). On top of the ages and extreme stellar densities, the Clumpy galaxies could then play an important role in main observational constrain on the formation of globu- forming and altering the star cluster populations of disc lar clusters is the bi-modal colour distribution observed galaxies, but their exact effects remain to be established. in some (but not all) massive galaxies (Zinn, 1985; Geb- hardt and Kissler-Patig, 1999; Larsen et al., 2001; Peng et al., 2006), which is often translated into a metallic- 8. Globular clusters: formation at high redshift ity bi-modality (but see Yoon et al., 2006, on the non- linearity of the colour-metallicity relations). 8.1. Uncertainties on the ages, and bi-modalities In the Milky Way, the blue clusters (metal poor, with Precise age determination is made difficult by the a distribution of [Fe/H] peaking at -1.5, and with no age-metallicity degeneracy (Worthey, 1994), but the cluster below [Fe/H] = −2.5, Harris 1996) are prefer- bulk of the Milky Way globular clusters is estimated to entially found in the halo. They do not yield any struc- have formed around z ≈ 3 – 6, i.e. ≈ 11.5 – 12.5 Gyr ago tured kinematics as a population, indicating that they (see e.g. Leaman et al., 2013; Forbes et al., 2015). In unlikely share a common origin. Conversely, red clus- addition to putting a lower limit on the age of the Uni- ters (metal rich, [Fe/H] peaking at -0.5) are associated verse (Krauss and Chaboyer, 2003) and the age of the with the Galactic disc with which they share the overall rotation motion. The bi-modality aspect of their colour and metallicity distributions (i.e. with a deficit of clus- 4 The opposite conclusion has been reached by Oklopciˇ c´ et al. ters between two peaks) suggests two distinct modes of (2017). However, their galaxies, due to a too early star formation (as it is often the case in cosmological simulations), yield a gas frac- formation. Forbes et al. (2015) estimated that the metal- tion lower than expected at the redshift examined (≈ 20 – 30% instead rich clusters are, on average, about 1 Gyr younger than of ≈ 50 – 60% of the baryonic mass at z ∼ 2, see Daddi et al. 2010a; their blue counterparts. Despite uncertainties, the bi- Tacconi et al. 2010). Therefore, the gas overdensities they study are modality is in place before z ≈ 2, possibly already at not massive clumps resulting from violent disc instabilities, but rather portions of spiral arms more susceptible to destruction by shear and z ≈ 3 – 4, i.e. 11 – 12 Gyr ago (Dotter et al., 2011; Lea- feedback. The differences between these two types of structures and man et al., 2013). their survivability depend on their nature and has been proven to be (in In our neighbour Andromeda, the colour distribution this specific case) independent of numerical methods (see Bournaud is rather uni-modal, yet spanning a comparable range et al., in preparation, who confirms that the low gas fraction “clumps” of Oklopciˇ c´ et al. 2017 are indeed short-lived, while giant clumps in (Caldwell and Romanowsky, 2016). Therefore, the for- higher gas fraction galaxies survive the effects of feedback). mation of globular clusters is strongly dependent on 17 galaxy evolution, such that galaxies with different as- poor, Universe as before. These clusters (and other sembly histories would yields different distributions of stars) would then chemically enrich their still gas-rich globulars. The solution of this puzzle is thus lying host galaxies, that eventually merge. Note that merg- in galaxy evolution. Several scenarios have been pro- ers are naturally more frequent at high redshift than in posed. the present-day Universe, because of its density going as (1 + z)3. During the merger of gas-rich galaxies, a 8.2. Formation in dark matter halos starburst activity is triggered (recall Section 6), and by analogy with observed local mergers like the Antennae, In their scenario, Peebles and Dicke (1968) evoke the massive clusters would thus form from the enriched ma- formation of globular clusters at very high redshift in terial, and populate the red group. It is however im- low metallicity medium, even before the formation of portant to keep in mind that the efficiency of mergers galaxies themselves. When the Universe has expanded to trigger starbursts (and thus, likely, the formation of enough such that primordial (metal-free) gas in over- globulars) decreases with redshift, as observed (Loft- densities reaches ∼ 4000 K and ∼ 104 cm−3, the Jeans house et al., 2017) and modelled (Perret et al., 2014; length and mass setting the lower limit for instabili- Fensch et al., 2017). ties are of the order of the typical sizes and masses of With such a scenario, the relative amplitude of the globular clusters (see also Fall and Rees, 1985; Bromm blue and red peaks in the colour distribution, and their and Clarke, 2002; Katz and Ricotti, 2014; Kimm et al., potential separation by a deficit of cluster at interme- 2016; Popa et al., 2016; Hirano et al., 2017; Penarrubia˜ diate colours can easily be regulated with the merger et al., 2017; Kim et al., 2018). epochs and their cluster formation efficiency. Yet, be- Variants of this idea have proposed to consider the cause there is a priori no particular reason for the merger triggering of star formation in otherwise stable halos by activity to pause at some epoch, one would expect that ionisation shock fronts (Cen, 2001), or by the merger of most galaxies yield a rather continuous, uni-modal dis- two or more small gas-rich halos (Trenti et al., 2015). tribution, making the Milky Way an outlier. The effi- Additionally, Kimm et al. (2016) found that enrichment ciency of cluster formation however could be affected from the first generation of supernovae could lead to the by reionization making the merger history a highly inad- subsequent formation of other clusters in the same mini equate tracer of the cluster formation history. McLaugh- halo, and that these structures would eventually merge lin et al. (1994) noted that the number of globular clus- in the centre, thus leading to a wide spread in metallicity ters produced in such events would however be too low within the resulting globular. compared to the census in massive elliptical galaxies The gathering of gas and its collapse would happen (which are formed through repeated mergers). preferentially in potential wells (Diemand et al., 2005; Kravtsov and Gnedin (2005) and Shapiro et al. (2010) Boley et al., 2009), likely containing dark matter (see proposed that globulars form in gas-rich (but not neces- Ricotti et al. 2016, but also Naoz and Narayan 2014). sary clumpy) disc at very high redshift (z 3). Galaxy Thus, the absence of dark matter in globular clusters & mergers would then gather the physical conditions to (Section 3.3, Conroy et al. 2011) requires to invoke in- enhance the overall formation, and more specifically ternal dynamics (Baumgardt et al., 2008) and tides to the most massive globulars (Li and Gnedin, 2014; Li strip at least part of this component (Mashchenko and et al., 2017; Kim et al., 2018). In this case, the varia- Sills, 2005). Determining precisely the kinematics of tions in metallicity would be due to a spread in forma- the outer part of globulars, and inferring the presence of tion epochs. In their complementary scenario, Li and a dark component is thus key to validate this idea. Gnedin (2014) invoked late major mergers to reproduce Such scenarios do not account for the observed bi- the metallicity spread observed in the massive elliptical modalities, and implies a rather early formation channel galaxies of the Virgo cluster. Such a formation channel with respect to the (however uncertain) age estimates. however is tailored to violent environments involving Therefore, this calls for alternative, or at least comple- recent mergers, and not that of the Milky Way which has mentary, formation scenarios. not experienced such an event since z ≈ 2 (e.g. Ruchti et al., 2015). 8.3. Formation in wet mergers Schweizer (1987) and Ashman and Zepf (1992) pro- 8.4. Multiphase collapse posed a formation channel in two steps, thus accounting An alternative scenario is that of Forbes et al. (1997) for an age difference between the blue and red popula- in which, again, metal-poor clusters form during the tions. The blue clusters would form in the early, metal collapse of the proto-galaxies. This phase ends early in 18 the formation of galaxy, possibly with the reionization population remain uncertain, due to the difficulty of dis- of the Universe that quenches star formation (Beasley entangling the contribution of each satellite, and the et al., 2002). Another phase starts when the formation of large uncertainties on the number (and epoch) of accre- the galactic discs make the ISM dense enough to resume tion events the main galaxy has encountered (e.g. van the formation of globulars, this time from enriched ma- den Bergh, 2000; Mackey and Gilmore, 2004). terial and thus making the metal-rich population. The In such a scenario, there is potentially no age differ- second phase could happen in the early, gas-rich and ence between the two populations, but the details de- turbulent clumpy discs (recall Section 7). Such a mech- pend on the enrichment histories of all galaxy progeni- anism naturally accounts for the spatial distributions and tors, and when they form its clusters along these histo- the kinematics of the two populations, as well as for a ries, which likely varies with galactic mass. clear bi-modality in their metallicity distribution, as ob- The cosmological simulation of the Milky Way of served in the Milky Way, but not in e.g. Andromeda. Renaud et al. (2017) confirms this hypothesis, arguing Age estimates from Forbes et al. (2015) indicate that that the details on the bi-modality would mostly depend the blue population has a mean age of 12.5 Gyr, cor- on the mass of the progenitors of the Milky Way, i.e. responding to formation at z ≈ 6, i.e. after (or near their ability to form stars and star clusters, and retain the end of) reionization. In addition, the red population the enriched gas (instead of launching outflows faster yields a mean age of 11.5 Gyr, i.e. a formation at z ≈ 3, than the escape velocity of the galaxy). In that sense, which is earlier than the predictions of the onset of disc low mass progenitors, including the young Milky Way formation in cosmological simulations (z ∼ 1.5 – 2.5, itself at high redshift, would only contribute to the blue, despite difficulties in correctly distributing the angular metal-poor population, while the red clusters would momentum, likely due to the implementations of a too mostly form in situ, but also in galaxies experiencing strong feedback, see discussions in Agertz et al. 2011 late major mergers with the Milky Way. Therefore, the and Renaud et al. 2017). If the disc of the Milky Way epoch of the last major merger would set the bi-modality is indeed not in place at the necessary epoch to form the (or its absence). However, the lack of resolution of red clusters, another formation mechanism of this pop- present-day cosmological simulations forbids us to con- ulation must be proposed. clude on the mass of the clusters formed. (Such simu- lations form “stellar particles” that, while they trace the 8.5. In situ formation and accretion global SFR relatively well, do not predict whether the formation occurs in massive clusters or not.5) As a re- Finally, Cotˆ e´ et al. (1998) and Tonini (2013) pro- sult, the relative importance of the red and blue popula- posed that red clusters formed in situ (i.e. in the Milky tions derived in these simulations remain hypothetical. Way itself), while the blue population has been inher- In conclusion, the lack of observational constraints, ited from other galaxies accreted onto the Milky Way of understanding of the physics of cluster formation at via major and minor mergers (and, to a lower extend, low redshift and how to extrapolate this to the early Uni- from the capture of clusters during a fly-by), but without verse leave the field of globular cluster formation full of necessarily triggering the associated starburst episodes. open questions. Some hints will soon be provided by The actual formation mechanisms are then borrowed Gaia on the assembly of the stellar populations of the from the above-mentioned ideas, i.e. formation of the Milky Way, and by the James Webb Space Telescope metal-rich clusters in high-redshift discs (Kravtsov and on the formation of the clusters themselves at relatively Gnedin, 2005; Prieto and Gnedin, 2008; Shapiro et al., high redshifts. 2010), possibly enhanced during mergers (Li et al., 2017), and formation of the metal-poor cluster in dwarf 9. Nuclear clusters and their connection with ultra- galaxies (Elmegreen et al., 2012). Note that in situ for- compact dwarf galaxies mation at very high redshift would contribute to the metal-poor population (Brodie et al., 2014). About 75% of spiral and dwarf elliptical galax- The accretion of satellite galaxies adds to the com- ies host a nuclear cluster in their centre (Cotˆ e´ et al., plexity of the distribution of globulars. Each accreted galaxy having its own formation history, it brings its 5A workaround would be to use a subgrid model calibrated on low- own population(s) of globular clusters, which then redshift observations to assess the mass of clusters formed, but this smooth out potential bi- or multi-modalities of the main would then bias the interpretation toward a universal formation mech- anism, thus neglecting the variations of cluster formation with small galaxy itself (see an example in Forbes and Bridges scale physics (. 0.1 – 1 pc) between low and high redshift, metal-rich 2010). The relative importances of in situ and accreted and metal-poor environments, etc. 19 2006; Seth et al., 2006; Neumayer and Walcher, 2012), and grow (Andersen et al., 2008; Capuzzo-Dolcetta and 4 – 8 which is usually denser (∼ 1 – 10 pc, 10 M , re- Miocchi, 2008; Antonini et al., 2012; Antonini, 2013). call Fig. 1) than globular clusters (Georgiev and Boker,¨ As gas inflows in the previous scenario, the infall of 2014), putting them amongst the densest stellar struc- clusters can also lead to a rotating nuclear cluster, for tures known. It was found that the fraction of early- at least some assembly histories (Tsatsi et al., 2017). type galaxies observed with a nuclear cluster reaches its Kinematics provide a partial way to disentangle the 9 maximum for stellar masses of ∼ 10 M , but decreases two scenarios. For instance, counter-rotating popula- with mass for smaller galaxies, and is truncated for the tions have been found in massive elliptical galaxies most massive hosts (Ferrarese et al., 2006; Pfeffer et al., (thus supporting the migration scenario, Seth et al., 2014). Most of these points are yet to be explained, but 2010; Lyubenova et al., 2013). In general however, both Antonini (2013) proposed that tidal disruption by super scenarios allow for complex stellar populations in the 8 massive black holes (> 10 M ) could destroy in-falling nuclear cluster, but the mixture of episodic and more clusters and prevent the assembly of nuclear clusters in continuous formation histories points toward a combi- massive galaxies (see also Arca-Sedda et al., 2016). nation of both channels (Lyubenova et al., 2013). Thus, A number of scaling relations have been empirically it is likely that these two scenarios both participate to the established between the mass of the nuclear clusters building-up of nuclear clusters (den Brok et al., 2014; and the luminosity of the galaxy, the mass or the ve- Cole and Debattista, 2016). Models estimate that 50 locity dispersion of the galactic bulge (Ferrarese et al., to 80% of the stars in nuclear clusters are formed in 2006; Rossa et al., 2006; Wehner and Harris, 2006; Gra- situ, while the rest has first migrated from larger galac- ham, 2012; Scott and Graham, 2013; Georgiev et al., tic radii before being accreted (Hartmann et al., 2011; 2016). Therefore, the formation and evolution of nu- Antonini et al., 2015). clear clusters is connected to that of their host galaxy In gas rich galaxies, since the densest clusters could (e.g. Leigh et al., 2015), such that these objects can be have a sufficiently strong gravitational influence on the used to explore several aspects of galaxy evolution, es- nearby ISM to retain a gas reservoir around them, their pecially in the central part, the assembly of the bulge potential migration to the galactic centre would bring and physics related to super massive black holes. For in- gas with them. Following this idea, Guillard et al. stance, by measuring the metallicity distribution of stars (2016) proposed that the compression of this reservoir at members of the nuclear cluster of the Milky Way, Ryde the centre, and the further accretion of other gas would and Schultheis (2015) found similarities with that of the thus lead to a episode of in situ formation. Other clus- bulge stars and concluded that at least a fraction of the ters, with or without gas can also be accreted onto the nuclear cluster formed in a comparable fashion as the nuclear cluster. Such process is highly dependent on galactic bulge (see also Rich et al., 2017). the properties of the clusters formed in the galaxy, their masses, their abilities of convoy a gas reservoir with 9.1. Two and a half formation scenarios them, and their trajectories toward the galactic centre In the in situ scenario (Milosavljevic,´ 2004), the nu- (Guillard et al., in preparation). clear cluster forms in the galactic centre. The deep po- Details on the assembly of nuclear clusters affect tential well and the fact that the galactic centre con- their final properties, in term of stellar populations, stitutes the destination of most of the large-scale gas abundance spreads but also kinematics and morphology. transport mechanisms (torques from interacting galax- For instance, merging clusters triggers a redistribution ies, bar, spiral etc.) favour the fuelling of this region of the angular momentum which affects the flattening with gas from larger radii. When it accumulates, the of the resulting object. (Guillard et al. 2016 noted that gas becomes dense enough to form stars as a massive a nuclear cluster formed by the merger of two massive cluster. clusters would be significantly flatter than if the merger In the migration scenario (Tremaine et al., 1975), star event did not occur, see also Antonini et al. 2012; Tsatsi formation occurs elsewhere in the galaxy. Lacking the et al. 2017). physical peculiarities of the galactic centre, the clus- ter formed is usually less dense and less massive than 9.2. Ultra compact dwarf galaxies typical nuclear clusters. It then migrates to the centre, Being the densest and the most bound objects to their because of kpc-scale dynamics (spirals, bar, dynamical galaxies, nuclear clusters are the most robust structure friction, see Section 12), within a few ∼ 0.1 – 1 Gyr of their host. This has led Bassino et al. (1994) to (Chandrasekhar, 1943; Mo et al., 2010). Once in the propose that a nucleated dwarf galaxy undergoing tidal centre, the cluster can further accrete other clusters stripping would lose all its material except the nuclear 20 cluster, which would them form a UCD (see also Bekki evolution which shapes the clusters. In order to con- and Couch 2001; Drinkwater et al. 2003; Pfeffer and nect present-day observations to the conditions of clus- Baumgardt 2013). In such a case, the IMF of UCDs ter formation, it is necessary to understand the physical should be similar to that of NCs (and therefore to the mechanisms that alter the properties like the mass or the bulk of the globular clusters if the NCs assemble by size, and the very survival of clusters. merging globulars and thus share their IMF). In par- ticular, one would not expect to find top-heavy IMFs 10. Collisional systems in UCDs, contrary to the proposition of Dabringhausen et al. (2009) to explain their high mass-to-light ratios Because of the high densities of clusters, their stars (at a late stage of their evolution, once the massive stars interact frequently, which drives exchanges of energy have transformed into dark remnants). Thus, in the for- toward equipartition. This leads for instance to mass mation scenario of UCDs from tidal stripping of galac- segregation (i.e. the most massive stars being pref- tic nuclei, the UCDs should retain a dark matter halo erentially in the cluster centre, see e.g. White 1977; to account for the observed high mass-to-light ratios Bonnell and Davies 1998, but also Trenti and van der (see e.g. Has¸egan et al., 2005; Forbes et al., 2014, and Marel 2013 and Parker et al. 2016 for cautionary notes). Section 3.3). However, observations of low-mass X- The time needed to significantly decrease the imprint ray binaries (i.e. binaries made of a neutron star and a of a perturbation from energy equipartition of stellar low-mass companion), support the excess of dark rem- systems is commonly estimated through the relaxation nants from the evolution of massive stars, and thus a time. Spitzer (1987) provides a simple expression for top-heavy IMF (Dabringhausen et al., 2012). Hence, such a timescale, evaluated at the half-mass radius rh this suggests a different formation scenario for (at least for a system of N stars of mean mass m (i.e. ≈ 0.5 M some) UCDs and NCs (and thus globulars if NCs are for old objects) as mergers of clusters). Another possibility to explain !3/2 !1/2 the elevated mass-to-light ratio is to consider the pres- rh 1 M τ ≈ 0.17 Myr N1/2. (2) ence of a central massive black hole, accounting for rh 1 pc m a significant fraction of the total mass of their galac- Applying this expression to dense stellar systems like tic host (∼ 10 – 20%), as predicted by Mieske et al. star clusters (r ∼ 1 – 10 pc, N ∼ 103 – 6, m ≈ (2013) and supported by the recent detections of super- h 0.5 – 1 M ) shows their half-mass relaxation time is massive black holes in UCDs (Seth et al., 2014; Ahn generally shorter than their lifetime. This hence indi- et al., 2017). cates that stellar encounters have likely operated a sig- Among the alternatives, Mieske et al. (2002, 2012) nificant evolution of the internal properties of the clus- suggested that the UCDs constitute the high-mass end of ters since their formation, contrarily to less dense6 sys- the star cluster distributions, and that they form in com- tems like galaxies where the relaxation time is several parable environments as young massive clusters (e.g. in orders of magnitude longer. This has led Forbes and galaxy mergers, see Renaud et al. 2015a). Another op- Kroupa (2011) to propose to use such a difference to tion would be that the UCDs result from the merger of tell apart clusters from galaxies (recall Section 3.4). several star clusters in large cluster complexes (Kroupa, One of the consequences of the collisional nature of 1998; Fellhauer and Kroupa, 2002; Bruns¨ et al., 2011). clusters is the energy transport from the inner parts to It is now likely that, like the nuclear cluster, several the outskirts (i.e. a negative dynamical heat capacity), formation channels exist for the UCDs, with their rela- that results in the collapse of the central part. This tive importance depending on the context (Mieske et al., mechanism, known as core-collapse or gravothermal 2006; Brodie et al., 2011; Norris and Kannappan, 2011; catastrophe (Antonov, 1962; Lynden-Bell and Wood, Pfeffer et al., 2014). 1968; Goodman, 1987), would eventually lead to an in- finite central density if it was not stopped by the injec- tion of energy from binary stars, either primordial, or Part III formed by gravitational capture (Heggie and Hut, 2003). At first order, one can consider the internal physics Evolution of the clusters (mainly stellar evolution and stellar en- counters) as an engine pumping energy, mostly from the When interested in old clusters like globulars, one should not neglect the long-term, post-gaseous phase 6Note that the relaxation time does not solely depend on density. 21 core of the cluster outward (Henon,´ 1961). This trans- 11. Modelling dense stellar systems lates into the (overall) expansion of the cluster. Henon´ (1961, 1965) proposed a description of energy flows In this rather technical section, I present the methods in clusters at a constant rate per relaxation time. Ac- and tools that have been developed to study star cluster cording to Henon´ (1961), the central source of inter- evolution, and why addressing this gravity-only ques- nal energy driving the expansion does not need to be tion is not as simple as it looks at first sight. formally identified, but the hardening of binaries in the cluster’s core is a natural candidate (Giersz and Heggie, 11.1. A numerical challenge 1994; Baumgardt et al., 2002). Yet, number of works The challenge of studying the evolution of clusters in showed that clusters considered in isolation (i.e. when their environment is to couple their internal physics to neglecting the galactic context) yield very long disso- a description of the influence of the environment. The lution timescales, up to 1000 times their initial relax- most natural way to do so would be to model jointly a ation time (Giersz and Heggie, 1994; Baumgardt et al., cluster and its host galaxy(ies) and account for all grav- 2002). Therefore, the environment, through the tidal itational interactions between all mass elements (stars, interactions discussed below, is the main actor setting gas, dark matter). Taking all interactions into account the clusters’ lifetimes (Gieles and Baumgardt, 2008). is commonly done in galaxy and cosmological simula- Tides contribute to the energy increase of the cluster, tions, where tens of millions resolution elements (par- but mainly in the outer layers. At some point, stars ticles and/or cells) are now routinely modelled. To reach an energy threshold set by the tides and the rest speed-up the computation, the gravitational accelera- of the cluster, and can (potentially, see Section 13.2) es- tion of distant elements is approximated (see tree codes, cape. Consequently, the evolution of star clusters results particle-mesh, P3M, fast multipole methods and oth- from the interplay between internal dynamics and exter- ers, e.g. Barnes and Hut 1986, Couchman et al. 1995, nal perturbations (of cosmological and galactic nature). Dehnen 2000, among many others), and the potential It is therefore equally important to understand the inter- of each element is smoothed to avoid divergences of nal engine and its evolution along the cluster lifetime, the acceleration during close encounters (Aarseth, 1963; and the role of the environment. The different processes Dehnen, 2001; Athanassoula et al., 1998). involved and their relative roles vary with time, with the However, in dense, collisional, stellar systems like position inside the cluster, and the galactic context (see clusters, star-star interactions in binaries and multiples e.g. Lamers et al., 2010). have been shown to be an important source of the clus- ter’s internal evolution (e.g. Heggie, 1975; Lynden-Bell Accounting for both the internal and external physics and Eggleton, 1980) by driving an outward flow of en- of clusters is the main challenge for theoretical works in ergy (Henon,´ 1961). It is thus essential for any method this field, because of the multi-scale and multi-physics to include their effect, and thus to capture details as nature of the problem. Aspects such as stellar evolu- short as the orbital period of a binary, and as small tion, the formation, evolution and destruction of bina- as a stellar radius. When put in galactic context, this ries and multiple systems, star-star collisions, the dy- translates into a wide range of space- and time-scales, namics of the so-called dark remnants (brown dwarfs, making this problem virtually impossible to solve di- neutron stars, stellar mass and intermediate mass black rectly with present-day high performance computing re- holes, Strader et al. 2012; Lutzgendorf¨ et al. 2013; sources. To date, the solution adopted consists in using Mastrobuono-Battisti et al. 2014; Peuten et al. 2016), distinct, specialized, algorithms to describe the cluster as well as mass segregation (Webb and Vesperini, 2017) on the one hand, and its environment on the other, and and multiple populations (Charbonnel, 2016) all have an make then communicate. Even before considering the non-negligible impact on the internal evolution of clus- galactic environment, modelling collisional systems is ters, on the interpretation of the observations (e.g. mass a challenge by itself, which requires a large amount of to light ratios, Strader et al. 2011; Kimmig et al. 2015; algorithmic and engineering tours de force. Watkins et al. 2015, velocity dispersions, Baumgardt In the direct N-body approach, each of the N stars 2017 etc.) and deserve dedicated reviews. I recom- of a cluster is modelled as a particle which experiences mend exploring the reviews by Heggie and Hut (2003) the gravitational acceleration from the N − 1 others and Vesperini (2010), and I focus here on the influences (see Spitzer, 1987; Heggie and Hut, 2003; Binney and of the galactic and cosmological environment in mod- Tremaine, 2008, for more complete and detailed pre- ifying the properties of the clusters, and forming tidal sentations). This direct method is implemented for in- features. stance in NBODY6 (Aarseth, 2003) and PH4 (McMillan 22 et al., 2012). Short range interactions in the collisional regime are treated accurately through the use of regu- larization algorithms to resolve the motion of close en- counters with no critical modification of the timestep that would otherwise considerably slow down the sim- ulations (see Kustaanheimo and Stiefel 1965; Heggie 1973, and Mikkola 2008 for a review). The effect of the environment (i.e. the tidal field) can then be added to the equations of motion of every stars, as described in the next section. The complexity of the N-body prob- lem goes as O(N2), but this computational cost is signif- icantly reduced by the use of software (Spurzem, 1999; Wang et al., 2016) and/or hardware accelerations (Hut 106 Wang et al. and Makino, 1999; Makino et al., 2003; Nitadori and Heggie Sippel & Hurley Aarseth, 2012). Fig. 3 shows the progress of the cluster Shara & Hurley simulations, made possible by always improving com- 105 Baumgardt & Makino puter resources and techniques. Yet, modelling massive Makino clusters with the direct N-body method still represents a challenge (see Wang et al., 2016, on solving the million 104 Spurzem & Aarseth Aarseth & Heggie body problem), making it impossible to cover a wide pa- Inagaki rameter space. To circumvent this problem, other meth- 103 Terlevich ods are also commonly used. number of stars Aarseth Another approach consists in solving the Boltzmann Aarseth equation on a distribution function, with the addition of GPU multinode 102 Aarseth a term describing the close interactions between stars GPU GRAPE (so-called collisions) which follows a Markov process. von Hoerner von Hoerner vectorial CPU For this reason such method is refereed to as Fokker- 101 CPU Planck. The evolution of the distribution function is 1960 1970 1980 1990 2000 2010 2020 usually computed numerically in a discrete grid of en- date of publication ergy bins (Cohn, 1979). However, without knowledge of the position of stars, the treatment of galactic tides is restricted to the idealised cases of clusters on circular Figure 3: Progress of N-body simulations of clusters through the years. The original version of this figure was first presented in Heg- orbits. gie and Hut (2003), and it has been updated by Douglas Heggie (pri- The Monte-Carlo method is a hybrid of Fokker- vate communication). Only works reaching the late evolution stage Planck and N-body: instead of evolving the distribution of clusters are shown (i.e. von Hoerner, 1960, 1963; Aarseth, 1966, functions on an energy grid, they are sampled by tracer 1968, 1973; Terlevich, 1980; Inagaki, 1986; Aarseth and Heggie, 1993; Spurzem and Aarseth, 1996; Makino, 1996; Baumgardt and particles evolving in phase-space. Such method is im- Makino, 2003; Shara and Hurley, 2006; Sippel and Hurley, 2013; plemented for example in the MOCCA code (Giersz et al., Heggie, 2014; Wang et al., 2016). Rapid progresses have been made 2008). This approach leads to a substantial speed-up by upgrading the technology used, especially with special-purpose compared to N-body method, which allows for the study hardware (e.g. GRAPE, Makino and Taiji 1998), or existing tech- nology (e.g. GPU, Nitadori and Aarseth 2012). of massive clusters (see e.g. an application in Heggie and Giersz, 2014). However, it is limited to spherically symmetric problems and thus cannot account for the anisotropic nature of tides. Note however, the contribu- tion of the RAGA code (Vasiliev, 2014) which allows for arbitrary geometries by implementing relaxation with local diffusion coefficients in velocity, and thus which could include a precise treatment of complex tides. In addition, Dehnen (2014) proposes a fast multipole method to integrate the motion of individual stars and reach the accuracy of collisional algorithms. By dis- tributing particles in cells, computing cell-cell interac- 23 tions and then approximating the force on particles us- also on its tidal debris (see Section 13.3). Alternatively ing Taylor expansions, this method achieves unprece- to these simulations of clusters, the focus can be shifted dented performances of O(N0.87). Yet, the lack of reg- to the galaxy, by including a parameterization of star ularisation considerably slows down the code execution cluster evolution in galaxy simulations (Kruijssen et al., when treating dense, collisional systems. 2011; Matsui et al., 2012; Brockamp et al., 2014). More Finally, semi-analytical methods solve a handful of detailed prescriptions of the cluster-galaxy connection coupled differential equations describing, at first or- have been proposed in the last years, opening the field der, the global evolution of the cluster (Ambartsum- to the study of clusters in cosmological context. ian, 1938; Chandrasekhar, 1942; King, 1958; Henon,´ These methods couple simulations of star clusters 1961; Heggie, 1975; Lee and Ostriker, 1987; Hut et al., with simulations of galaxies. One possibility is to ex- 1992; Gieles et al., 2011). An example of this is the tract the gravitational field along one orbit in the galaxy code EMACSS (Alexander and Gieles, 2012; Alexander or cosmological volume and to then pass this informa- et al., 2014; Gieles et al., 2014) which solves for the tion (e.g. through the form of the tidal tensor) to a mass, the half-mass radius and the energy of the clus- star cluster simulation (see the NBODY6tt code, Re- ter. The mathematical complexity of these problems naud et al. 2011, based on NBODY6, Aarseth 2003). requires a numerical solver but this approach remains Doing so allows for time-varying, complex galaxies to doubtlessly the fastest. To estimate the mass-loss of be considered (e.g. mergers, Renaud and Gieles 2013, clusters, EMACSS is calibrated using the results of N- satellite galaxies, Bianchini et al. 2015, full cosmolog- body simulations for clusters on circular orbits around ical context, Rieder et al. 2013). However, the orbit in galaxies, perfectly suited in the cases of constant tides, the galaxy must be assumed beforehand, which neglects or when the tidal field changes adiabatically with re- the non-trivial effect of dynamical friction (see Petts spect to the cluster internal evolution (and when tides et al., 2015). This limitation disappears when adopt- can be reasonably approximated by a piecewise constant ing the option of evolving the two simulations simul- function, see an example in Renaud and Gieles 2015b). taneously, by making both codes communicate (see the However, this method does not retrieve the mass-loss BRIDGE code, Fujii et al. 2007, and its extension in derived from N-body methods in non-constant tides. AMUSE, Pelupessy et al. 2013). However, the different timescales involved on the galactic and cluster sides im- 11.2. Accounting for the environment ply that a numerically efficient coupling is only reached A very large body of work has been done using the in (sub-)parsec resolution galaxy simulations (i.e. when above-mentioned methods to study e.g. the intrinsic both timesteps become comparable), of which cost lim- evolution of clusters, of their stellar populations and the its these runs to short periods of time. Such approach is effect of (potential) intermediate-mass black holes. In thus particularly well suited in simulating the early life their large majority, these studies ignore the effect of the of star cluster (in particular when accounting for hydro- environment on the cluster to focus on internal dynam- dynamics), but is less optimum for Gyr-long modelling. ics. However, the environment plays a paramount role in the questions of e.g., long term evolution, mass-loss, 12. Dynamical friction kinematics, especially in the outermost layers of clus- ters. The simplest and historical approach accounting When orbiting the dense regions of galaxies, star for this consists in considering clusters on fixed circu- clusters experience a loss of their orbital energy caused lar or elliptical orbits in an analytical representation of by dynamical friction. Such a modification of their tra- the galactic potential, where the tidal force can be ex- jectory does not directly impact their properties like pressed analytically and added to the force every star mass and size, but brings clusters in different, usually experiences from the rest of the cluster (e.g. Chernoff stronger, tidal fields that alter these properties. and Weinberg 1990 and Gnedin and Ostriker 1997 with Dynamical friction is the gravitational drag induced the Fokker-Planck method, Vesperini and Heggie 1997 by the constituents of a dense medium in which a mas- and Portegies Zwart et al. 1998 with the N-body way, sive object moves. The object, a cluster or a satellite and Giersz 2001 with the Monte Carlo approach). To galaxy for instance, attracts constituents of the medium, go further, the galactic potential can also be used to in- like stars, gas or dark matter, toward its position. How- tegrate the orbit of the cluster and compute the relevant ever, because of its motion, the medium accumulates tidal forces, thus allowing for non-fixed orbits (Renaud behind the cluster (i.e. at the position it occupied mo- and Gieles, 2015a). Such a method provides the galac- ments earlier). Such accumulation constitutes an over- tic acceleration not only at the position of the cluster, but density in the medium of which gravitational force on 24 the cluster drags it against its motion and thus slows most massive clusters (by merger with the galactic cen- it down (Chandrasekhar, 1943). This effect is for ex- tre like nuclear clusters or the bulge), typically over a ample responsible for the transfer of orbital energy of few Gyr timescale. interacting galaxies which, under some circumstances, leads to their merger and coalescence (Duc and Renaud, 2013). It can also make clusters spiral in toward galactic 13. Adiabatic, secular tides centres, where strong tides accelerate their dissolution Once the gas has been removed from the cluster vol- and help forming galactic nuclei or nuclear clusters (see ume, the only interaction between the cluster and its en- Tremaine et al., 1975; Capriotti and Hawley, 1996; Lotz vironment is gravitational. This effect is dual: global et al., 2001; Arca-Sedda and Capuzzo-Dolcetta, 2014, and differential. In its global flavour, this interaction and Section 9). Furthermore, Fellhauer and Lin (2007) sets the trajectory of the cluster as a whole, and does noted that the stars lost by the cluster could contribute not directly affect its evolution. A notable exception is to the medium, thus participating in dynamical friction cluster-cluster interactions and mergers, which mostly and accelerating the process. occur in galactic centres, and participate in building nu- From this simple explanation, one can deduce dy- clear clusters (recall Section 9). namical friction is efficient for massive clusters, in The differential effect however directly affects the dense media, at low velocity, which is depicted in Chan- evolution of the clusters, by inducing an acceleration of drasekhar (1943) formula. His formalism lacks several which intensity and direction varies across the cluster. physical aspects like the self-gravity of the medium and Such differential forces are tides, and encompass several resonant interactions (Inoue, 2009). Yet, it is remark- aspects, usually sorted according to their timescales. In ably accurate at predicting the results of simulations the rapid, impulsive fashion, the tidal effects are consid- in many configurations (e.g. Tremaine and Weinberg, ered as shocks (see Section 14). When the gravitational 1984; Weinberg, 1986; Hashimoto et al., 2003). In some influence of its environment affects a cluster in a slower, other cases however, Chandrasekhar’s framework is in- longer manner with respect to its internal evolution, one complete. This is for instance the case in uniform den- speaks of adiabatic tides. Details on the evolution of the sity media like the cores of galaxies (Read et al., 2006; cluster properties have been intensively studied over the Goerdt et al., 2010; Petts et al., 2015), and when the last decades, in particular with N-body simulations. cluster becomes of comparable mass as the surrounding medium, like in galactic centres (Gualandris and Mer- ritt, 2008). 13.1. Tidal radii and Jacobi surface Modelling precisely dynamical friction can thus re- The tidal radius is a useful quantity to describe the quires a self-consistent treatment of the medium, which strength of the tidal field on a given object. It is unfortu- is often too numerically expensive. The alternative is nately rarely explicitly defined in the literature, although to use semi-analytic models to mimic the effects of dy- it can refer to several, distinct, quantities. It seems that namical friction as accurately as possible (see examples a major source of confusion came from the use of the in Just and Penarrubia,˜ 2005; Just et al., 2011; Arca- term “tidal” to designate the outermost radius of an ob- Sedda et al., 2015). In this context, Petts et al. (2016) served cluster. The tidal radius appears as a parame- proposed a complement to Chandrasekhar’s formula to ter of the King (1966) density profile, as the radius at reproduce the inefficiency of dynamical friction in the which the density drops (see also Michie, 1963; Michie inner parts of galaxies, and in particular in cored dwarfs and Bodenheimer, 1963; King, 1981). Such radius can (the so-called core-stalling phenomenon). This was then exist for clusters in isolation and is, most of the time, applied to clusters orbiting dwarf galaxies to discrimi- independent of tides, as shown observationally by e.g. nate between cored and cuspy dark matter profiles (Con- Odenkirchen et al. (1997). Even if leaving this confu- tenta et al. 2017, see also Hernandez and Gilmore 1998; sion aside, the tidal radius remains an ill-defined con- Sanchez-Salcedo´ et al. 2006; Goerdt et al. 2006; Cole cept. et al. 2012; Amorisco 2017). In a first definition, one can delineate the volume It is worth noticing that, in the Chandrasekhar’s for- of influence of a cluster as the sphere centred on the malism, the intensity of the drag force varies with the cluster and going through the Lagrange point L1, i.e. square of the mass of the cluster, meaning that massive the point between the cluster and the galaxy, where clusters are more prone to dynamical friction than their the gravitational force from the cluster exactly balances lower mass counterparts. Dynamical friction is thus that of the galaxy. The radius of this sphere is called one of the rare mechanisms preferentially destroying the the tidal (or Jacobi) radius. It evolves with the cluster 25 (when it grows, shrinks, moves to volumes of strong or shape of the local galactic potential. This implies that, weak tides), and it is thus possible that stars remain well for a given tidal radius, different galaxies can lead to dif- within this radius, such that it does not mark any visible ferent Jacobi surfaces, and thus to different tidal accel- truncation in the stellar distribution. erations on cluster members. Tanikawa and Fukushige In this domain, it is convenient and common to adopt (2010) illustrated this point by studying the mass-loss the tidal approximation, which consists in assuming that rate of clusters in galaxies with a power-law density pro- the distance between the cluster and the galaxy is much file, varying the power-law index, but keeping the tidal larger than that between the cluster and its stars. Do- radius unchanged (which implies to change the galac- ing so allows us to linearize (with a first order Taylor tocentric distance R). They found that shallow galac- expansion) the expressions of the galactic gravitational tic profiles (i.e. flattened Jacobi surfaces) induce slow forces on stars and provides an analytical expressions in mass-loss. This can be understood when considering several (idealised but illustrative) cases. We will do so the conditions required for a star to leave the cluster, in the following, keeping in mind that this formalism is which is itself a rather complicated topic, as discussed not suited for clusters near steep features in the galactic e.g. in Ross et al. (1997) and in the next section. potential (e.g. the galactic centre), nor for stars far from In compressive tides, the mathematical derivation of their cluster. a tidal radius (using any of the above expression) leads When considering the cluster and the galaxy as points rt to be the cubic root of a negative quantity, reflect- of respective masses m and M separated by a distance R, ing the fact that there is no point where the force from the tidal radius reads the galaxy balances that from the cluster since they both  m 1/3 point in the same direction. The tidal radius (and the La- r = R , (3) grange points, and the Jacobi surface) cannot be defined. t, point 2M Yet, the acceleration on stars induced by the compres- (von Hoerner, 1957; Spitzer, 1987). Variants of this sive tides still speeds-up the mass-loss of clusters, with (purely gravitational) expression account for the cen- respect to isolated clusters (Renaud et al., 2011, see also trifugal force of clusters on circular orbits, which then the compressive shocks discussed in Section 14). brings the equilibrium point toward the cluster: Therefore, the tidal radius is an ill-defined quantity,  1/3 that does not directly relate to the evolution of clusters, 0 m rt, point = R , (4) but which remains, to some extend, convenient to illus- 3M trate some tidal effects. A more mathematically rig- (King, 1962), which can also be generalized to circular orous and more general quantity to gauge tides could orbits around any spherically symmetric galactic poten- be the strength of the tidal field along its main axis of tial when introducing the orbital angular frequency Ω action (which is the maximum eigenvalue of the tidal and the local galactic gravitational potential φ tensor, see the Appendix A of Renaud et al., 2017, for

 1/3 details, and the connection to the tidal radius), but this  Gm  would then neglect the changes of the orientation of the r0 =   , (5) t  2  Ω2 − ∂ φ  tidal field (i.e. the changes in the eigenbase of the ten- ∂R2 sor), which necessarily slow down the mass-loss due (see the Appendix A of Renaud et al. 2017 for a more to a mismatch between the escape funnels (around the detailed explanation of the different expressions and the Largange point) and the directions of the accelerations corresponding assumptions). the stars experience before getting there. However, the concept of tidal radius is a shortcut of- ten leading to oversimplifications. For instance, the vol- 13.2. Potential escapers and kinematics in the outer re- ume of influence of the cluster is not a sphere but rather gions a flatten “lemon-like” shape, defined by the equipoten- To escape its cluster, a star needs to cross the gravita- tial surface going through L1, called the Jacobi surface tional barrier between the potential wells of the cluster (or the Roche surface, see Renaud et al. 2011 for its and the galaxy. A necessary but not sufficient condi- equation). The deviation from spherical symmetry in- tion is that the star has a higher (more positive) energy duce by tides can have a comparable effect on the shape than that of the Jacobi surface (sometimes called criti- of a cluster as a degree of intrinsic rotation (Bertin and cal energy). However, even in this case, a star is likely Varri, 2008; Varri and Bertin, 2009). The Jacobi sur- to remain in the cluster for a significant period of time. face is parameterised by the tidal radius, and its flatten- Fig. 4 shows diagrams in space and energy of such con- ing along the other two axes that depends on the exact figuration, with the example of a star having 0.9 times 26 the (negative) energy of the Jacobi surface. Thanks to this excess of energy, the star can travel further than the Jacobi surface. However, to escape, it also needs to ac- tually leave the volume of the cluster, i.e. move through the openings in its own iso-energy surface, near L1 or L2. Only then, the star will definitely leave the cluster (assuming the tides are constant). The size and shape of these openings depend on the energy excess and the 1.0 local galactic potential (see Renaud et al., 2011, for an analytical derivation). 0.5 Therefore, even with an energy higher than the Ja- cobi energy, stars are trapped in the cluster until they t

r reach these openings. Such stars are called potential es- / L1 L2 y 0.0 capers, as described by Henon´ (1969, see also Henon´ 1970, Fukushige and Heggie 2000, Baumgardt 2001, 0.5 Daniel et al. 2017). Depending on the motion of the stars with respect to the (possibly varying) positions of the Lagrange points, and their energies (i.e. the size of 1.0 the openings), the period spent in this phase can be as long as the cluster lifetime (i.e. possibly longer than a 1.0 Hubble time), which implies that a significant fraction 0.5 of the material within the volume of the cluster can be in the potential escaper phase (10 to 35%, and possibly 0.0 higher in massive clusters, see e.g. Baumgardt, 2001; Just et al., 2009). J E The presence of potential escapers affects the dynam- 0.5 potential energy E ics and kinematics of clusters, and have been proposed Tidal radius rt 1.0 Jacobi energy EJ by Kupper¨ et al. (2010a, see also Claydon et al. 2017) 0.9EJ to explain the observed flattening of the velocity dis- bound stars 1.5 persion profile in the outermost layers of several glob- potential escapers escapers, galaxy ular clusters (Drukier et al., 1998; Sollima et al., 2009; 2.0 Scarpa and Falomo, 2010; Lane et al., 2010; Da Costa, 1.5 1.0 0.5 0.0 0.5 1.0 1.5 2012; Baumgardt, 2017). The velocity anisotropies are x/rt also affected by this population (Bianchini et al., 2017). Note however that, the net gravitational acceleration Figure 4: Definition of the potential escapers in a geometrical diagram is usually weak in these outer regions and could even, (top) and an energy plane (bottom), illustrated with a point-mass clus- for some clusters, be below the critical acceleration ter (set at x = y = 0) experiencing tides from a point-mass galaxy at ∼ −10 −2 x < 0, y = 0. (An inertial reference frame is considered to compute a0 ( 10 m s ) of the modified gravitation frame- the energy terms.) The black line represents the potential energy from work MOND (Milgrom, 1983). In this regime, an extra the cluster and the tides: stars cannot have less energy. Outside the MONDian acceleration is added by the modification of tidal radius (rJ, dashed red line) is the galaxy (shaded grey area), in- the law of gravity to that from the baryonic distribution. cluding the stars that have escaped the cluster. Inside, one finds bound stars (shaded blue area) with an energy lower than the Jacobi energy Such an additional contribution would then account for EJ, and potential escapers (shaded red area) with an energy higher. As the observed velocity dispersions (Scarpa et al., 2007; an example, the orange line marks 0.9 times the Jacobi energy. A star Hernandez et al., 2013). An alternative theory is that with this energy inside the tidal radius (i.e. in the hatched area) is a globulars at large galactic radii could have retained their potential escaper. own dark matter halo (Ibata et al., 2013b; Penarrubia˜ et al., 2017), remnant of their formation in mini-halos at high redshift (Peebles and Dicke, 1968, recall Sec- tion 8.2). This component would thus have the expected effect in increasing the outer velocity dispersion. In this field, Gaia will soon provide important insights by de- tecting and estimating kinematic properties of stars at 27 large distances from the centre of their cluster. originating from stellar systems on a circular or eccen- tric orbit around their host galaxy (Just et al., 2009; 13.3. Stellar streams Kupper¨ et al., 2010b; Lane et al., 2012; Mastrobuono- Battisti et al., 2012). On the other hand, the morphology After leaving the cluster, stars can be used to probe and structure of streams is also affected by the galactic the structure and mass distribution of the galaxy. Be- environment, and is thus used to probe this environment cause of the shape of the equipotential surfaces around and its evolution. the cluster, most stars leave by flying by the Lagrange Note that streams have not been detected around points L1 (between the cluster and the galaxy) and L2 many clusters in the Milky Way, even when experienc- (on the opposite side, see Fig. 4). Those escaping ing strong enough tides to induce mass-loss. The reason through L1 are thus at smaller galactic radii than the for this could be detection limits: stars in streams origi- cluster such that, for a given kinetic energy, their orbital nate from the outer layers of the clusters which, because motion around the galaxy is faster than that of the clus- of mass segregation (see Section 10), mainly encompass ter. The exact opposite is found for the stars escaping low-mass, faint stars (see Koch et al., 2004, in the case from L2. As a consequence, the tidal debris forms two of Palomar-5). Deeper observations of the outermost tails: one leading (connected to the cluster through L1) regions of clusters, with the necessary corrections for and one trailing (from L2), the two making a character- background contamination, could thus reveal faint tidal istic “S”-shape. Apart from these small differences, the tails. Furthermore, depending on its age and mass-loss tails initially have a similar orbit than their progenitor of the cluster (or dwarf galaxy) making it, a stream can cluster around the galaxy (Johnston, 1998), and remain be observed after the final dissolution of its progenitor. coherent structures for several revolutions (Helmi et al., This is for instance the case of the Ophiuchus (Sesar 1999). Such pairs of tails form in any tidally interacting et al., 2015) and Phoenix streams (Balbinot et al., 2016). systems (clusters, galaxies etc.), but their mass, elonga- Streams can be used to infer some properties of their tion and even detectability depend on several properties progenitors. In particular, tidal stripping is less efficient (see e.g. Balbinot and Gieles, 2018), like the masses on a dwarf galaxy with a cored dark matter halo than involved, the orbit and the velocity distribution of the on a cuspy one (Penarrubia˜ et al. 2010, see also Read stars (e.g. the spin-orbit coupling, see Duc and Renaud et al. 2016 for a review on the core/cusp problem and 2013). For instance, in a pair of interacting galaxies, the the insights gained when accounting for detailed bary- tidal bridges connecting the two members are the lead- onic physics in cosmological simulations). Thus, the ing tails of each galaxy (Toomre and Toomre, 1972). resulting stream (as being directly linked to the mass- When originating from a cluster or a dwarf galaxy, loss induced by tides) traces the dark matter distribution these tails are often called streams. Number of them of the progenitor (Errani et al., 2015). However, streams have been detected around galaxies in the Local group, are mostly used to map the gravitational potential of the like the Sagittarius, Orphan and Monoceros streams host galaxy, in particular in the Milky Way, by infer- (Ibata et al., 2001; Newberg et al., 2002; Yanny et al., ring the properties of the dark halo on larger scales than 2003; Grillmair, 2006; Belokurov et al., 2006), espe- what stellar kinematics and the neutral gas trace (Sack- cially in the recent years in all-sky surveys like DES ett, 1999). (The Dark Energy Survey Collaboration, 2005), SDSS Encounters with sub-structures alter the morphology (Ahn et al., 2012) and VST ATLAS (Shanks et al., and the kinematics of the streams, making them dynam- 2015). More are expected to be discovered in the ically hotter (Ibata et al., 2002; Johnston et al., 2002). near future with data releases from the GAIA mission. Relying on this, streams are used to probe the num- These surveys are deep enough to unambiguously de- ber and size distribution of dark-matter sub-structures, 6 – 7 tect streams, but characterising their morphology and down to ∼ 10 M (Erkal et al., 2016), as a com- the potential presence of sub-structures like gaps along plement to gravitational lensing methods (e.g. Mao and them is more disputed (see below). On the one hand, Schneider, 1998; Vegetti and Koopmans, 2009). Sev- streams intrinsically host variations in their densities, as eral works reported the presence of gaps along streams shown by early numerical experiments by Combes et al. (e.g. Odenkirchen et al., 2001, 2003; Grillmair, 2006), (1999). Kupper¨ et al. (2008) explained that the motions which would result from local tidal disruption caused of stars along streams lead to orbital crowding at the by sub-structures in the galactic potential (e.g. Carlberg epicycles (i.e. a slowing down of the stars at a fixed po- et al., 2011, 2012; Sanders et al., 2016). Sub-halos have sition, see also Capuzzo Dolcetta et al. 2005). This then been proposed as the main suspects since the typical translates into regularly spaced overdensities in streams sizes of gaps match the predictions from the cold dark 28 matter framework (Yoon et al. 2011; Ngan and Carlberg duces a so-called external field effect (Wu et al., 2010), 2014; Bovy et al. 2017, but see Mastrobuono-Battisti generates an asymmetric net potential: the potential et al. 2012 who excluded this hypothesis for the inner- around L1 is raised with respect to that at L2, by the most gaps). The main objective of present-day studies is equivalent of a negative density contribution on top of thus to establish how structures in streams depend on the that of the cluster and the galaxy, which thus as a repul- properties of the galactic halo. In that respect, Pearson sive effect. Such configuration reproduces the observa- et al. (2015) showed that a triaxial halo leads to fanning tions of asymmetric streams, with the leading tail being of the stream at large distance from the progenitor. Fur- significantly shorter than the trailing one (Odenkirchen thermore, Sandford et al. (2017) conducted a numerical et al., 2001; Ibata et al., 2017), as well as the high veloc- surveys to highlight the strong sensitivity of the thin- ity dispersion in the cluster itself (recall Section 13.2). ness of streams, their symmetry and the regularity of Reproducing numerically a real stream remains chal- their morphology on the galactic halo shape. Future de- lenging, due to the size of the parameter space to be ex- tections of streams (in particular at distances where the plored (orbit, galactic potential, initial conditions of the dark matter dominates the mass distribution) could then progenitor) and the non-uniqueness of the model. It also be matched by numerical models to map the potential of requires a proper description of the (evolving) mass-loss the Galaxy, and reconstruct its assembly. rate of the progenitor which, in a case of collisional sys- Yet, deep observations have recently discredited the tems implies to account for two-body encounters (recall existence of gaps in the Palomar-5 stream (Ibata et al. Section 11.1). Yet, some level of simplification is usu- 2016, see also Thomas et al. 2016) which, if general- ally adopted, like assuming a flux of stars from the La- ized to other streams, would have implications on the grange points of the progenitor centre (i.e. a constant 6 – 9 number of small sub-halos (∼ 10 M ) in the Galaxy. mass-loss rate for the progenitor). This allows simula- Garrison-Kimmel et al. (2017) proposed that galactic tions to e.g. predict future evolution of observed streams tides can dissolve such halos, and would even deplete (Dehnen et al., 2004), or estimate local properties of the inner Galaxy (. 15 kpc) from them, when account- the Milky Way potential (Kupper¨ et al., 2015). Recent ing for the tidal effects from the stellar disc (which is models have started to account for the time-evolution ignored when using dark-matter only cosmological sim- of galactic potential (Carlberg, 2017), which introduces ulations). a complex, yet necessary, additional degree of freedom Theoretically, gaps in streams can also result from in- when using streams to retrace the assembly history of teractions with dense baryonic structures (see also Sec- the Galaxy. tion 14). Using simulations, Amorisco et al. (2016) showed that encounters between a stream and galactic 13.4. Evolution in a fixed, isolated galaxy giant molecular clouds would form gaps, comparable Most of the numerical studies on star cluster evolu- to observed ones. Furthermore, Pearson et al. (2017) tion taking the galactic environment into account as- demonstrated that tidal torques induced by the passage sumed a fixed, constant galactic potential in which the of the Galactic bar could lead to similar results. There- cluster evolves, usually on a circular or an elliptical or- fore, the link between gaps in streams and the dark sub- bit. While these strong assumptions were first justified structures is less direct than previously thought, spe- by technical and methodological limitations, they re- cially for progenitors of which orbits pass(ed) by the re- main valid with present-day resources when considering gion of the dense disc (. 15 kpc) and the bar (. 3 kpc). that most of the growth of the Milky Way (in mass, size It has also been found that different dark matter par- and in term of the formation of structures like the discs) ticles, that make different halo shapes, would change occurred before redshift unity (see e.g. Behroozi et al., the appearance of streams (see examples for several 2013), i.e. in the first 5 Gyr of the Universe. Adopting dark matter flavours in Dubinski and Carlberg, 1991; an idealised description of the Galaxy has allowed for a Yoshida et al., 2000; Dave´ et al., 2001). Streams are better understanding of cluster evolution, on which re- thus of prime importance to determine the very nature cent studies in more complex and realistic frameworks of dark matter, and its distribution in the outer volumes rely. of galaxies (Koposov et al., 2010; Gibbons et al., 2014), Following the analytical study of Henon´ (1961), and also to test and constrain alternative paradigms. For Hayli (1970) presented the first numerical models of instance, Thomas et al. (2017a,b) modelled the Sagit- star clusters in an external potential mimicking a galaxy. tarius and Palomar 5 streams in the context of modified Both established that introducing a galactic potential Newtonian dynamics (MOND, Milgrom 1983). They creates the saddle points and the general features shown showed that the non-linear nature of MOND, which in- in Fig. 4, which leads to an accelerated mass-loss of 29 the cluster with respect to isolation, as described above. tidal radius (recall Section 13.1). Indeed stars experi- Since then, studies on this topic have aimed at deriv- ence a three-dimensional acceleration from the galaxy, ing how this mass-loss depends on the galaxy, the clus- which is not necessary aligned with the cluster-galaxy ter, and its orbit (e.g. Lamers et al., 2010; Madrid et al., axis, i.e. not always toward the Lagrange points. So 2017). The commonly adopted approach is to split the far, the proposed analytical derivations of a dissolution problem into individual mechanism, then to considers timescale still fail at reproducing the simulations results several combinations of them representative of specific (Fukushige and Heggie, 2000; Tanikawa and Fukushige, phases of galaxy evolution, and finally to consider the 2010; Renaud et al., 2011; Claydon et al., 2017). Re- full cosmological context accounting for the formation cently, Daniel et al. (2017) used simulations to build and evolution of the host galaxy(ies). a dynamical model for potential escapers (based on a Woolley 1954 model). Such a model opens new per- 13.4.1. Constant tides: circular orbits spectives in understanding the mass-loss rate of clusters In the absence of tides, the internal dynamics of a and the kinematics in their outer regions. cluster drives its expansion (Henon,´ 1965, recall Sec- Furthermore, the escape condition depends on the en- tion 10). However, when considering the galactic en- ergy of the star and is thus different for all of them. vironment, the tides slow down this expansion. By in- Keenan and Innanen (1975) illustrated this by show- troducing a self-similar model of clusters which con- ing that stars on prograde orbits (i.e. with an orbital serves their average density, Henon´ (1961) showed that angular momentum vector roughly aligned with that of the mass-loss (accelerated by the tides) must be asso- the cluster around the galaxy) escape more easily than ciated with the contraction of the cluster. In other oth- stars on retrograde orbits (see also Fukushige and Heg- ers words, escaping stars remove (negative) energy from gie, 2000; Read et al., 2006; Ernst et al., 2007). This the cluster which must therefore re-adjust. Gieles et al. can be understood when considering the centrifugal ef- (2011) pushed forward the formalism of Henon´ and fect on these stars, which provides an additional positive showed that clusters in constant tides experience expan- (respectively negative) contribution to their energy with sion for about half of their life, followed by contrac- respect to that of the cluster in the case of prograde (re- tion. Such a behaviour has been noted in a very large spectively retrograde) orbits7. Over a long evolution, number of simulations. In such models, the evolution this phenomenon results in the depletion of the star on of quantities is normalised to characteristic times like prograde orbits, and thus induces a net retrograde spin the half-mass relaxation time and the dissolution time, of clusters (Tiongco et al., 2016). This is particularity which thus remain to be estimated. noticed in the outer layers, more sensitive to tides, and Early works showed that the dissolution time of clus- less to possible disruption by two-body encounters than ters on circular orbits is rather insensitive to their initial in the central regions (Baumgardt and Makino, 2003). structure (e.g. its concentration), but is mostly depen- The spin-orbit coupling of the cluster is thus an impor- dent on their initial mass and the strength of the tides, tant parameter when studying its evolution and mass- characterised by the galactic-centric distance and the loss. Simulations are however often initialized in the so- orbital velocity (Chernoff and Weinberg, 1990; Taka- called phase-lock (also known as tidal locking) setup, hashi and Portegies Zwart, 2000; Vesperini and Heggie, such that the Jacobi surface and the Lagrange points 1997; Aarseth and Heggie, 1998). Fukushige and Heg- are fixed in the (spinning) reference frame of the cluster gie (2000) then combined simulations and analytical (e.g. Heggie and Hut, 2003). Despite this, the asym- derivations to establish an expression for the dissolution metry of the tidal effects induces differential rotation in time of clusters on circular orbits around point-mass clusters as observed (e.g. van de Ven et al., 2006) and galaxies, as a function of the excess of energy of stars modelled (Boily and Pichon, 2001; Theis, 2002; Ves- above the critical energy. They highlighted the compli- perini et al., 2014). This is strongly enhanced around cation introduced by potential escapers that delay the core-collapse when the Coriolis effect alters the inward mass-loss in a yet not fully understood manner (recall (in the contraction phase, respectively outward in the Section 13.2). Furthermore, the approach of Tanikawa re-expansion phase) radial motions of stars by making and Fukushige (2010) of setting the tidal radius and them prograde (respectively retrograde). Vesperini et al. varying the slope of the galactic potential (which im- (2014) found that the layers close to 1 – 2 times the plies playing with the galactocentric distance) showed the importance of the three-dimensional aspect of tides. 7We note that the same situation if found in interacting disc galax- This further underlines that, even for clusters on circular ies: prograde galaxies yield long tidal tails, while retrograde ones only orbits, the evolution is not well represented by the sole undergo mild disruptions (Duc and Renaud, 2013). 30 half mass radius rotate significantly faster than others. pers constitute a population buffer, filled from the bound Therefore, the potential initial rotation of clusters inher- material (by relaxation and tidal acceleration), and emp- ited from their formation phase is disturbed and possibly tied through the openings around the Lagrange points erased by the tidal field. of which size (and position, see below) varies along the cluster orbit. Hence, mass-loss only occurs when the 13.4.2. Time-varying tides: the example of eccentric emptying mechanism is active (i.e. strong tides), and orbits when the buffer actually contains stars to be ejected. The non-spherically symmetric potential of galax- Because these two conditions do not share the same ies implies that most clusters are on non-circular or- timescale, it is likely that they are not always simultane- bits (Dinescu et al., 1999; Casetti-Dinescu et al., 2013). ously met in time-varying tidal fields. Therefore, this requires studying their evolution in time- The evolution of the mass of a cluster on an ec- varying tides and understanding how it differs from the centric orbit exhibits the classical “stair-case” func- constant tidal field of circular orbits. The first, simplest tional form, with steep mass-loss just before pericen- step in this direction is to keep the galaxy spherically tre, and slower evolution elsewhere (e.g. Baumgardt and symmetric, but to set the clusters on eccentric orbits. Makino, 2003). Depending on how cluster membership The complexity of the problem of time-varying tides is defined (e.g. within the tidal radius or below some lies on the delay between the tidal acceleration under- energy threshold), the mass of a cluster can slightly in- gone by stars and their actual escape from the cluster. crease after the pericentre, when stars that recently es- Let’s consider a cluster on an eccentric orbit, approach- caped are re-captured by the cluster as its Jacobi volume ing its pericentre. Because of tidal acceleration and two- increases. body relaxation, the cluster expands and the energy of An additional complexity arises when considering the its stars in the outer regions increases (on average). In motion of the Lagrange points in the cluster reference the same time, the Jacobi surface shrinks. Close to peri- frame. On eccentric orbits, this reference frame is non- centre, some of the stars yield an energy above that of inertial, and accelerated with respect to an inertial one the critical one, and become potential escapers (recall (e.g. that of the galaxy). Therefore, the Euler effect Section 13.2). Some of them escape rapidly (i.e. within must be taken into account, on top of the centrifugal a crossing time), while others do not reach the openings and Coriolis effects already present in the case of cir- in their Jacobi surface and are trapped in the cluster for cular orbits. By altering the positions where the force a longer period of time. from the cluster balances that from the environment, it The tidal configuration is symmetric on both the ap- induces a motion of the Lagrange points, as oscillations proaching and receding side of the pericentre such that, around the galaxy-cluster axis (see e.g. Renaud et al., if one neglects the mass-loss the cluster experiences 2011). Therefore, stars that are tidally accelerated to- around pericentre, both the tidal radius and the Jacobi ward a Lagrange point at time t could well face a higher surface are identical before and after the pericentre. Yet, potential barrier when arriving at this position at t + dt, when moving away from pericentre the cluster has al- as the Lagrange point has moved away. This increases ready lost the stars susceptible to escape: the high en- the population of potential escapers and lower the mass- ergy population has already been almost completely de- loss rate, for a given tidal radius. pleted. The only remaining stars to be stripped are the Although inferring the instantaneous mass-loss of a potential escapers trapped since before the pericentre, cluster on eccentric orbits is analytically involved8, em- and those which have gained energy since then and thus pirical relations have been established to express the recently became potential escapers. As a consequence, total lifetime of clusters, using numerical experiments. despite symmetric tides, the mass-loss is highly non- Baumgardt and Makino (2003) found that the lifetime of symmetric with respect to the pericentre passage. Fur- a cluster on an eccentric orbit is the same as that of the thermore, when receding from the galaxy, the cluster cluster on a circular orbit passing through the pericentre, yields a growing Jacobi surface and can potentially re- but corrected by a linear factor of the eccentricity (see capture stars previously lost, if they fall back in the clus- also Webb et al., 2014a). Cai et al. (2016) then found ter potential well (e.g. Webb et al., 2013). (It is how- that it exists a circular orbit on which the evolution of ever likely that such stars will be among the firsts to the cluster (in terms of mass and half-mass radius) is escape again, due to their high energy, unless they get the same than on an eccentric orbit when averaged over trapped in the potential escaper phase, or lose energy through two-body encounters.) One can conveniently 8Attempts are currently being made, following the perturbation picture this process by considering that potential esca- approach from the circular case (Bar-Or et al, in preparation). 31 the long term (i.e. for longer than the orbital period). Martinez-Medina et al. (2017) showed that tidal com- They found the corresponding orbital radius by trial and pression generated by local cores in the overlapping po- error, and noted it is neither the semi-major axis of the tentials of the bar and spiral arms (i.e. mainly at the eccentric orbit, nor the average orbital radius (see also tips of the bar) play a protective role on clusters (see Brosche et al., 1999; Webb et al., 2014a). Cai et al. also Renaud et al., 2015a, for considerations on cluster (2016) concluded that the evolution of a cluster on an formation in these areas). They found that clusters in eccentric orbit can be approximated with that of a clus- such regimes could survive twice longer than if outside ter on a circular orbit with the same dissolution time, of these structures, at a given galactic radius. when averaging over the entire cluster lifetime. This Yet, these studies neglected the formation, evolution, conclusion can however not be extended to the cases of possible destruction and re-formation of the bar itself non-periodic tides. (see e.g. Cole and Weinberg 2002), and the evolution of its strength, which would have required a detailed pre- 13.4.3. In non-axisymmetric, barred potentials scription of the galactic evolution (in particular through Moving away from spherically symmetric galactic galaxy mergers and the accretion of gas, see Kraljic potentials, several works explored the effect of the et al. 2012). galactic bar on the evolution of clusters. (The role of the disc and spiral arms is discussed in Section 14.) The mass-loss (or even dissolution) of clusters in these vol- 14. Tidal shocks umes participates in assembling the galactic bulge. To As opposed to adiabatic tides, short-lived perturba- understand how galaxies like the Milky Way form, it is tions (with respect to the internal crossing time of clus- therefore important to pin down where the various stel- ters) fall into the shock regime. Encounters with nearby lar populations observed originate from, whether they dense stellar and/or gaseous structures generate such formed in situ or not, when they reached the bulge, and tidal shocks. This happens when clusters cross a galac- what is the role of the bar in such process. tic disc, fly by a spiral arm, or interact with dense gas The first order effect of a non-axisymmetric potential clouds. During such events, the impulsive perturbation is to modify the orbits of clusters. Pichardo et al. (2004) rapidly and strongly accelerates the stars of clusters, showed the presence of a bar induces dispersion in the shrinks their Jacobi surface and can strip the least bound orbital energy and angular momentum of clusters, that layers. This can severely damage or even destroy clus- could explain the observed distribution of prograde and ters, especially the fragile ones, at the low-mass end of retrograde orbits (see also Moreno et al., 2014). On top their mass function. of such aspects, a bar contributes to the tidal field expe- rienced by cluster and alters their evolution. Berentzen and Athanassoula (2012) found that the dissolution time 14.1. Disc crossing of a cluster in a barred potential is mainly set by the av- Some star clusters being observed at high galactic lat- erage tidal forcing along the orbit. They showed the bar itudes, their orbits must be inclined with respect to the induces periodic expansion and contraction of the tidal plane of the disc. This naturally implies that clusters debris, making characteristic morphologies of under- fly through the disc and experience then a short but in- and over-densities that correlates with the orbital period tense tidal acceleration. Ostriker et al. (1972, see also (see also Section 13.3). Going further (mainly by in- Spitzer 1987) derived the vertical acceleration induced cluding collisional aspects in their simulations), Rossi by a disc (modelled as an infinite plane) on a cluster and Hurley (2015a) highlighted the importance of the moving perpendicularly toward it. They noted that the orbital type in the evolution of clusters within the zone vertical tidal forces experienced by stars of such a clus- of influence of the bar (i.e. . 4 kpc, as they found little ter point toward the cluster centre. Hence, one speaks effect at larger radii). They found that clusters on non- of compressive tidal shocks, which increase the acceler- chaotic orbits (i.e. co-rotating, anti-rotating, or with a ations of stars (in the cluster reference frame), and thus time-varying rotation with respect to the bar) only ex- constitute a source of (dynamical) heating for the clus- perience mild perturbations when passing by the bar, ter (see also D’Onghia et al., 2010, in the context of and not a dramatic change in their mass-loss timescale. destroying dark matter sub-structures). However, clusters on chaotic orbits undergo strong dis- In this formalism, one classically neglects the high or- ruptions at their pericentre passages. This quickly de- der moments in the energy input by the shock, and focus pletes this family of orbits such that Rossi and Hurley on the first order h∆Ei (i.e. energy shift). However, the (2015a) suggested that such clusters should be rare. quadratic moment h(∆E)2i (i.e. energy diffusion) can be 32 as important as the first order term in the outskirts of the below) the disc plane. Furthermore, spirals can have a clusters, and even more at smaller radii (Aguilar et al., strong tidal effect on clusters. 1988; Kundic and Ostriker, 1995). In fact, stars with Gieles et al. (2007) showed that the heating induced an energy close the critical energy could gain energy by the slow passage of a spiral arm (with respect to the through the energy dispersion term, that could make cluster crossing time, i.e. in the adiabatic regime), is them unbound (see also Madrid et al., 2014). The sec- mostly damped. Thus, the broad (stellar) component of ond order term could thus govern cluster mass-loss (see a spiral arm has little net effect on clusters crossing it. also a detailed mathematical derivation in Kundic and However, encounters with the thinner dense gas compo- Ostriker 1995, and an early discussion in Spitzer and nent of spirals lie in the impulsive regime, which could Chevalier 1973). On top of accelerating the dissolu- potentially severely damage the clusters. Yet, Gieles tion of clusters (Gnedin and Ostriker, 1997; Vesperini et al. (2007) noted that most of the energy gain is con- and Heggie, 1997; Gnedin et al., 1999a), the contribu- verted to high velocity, but only for a few stars in the tion of shocks to the energy of stars can also speed-up outer layers of the cluster, and thus does not induce an its internal evolution, for instance toward core-collapse important mass-loss. Furthermore, they showed that the (Chernoff et al., 1986; Gnedin et al., 1999b). effect of subsequent passages of an arm depends on the Furthermore, because the energy gain from tidal period between successive encounters, which is thus a shocks depends on the position in the galaxy and the dis- function of the galactocentric radius (with respect to co- tance to the cluster centre (and thus the mass of the in- rotation). All together, spiral arms do not constitute a dividual stars when the cluster is mass-segregated), disc major source of perturbation of massive clusters. shocking could explain the observed relation between the position of clusters and the slope of their (present- 14.3. High speed encounters with clouds day) stellar mass function (Stiavelli et al., 1991, see In addition to compressive shocks due to disc cross- also Capaccioli et al. 1993; Djorgovski et al. 1993). ing and the passage of spiral arms, clusters can also ex- Disc shocking tends to flatten the stellar mass-function perience shocks when encountering local, dense sources of cluster, by depleting preferentially the low-mass end of gravitation like giant molecular clouds. Such encoun- (Vesperini and Heggie, 1997). ters are expected to be a major player in altering the By varying the orbital inclination, Martinez-Medina clusters’ properties (or even dissolving them) at high et al. (2017) showed, on the one hand, that clusters with redshift, before the galactic disc is fully formed, and a low maximum orbital latitude (i.e. staying near the when the baryonic content of galaxies is dominated by disc) cross the disc at a lower speed than those on a gas, more turbulent than in the Local Universe (Gen- more inclined orbit. This results in a lower dynamical zel et al., 2006), and possibly in the form of massive, heating, a weaker tidal shock, and thus a longer clus- dense clumps (at z ∼ 2 – 3, see e.g. Tacconi et al. 2010; ter lifetime (see also Webb et al., 2014b). On the other Zanella et al. 2015). hand, clusters with a high orbital inclination cross the Spitzer (1958) considered interactions between a disc less frequently, which also results in a longer life- cluster and several clouds along its orbit (in the impul- time. By combining these two effects, Martinez-Medina sive regime, but also with an extension to slow encoun- et al. (2017) found that clusters with a maximum alti- ters), and found analytically that the tidal heating in- tude of 600 pc above the disc plane yield the shortest duced by such shocks leads to disruption on a timescale lifetime. At large galactic distances (i.e. typically fur- that scales with the cluster density. This was later con- ther than the dense stellar disc), (Webb et al., 2014b). firmed with simulations by Gnedin and Ostriker (1999) Orbital inclination, and thus the origin of clusters and and Gieles et al. (2006), who also found that the dis- the process that put them at high latitude are thus key solution time due to GMC encounters is several times for their long-term survival. shorter than that from typical adiabatic tides. Wielen (1985) showed however that low-mass gas structures 14.2. Passages of spiral arms have a limited impact on star clusters, and that only passages of giant molecular clouds significantly affect On top of playing a role on their formation (recall them. Section 5.3), spiral arms can also alter the orbits of The dense gas clouds being in the Galactic disc, clusters. For instance, Martinez-Medina et al. (2016) cloud-clusters encounters have been proposed as a showed that the passage of a spiral arm could lift the driver of the disruption of low-mass clusters there, and clusters to high galactic altitude and contribute to es- thus to explain the dearth of old open clusters in the tablishing their observed thick distribution above (and Milky Way disc (Wielen, 1985; Terlevich, 1987). 33 Elmegreen and Hunter (2010) proposed that, because of the hierarchical structure of the ISM (particularly at high redshift), the tidal field felt by a young cluster is dominated by local gaseous sub-structures in the cloud complex where it forms. This situation endures until the cluster drifts away from the complex, i.e. for about 100 Myr, or until the gas structure is dissolved. Dur- ing this time, shocks with dense gas structures partici- pate in shaping the clusters, and can help reproducing the observed mass-age distributions. Elmegreen (2010) pushed this argument further to propose that this mecha- nism would drive the evolution of an initial cluster mass function of the Schechter type (recall Section 4), into the 1.0 observed present-day log-normal shape (see also Krui- jssen, 2015). 0.8 Gieles and Renaud (2016) noted that repeated shocks shock-dominated evolution do not yield a cumulative effect: tidal shocks are self- limited (see also Gnedin et al., 1999b). While the first 0.6 ) ] shock does potentially significantly affect the clusters c p [

properties by stripping the outermost stars as discussed h r (

g 0.4 above, it takes the cluster a long time to replenish these o l layers (through two-body relaxation and adiabatic tidal relaxation-dominated evolution acceleration). In the mean time, it is little sensitive to repeated shocks, and its evolution is thus mainly driven 0.2 by secular processes. As a consequence, clusters that 3 are not fully destroyed by a shock are likely to sur- h = 30 M pc 0.0 1/9 vive subsequent shocks of similar (and lower) ampli- rh M tude, over a period of a few relaxation times, until they 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 expand again. The result of combining shocks and re- log(M [M ]) laxation is an equilibrium in the mass-size relation, as shown in Fig. 5. Gieles and Renaud (2016) showed that Figure 5: Adaptation of the figure 3 of Gieles and Renaud (2016), clusters denser than this equilibrium lie in a relaxation- showing evolutionary tracks in the mass-size diagram of clusters sub- dominated regime that makes them expand, while those ject to repeated shocks with giant molecular clouds in a Milky Way- less dense are more sensitive to tides that reduce their like ISM. The formalism and all numerical values are identical to that of Gieles and Renaud (2016). The solid black line marks the equilib- mass and their size, until they reach the equilibrium. 1/9 rium relation between the half-mass radius and the mass: rh ∝ M . The weak dependence of the size on the mass could ex- The dots represents clusters of different masses but same half-mass plain the observations of a typical radius of ≈ 3 pc for density. Clusters initially denser than the equilibrium relation lie in the low-mass and young clusters (i.e. before the secular the relaxation-dominated regime. Internal dynamics first make them expand, toward the equilibrium relation. Conversely, clusters initially tides dominate their evolution, Larsen 2004). less dense than the equilibrium are more sensitive to tidal harassment, Gnedin and Ostriker (1997) modified the formalism which truncates them and thus reduces their size, toward the equilib- of Spitzer (1958) of point-mass clouds to consider ex- rium relation. tended objects to study tidal shocks induced by the galactic bulge (see also Aguilar and White, 1985; Wein- berg, 1994). They concluded that, when integrating in- ternal evolution, secular tides and shocks, the dissolu- tion time of clusters is such that most clusters have al- ready dissolved, and that their remnants account for a significant fraction of the old stellar components of the Galaxy (i.e. the bulge and the halo). In all the above-mentioned studies, the importance of tidal shocks has been established in the context of the present-day Milky Way. However, changes in the prop- 34 erties of the ISM (in particular the distribution and den- (Deason et al., 2013; Ruchti et al., 2014), these con- sity of molecular clouds), the formation of the disc, the clusions can be applied to clusters over this period in bulge and the spiral arms must be taken into account this galaxy, but the evolution of the baryonic structures when inferring the evolution of clusters at high redshift. must be accounted for in the case of clusters orbiting Unfortunately, there are little observational constraints in the innermost regions. In particular, the formation of on this, and the exact role of tidal shocks in shaping the the bar and the evolution of its strength could alter the cluster population is yet to be established. evolution of clusters (recall Section 13.4.3). The same could be concluded about the formation of (transients, see e.g. Roskarˇ et al. 2012) spiral arms, but Gieles 15. Evolution in cosmological context et al. (2007) showed the inefficiency of spirals in alter- ing clusters (Section 14.2). Furthermore, the disc(s) are When studying the evolution of old clusters ( & supposedly formed at z ≈ 2 (Wyse, 2001), i.e. close to 6 – 8 Gyr), it becomes crucial to account for the evolu- the end of the period during which the growth is domi- tion of their context. Studying the dynamics of clusters nated by major mergers, i.e. in a violent, non-adiabatic for several Gyr in a static galactic potential, as repre- phase (Renaud et al., 2017). The evolution of the impor- sentative of a present-day galaxy it can be, misses the tance of disc crossings thus remains difficult to address important aspects of galaxy evolution and the variations without a clear understanding of the processes forming of the tidal field it induces. The formation and evolu- the stellar and gas discs (Mo et al., 1998; de la Fuente tion of galactic substructures (bars, spirals), the secu- Marcos and de la Fuente Marcos, 2009). lar growth of the galaxy itself, interactions and mergers Furthermore, the mechanism of radial migration with neighbour galaxies and satellites must be consid- within the disc (induced by the bar and the spirals, ered. To date, these aspects have been mostly ignored, Minchev and Famaey 2010; Aumer et al. 2017) is ex- but numerical methods allowing to couple galaxy and pected to change the tidal field of the Galaxy, but also clusters simulations have recently opened new possibil- to modify the orbits of the disc clusters (e.g. Grand ities (recall Section 11.2). et al., 2012; Martinez-Medina et al., 2017) and possi- bly speed up or slow down their dissolution (Fujii and 15.1. Adiabatic growth Baba, 2012), depending on the origin and destination of The first step when accounting the evolution of the their migration. environment of star clusters is to consider the adiabatic, secular evolution of the galaxy itself. This encompasses 15.2. Accretion of satellite galaxies its growth form the accretion of dark matter and gas, the conversion of this gas into stars, the (possible) for- On top of the adiabatic growth, galaxy also ac- mation of the discs, the bulge, the spiral arms, the bar, crete material in impulsive ways, by merging with their the stellar and the gaseous halo etc. Numerical and neighbours. Let’s consider a star cluster in a dwarf observational studies on galaxy formation provide con- galaxy, which approaches a bigger galaxy and becomes straints on these aspects, but their exact causes, when one of its satellites. Dynamical friction induced by the these events occurs, and at which speed remain poorly dark matter halo, the hot gas corona and any dense com- known (see a review in Freeman and Bland-Hawthorn, ponent of the main galaxy makes the dwarf lose or- 2002). Furthermore, they vary from galaxy to galaxy bital energy and spiral in (Section 12). In the process, and thus, modelling them to study the populations of the dwarf is tidally stripped by the main and the clus- clusters requires exploring a large parameter space. ter eventually escapes the dwarf (or the dwarf dissolves Renaud and Gieles (2015b) considered the adiabatic completely). The cluster then experiences a transition growth of the dark matter halo (in mass and in size) from (mainly) the tidal field of the dwarf to (mainly) across cosmic time (neglecting the baryonic compo- that of the main galaxy. Before the accretion, the ve- nents) to infer how such a time-varying tidal field af- locities of the stars of the cluster are set according to fects star clusters. They found that, because most of the the potential of the dwarf. Equilibrium is not necessar- growth of the galaxy is done at high redshift (z & 1 – 2), ily reached, but for simplicity, we consider here that the the details of the growth have close to none influence cluster is virialised. The transition to a different tidal on clusters and their tidal debris, and that only the fi- field implies that the velocity of the stars are likely not nal state of the halo sets their properties (mass, size, adequate to the new potential to maintain virialisation, morphology) at z = 0. Since the Milky Way has not and the cluster needs to adjust to its new environment experienced any major merger over the last 6 – 9 Gyr (see Hills 1980; Boily and Kroupa 2003a,b; Baumgardt 35 and Kroupa 2007 for a comparable formalism but in a (Renaud et al., 2008, 2009). Such compression is how- different context). ever short-lived (30 – 100 Myr, Renaud et al. 2009) and This situation is particularly interesting when the is thus not expected to significantly alter the evolution cluster initially lies in the core of the dwarf, i.e. in com- of pre-existing star clusters (but see Section 6 on the pressive tides. Such a tidal configuration increases the enhancement of their formation). On the other hand, energy of the stars while limiting the mass-loss of the galactic interactions significantly alter the orbits of their cluster (with respect to classical extensive tides, see Re- constituents, which indirectly modify the tidal fields in naud et al. 2011). When switching to the extensive tidal which they evolve. The evolution of clusters in interact- field of the main galaxy, stars likely yield an excess of ing galaxies is thus non-trivial, and highly depends on kinetic energy with respect to the new gravitational po- their position within the system. tential, i.e. the cluster is super-virialised and it expands Renaud and Gieles (2013) conducted a survey of star (Bianchini et al. 2015; Webb et al. 2017, see also Mi- cluster simulations, imposing different tidal fields de- holics et al. 2014, 2016 for a comparable evolution but rived from a simulation of a merger. They found that without compressive tides). Such an expansion could the collisions themselves have little effects on the exist- possibly explain the properties of the extended clusters ing clusters, because the tides are too short-lived and too (also known as faint fuzzy clusters, recall Fig. 1). weak to affect dense clusters. However, the modifica- This expansion is much faster than one a cluster tions of the clusters’ orbits significantly alter their long- which would have always been in the tidal field of the term evolution. For instance, clusters ejected into the main could experience, and likely makes the cluster fill- galactic tidal debris experience much weaker tides than ing its Jacobi volume (or, in other terms, be tidally- before the interaction, which increases their lifetime. limited). Therefore, the cluster from the dwarf could On the contrary, clusters dragged toward the galactic be more extended than one from the main galaxy if the central regions lose mass faster after the interaction than latter is not (yet) tidally limited (Webb and Vesperini, before. Because the tidal debris represent a small frac- 2017), i.e. if it is still in the expansion phase of its tion of the total mass of a merger (and originate from evolution (Baumgardt et al., 2010; Gieles et al., 2011). low density environments in the progenitors), the latter However, if the cluster from the main were already in case is much more frequent than the former. the tidally-limited phase, the one from the dwarf would A full description of the tides, especially the shocks have a comparable size, set by the Jacobi surface (Bian- induced by small, dense gas structures requires pc- chini et al., 2015; Miholics et al., 2014, 2016). An- resolution (e.g. on the softening of the gravitational po- other mechanism, likely linked to the formation of these tential), which was not the case in Renaud and Gieles clusters (e.g. primordial binaries and/or gas expulsion), (2013), and remains rare in modern galaxy simulations. should then be invoked to explain observational data As a consequence, the tidal shocks were not properly (see e.g. Zonoozi et al., 2011, 2014; Leigh et al., 2013, accounted for. It is however likely that the collision of 2015). In any case, the transition from compressive to two galaxies increases the number of shocks, including extensive tides induces an important mass-loss, and pos- disc crossing (due to their relative inclination), and the sibly the dissolution of the cluster. relative velocity of clusters with respect to dense ma- terial, which increases the energy input (recall the for- 15.3. Interacting galaxies and major mergers malism of Ostriker et al., 1972, see also Section 14). Interactions and mergers between galaxies of compa- The re-distribution of orbits during interaction events rable masses induce a number of transformations of the has been proposed by Kravtsov and Gnedin (2005) and properties of the progenitors and of their cluster pop- Elmegreen (2010) as an efficient way to send clusters to ulations. One effect is to efficiently form clusters, in weak tidal environments (e.g. at high altitudes above the particular massive ones, during the possible starburst disc remnant), and ease the disruptions they undergo. episode(s) such encounters trigger (recall Section 6). This would preserve a fraction of the low-mass clus- The role galaxy encounters play on the evolution of ters, while the most massive ones are supposedly dense existing clusters is dual. On the one hand, galactic tides, enough to survive most of the tidal harassment. inflows, outflows, shocks and mergers modify the mass distribution of the galaxies (see a review in Duc and 15.4. Full cosmological context Renaud, 2013), and thus change their tidal fields. In To establish the relative importance of all the above- particular, many (if not all) interactions trigger a sig- mentioned mechanisms on the evolution of star clus- nificant enhancement of compressive tides, over large ters, one must study the global cosmological context of volumes, thus changing the very nature of the tidal field their evolution, i.e. the full formation and evolution of 36 the galaxy(ies) in which they evolve. Such a holistic and Gnedin, in preparation). Yet, it remains difficult to approach requires to resolve pc-size perturbations (e.g. disentangle signal from numerical artefacts, especially dense clouds) over cosmological volumes, which is still because of the use of smoothing kernel modifying the out-of-reach of large-volume simulations, but can be ob- local gravitational field, and thus setting a minimum tained in zoom-in setups. Present-day works thus focus scale to derive gravitation-related quantities. on either a large-volume with a statistical diversity of In conclusion, the secular tidal histories of clusters galaxies, or on a narrower space but at higher resolution are rather well understood now, despite some uncertain- (see e.g. Kravtsov and Gnedin, 2005; Li et al., 2017; ties on e.g. disc formation. However, the early evolution Kim et al., 2018, in the context of globular cluster for- of clusters and the dissolution of the least resistant ones mation). is yet to be quantified. This suffers from the lack of Rieder et al. (2013) derived the evolution of tidal observational constraints on galaxy formation and their fields along several orbits in dark matter only simula- ISM at high redshift, and on the formation of the clus- tions describing two different assembly histories (i.e. ters themselves. Although the physical processes de- different merger trees) of Milky Way size halos. They scribed above are rather well understood individually, then followed the evolution of clusters on these selected their non-linear combination and its exact effects on orbits and found rather smooth, continuous mass-losses, clusters is still to be established. with no abrupt changes (except those from pericentre passages on eccentric orbits, recall Section 13.4.2), but still accelerated by major merger events. Since the mass Part IV of the host galaxy directly influences cluster mass-loss, a Milky Way assembled by many minor mergers inherits Open questions clusters that have experienced (on average) less severe mass-loss than one assembled by a few major mergers. Most of the still open questions in this field originate Therefore, despite forming comparable halos at z = 0, from our lack of understanding of the relative impor- Rieder et al. (2013) found that their two assembly his- tance of formation and evolution processes, in particular tories have different imprints on clusters. for old objects like globulars. This nature versus nurture Renaud et al. (2017) included the baryonic compo- puzzle hinders our attempts to connect well-observed nent and a prescription for star formation in their sim- present-day properties and objects with their equivalent ulation of the assembly of the Milky Way and also in the early Universe. I identify below, in no particular tracked the evolution of the tidal field. Without mod- order, a few open questions for which the community elling the clusters themselves, they found that clusters has started to provide answers, with various degrees of accreted by the Milky Way (i.e. formed in another completeness. galaxy) experience (on average) weaker tides than those formed in situ, but that the overall behaviour these evo- lutions remain similar. They are marked by phases 16. Are YMCs young globular clusters? in galaxy evolution like the early assembly by major mergers, the formation of the disc, and the late adia- The similar masses between YMCs formed in the batic growth. On average, all types of clusters experi- present-day Universe (e.g. in starbusting galaxy merg- ence tides significantly weaker in the past, even those ers, recall Section 6) and the globular clusters formed in formed in the central parts of dwarf satellites. The the early Universe (Section 8) suggest that the buildings strength of the tidal field average over the cluster popu- of two populations share a number of physical prop- lation increases the most when the galactic disc forms, erties and mechanisms. Yet, the question “Are YMCs that the average galactocentric radius of cluster becomes young globular clusters?” can be interpreted in different smaller (due to more clusters forming in the inner re- ways. gions of a massive galaxy). But again, the lack of nu- First, one could rephrase it as “will the present-day merical resolution (on the softening length of the gravi- YMCs observed in ∼ 12 Gyr from now resemble the tational force) forbids to infer the effects of small-scale present-day globulars today?”. Assuming that YMCs do structures and thus the tidal shocks. resemble young globulars (which is a strong assumption The most recent simulations can reach resolutions of discussed below), the answer only depends on the evo- a few parsecs, meaning that the gravitational effects of lution of the two populations, and is therefore clearly external molecular clouds can be (barely) reproduced negative. Indeed, the early Universe hosts physical con- and that (some) tidal shocks can be accounted for (Li ditions significantly different than the present-day and 37 the future Universe, in term of density, galaxy interac- metallicity detected in Milky Way globulars ([Fe/H] = tion and merger rate, gas content, etc. Therefore, the -2.5, Harris 1996), of which origin remains to be under- tidal effects (impulsive and secular) the YMCs will ex- stood. perience in the next few Gyr are necessarily different Furthermore, turbulence is much stronger at high red- than those experienced by the globular clusters in the shift, due to the larger amount of gas and the repeated past few Gyr. First, the consumption of gas for star pumping by kpc-scale phenomena (interactions, merg- formation will progressively decrease the importance of ers and gas accretion, Wise and Abel 2007; Forster¨ tidal shocks (Section 14). Second, the expansion of the Schreiber et al. 2009; Swinbank et al. 2011). However, Universe will continue to slow down the merger rate and the absence of measurement of turbulence of the dense thus the increased mass-loss of the majority of clusters gas at very high redshift (z & 3) makes it difficult to (Section 15.3). Finally, the fact that galaxies are and assess whether the formation sites of the first globular will be more massive than before implies that their (av- clusters are strongly more turbulent than that of YMCs erage) tidal field will be stronger than it has ever been in galaxy mergers for instance. With such variations of (see an illustration in Renaud et al., 2017, their figure metallicity and turbulence, one could question the uni- 10). The combination of these factors make unlikely versality of the stellar IMF, which is of prime impor- that the environment-induced mass-loss will proceed tance for many aspects of astrophysics, and in particular the same way in the next 12 Gyr than it has in the last for the internal dynamics of clusters (segregation, bina- 12 Gyr. This statement must however be balanced by rity, stellar mass-loss), the formation and evolution of the fact that the most massive clusters are only weakly binary stars, the feedback processes (in term of intensity affected by tides, and thus the differences between the and timing, due to possible delays for supernovae, see two populations (YMCs and globulars) after ∼ 12 Gyr Walcher et al. 2016; Zapartas et al. 2017) and thus the of evolution could be moderate. subsequent star formation, gas removal and early shock- Another way to interpret the question above- driven evolution of the clusters. Analytical models sug- mentioned could be to take the opposite perspective, gest that a compression-dominated turbulence (like that i.e. “were the present-day globulars 12 Gyr ago resem- measured in mergers, Renaud et al. 2014b, and expected bling the present-day YMCs today?”. The interest of to be stronger and longer-lived in the early Universe, doing so is seeking whether one can observed the rela- Renaud et al. 2017) would lead to a bottom-heavy IMF tively nearby YMCs to better understand the formation (Chabrier et al., 2014). Yet, how such a turbulence cas- of globular clusters, which remains out of reach obser- cades and decays from its galactic injection scale down vationally (but see the recent use of gravitational lensing to that of the pre-stellar cores is still unknown (due to to probe young globulars at redshifts up to 6, Vanzella the enormous range of scales that must be captured self- et al. 2017a,b). In this case, our lack of understand- consistently), and it remains possible that some mech- ing of the process of star cluster formation, in particular anism(s) like early feedback or magnetic fields could for the most massive objects makes this question more erase the imprint of the large scales in the sub-parsec complicated than the previous one. regime. Despite uncertainties, the observed variations Some of the differences between the formation sites in massive elliptical galaxies and dwarfs tend to show of massive clusters at high redshift and in the Local Uni- that the environment does alter the IMF (see e.g. Cap- verse are the chemical composition (including metallic- pellari et al., 2012; Geha et al., 2013; Dib et al., 2017). ity and dust content) and possibly the turbulence. Clark Remains the early evolution of these clusters. The et al. (2008) suggested that dust cooling dominates over self-limited aspect of tidal shocks and the strong depen- −5 H2 at extremely low metallicities (. 10 Z ) and high dence on the properties of the ISM (Gieles and Renaud, 12 −3 densities (& 10 cm , see also Dopcke et al. 2013). 2016, see also Section 14) suggests yet another evolu- Note however that this in not the case in lower densities tion in the dense, gas-rich, clumpy ISM in which globu- (see a discussion in Glover and Clark, 2014). Therefore, lar formed, compared to the significantly smoother and it seems critical to distinguish the formation of the cloud more diffuse ISM of the local Universe, even in the for- (for which cooling at low densities matters) from the mation sites of YMCs. Uncertainties on the density of formation of the cluster itself (at much higher densities). gas structures and their velocity dispersions leave the Different cooling could result in different fragmentation problem without a clear conclusion. and hierarchical structure of the star forming regions, The importance of tides on shaping the cluster mass- through e.g. the formation of sub-clusters that would function has been demonstrated qualitatively and quan- eventually merge to build a massive system. The asso- titatively (e.g. Elmegreen and Hunter, 2010; Rossi and ciated physics likely relates to the observed minimum Hurley, 2015b; Gieles and Renaud, 2016), but as men- 38 tioned before, the physical conditions of cluster forma- erased for old objects. Furthermore, the rate of erase- tion at high redshift are not fully known, and it is pos- ment would depend on the collisional nature of the sys- sible that the initial mass function of present-day glob- tem, and would then be different in dense globular clus- ulars differs from that of present-day YMCs. ter and in extended galaxies, making a common zone of All the points mentioned above seems to indicate dif- avoidance rather unlikely. Another problem of this the- ferent formation and evolution schemes for the globular ory is that objects formed in low-metallicity ISM would clusters and the YMCs, but many uncertainties and the be surrounded by an optically thinner medium, making lack of quantitative description still forbids to conclude. radiative pressure less efficient. Probing the early stages of cluster evolution at high red- Furthermore, the formation mechanisms, epochs, en- shift with e.g. the forthcoming James Webb Space Tele- vironments and timescales of these objects differ, mak- scope will certainly provide insights on this question. ing the existence of such a common limit somewhat sur- prising. For instance, the formation of a globular cluster takes place in a small region of space and in a relatively 17. Zone of avoidance short period of time. On the contrary, that of massive elliptical galaxies results from the successive assembly Mass - size diagrams of stellar systems show what of galaxies with different sizes and masses, formed in Fig. 1 already hints in the luminosity - size plane: no different regions and at different epochs, thus in con- stellar systems (galaxies and star clusters) is found be- trast with the rapid, in-situ formation mode of globular low an oblique line setting a maximum mass as a func- clusters. As a result, one would not expect these two tion of projected size. This corresponds to a maxi- processes of in-situ formation and assembly would gen- mum surface density of ∼ 105 M pc−2, which defines erate a common zone of avoidance (Norris et al., 2014). the boundary of the so-called zone of avoidance9 (see The same line of arguments holds for nuclear clusters, e.g. Norris et al., 2014), irrespective of the type of which also (at least partly) assemble by merging pre- object considered (globular clusters, nuclear clusters, existing structures (see Section 9). ultra-compact dwarfs, compact elliptical galaxies, ellip- To date, the physical origin(s) of the zone of avoid- tical galaxies). The existence of the common boundary ance and, most importantly, the reason making it rel- in objects ranging from central ellipticals to compact evant for a broad range of objects with different for- globulars tends to suggest a common physical process mation mechanisms and epochs, remains unidentified, setting such an upper surface density limit. Hopkins which further demonstrates our lack of understanding et al. (2010) underlined that the limit is detected in sur- of cluster and galaxy formation. face density and not volume density (because its spans a wide diversity of objects), which already provides a constrain on the physical process(es) responsible for it. 18. Maximum cluster mass and star formation rate Naturally, the first idea coming to mind invokes reg- ulation at formation. Using a timescale argument, Hop- The observational evidence that starbursting mergers kins et al. (2010) noted that feedback from supernovae host the formation of massive star clusters (recall Sec- go off too late to set the maximum surface density. tion 6) tends to suggest a relation between the galactic However, wind and radiation from massive stars are SFR and the mass of the most massive young cluster active early enough to account for the observed zone (Kravtsov and Gnedin, 2005; Whitmore et al., 2010). of avoidance in young objects. The radiative pressure Larsen (2002) conducted an analysis of cluster lumi- from young massive stars on a dusty medium (i.e. opti- nosity functions confirming this idea, and in particular cally thick) would then be sufficient to balance collapse in galaxies with lower SFRs than starbursts (see also and could possibly set the maximum surface density. Larsen and Richtler, 2000; Billett et al., 2002; Larsen, As noted by the authors, this explanation suffers from 2004; Cantiello et al., 2009). Johnson et al. (2017) also several shortcuts: first, dense systems experience relax- reported a comparable relation, more precisely between ation that erase within a few relaxation time any struc- the surface density of SFR (ΣSFR) and the characteris- ture, in particular in the densest regions. Therefore, if tic truncation mass of the Schechter mass function (Mc, set at formation, the maximum surface density will be see equation (1)), i.e. a proxy for the mass of the most massive cluster (recall Section 4)10. They considered

9not to be confused with the region behind the Milky Way centre, where observations suffer from strong extinction (Kraan-Korteweg 10Recall however that the high-mass end of the cluster mass func- and Lahav, 2000). tion suffers from low number statistics (see e.g. Chandar et al., 2017). 39 a sample of four galaxies: M 31, M 83, M 51 and the cluster formation rates (see their figure 11). The reason Antennae, i.e. a green-valley galaxy with a low SFR, for this is again the differences in the physical processes a grand-design star forming spiral, a minor merger and setting the organisation of the ISM and triggering the a starbursting major merger, and found a power-law of starburst (tidal and turbulent compression, shocks and index close to unity connecting the two quantities (see nuclear inflows, recall Section 6). their figure 6). The diversity of galaxies considered then Although these points do not inform us directly about suggests that the physical process(es) setting the mass of the most massive cluster per se, it would be rather sur- the most massive cluster depends on the global star for- prising that masses of the clusters formed in the two sets mation activity, but not on the physical conditions set- of physical conditions (i.e. the two galactic types, with ting star formation itself. For instance, the existence of two different ISM structures) would be the same, as two regimes of star formation (discs and mergers, see the relation observed by Johnson et al. (2017) suggests. e.g. Daddi et al. 2010b and Genzel et al. 2010) demon- Solving this paradox requires a better understanding of strates that a given SFR can have different physical ori- the shortcuts followed in the reasoning above. In par- gins, which thus questions the origin of the relation ob- ticular, does the most massive cluster form in the most served. massive gas clump? Do, for some reason, the changes in Simulations and analytical models show that these the structure of the ISM only affect the bulk of the star two regimes originates from a different conversion from forming regions and not the most massive ones? The the local 3D gas volume density (. 100 pc, more or less hierarchy of the ISM and the potential fragmentation of directly linked with the SFR), to the large scale, galactic clouds into sub-structures should be taken into account, surface density of gas (& 100 pc, see Renaud et al. 2012 which requires a scale-dependent description of insta- and Renaud et al. 2014b). For a given surface density of bilities accounting for external effects (tides, shear etc., gas, a starbursting merger galaxy yield a different turbu- see e.g. Jog 2013, 2014). lence, which compresses its gas more efficiently11 into A convenient but unsatisfactory, solution lies in the dense clumps (i.e. potential sites for cluster formation), observational errors reported by Johnson et al. (2017), which in turn creates an excess of star formation, with which allows for different scaling relations between iso- respect to an isolated galaxy. The distributions of gas lated galaxies and mergers. A larger galaxy sample is densities (i.e. the very organisation of the ISM) is thus needed to conclude on this point. If the relation is con- not directly linked to the large scale surface density of firmed, it would then be interesting to understand how gas. it would behave at high redshift, at the formation epoch Let’s now change our perspective and consider for in- of globulars, in line with questions raised in Section 16, stance a gas-rich disc galaxy at z ∼ 2 yielding a SFR of but observational limitations makes this task out of our −1 −1 −2 ∼ 50 M yr , with ΣSFR ≈ 0.3 M yr kpc . Such current and short term reach. a galaxy is comparable to those in the BzK sample of Tacconi et al. (2010). Now, let’s consider a local star- 19. The halo mass - cluster mass relation bursting merger, with the same SFR and ΣSFR, which could be considered as a LIRG (Kennicutt, 1998). Be- cause the high redshift galaxy lies in the normal star Estimating the total mass of a galaxy (i.e. includ- formation mode while the merger is in starburst mode, ing the dark matter component) is key in many dynam- the efficiencies of their star formation (i.e. the inverse of ical studies, but represents an observational challenge. their gas reservoir depletion time) are not the same: for Usual techniques are based on rotation curves (e.g. So- fue and Rubin, 2001), the detection of X-ray gas in giant the same ΣSFR, the LIRG can have a surface density of gas (typically) one order of magnitude lower than that elliptical galaxies (e.g. O’Sullivan and Ponman, 2004), of the BzK (Daddi et al., 2010b). (See also Bournaud the kinematics of clusters (e.g. Romanowsky et al., et al. 2015 showing different physical triggers leading 2009) and stellar streams (e.g. Kupper¨ et al., 2015), and weak and strong lensing of background objects (Hoek- to comparable SFRs and ΣSFR’s in clumpy discs and in mergers.) stra et al., 2005; Ferreras et al., 2005; Mandelbaum Furthermore, in the merger simulation of Renaud et al., 2006). However, each of these approaches is often et al. (2015a), the two consecutive galaxy interactions12 restricted to a particular galaxy type, which introduces different biases. yield comparable SFRs and ΣSFR’s, but very different

11Efficiency should not be confused with rate! cause of dynamical friction experienced during the first) is more pen- 12The first passage is rather distant and rapid, while the second (be- etrating and marks the onset of final coalescence. 40 quency was universal, before universality got erased by 15 evolution of clusters in a diversity of galactic environ- ments, see Zaritsky et al. 2016; Elmegreen et al. 2017.) 14 This derives from the fact that the stellar mass does not

) correlates with the total mass across galactic types, but ] 13 remains intriguing as one would expect a tighter rela- M [

s tion between baryonic components than between clus- s

a 12

m ters and dark matter. The physical reason remains to be

o l

a understood. h

11

y In this context, the enhanced formation of massive

x Galaxy cluster a l

a cD clusters in galaxy mergers (see Section 6) raises a prob- g ( 10 E g lem. Let’s consider the merger of two equal-mass gas- o l dE rich galaxies. It is reasonable to neglect the galactic 9 S0 S mass-loss due to tidal ejection (as most of the tidal de- MW and M31 bris will eventually fall back onto the remnant), and 8 Local group dwarfs thus to consider that galactic mass is doubled during 5 6 7 8 9 10 the merger process. However, the physical conditions log(mass of globular cluster population [M ]) encountered during the interaction lead to the forma- tion of massive clusters, which add up to the existing Figure 6: Correlation between the mass of the globular cluster popula- population of clusters from the two progenitors which tion and the halo mass, over a range of galaxy types: galaxy clusters, thus more than doubles. As a result, the cluster mass- central dominants in galaxy clusters (cD), ellipticals (E), dwarf el- halo mass relation should break, in particular in central lipticals (dE), lenticulars (S0), spirals (S), the Milky Way (MW) and galaxies of clusters which have likely experienced (at Andromeda (M31), and a few local dwarf galaxies. The line repre- sents a power-law of slope unity. Data points from Spitler and Forbes least) one merger event. This paradox can be solved in (2009). two ways. The first relies on cluster evolution: by reducing the mass of cluster system by the same amount that what Spitler and Forbes (2009) proposed a method to de- is created, the galaxy would stay on the relation. The rive the total mass of the galaxy from its globular cluster increased mass of the merger remnant relative to that population. They showed empirically that the mass of of the progenitors does strengthen the tidal field expe- the population scales linearly with the total mass of the rience by the clusters and accelerate their dissolution. galaxy (including baryons and dark matter), almost in- This enhanced mass-loss of clusters in mergers has been dependently of the galactic mass or morphological type: noted numerically, due to more frequent tidal shocks, ≈ 0.007% of the total galactic mass is found in the form and from the secular, long term tidal field (recall Sec- of globular clusters (see Fig. 6 and also Blakeslee et al. tion 15.3). This effect has a more dramatic impact for 1997, Harris et al. 2017). However, a deviation from this low-density and/or low-mass clusters which are more relation is observed at the low mass end (galactic mass sensitive to tidal harassment than their massive counter- 9 . 10 M ), i.e. in the regime of dwarf ellipticals. The parts. Thus, a large number of small clusters should be reason for this is likely due to small number statistics, destroyed to balance the formation of YMCs during the as these galaxies counts only a handful of globular clus- interaction events. 13 ters . The relation is otherwise remarkably tight for The second option to solve the paradox is based on more massive galaxies, indicating a strong link between cluster formation. The above-mentioned problem only the global properties of the galaxies and the formation holds if one considers that the young massive clusters and evolution of its globular clusters. are the analogues, or the progenitors, of globular clus- Note that no such relation exists between the mass ters. If one lifts this hypothesis (see Section 16), then of the globular cluster population and the stellar mass the ageing massive clusters could not resemble present- of the galaxy. (It however might have been the case day globulars, and thus would not be accounted for in at the epoch of formation, if the cluster specific fre- the mass of the globular cluster population. A merger would then have a different impact on the cluster mass, such that the cluster mass - halo mass relation could still 13The authors argued that the discrepancy noted for the local dwarfs originates from uncertainties in the conversion of the measured veloc- hold. ity dispersions to total mass. Exploring these two options requires a precise quan- 41 titative modelling of both the formation and destruc- redshift. For instance, Robertson et al. (2015) showed tion mechanisms of clusters. Yet, even once the effects that early star formation (including population III stars) of galaxy evolution on this relation studied and under- could play an important role. Such scenarios suffer from stood, it will remain to explain the very origin of the uncertainties on the escape fraction, which is only a relation between a baryonic component and the mass of few percent in massive systems (Grazian et al., 2016; the dark matter halo. The solution is likely hidden in Vasei et al., 2016), but significantly the higher in low galaxy formation, including cooling processes. mass galaxies (Vanzella et al., 2016; Bian et al., 2017). On top of tracing the mass of the galaxies, globular Therefore, yielding an higher escape fraction and being clusters are also used to map the dark halos at large more numerous, dwarf galaxies could also be a strong radii, i.e. where the stellar component is too faint (e.g. driver. Weisz and Boylan-Kolchin (2017) showed that Pota et al., 2013; Kartha et al., 2016). Along those lines, most of the stars in low mass galaxies are formed early, Hudson and Robison (2017) and Forbes (2017) noted a and could thus participate in the re-ionising flux. The correlation between the size of the cluster population limitation here is the ability of the lowest-mass systems and the virial radius of the galaxy. There is an ongoing to actually form stars efficiently, which supposedly sets debate on the nature of the scaling but, in any case, one a mass threshold to participate in the reionization, and can ask the same questions as for the mass relation, on is yet to be established. the origin of this relation and why it apparently remains Following this idea, it appears that young globular insensitive to galaxy evolution. clusters formed as early as z & 6 (Katz and Ricotti, 2013) could also provide a sufficient ultra-violet flux to 20. The role of globular clusters in reionization be major actors in the reionization process. Boylan- Kolchin (2017) estimated the evolution of the cluster Reionization of the Universe is a major step of galaxy luminosity function by adopting a star cluster forma- evolution as it alters the star formation activity and can tion history and assuming that the shape of the lumi- even quench it in the most fragile, low-mass galax- nosity function remains the same for a given population ies (e.g. Brown et al., 2014). It has been invoked for of clusters. This assumption is motivated by our lack instance to solve the so-called missing satellite prob- of understanding of the dissolution processes affecting lem (Moore et al., 1999; Bullock, 2010), i.e. the over- clusters (and their dependence on mass and luminosity, production of halos populated with stars in cosmo- as discussed before). Yet, even with these conservative logical simulations, with respect to the observation of arguments and others (like the universality of the IMF), nearby dwarfs. Boylan-Kolchin (2017) found that young globular clus- Despite its importance, the lack of strong observa- ters would emit a sufficient ultra-violet flux to reion- tional constraints on reionization hinders our under- ize the Universe. By better understanding the condi- standing of the physics driving it. Simulations also can- tion of globular cluster formation (including its epoch), not provide a definite answer, due to the huge range of of their multiple stellar populations, with more precise scales to cover. It is indeed crucial that ionising stars constrains on their dissolution, and finally with a de- and the medium in which the flux propagates and is scription of the ISM in which their ionising flux prop- absorbed are captured together. At present-day resolu- agates, one could quantify their role and date reion- tions, one has to rely on sub-grid recipes to describe ization. Such prediction should then be confronted to the star formation process, the emission of the ionising observational data, in particular in the deep fields of flux (and the other feedback mechanisms), and its abil- the Hubble Space telescope (HST), and the forthcom- ity to propagate in the ISM. This escape fraction is thus ing James Web Space Telescope. largely linked to the porosity of the ISM which can only As often, it is quite possible that the net ionising flux be described by capturing its turbulence. results from the combination of several types of sources. In particular, the sources of ionising photons have not In this picture, it is for instance not clear whether dwarf been clearly identified. Naturally, one first considers galaxies are the main drivers of the reionization, its vic- powerful sources like active galactic nuclei. However, tim, or both. While HST observations provides con- such objects are rare before z ∼ 3 (Ricci et al., 2017), straints on the ultra-violet flux at the end of the reioniza- and despite being powerful, they might not be sufficient tion era (Finkelstein et al., 2015; Bouwens et al., 2015), to produce (alone) the flux needed for re-ionising the the escape fraction remains largely uncertain (see a dis- Universe (Hassan et al., 2018). The next candidates cussion in Leitet et al., 2013). This situation will im- are massive star forming galaxies (Barkana and Loeb, prove with JWST that will push further the boundaries 2001; Loeb and Barkana, 2001), more numerous at high in term of redshift and luminosity, but the faintest galax- 42 ies will still remain out of reach. review. I acknowledge support from the European Re- search Council through grant ERC-StG-335936, and the Knut and Alice Wallenberg Foundation. The prepa- 21. And many more ... ration of this contribution made intensive use of the NASA’s Astrophysics Data System, as illustrated below. Only a few open questions are highlighted above, but many more deserve deep investigations. In no particular order, one could mention: References • What sets the initial binary fraction, and is it the same at high redshift? References

• How do extended (faint-fuzzy) clusters form? Aarseth, S., 1968. Dynamical evolution of simulated N-Body systems. Vol. 3 of Bull. Astron. p. 105. • Aarseth, S. J., 1963. Dynamical evolution of clusters of galaxies, I. How to explain the high specific frequency of clus- MNRAS126, 223. ters in dwarf galaxies? Aarseth, S. J., 1966. Dynamical evolution of clusters of galaxies, II. MNRAS132, 35. • How do multiple stellar populations form? Aarseth, S. J., 1973. Computer simulations of star cluster dynamics. Vistas in Astronomy 15, 13–37. • Do intermediate mass black holes exist in clusters, Aarseth, S. J., Nov. 2003. Gravitational N-Body Simulations. Cam- and how do they relate to stellar-mass and super bridge University Press, November 2003. Aarseth, S. J., Heggie, D. C., Jan. 1993. A 6000-Body Simulation massive black holes? with Primordial Binaries. In: Smith, G. H., Brodie, J. P. (Eds.), The Globular Cluster-Galaxy Connection. Vol. 48 of Astronomical • and many more. Society of the Pacific Conference Series. p. 701. Aarseth, S. J., Heggie, D. C., Jul. 1998. Basic N-body modelling of These open questions, as well as all the topics addressed the evolution of globular clusters - I. Time scaling. MNRAS297, in this review show that the previously distinct fields 794–806. Adamo, A., Bastian, N., Nov. 2015. The Lifecycle of Clusters in of star clusters and galaxies, but also stellar evolution Galaxies. ArXiv e-prints. and cosmology are increasingly getting closer to each Adamo, A., Ryon, J. E., Messa, M., Kim, H., Grasha, K., Cook, D. O., others, and might well be on the verge of merging in a Calzetti, D., Lee, J. C., et al., Jun. 2017. Legacy ExtraGalactic UV few years. The huge and diverse sample of multi-scale Survey with The Hubble Space Telescope: Stellar Cluster Catalogs and First Insights Into Cluster Formation and Evolution in NGC and multi-physics topics will keep many astrophysics 628. ApJ841, 131. busy and excited for a long time. Agertz, O., Kravtsov, A. V., May 2015. On the Interplay between Star Formation and Feedback in Galaxy Formation Simulations. ApJ804, 18. Acknowledgements Agertz, O., Lake, G., Teyssier, R., Moore, B., Mayer, L., Romeo, A. B., Jan. 2009. Large-scale galactic turbulence: can self-gravity It is a pleasure to thank colleagues and friends drive the observed HI velocity dispersions? MNRAS392, 294– 308. for many discussions on the topics of this review Agertz, O., Teyssier, R., Moore, B., Jan. 2011. The formation of disc over the years, and for pointing me to useful works galaxies in a ΛCDM universe. MNRAS410, 1391–1408. and datasets. I am particularly grateful to Oscar Aguilar, L., Hut, P., Ostriker, J. P., Dec. 1988. On the evolution of globular cluster systems. I - Present characteristics and rate of de- Agertz, Mark Gieles, Fred´ eric´ Bournaud, Christian struction in our Galaxy. ApJ335, 720–747. Boily, Eduardo Balbinot, Michelle Collins, Francoise Aguilar, L. A., White, S. D. M., Aug. 1985. Tidal interactions between Combes, Filippo Contenta, Maxime Delorme, Pierre- spherical galaxies. ApJ295, 374–387. Alain Duc, Bruce Elmegreem, Eric Emsellem, Jeremy Ahn, C. P., Alexandroff, R., Allende Prieto, C., Anderson, S. F., An- derton, T., Andrews, B. H., Aubourg, E.,´ Bailey, S., Balbinot, E., Fensch, Duncan Forbes, Morgan Fouesneau, Iskren Barnes, R., et al., Dec. 2012. The Ninth Data Release of the Sloan Georgiev, Nicolas Guillard, Oleg Gnedin, Douglas Heg- Digital Sky Survey: First Spectroscopic Data from the SDSS-III gie, Michael Hilker, Nathan Leigh, Nicolas Martin, Baryon Oscillation Spectroscopic Survey. ApJS203, 21. Alessandra Mastrobuono-Battisti, Ramon Rey-Raposo, Ahn, C. P., Seth, A. C., den Brok, M., Strader, J., Baumgardt, H., van den Bosch, R., Chilingarian, I., Frank, M., Hilker, M., Mc- Aaron Romanowsky, Anna Sippel, Enrico Vesperini and Dermid, R., Mieske, S., Romanowsky, A. J., Spitler, L., Brodie, J., Karina Voggel. I also thank the members of the as- Neumayer, N., Walsh, J. L., Apr. 2017. Detection of Supermassive trophysics groups at the Universities of Lund (Sweden) Black Holes in Two Virgo Ultracompact Dwarf Galaxies. ApJ839, and Surrey (UK) for numerous interesting, stimulating 72. Alexander, P. E. R., Gieles, M., Jun. 2012. A prescription and fast and motivating discussions. I am indebted to the edi- code for the long-term evolution of star clusters. MNRAS422, tor, Cathie Clarke, for her kind invitation to write this 3415–3432. 43 Alexander, P. E. R., Gieles, M., Lamers, H. J. G. L. M., Baumgardt, Balbinot, E., Yanny, B., Li, T. S., Santiago, B., Marshall, J. L., Fin- H., Aug. 2014. A prescription and fast code for the long-term evo- ley, D. A., Pieres, A., Abbott, T. M. C., Abdalla, F. B., Allam, S., lution of star clusters - III. Unequal masses and stellar evolution. Benoit-Levy,´ A., Bernstein, G. M., Bertin, E., Brooks, D., Burke, MNRAS442, 1265–1285. D. L., Carnero Rosell, A., Carrasco Kind, M., Carretero, J., Cunha, Ambartsumian, V. A., 1938. Ann. Len. State Univ. 22, 19. C. E., da Costa, L. N., DePoy, D. L., Desai, S., Diehl, H. T., Amorisco, N. C., Jul. 2017. Deadly Dark Matter Cusps versus Faint Doel, P., Estrada, J., Flaugher, B., Frieman, J., Gerdes, D. W., and Extended Star Clusters: Eridanus II and Andromeda XXV. Gruen, D., Gruendl, R. A., Honscheid, K., James, D. J., Kuehn, ApJ844, 64. K., Kuropatkin, N., Lahav, O., March, M., Martini, P., Miquel, Amorisco, N. C., Gomez,´ F. A., Vegetti, S., White, S. D. M., Nov. R., Nichol, R. C., Ogando, R., Romer, A. K., Sanchez, E., Schub- 2016. Gaps in globular cluster streams: giant molecular clouds can nell, M., Sevilla-Noarbe, I., Smith, R. C., Soares-Santos, M., So- cause them too. MNRAS463, L17–L21. breira, F., Suchyta, E., Tarle, G., Thomas, D., Tucker, D., Walker, Anathpindika, S. V., Jul. 2010. Collision between dissimilar clouds: A. R., DES Collaboration, Mar. 2016. The Phoenix Stream: A stability of the bow-shock, and the formation of pre-stellar cores. Cold Stream in the Southern Hemisphere. ApJ820, 58. MNRAS405, 1431–1443. Banerjee, S., Kroupa, P., Feb. 2015. The formation of NGC 3603 Anders, P., de Grijs, R., Fritze-v. Alvensleben, U., Bissantz, N., Jan. young starburst cluster: ‘prompt’ hierarchical assembly or mono- 2004. Star cluster formation and evolution in the dwarf starburst lithic starburst? MNRAS447, 728–746. galaxy NGC 1569. MNRAS347, 17–28. Barkana, R., Loeb, A., Jul. 2001. In the beginning: the first sources of Andersen, D. R., Walcher, C. J., Boker,¨ T., Ho, L. C., van der Marel, light and the reionization of the universe. Phys. Rep.349, 125–238. R. P., Rix, H.-W., Shields, J. C., Dec. 2008. Caught in Formation: Barnes, A. T., Longmore, S. N., Battersby, C., Bally, J., Kruijssen, The Nuclear-Cluster-to-Be in NGC 2139. ApJ688, 990–999. J. M. D., Henshaw, J. D., Walker, D. L., Aug. 2017. Star formation Andre,´ P., Di Francesco, J., Ward-Thompson, D., Inutsuka, S.-I., Pu- rates and efficiencies in the Galactic Centre. MNRAS469, 2263– dritz, R. E., Pineda, J. E., 2014. From Filamentary Networks to 2285. Dense Cores in Molecular Clouds: Toward a New Paradigm for Barnes, J., Hut, P., Dec. 1986. A hierarchical O(N log N) force- Star Formation. Protostars and Planets VI, 27–51. calculation algorithm. Nature324, 446–449. Antonini, F., Jan. 2013. Origin and Growth of Nuclear Star Clusters Barnes, J. E., Hernquist, L., Nov. 1996. Transformations of Galaxies. around Massive Black Holes. ApJ763, 62. II. Gasdynamics in Merging Disk Galaxies. ApJ471, 115. Antonini, F., Barausse, E., Silk, J., Oct. 2015. The Coevolution of Nu- Barnes, J. E., Hernquist, L. E., Apr. 1991. Fueling starburst galaxies clear Star Clusters, Massive Black Holes, and Their Host Galaxies. with gas-rich mergers. ApJ370, L65–L68. ApJ812, 72. Bassino, L. P., Muzzio, J. C., Rabolli, M., Aug. 1994. Are globular Antonini, F., Capuzzo-Dolcetta, R., Mastrobuono-Battisti, A., Mer- clusters the nuclei of cannibalized dwarf galaxies? ApJ431, 634– ritt, D., May 2012. Dissipationless Formation and Evolution of the 639. Milky Way Nuclear Star Cluster. ApJ750, 111. Bastian, N., Adamo, A., Gieles, M., Silva-Villa, E., Lamers, Antonov, V. A., 1962. Solution of the problem of stability of stellar H. J. G. L. M., Larsen, S. S., Smith, L. J., Konstantopoulos, system Emden’s density law and the spherical distribution of ve- I. S., Zackrisson, E., Jan. 2012. Stellar clusters in M83: forma- locities. tion, evolution, disruption and the influence of the environment. Arca-Sedda, M., Capuzzo-Dolcetta, R., Nov. 2014. The globular clus- MNRAS419, 2606–2622. ter migratory origin of nuclear star clusters. MNRAS444, 3738– Bastian, N., Covey, K. R., Meyer, M. R., Sep. 2010. A Univer- 3755. sal Stellar Initial Mass Function? A Critical Look at Variations. Arca-Sedda, M., Capuzzo-Dolcetta, R., Oct. 2017. The MEGaN ARA&A48, 339–389. project - I. Missing formation of massive nuclear clusters and tidal Bastian, N., Gieles, M., Goodwin, S. P., Trancho, G., Smith, L. J., disruption events by star clusters-massive black hole interactions. Konstantopoulos, I., Efremov, Y., Sep. 2008. The early expansion MNRAS471, 478–490. of cluster cores. MNRAS389, 223–230. Arca-Sedda, M., Capuzzo-Dolcetta, R., Antonini, F., Seth, A., Jun. Bastian, N., Gieles, M., Lamers, H. J. G. L. M., Scheepmaker, R. A., 2015. Henize 2-10: The Ongoing Formation of a Nuclear Star de Grijs, R., Mar. 2005. The star cluster population of M 51. II. Age Cluster around a Massive Black Hole. ApJ806, 220. distribution and relations among the derived parameters. A&A431, Arca-Sedda, M., Capuzzo-Dolcetta, R., Spera, M., Mar. 2016. The 905–924. dearth of nuclear star clusters in bright galaxies. MNRAS456, Bastian, N., Goodwin, S. P., Jun. 2006. Evidence for the strong effect 2457–2466. of gas removal on the internal dynamics of young stellar clusters. Arp, H., Sandage, A., Jul. 1985. Spectra of the two brightest objects MNRAS369, L9–L13. in the amorphous galaxy NGC 1569 - Superluminous young star Bastian, N., Trancho, G., Konstantopoulos, I. S., Miller, B. W., Aug. clusters-or stars in a nearby peculiar galaxy? AJ90, 1163–1171. 2009. Gemini Spectroscopic Survey of Young Star Clusters in Ashman, K. M., Zepf, S. E., Jan. 1992. The formation of globular Merging/Interacting Galaxies. III. The Antennae. ApJ701, 607– clusters in merging and interacting galaxies. ApJ384, 50–61. 619. Ashman, K. M., Zepf, S. E., May 1998. Globular Cluster Systems. Baugh, C. M., Dec. 2006. A primer on hierarchical galaxy formation: Ashman, K. M., Zepf, S. E., Oct. 2001. Some Constraints on the For- the semi-analytical approach. Reports on Progress in Physics 69, mation of Globular Clusters. AJ122, 1888–1895. 3101–3156. Athanassoula, E., Bosma, A., Lambert, J.-C., Makino, J., Feb. 1998. Baumgardt, H., Aug. 2001. Scaling of N-body calculations. MN- Performance and accuracy of a GRAPE-3 system for collisionless RAS325, 1323–1331. N-body simulations. MNRAS293, 369–380. Baumgardt, H., Jan. 2017. N -body modelling of globular clusters: Aumer, M., Binney, J., Schonrich,¨ R., Sep. 2017. Migration and kine- masses, mass-to-light ratios and intermediate-mass black holes. matics in growing disc galaxies with thin and thick discs. MN- MNRAS464, 2174–2202. RAS470, 3685–3706. Baumgardt, H., Hut, P., Heggie, D. C., Nov. 2002. Long-term evolu- Balbinot, E., Gieles, M., Feb. 2018. The devil is in the tails: the role of tion of isolated N-body systems. MNRAS336, 1069–1081. globular cluster mass evolution on stream properties. MNRAS474, Baumgardt, H., Kroupa, P., Oct. 2007. A comprehensive set of simu- 2479–2492. lations studying the influence of gas expulsion on star cluster evo-

44 lution. MNRAS380, 1589–1598. Bianchini, P., Sills, A., Miholics, M., Oct. 2017. Characterization of Baumgardt, H., Kroupa, P., Parmentier, G., Mar. 2008. The influence the velocity anisotropy of accreted globular clusters. MNRAS471, of residual gas expulsion on the evolution of the Galactic globu- 1181–1191. lar cluster system and the origin of the Population II halo. MN- Bianchini, P., Varri, A. L., Bertin, G., Zocchi, A., Jul. 2013. Rotating RAS384, 1231–1241. Globular Clusters. ApJ772, 67. Baumgardt, H., Makino, J., Mar. 2003. Dynamical evolution of star Bik, A., Lamers, H. J. G. L. M., Bastian, N., Panagia, N., Romaniello, clusters in tidal fields. MNRAS340, 227–246. M., Jan. 2003. Clusters in the inner spiral arms of M 51: The clus- Baumgardt, H., Mieske, S., Dec. 2008. High mass-to-light ratios of ter IMF and the formation history. A&A397, 473–486. ultra-compact dwarf galaxies - evidence for dark matter? MN- Billett, O. H., Hunter, D. A., Elmegreen, B. G., Mar. 2002. Compact RAS391, 942–948. Star Clusters in Nearby Dwarf Irregular Galaxies. AJ123, 1454– Baumgardt, H., Parmentier, G., Anders, P., Grebel, E. K., Mar. 2013. 1475. The star cluster formation history of the LMC. MNRAS430, 676– Binney, J., Tremaine, S., 2008. Galactic Dynamics: Second Edition. 685. Princeton University Press. Baumgardt, H., Parmentier, G., Gieles, M., Vesperini, E., Jan. 2010. Blakeslee, J. P., Tonry, J. L., Metzger, M. R., Aug. 1997. Globular Evidence for two populations of Galactic globular clusters from the Clusters in 19 Northern Abell Clusters. AJ114, 482–506. ratio of their half-mass to Jacobi radii. MNRAS401, 1832–1838. Blitz, L., Bazell, D., Desert, F. X., Mar. 1990. Molecular clouds with- Beasley, M. A., Baugh, C. M., Forbes, D. A., Sharples, R. M., Frenk, out detectable CO. ApJ352, L13–L16. C. S., Jun. 2002. On the formation of globular cluster systems in a Boily, C. M., Kroupa, P., Jan. 2003a. The impact of mass loss on star hierarchical Universe. MNRAS333, 383–399. cluster formation - I. Analytical results. MNRAS338, 665–672. Beccari, G., Bellazzini, M., Magrini, L., Coccato, L., Cresci, G., Fra- Boily, C. M., Kroupa, P., Jan. 2003b. The impact of mass loss on star ternali, F., de Zeeuw, P. T., Husemann, B., Ibata, R., Battaglia, cluster formation - II. Numerical N-body integration and further G., Martin, N., Testa, V., Perina, S., Correnti, M., Feb. 2017. A applications. MNRAS338, 673–686. very dark stellar system lost in Virgo: kinematics and metallicity Boily, C. M., Pichon, C., 2001. The Impact of Tides on Cluster For- of SECCO 1 with MUSE. MNRAS465, 2189–2197. mation. In: Deiters, S., Fuchs, B., Just, A., Spurzem, R., Wielen, Behroozi, P. S., Wechsler, R. H., Conroy, C., Jun. 2013. The Average R. (Eds.), Dynamics of Star Clusters and the Milky Way. Vol. 228 Star Formation Histories of Galaxies in Dark Matter Halos from z of Astronomical Society of the Pacific Conference Series. p. 392. = 0-8. ApJ770, 57. Bolatto, A. D., Leroy, A. K., Rosolowsky, E., Walter, F., Blitz, L., Oct. Bekki, K., Couch, W. J., Aug. 2001. Potential Formation Sites of 2008. The Resolved Properties of Extragalactic Giant Molecular Super-Star Clusters in Ultraluminous Infrared Galaxies. ApJ557, Clouds. ApJ686, 948–965. L19–L22. Boley, A. C., Lake, G., Read, J., Teyssier, R., Nov. 2009. Globular Bekki, K., Freeman, K. C., Dec. 2003. Formation of ω Centauri from Cluster Formation Within a Cosmological Context. ApJ706, L192– an ancient nucleated dwarf galaxy in the young Galactic disc. MN- L196. RAS346, L11–L15. Bonnell, I. A., Bate, M. R., Vine, S. G., Aug. 2003. The hierarchical Bekki, K., Yong, D., Jan. 2012. On the origin of the stellar halo and formation of a stellar cluster. MNRAS343, 413–418. multiple stellar populations in the globular cluster NGC 1851. MN- Bonnell, I. A., Davies, M. B., Apr. 1998. Mass segregation in young RAS419, 2063–2076. stellar clusters. MNRAS295, 691. Belokurov, V., Evans, N. W., Irwin, M. J., Hewett, P. C., Wilkinson, Bournaud, F., Jun. 2010. Star Formation and Structure Formation in M. I., Jan. 2006. The Discovery of Tidal Tails around the Globular Galaxy Interactions and Mergers. In: B. Smith, J. Higdon, S. Hig- Cluster NGC 5466. ApJ637, L29–L32. don, & N. Bastian (Ed.), Galaxy Wars: Stellar Populations and Belokurov, V., Zucker, D. B., Evans, N. W., Kleyna, J. T., Koposov, Star Formation in Interacting Galaxies. Vol. 423 of Astronomical S., Hodgkin, S. T., Irwin, M. J., Gilmore, G., Wilkinson, M. I., Society of the Pacific Conference Series. pp. 177–+. Fellhauer, M., Bramich, D. M., Hewett, P. C., Vidrih, S., De Jong, Bournaud, F., Daddi, E., Weiß, A., Renaud, F., Mastropietro, C., J. T. A., Smith, J. A., Rix, H.-W., Bell, E. F., Wyse, R. F. G., New- Teyssier, R., Mar. 2015. Modeling CO emission from hydrody- berg, H. J., Mayeur, P. A., Yanny, B., Rockosi, C. M., Gnedin, namic simulations of nearby spirals, starbursting mergers, and O. Y., Schneider, D. P., Beers, T. C., Barentine, J. C., Brewington, high-redshift galaxies. A&A575, A56. H., Brinkmann, J., Harvanek, M., Kleinman, S. J., Krzesinski, J., Bournaud, F., Elmegreen, B. G., Apr. 2009. Unstable Disks at High Long, D., Nitta, A., Snedden, S. A., Jan. 2007. Cats and Dogs, Hair Redshift: Evidence for Smooth Accretion in Galaxy Formation. and a Hero: A Quintet of New Milky Way Companions. ApJ654, ApJ694, L158–L161. 897–906. Bournaud, F., Powell, L. C., Chapon, D., Teyssier, R., Aug. 2011. Star Benincasa, S. M., Tasker, E. J., Pudritz, R. E., Wadsley, J., Oct. 2013. formation in galaxy mergers: ISM turbulence, dense gas excess, Giant Molecular Cloud Formation in Disk Galaxies: Characteriz- and scaling relations for disks and starbusts. In: IAU Symp. 271, ing Simulated versus Observed Cloud Catalogs. ApJ776, 23. Star Formation in Galaxy Mergers: ISM Turbulence, Dense Gas Berentzen, I., Athanassoula, E., Feb. 2012. Star cluster evolution in Excess, and Scaling Relations for Disks and Starbursts, ed. N. H. barred disc galaxies - I. Planar periodic orbits. MNRAS419, 3244– Brummell, A. S. Brun, M. S. Miesch, & Y. Ponty (Cambridge: 3257. Cambridge Univ. Press). Vol. 271 of IAU Symposium. p. 160. Bertin, G., Romeo, A. B., Apr. 1988. Global spiral modes in stellar Boutloukos, S. G., Lamers, H. J. G. L. M., Jan. 2003. Star cluster for- disks containing gas. A&A195, 105–113. mation and disruption time-scales - I. An empirical determination Bertin, G., Varri, A. L., Dec. 2008. The Construction of Nonspherical of the disruption time of star clusters in four galaxies. MNRAS338, Models of Quasi-Relaxed Stellar Systems. ApJ689, 1005–1019. 717–732. Bian, F., Fan, X., McGreer, I., Cai, Z., Jiang, L., Mar. 2017. High Bouwens, R. J., Illingworth, G. D., Oesch, P. A., Caruana, J., Hol- Lyman Continuum Escape Fraction in a Lensed Young Compact werda, B., Smit, R., Wilkins, S., Oct. 2015. Reionization After Dwarf Galaxy at z = 2.5. ApJ837, L12. Planck: The Derived Growth of the Cosmic Ionizing Emissivity Bianchini, P., Renaud, F., Gieles, M., Varri, A. L., Feb. 2015. The Now Matches the Growth of the Galaxy UV Luminosity Density. inefficiency of satellite accretion in forming extended star clusters. ApJ811, 140. MNRAS447, L40–L44. Bovy, J., Erkal, D., Sanders, J. L., Apr. 2017. Linear perturbation

45 theory for tidal streams and the small-scale CDM power spectrum. Cappellari, M., McDermid, R. M., Alatalo, K., Blitz, L., Bois, M., MNRAS466, 628–668. Bournaud, F., Bureau, M., Crocker, A. F., Davies, R. L., Davis, Boylan-Kolchin, M., Oct. 2017. The Little Engines That Could? T. A., de Zeeuw, P. T., Duc, P.-A., Emsellem, E., Khochfar, S., Kra- Globular Clusters Contribute Significantly to Reionization-era Star jnovic,´ D., Kuntschner, H., Lablanche, P.-Y., Morganti, R., Naab, Formation. ArXiv e-prints. T., Oosterloo, T., Sarzi, M., Scott, N., Serra, P., Weijmans, A.-M., Bragaglia, A., Sneden, C., Carretta, E., Gratton, R. G., Lucatello, S., Young, L. M., Apr. 2012. Systematic variation of the stellar initial Bernath, P. F., Brooke, J. S. A., Ram, R. S., Nov. 2014. Searching mass function in early-type galaxies. Nature484, 485–488. for Chemical Signatures of Multiple Stellar Populations in the Old, Capriotti, E. R., Hawley, S. L., Jun. 1996. Evaporation, Tidal Dis- Massive Open Cluster NGC 6791. ApJ796, 68. ruption, and Orbital Decay of Star Clusters in a Galactic Halo. Bressert, E., Bastian, N., Gutermuth, R., Megeath, S. T., Allen, L., ApJ464, 765. Evans, II, N. J., Rebull, L. M., Hatchell, J., Johnstone, D., Bourke, Capuzzo Dolcetta, R., Di Matteo, P., Miocchi, P., Apr. 2005. Forma- T. L., Cieza, L. A., Harvey, P. M., Merin, B., Ray, T. P., Tothill, tion and Evolution of Clumpy Tidal Tails around Globular Clus- N. F. H., Nov. 2010. The spatial distribution of star formation in ters. AJ129, 1906–1921. the solar neighbourhood: do all stars form in dense clusters? MN- Capuzzo-Dolcetta, R., Miocchi, P., Jul. 2008. Self-consistent simula- RAS409, L54–L58. tions of nuclear cluster formation through globular cluster orbital Brockamp, M., Kupper,¨ A. H. W., Thies, I., Baumgardt, H., Kroupa, decay and merging. MNRAS388, L69–L73. P., Jun. 2014. Erosion of globular cluster systems: the influence Carlberg, R. G., Mar. 2017. Star Streams and the Assembly History of of radial anisotropy, central black holes and dynamical friction. the Galaxy. ApJ838, 39. MNRAS441, 150–171. Carlberg, R. G., Grillmair, C. J., Hetherington, N., Nov. 2012. The Pal Brodie, J. P., Romanowsky, A. J., Strader, J., Forbes, D. A., Dec. 5 Star Stream Gaps. ApJ760, 75. 2011. The Relationships among Compact Stellar Systems: A Fresh Carlberg, R. G., Richer, H. B., McConnachie, A. W., Irwin, M., Ibata, View of Ultracompact Dwarfs. AJ142, 199. R. A., Dotter, A. L., Chapman, S., Fardal, M., Ferguson, A. M. N., Brodie, J. P., Romanowsky, A. J., Strader, J., Forbes, D. A., Foster, Lewis, G. F., Navarro, J. F., Puzia, T. H., Valls-Gabaud, D., Apr. C., Jennings, Z. G., Pastorello, N., Pota, V., Usher, C., Blom, C., 2011. Density Variations in the NW Star Stream of M31. ApJ731, Kader, J., Roediger, J. C., Spitler, L. R., Villaume, A., Arnold, 124. J. A., Kartha, S. S., Woodley, K. A., Nov. 2014. The SAGES Carretta, E., Bragaglia, A., Gratton, R. G., D’Orazi, V., Lucatello, S., Legacy Unifying Globulars and GalaxieS Survey (SLUGGS): Sollima, A., Jan. 2014. Terzan 8: a Sagittarius-flavoured globular Sample Definition, Methods, and Initial Results. ApJ796, 52. cluster. A&A561, A87. Brodie, J. P., Strader, J., Sep. 2006. Extragalactic Globular Clusters Carretta, E., Bragaglia, A., Gratton, R. G., Lucatello, S., Catanzaro, and Galaxy Formation. ARA&A44, 193–267. G., Leone, F., Bellazzini, M., Claudi, R., D’Orazi, V., Momany, Y., Bromm, V., Clarke, C. J., Feb. 2002. The Formation of the First Glob- Ortolani, S., Pancino, E., Piotto, G., Recio-Blanco, A., Sabbi, E., ular Clusters in Dwarf Galaxies before the Epoch of Reionization. Oct. 2009. Na-O anticorrelation and HB. VII. The chemical com- ApJ566, L1–L4. position of first and second-generation stars in 15 globular clusters Brosche, P., Odenkirchen, M., Geffert, M., Mar. 1999. Instantaneous from GIRAFFE spectra. A&A505, 117–138. and average tidal radii of globular clusters. New Astronomy 4, Carretta, E., Bragaglia, A., Gratton, R. G., Recio-Blanco, A., Lu- 133–139. catello, S., D’Orazi, V., Cassisi, S., Jun. 2010. Properties of stellar Brown, T. M., Tumlinson, J., Geha, M., Simon, J. D., Vargas, L. C., generations in globular clusters and relations with global parame- VandenBerg, D. A., Kirby, E. N., Kalirai, J. S., Avila, R. J., Gen- ters. A&A516, A55. naro, M., Ferguson, H. C., Munoz,˜ R. R., Guhathakurta, P., Ren- Casetti-Dinescu, D. I., Girard, T. M., J´ılkova,´ L., van Altena, W. F., zini, A., Dec. 2014. The Quenching of the Ultra-faint Dwarf Galax- Podesta,´ F., Lopez,´ C. E., Aug. 2013. Space Velocities of Southern ies in the Reionization Era. ApJ796, 91. Globular Clusters. VII. NGC 6397, NGC 6626 (M28), and NGC Bruns,¨ R. C., Kroupa, P., Fellhauer, M., Metz, M., Assmann, P., May 6656 (M22). AJ146, 33. 2011. A parametric study on the formation of extended star clusters Caso, J. P., Bassino, L. P., Richtler, T., Calderon,´ J. P., Smith Castelli, and ultra-compact dwarf galaxies. A&A529, A138. A. V., Jul. 2014. Ultracompact dwarfs around NGC 3268. MN- Bullock, J. S., Sep. 2010. Notes on the Missing Satellites Problem. RAS442, 891–899. ArXiv e-prints. Cen, R., Oct. 2001. Synchronized Formation of Subgalactic Systems Bundy, K., Bershady, M. A., Law, D. R., Yan, R., Drory, N., Mac- at Cosmological Reionization: Origin of Halo Globular Clusters. Donald, N., Wake, D. A., Cherinka, B., et al., Jan. 2015. Overview ApJ560, 592–598. of the SDSS-IV MaNGA Survey: Mapping nearby Galaxies at Chabrier, G., Hennebelle, P., Charlot, S., Dec. 2014. Variations of the Apache Point Observatory. ApJ798, 7. Stellar Initial Mass Function in the Progenitors of Massive Early- Burkhart, B., Lazarian, A., Ossenkopf, V., Stutzki, J., Jul. 2013. The type Galaxies and in Extreme Starburst Environments. ApJ796, 75. Turbulence Power Spectrum in Optically Thick Interstellar Clouds. Chandar, R., Fall, S. M., Whitmore, B. C., Mar. 2010. New Tests ApJ771, 123. for Disruption Mechanisms of Star Clusters: The Large and Small C¸alıs¸kan, S¸., Christlieb, N., Grebel, E. K., Jan. 2012. Abundance anal- Magellanic Clouds. ApJ711, 1263–1279. ysis of the outer halo globular cluster Palomar 14. A&A537, A83. Chandar, R., Fall, S. M., Whitmore, B. C., Mulia, A. J., Nov. 2017. Cai, M. X., Gieles, M., Heggie, D. C., Varri, A. L., Jan. 2016. Evolu- The Fraction of Stars That Form in Clusters in Different Galaxies. tion of star clusters on eccentric orbits. MNRAS455, 596–602. ApJ849, 128. Caldwell, N., Romanowsky, A. J., Jun. 2016. Star Clusters in M31. Chandar, R., Whitmore, B. C., Calzetti, D., Di Nino, D., Kennicutt, VII. Global Kinematics and Metallicity Subpopulations of the R. C., Regan, M., Schinnerer, E., Feb. 2011. New Constraints on Globular Clusters. ApJ824, 42. Mass-dependent Disruption of Star Clusters in M51. ApJ727, 88. Cantiello, M., Brocato, E., Blakeslee, J. P., Aug. 2009. The star cluster Chandrasekhar, S., 1942. Principles of stellar dynamics. Physical Sci- population of the spiral galaxy NGC 3370. A&A503, 87–101. ences Data. Capaccioli, M., Piotto, G., Stiavelli, M., Apr. 1993. Disc Shocking Chandrasekhar, S., Mar. 1943. Dynamical Friction. I. General Con- and the Mass Function of Galactic Globular Clusters. MNRAS261, siderations: the Coefficient of Dynamical Friction. ApJ97, 255. 819. Charbonnel, C., Nov. 2016. Multiple Stellar Populations and Their

46 Evolution in Globular Clusters: A Nucleosynthesis Perspective. In: arty, L., Richards, S., Goodwin, M., et al., Mar. 2012. The Sydney- Moraux, E., Lebreton, Y., Charbonnel, C. (Eds.), EAS Publications AAO Multi-object Integral field spectrograph. MNRAS421, 872– Series. Vol. 80 of EAS Publications Series. pp. 177–226. 893. Chernoff, D. F., Kochanek, C. S., Shapiro, S. L., Oct. 1986. Tidal Da Costa, G. S., May 2012. The Dynamics of the Outer Parts of ω heating of globular clusters. ApJ309, 183–209. Centauri. ApJ751, 6. Chernoff, D. F., Weinberg, M. D., Mar. 1990. Evolution of globular Dabringhausen, J., Hilker, M., Kroupa, P., May 2008. From star clus- clusters in the Galaxy. ApJ351, 121–156. ters to dwarf galaxies: the properties of dynamically hot stellar Chiaberge, M., Gilli, R., Lotz, J. M., Norman, C., Jun. 2015. Radio systems. MNRAS386, 864–886. Loud AGNs are Mergers. ApJ806, 147. Dabringhausen, J., Kroupa, P., Baumgardt, H., Apr. 2009. A top- Chien, L.-H., Barnes, J. E., Sep. 2010. Dynamically driven star for- heavy stellar initial mass function in starbursts as an explanation mation in models of NGC 7252. MNRAS407, 43–54. for the high mass-to-light ratios of ultra-compact dwarf galaxies. Clark, P. C., Glover, S. C. O., Klessen, R. S., Jan. 2008. The First MNRAS394, 1529–1543. Stellar Cluster. ApJ672, 757–764. Dabringhausen, J., Kroupa, P., Pflamm-Altenburg, J., Mieske, S., Mar. Clarkson, W. I., Ghez, A. M., Morris, M. R., Lu, J. R., Stolte, A., 2012. Low-mass X-Ray Binaries Indicate a Top-heavy Stellar Ini- McCrady, N., Do, T., Yelda, S., Jun. 2012. Proper Motions of the tial Mass Function in Ultracompact Dwarf Galaxies. ApJ747, 72. Arches Cluster with Keck Laser Guide Star Adaptive Optics: The Daddi, E., Bournaud, F., Walter, F., Dannerbauer, H., Carilli, C. L., First Kinematic Mass Measurement of the Arches. ApJ751, 132. Dickinson, M., Elbaz, D., Morrison, G. E., et al., Apr. 2010a. Very Claydon, I., Gieles, M., Zocchi, A., May 2017. The properties of en- High Gas Fractions and Extended Gas Reservoirs in z = 1.5 Disk ergetically unbound stars in stellar clusters. MNRAS466, 3937– Galaxies. ApJ713, 686–707. 3950. Daddi, E., Elbaz, D., Walter, F., Bournaud, F., Salmi, F., Carilli, C., Cohen, J. G., Jul. 1978. Abundances in globular cluster red giants. I - Dannerbauer, H., Dickinson, M., et al., May 2010b. Different Star M3 and M13. ApJ223, 487–508. Formation Laws for Disks Versus Starbursts at Low and High Red- Cohen, J. G., Christlieb, N., McWilliam, A., Shectman, S., Thomp- shifts. ApJ714, L118–L122. son, I., Wasserburg, G. J., Ivans, I., Dehn, M., Karlsson, T., Melen- Dale, J. E., Oct. 2015. The modelling of feedback in star formation dez, J., Sep. 2004. Abundances In Very Metal-Poor Dwarf Stars. simulations. New A Rev.68, 1–33. ApJ612, 1107–1135. Dale, J. E., Ercolano, B., Bonnell, I. A., Jul. 2012. Ionizing feedback Cohn, H., Dec. 1979. Numerical integration of the Fokker-Planck from massive stars in massive clusters - II. Disruption of bound equation and the evolution of star clusters. ApJ234, 1036–1053. clusters by photoionization. MNRAS424, 377–392. Cole, A. A., Weinberg, M. D., Jul. 2002. An Upper Limit to the Age Dame, T. M., Hartmann, D., Thaddeus, P., Feb. 2001. The Milky Way of the Galactic Bar. ApJ574, L43–L46. in Molecular Clouds: A New Complete CO Survey. ApJ547, 792– Cole, D. R., Debattista, V. P., 2016. Nuclear Star Clusters and Bulges. 813. Galactic Bulges 418, 107. Daniel, K. J., Heggie, D. C., Varri, A. L., Jun. 2017. An approxi- Cole, D. R., Dehnen, W., Read, J. I., Wilkinson, M. I., Oct. 2012. The mate analytic model of a star cluster with potential escapers. MN- mass distribution of the Fornax dSph: constraints from its globular RAS468, 1453–1473. cluster distribution. MNRAS426, 601–613. Dave,´ R., Spergel, D. N., Steinhardt, P. J., Wandelt, B. D., Feb. 2001. Colombo, D., Hughes, A., Schinnerer, E., Meidt, S. E., Leroy, A. K., Halo Properties in Cosmological Simulations of Self-interacting Pety, J., Dobbs, C. L., Garc´ıa-Burillo, S., Dumas, G., Thompson, Cold Dark Matter. ApJ547, 574–589. T. A., Schuster, K. F., Kramer, C., Mar. 2014. The PdBI Arcsecond Davies, B., Bastian, N., Gieles, M., Seth, A. C., Mengel, S., Kon- Whirlpool Survey (PAWS): Environmental Dependence of Giant stantopoulos, I. S., Feb. 2011. GLIMPSE-CO1: the most massive Molecular Cloud Properties in M51. ApJ784, 3. intermediate-age stellar cluster in the Galaxy. MNRAS411, 1386– Combes, F., Leon, S., Meylan, G., Dec. 1999. N-body simulations of 1394. globular cluster tides. A&A352, 149–162. Davies, B., de La Fuente, D., Najarro, F., Hinton, J. A., Trombley, Conroy, C., Loeb, A., Spergel, D. N., Nov. 2011. Evidence against C., Figer, D. F., Puga, E., Jan. 2012. A newly discovered young Dark Matter Halos Surrounding the Globular Clusters MGC1 and massive star cluster at the far end of the Galactic Bar. MNRAS419, NGC 2419. ApJ741, 72. 1860–1870. Contenta, F., Gieles, M., Balbinot, E., Collins, M. L. M., Apr. 2017. Davies, M. B., Amaro-Seoane, P., Bassa, C., Dale, J., De Angeli, The contribution of dissolving star clusters to the population of F., Freitag, M., Kroupa, P., Mackey, D., Miller, M. C., Portegies ultra faint objects in the outer halo of the Milky Way. MNRAS466, Zwart, S., Dec. 2006. The MODEST questions: Challenges and 1741–1756. future directions in stellar cluster research. New A12, 201–214. Cotˆ e,´ P., Marzke, R. O., West, M. J., Jul. 1998. The Formation of Giant de Grijs, R., Anders, P., Feb. 2006. How well do we know the age and Elliptical Galaxies and Their Globular Cluster Systems. ApJ501, mass distributions of the star cluster system in the Large Magel- 554–570. lanic Cloud? MNRAS366, 295–307. Cotˆ e,´ P., Piatek, S., Ferrarese, L., Jordan,´ A., Merritt, D., Peng, E. W., de Grijs, R., Lee, J. T., Clemencia Mora Herrera, M., Fritze-v. Al- Has¸egan, M., Blakeslee, J. P., Mei, S., West, M. J., Milosavljevic,´ vensleben, U., Anders, P., Feb. 2003. Stellar populations and star M., Tonry, J. L., Jul. 2006. The ACS Virgo Cluster Survey. VIII. cluster formation in interacting galaxies with the Advanced Cam- The Nuclei of Early-Type Galaxies. ApJS165, 57–94. era for Surveys. New A8, 155–171. Couchman, H. M. P., Thomas, P. A., Pearce, F. R., Oct. 1995. Hydra: de la Fuente Marcos, R., de la Fuente Marcos, C., Jul. 2009. Hierar- an Adaptive-Mesh Implementation of P 3M-SPH. ApJ452, 797. chical Star Formation in the Milky Way Disk. ApJ700, 436–446. Cowie, L. L., Songaila, A., Hu, E. M., Cohen, J. G., Sep. 1996. De Marchi, G., Paresce, F., Panagia, N., Beccari, G., Spezzi, L., Siri- New Insight on Galaxy Formation and Evolution From Keck Spec- anni, M., Andersen, M., Mutchler, M., Balick, B., Dopita, M. A., troscopy of the Hawaii Deep Fields. AJ112, 839. Frogel, J. A., Whitmore, B. C., Bond, H., Calzetti, D., Carollo, Cox, T. J., Jonsson, P., Primack, J. R., Somerville, R. S., Dec. C. M., Disney, M. J., Hall, D. N. B., Holtzman, J. A., Kimble, 2006. Feedback in simulations of disc-galaxy major mergers. MN- R. A., McCarthy, P. J., O’Connell, R. W., Saha, A., Silk, J. I., RAS373, 1013–1038. Trauger, J. T., Walker, A. R., Windhorst, R. A., Young, E. T., Sep. Croom, S. M., Lawrence, J. S., Bland-Hawthorn, J., Bryant, J. J., Fog- 2011. Star Formation in 30 Doradus. ApJ739, 27.

47 Deason, A. J., Belokurov, V., Evans, N. W., Johnston, K. V., Feb. structure Depletion in the Milky Way Halo by the Disk. ApJ709, 2013. Broken and Unbroken: The Milky Way and M31 Stellar Ha- 1138–1147. los. ApJ763, 113. Donovan Meyer, J., Koda, J., Momose, R., Mooney, T., Egusa, F., Dehnen, W., Jun. 2000. A Very Fast and Momentum-conserving Tree Carty, M., Kennicutt, R., Kuno, N., Rebolledo, D., Sawada, T., Code. ApJ536, L39–L42. Scoville, N., Wong, T., Aug. 2013. Resolved Giant Molecular Dehnen, W., Jun. 2001. Towards optimal softening in three- Clouds in Nearby Spiral Galaxies: Insights from the CANON CO dimensional N-body codes - I. Minimizing the force error. MN- (1-0) Survey. ApJ772, 107. RAS324, 273–291. Dopcke, G., Glover, S. C. O., Clark, P. C., Klessen, R. S., Apr. 2013. Dehnen, W., Dec. 2014. A fast multipole method for stellar dynamics. On the Initial Mass Function of Low-metallicity Stars: The Impor- Computational Astrophysics and Cosmology 1, 1. tance of Dust Cooling. ApJ766, 103. Dehnen, W., Odenkirchen, M., Grebel, E. K., Rix, H.-W., May 2004. Dotter, A., Sarajedini, A., Anderson, J., Sep. 2011. Globular Clus- Modeling the Disruption of the Globular Cluster Palomar 5 by ters in the Outer Galactic Halo: New Hubble Space Tele- Galactic Tides. AJ127, 2753–2770. scope/Advanced Camera for Surveys Imaging of Six Globular Dekel, A., Devor, J., Hetzroni, G., May 2003. Galactic halo cusp-core: Clusters and the Galactic Globular Cluster Age-metallicity Rela- tidal compression in mergers. MNRAS341, 326–342. tion. ApJ738, 74. Dekel, A., Sari, R., Ceverino, D., Sep. 2009. Formation of Mas- Drinkwater, M. J., Gregg, M. D., Hilker, M., Bekki, K., Couch, W. J., sive Galaxies at High Redshift: Cold Streams, Clumpy Disks, and Ferguson, H. C., Jones, J. B., Phillipps, S., May 2003. A class of Compact Spheroids. ApJ703, 785–801. compact dwarf galaxies from disruptive processes in galaxy clus- den Brok, M., Peletier, R. F., Seth, A., Balcells, M., Dominguez, L., ters. Nature423, 519–521. Graham, A. W., Carter, D., Erwin, P., Ferguson, H. C., Goudfrooij, Drukier, G. A., Slavin, S. D., Cohn, H. N., Lugger, P. M., Berrington, P., Guzman,´ R., Hoyos, C., Jogee, S., Lucey, J., Phillipps, S., R. C., Murphy, B. W., Seitzer, P. O., Feb. 1998. Global kinematics Puzia, T., Valentijn, E., Kleijn, G. V., Weinzirl, T., Dec. 2014. The of the globular cluster M15. AJ115, 708. HST/ACS Coma Cluster Survey - X. Nuclear star clusters in low- Dubinski, J., Carlberg, R. G., Sep. 1991. The structure of cold dark mass early-type galaxies: scaling relations. MNRAS445, 2385– matter halos. ApJ378, 496–503. 2403. Duc, P.-A., Bournaud, F., Masset, F., Dec. 2004. A top-down scenario Denisenkov, P. A., Denisenkova, S. N., Jul. 1990. Correlation Be- for the formation of massive Tidal Dwarf Galaxies. A&A427, 803– tween the Abundances of NA and the CNO Elements in Red Giants 814. in Omega-Centauri. Soviet Astronomy Letters 16, 275. Duc, P.-A., Mirabel, I. F., 1999. Tidal Dwarf Galaxies. In: Barnes, Dessauges-Zavadsky, M., Schaerer, D., Cava, A., Mayer, L., Tam- J. E., Sanders, D. B. (Eds.), Galaxy Interactions at Low and High burello, V., Feb. 2017. On the Stellar Masses of Giant Clumps in Redshift. Vol. 186 of IAU Symposium. p. 61. Distant Star-forming Galaxies. ApJ836, L22. Duc, P.-A., Paudel, S., McDermid, R. M., Cuillandre, J.-C., Serra, P., Di Matteo, P., Combes, F., Melchior, A., Semelin, B., Jun. 2007. Star Bournaud, F., Cappellari, M., Emsellem, E., May 2014. Identifica- formation efficiency in galaxy interactions and mergers: a statisti- tion of old tidal dwarfs near early-type galaxies from deep imaging cal study. A&A468, 61–81. and H I observations. MNRAS440, 1458–1469. Di Matteo, T., Springel, V., Hernquist, L., Feb. 2005. Energy input Duc, P.-A., Renaud, F., 2013. Tides in Colliding Galaxies. In: from quasars regulates the growth and activity of black holes and Souchay, J., Mathis, S., Tokieda, T. (Eds.), Lecture Notes in their host galaxies. Nature433, 604–607. Physics, Berlin Springer Verlag. Vol. 861. p. 327. Dib, S., Gutkin, J., Brandner, W., Basu, S., Dec. 2013. Feedback- Eden, D. J., Moore, T. J. T., Morgan, L. K., Thompson, M. A., regulated star formation - II. Dual constraints on the SFE and the Urquhart, J. S., May 2013. Star formation in Galactic spiral arms age spread of stars in massive clusters. MNRAS436, 3727–3740. and the interarm regions. MNRAS431, 1587–1595. Dib, S., Helou, G., Moore, T. J. T., Urquhart, J. S., Dariush, A., Oct. Eden, D. J., Moore, T. J. T., Plume, R., Morgan, L. K., Jun. 2012. 2012. The Lesser Role of Shear in Galactic Star Formation: Insight Star formation towards the Scutum tangent region and the effects from the Galactic Ring Survey. ApJ758, 125. of Galactic environment. MNRAS422, 3178–3188. Dib, S., Schmeja, S., Hony, S., Jan. 2017. Massive stars reveal vari- Einsel, C., Spurzem, R., Jan. 1999. Dynamical evolution of rotating ations of the stellar initial mass function in the Milky Way stellar stellar systems - I. Pre-collapse, equal-mass system. MNRAS302, clusters. MNRAS464, 1738–1752. 81–95. Diemand, J., Madau, P., Moore, B., Dec. 2005. The distribution and Ellison, S. L., Patton, D. R., Simard, L., McConnachie, A. W., May kinematics of early high-σ peaks in present-day haloes: impli- 2008. Galaxy Pairs in the Sloan Digital Sky Survey. I. Star Forma- cations for rare objects and old stellar populations. MNRAS364, tion, Active Galactic Nucleus Fraction, and the Mass-Metallicity 367–383. Relation. AJ135, 1877–1899. Dinescu, D. I., Girard, T. M., van Altena, W. F., Apr. 1999. Space Elmegreen, B. G., Aug. 1995. An effective Q parameter for two-fluid Velocities of Globular Clusters. III. Cluster Orbits and Halo Sub- instabilities in spiral galaxies. MNRAS275, 944–950. structure. AJ117, 1792–1815. Elmegreen, B. G., Jan. 2008. Variations in Stellar Clustering with Djorgovski, S., Piotto, G., Capaccioli, M., Jun. 1993. What deter- Environment: Dispersed Star Formation and the Origin of Faint mines the stellar mass functions in globular clusters? AJ105, Fuzzies. ApJ672, 1006–1012. 2148–2157. Elmegreen, B. G., Apr. 2010. The Globular Cluster Mass Function as Dobbs, C. L., Bonnell, I. A., Apr. 2006. Spurs and feathering in spiral a Remnant of Violent Birth. ApJ712, L184–L188. galaxies. MNRAS367, 873–878. Elmegreen, B. G., 2015. Formation of stars and clusters over cosmo- Dobbs, C. L., Burkert, A., Pringle, J. E., Jun. 2011. Why are most logical time. In: Freeman, K., Elmegreen, B., Block, D., Woolway, molecular clouds not gravitationally bound? MNRAS413, 2935– M. (Eds.), Lessons from the Local Group: A Conference in Hon- 2942. our of David Block and Bruce Elmegreen. Springer International Dobbs, C. L., Pringle, J. E., Burkert, A., Sep. 2012. Giant molecu- Publishing. p. 477. lar clouds: what are they made from, and how do they get there? Elmegreen, B. G., Efremov, Y. N., May 1997. A Universal Forma- MNRAS425, 2157–2168. tion Mechanism for Open and Globular Clusters in Turbulent Gas. D’Onghia, E., Springel, V., Hernquist, L., Keres, D., Feb. 2010. Sub- ApJ480, 235.

48 Elmegreen, B. G., Efremov, Y. N., Larsen, S., Jun. 2000. A Young internal evolution. MNRAS375, 604–614. Globular Cluster in the Galaxy NGC 6946. ApJ535, 748–758. Fensch, J., Renaud, F., Bournaud, F., Duc, P.-A., Agertz, O., Amram, Elmegreen, B. G., Elmegreen, D. M., Apr. 1983. Regular strings of H P., Combes, F., Di Matteo, P., et al., Feb. 2017. High-redshift major II regions and superclouds in spiral galaxies - Clues to the origin mergers weakly enhance star formation. MNRAS465, 1934–1949. of cloudy structure. MNRAS203, 31–45. Ferrarese, L., Cotˆ e,´ P., Jordan,´ A., Peng, E. W., Blakeslee, J. P., Piatek, Elmegreen, B. G., Elmegreen, D. M., Jul. 2005. Stellar Populations S., Mei, S., Merritt, D., Milosavljevic,´ M., Tonry, J. L., West, M. J., in 10 Clump-Cluster Galaxies of the Hubble Ultra Deep Field. Jun. 2006. The ACS Virgo Cluster Survey. VI. Isophotal Analysis ApJ627, 632–646. and the Structure of Early-Type Galaxies. ApJS164, 334–434. Elmegreen, B. G., Falgarone, E., Nov. 1996. A Fractal Origin for the Ferreras, I., Saha, P., Williams, L. L. R., Apr. 2005. Stellar and Total Mass Spectrum of Interstellar Clouds. ApJ471, 816. Mass in Early-Type Lensing Galaxies. ApJ623, L5–L8. Elmegreen, B. G., Hunter, D. A., Mar. 2010. On the Disruption of Star Ferriere,` K., Gillard, W., Jean, P., May 2007. Spatial distribution of Clusters in a Hierarchical Interstellar Medium. ApJ712, 604–623. interstellar gas in the innermost 3 kpc of our galaxy. A&A467, Elmegreen, B. G., Kaufman, M., Bournaud, F., Elmegreen, D. M., 611–627. Struck, C., Brinks, E., Juneau, S., May 2016. High Star Forma- Field, G. B., Blackman, E. G., Keto, E. R., Sep. 2011. Does external tion Rates in Turbulent Atomic-dominated Gas in the Interacting pressure explain recent results for molecular clouds? MNRAS416, Galaxies IC 2163 and NGC 2207. ApJ823, 26. 710–714. Elmegreen, B. G., Malhotra, S., Rhoads, J., Sep. 2012. Formation Figer, D. F., Najarro, F., Gilmore, D., Morris, M., Kim, S. S., Serabyn, of Metal-poor Globular Clusters in Lyα Emitting Galaxies in the E., McLean, I. S., Gilbert, A. M., Graham, J. R., Larkin, J. E., Early Universe. ApJ757, 9. Levenson, N. A., Teplitz, H. I., Dec. 2002. Massive Stars in the Elmegreen, D. M., Elmegreen, B. G., Jun. 1995. Inner two-arm sym- Arches Cluster. ApJ581, 258–275. metry in spiral galaxies. ApJ445, 591–598. Finkelstein, S. L., Ryan, Jr., R. E., Papovich, C., Dickinson, M., Song, Elmegreen, D. M., Elmegreen, B. G., Kaufman, M., Brinks, E., M., Somerville, R. S., Ferguson, H. C., Salmon, B., Giavalisco, Struck, C., Bournaud, F., Sheth, K., Juneau, S., May 2017. ALMA M., Koekemoer, A. M., Ashby, M. L. N., Behroozi, P., Castellano, CO Clouds and Young Star Complexes in the Interacting Galaxies M., Dunlop, J. S., Faber, S. M., Fazio, G. G., Fontana, A., Gro- IC 2163 and NGC 2207. ApJ841, 43. gin, N. A., Hathi, N., Jaacks, J., Kocevski, D. D., Livermore, R., Elmegreen, D. M., Elmegreen, B. G., Ravindranath, S., Coe, D. A., McLure, R. J., Merlin, E., Mobasher, B., Newman, J. A., Rafel- Apr. 2007. Resolved Galaxies in the Hubble Ultra Deep Field: Star ski, M., Tilvi, V., Willner, S. P., Sep. 2015. The Evolution of the Formation in Disks at High Redshift. ApJ658, 763–777. Galaxy Rest-frame Ultraviolet Luminosity Function over the First Emsellem, E., Renaud, F., Bournaud, F., Elmegreen, B., Combes, F., Two Billion Years. ApJ810, 71. Gabor, J. M., Jan. 2015. The interplay between a galactic bar and a Forbes, D. A., Dec. 2005. Bimodal Galaxies and Bimodality in Glob- supermassive black hole: nuclear fuelling in a subparsec resolution ular Cluster Systems. ApJ635, L137–L140. galaxy simulation. MNRAS446, 2468–2482. Forbes, D. A., Nov. 2017. How large are the globular cluster systems Erkal, D., Belokurov, V., Bovy, J., Sanders, J. L., Nov. 2016. The of early-type galaxies and do they scale with galaxy halo proper- number and size of subhalo-induced gaps in stellar streams. MN- ties? MNRAS472, L104–L108. RAS463, 102–119. Forbes, D. A., Bridges, T., May 2010. Accreted versus in situ Milky Ernst, A., Glaschke, P., Fiestas, J., Just, A., Spurzem, R., May 2007. Way globular clusters. MNRAS404, 1203–1214. N-body models of rotating globular clusters. MNRAS377, 465– Forbes, D. A., Brodie, J. P., Grillmair, C. J., May 1997. On the Origin 479. of Globular Clusters in Elliptical and cD Galaxies. AJ113, 1652. Errani, R., Penarrubia,˜ J., Tormen, G., Apr. 2015. Constraining the Forbes, D. A., Kroupa, P., Mar. 2011. What Is a Galaxy? Cast Your distribution of dark matter in dwarf spheroidal galaxies with stellar Vote Here. PASA28, 77–82. tidal streams. MNRAS449, L46–L50. Forbes, D. A., Norris, M. A., Strader, J., Romanowsky, A. J., Pota, V., Espinoza, P., Selman, F. J., Melnick, J., Jul. 2009. The massive star Kannappan, S. J., Brodie, J. P., Huxor, A., Nov. 2014. The AIMSS initial mass function of the Arches cluster. A&A501, 563–583. Project II: dynamical-to-stellar mass ratios across the star cluster- Evstigneeva, E. A., Gregg, M. D., Drinkwater, M. J., Hilker, M., Apr. galaxy divide. MNRAS444, 2993–3003. 2007. Internal Properties of Ultracompact Dwarf Galaxies in the Forbes, D. A., Pastorello, N., Romanowsky, A. J., Usher, C., Brodie, Virgo Cluster. AJ133, 1722–1740. J. P., Strader, J., Sep. 2015. The SLUGGS survey: inferring the Faesi, C. M., Lada, C. J., Forbrich, J., Apr. 2016. Resolving Giant formation epochs of metal-poor and metal-rich globular clusters. Molecular Clouds in NGC 300: A First Look with the Submil- MNRAS452, 1045–1051. limeter Array. ApJ821, 125. Forbes, D. A., Pota, V., Usher, C., Strader, J., Romanowsky, A. J., Fall, S. M., Rees, M. J., Nov. 1985. A theory for the origin of globular Brodie, J. P., Arnold, J. A., Spitler, L. R., Aug. 2013. Filling the clusters. ApJ298, 18–26. gap: a new class of old star cluster? MNRAS435, L6–L10. Faucher-Giguere,` C.-A., Quataert, E., Hopkins, P. F., Aug. 2013. Forster¨ Schreiber, N. M., Genzel, R., Bouche,´ N., Cresci, G., Davies, Feedback-regulated star formation in molecular clouds and galac- R., Buschkamp, P., Shapiro, K., Tacconi, L. J., et al., Dec. 2009. tic discs. MNRAS433, 1970–1990. The SINS Survey: SINFONI Integral Field Spectroscopy of z ˜ 2 Federrath, C., Roman-Duval, J., Klessen, R. S., Schmidt, W., Mac Star-forming Galaxies. ApJ706, 1364–1428. Low, M.-M., Mar. 2010. Comparing the statistics of interstellar tur- Fouesneau, M., Johnson, L. C., Weisz, D. R., Dalcanton, J. J., Bell, bulence in simulations and observations. Solenoidal versus com- E. F., Bianchi, L., Caldwell, N., Gouliermis, D. A., Guhathakurta, pressive turbulence forcing. A&A512, A81. P., Kalirai, J., Larsen, S. S., Rix, H.-W., Seth, A. C., Skillman, Fellhauer, M., Kroupa, P., Oct. 2002. The Possible Origin of the Faint E. D., Williams, B. F., May 2014. The Panchromatic Hubble An- Fuzzy Star Clusters in NGC 1023. AJ124, 2006–2011. dromeda Treasury. V. Ages and Masses of the Year 1 Stellar Clus- Fellhauer, M., Kroupa, P., May 2005. A possible formation scenario ters. ApJ786, 117. for the ultramassive cluster W3 in NGC 7252. MNRAS359, 223– Fouesneau, M., Lanc¸on, A., Oct. 2010. Accounting for stochastic fluc- 227. tuations when analysing the integrated light of star clusters. I. First Fellhauer, M., Lin, D. N. C., Feb. 2007. The influence of mass-loss systematics. A&A521, A22+. from a star cluster on its dynamical friction - I. Clusters without Foyle, K., Natale, G., Wilson, C. D., Popescu, C. C., Baes, M., Bendo,

49 G. J., Boquien, M., Boselli, A., Cooray, A., Cormier, D., De dance Distribution in NGC 6791: The First Open(?) Cluster with Looze, I., Fischera, J., Karczewski, O. Ł., Lebouteiller, V., Mad- Multiple Populations. ApJ756, L40. den, S., Pereira-Santaella, M., Smith, M. W. L., Spinoglio, L., Genzel, R., Burkert, A., Bouche,´ N., Cresci, G., Forster¨ Schreiber, Tuffs, R. J., Jul. 2013. Star formation and dust heating in the FIR N. M., Shapley, A., Shapiro, K., Tacconi, L. J., Buschkamp, P., bright sources of M83. MNRAS432, 2182–2207. Cimatti, A., Daddi, E., Davies, R., Eisenhauer, F., Erb, D. K., Foyle, K., Rix, H.-W., Walter, F., Leroy, A. K., Dec. 2010. Arm and Genel, S., Gerhard, O., Hicks, E., Lutz, D., Naab, T., Ott, T., Ra- Interarm Star Formation in Spiral Galaxies. ApJ725, 534–541. bien, S., Renzini, A., Steidel, C. C., Sternberg, A., Lilly, S. J., Nov. Freeman, K., Bland-Hawthorn, J., 2002. The New Galaxy: Signatures 2008. From Rings to Bulges: Evidence for Rapid Secular Galaxy of Its Formation. ARA&A40, 487–537. Evolution at z ˜ 2 from Integral Field Spectroscopy in the SINS Freeman, K. C., Jan. 1993. Globular Clusters and Nucleated Dwarf Survey. ApJ687, 59–77. Ellipticals. In: Smith, G. H., Brodie, J. P. (Eds.), The Globular Genzel, R., Tacconi, L. J., Eisenhauer, F., Forster¨ Schreiber, N. M., Cluster-Galaxy Connection. Vol. 48 of Astronomical Society of the Cimatti, A., Daddi, E., Bouche,´ N., Davies, R., Lehnert, M. D., Pacific Conference Series. p. 608. Lutz, D., Nesvadba, N., Verma, A., Abuter, R., Shapiro, K., Stern- Freeman, P., Rosolowsky, E., Kruijssen, J. M. D., Bastian, N., Adamo, berg, A., Renzini, A., Kong, X., Arimoto, N., Mignoli, M., Aug. A., Jun. 2017. The varying mass distribution of molecular clouds 2006. The rapid formation of a large rotating disk galaxy three bil- across M83. MNRAS468, 1769–1781. lion years after the Big Bang. Nature442, 786–789. Fujii, M., Iwasawa, M., Funato, Y., Makino, J., Dec. 2007. BRIDGE: Genzel, R., Tacconi, L. J., Gracia-Carpio, J., Sternberg, A., Cooper, A Direct-Tree Hybrid N-Body Algorithm for Fully Self-Consistent M. C., Shapiro, K., Bolatto, A., et al., Oct. 2010. A study of the Simulations of Star Clusters and Their Parent Galaxies. PASJ59, gas-star formation relation over cosmic time. MNRAS407, 2091– 1095–. 2108. Fujii, M. S., Baba, J., Nov. 2012. Destruction of star clusters due to Georgiev, I. Y., Boker,¨ T., Jul. 2014. Nuclear star clusters in 228 spiral the radial migration in spiral galaxies. MNRAS427, L16–L20. galaxies in the HST/WFPC2 archive: catalogue and comparison to Fujii, M. S., Portegies Zwart, S., May 2015. The initial mass func- other stellar systems. MNRAS441, 3570–3590. tion of star clusters that form in turbulent molecular clouds. MN- Georgiev, I. Y., Boker,¨ T., Leigh, N., Lutzgendorf,¨ N., Neumayer, RAS449, 726–740. N., Apr. 2016. Masses and scaling relations for nuclear star clus- Fujimoto, Y., Bryan, G. L., Tasker, E. J., Habe, A., Simpson, C. M., ters, and their co-existence with central black holes. MNRAS457, Sep. 2016. GMC evolution in a barred spiral galaxy with star for- 2122–2138. mation and thermal feedback. MNRAS461, 1684–1700. Georgiev, I. Y., Hilker, M., Puzia, T. H., Goudfrooij, P., Baumgardt, Fujimoto, Y., Tasker, E. J., Wakayama, M., Habe, A., Mar. 2014. H., Jun. 2009. Globular cluster systems in nearby dwarf galaxies Do giant molecular clouds care about the galactic structure? MN- - II. Nuclear star clusters and their relation to massive Galactic RAS439, 936–953. globular clusters. MNRAS396, 1075–1085. Fukui, Y., Ohama, A., Hanaoka, N., Furukawa, N., Torii, K., Daw- Georgiev, I. Y., Puzia, T. H., Goudfrooij, P., Hilker, M., Aug. 2010. son, J. R., Mizuno, N., Hasegawa, K., et al., Jan. 2014. Molecu- Globular cluster systems in nearby dwarf galaxies - III. Formation lar Clouds toward the Super Star Cluster NGC 3603 Possible Ev- efficiencies of old globular clusters. MNRAS406, 1967–1984. idence for a Cloud-Cloud Collision in Triggering the Cluster For- Gibbons, S. L. J., Belokurov, V., Evans, N. W., Dec. 2014. ‘Skinny mation. ApJ780, 36. Milky Way please’, says Sagittarius. MNRAS445, 3788–3802. Fukushige, T., Heggie, D. C., Nov. 2000. The time-scale of escape Gieles, M., Apr. 2009. The early evolution of the star cluster mass from star clusters. MNRAS318, 753–761. function. MNRAS394, 2113–2126. Furukawa, N., Dawson, J. R., Ohama, A., Kawamura, A., Mizuno, Gieles, M., Alexander, P. E. R., Lamers, H. J. G. L. M., Baumgardt, N., Onishi, T., Fukui, Y., May 2009. Molecular Clouds Toward H., Jan. 2014. A prescription and fast code for the long-term evo- RCW49 and Westerlund 2: Evidence for Cluster Formation Trig- lution of star clusters - II. Unbalanced and core evolution. MN- gered by Cloud-Cloud Collision. ApJ696, L115–L119. RAS437, 916–929. Gao, Y., Wang, Q. D., Appleton, P. N., Lucas, R. A., Oct. 2003. Non- Gieles, M., Athanassoula, E., Portegies Zwart, S. F., Apr. 2007. The nuclear Hyper/Ultraluminous X-Ray Sources in the Starbursting effect of spiral arm passages on the evolution of stellar clusters. Cartwheel Ring Galaxy. ApJ596, L171–L174. MNRAS376, 809–819. Garrison-Kimmel, S., Wetzel, A., Bullock, J. S., Hopkins, P. F., Gieles, M., Baumgardt, H., Sep. 2008. Lifetimes of tidally limited star Boylan-Kolchin, M., Faucher-Giguere,` C.-A., Keres,ˇ D., Quataert, clusters with different radii. MNRAS389, L28–L32. E., Sanderson, R. E., Graus, A. S., Kelley, T., Oct. 2017. Not so Gieles, M., Heggie, D. C., Zhao, H., Jun. 2011. The life cycle of star lumpy after all: modelling the depletion of dark matter subhaloes clusters in a tidal field. MNRAS413, 2509–2524. by Milky Way-like galaxies. MNRAS471, 1709–1727. Gieles, M., Portegies Zwart, S. F., Jan. 2011. The distinction between Gavagnin, E., Bleuler, A., Rosdahl, J., Teyssier, R., Dec. 2017. Star star clusters and associations. MNRAS410, L6–L7. cluster formation in a turbulent molecular cloud self-regulated by Gieles, M., Portegies Zwart, S. F., Baumgardt, H., Athanassoula, E., photoionization feedback. MNRAS472, 4155–4172. Lamers, H. J. G. L. M., Sipior, M., Leenaarts, J., Sep. 2006. Star Gavagnin, E., Mapelli, M., Lake, G., Sep. 2016. A critical look at cluster disruption by giant molecular clouds. MNRAS371, 793– the merger scenario to explain multiple populations and rotation in 804. iron-complex globular clusters. MNRAS461, 1276–1287. Gieles, M., Renaud, F., Nov. 2016. If it does not kill them, it makes Gebhardt, K., Kissler-Patig, M., Oct. 1999. Globular Cluster Systems. them stronger: collisional evolution of star clusters with tidal I. V-I Color Distributions. AJ118, 1526–1541. shocks. MNRAS463, L103–L107. Geha, M., Brown, T. M., Tumlinson, J., Kalirai, J. S., Simon, J. D., Giersz, M., Jun. 2001. Monte Carlo simulations of star clusters - Kirby, E. N., VandenBerg, D. A., Munoz,˜ R. R., Avila, R. J., II. Tidally limited, multimass systems with stellar evolution. MN- Guhathakurta, P., Ferguson, H. C., Jul. 2013. The Stellar Initial RAS324, 218–230. Mass Function of Ultra-faint Dwarf Galaxies: Evidence for IMF Giersz, M., Heggie, D. C., May 1994. Statistics of N-Body Simu- Variations with Galactic Environment. ApJ771, 29. lations - Part One - Equal Masses Before Core Collapse. MN- Geisler, D., Villanova, S., Carraro, G., Pilachowski, C., Cummings, RAS268, 257–+. J., Johnson, C. I., Bresolin, F., Sep. 2012. The Unique Na:O Abun- Giersz, M., Heggie, D. C., Hurley, J. R., Jul. 2008. Monte Carlo sim-

50 ulations of star clusters - IV. Calibration of the Monte Carlo code ulations in globular clusters. Lessons learned from the Milky Way and comparison with observations for the open cluster M67. MN- globular clusters. A&A Rev.20, 50. RAS388, 429–443. Grazian, A., Giallongo, E., Gerbasi, R., Fiore, F., Fontana, A., Le Gilmore, G., Wilkinson, M. I., Wyse, R. F. G., Kleyna, J. T., Koch, A., Fevre,` O., Pentericci, L., Vanzella, E., et al., Jan. 2016. The Lyman Evans, N. W., Grebel, E. K., Jul. 2007. The Observed Properties of continuum escape fraction of galaxies at z = 3.3 in the VUDS- Dark Matter on Small Spatial Scales. ApJ663, 948–959. LBC/COSMOS field. A&A585, A48. Ginsburg, A., Walsh, A., Henkel, C., Jones, P. A., Cunningham, Grillmair, C. J., Jul. 2006. Detection of a 60deg-long Dwarf Galaxy M., Kauffmann, J., Pillai, T., Mills, E. A. C., Ott, J., Kruijssen, Debris Stream. ApJ645, L37–L40. J. M. D., Menten, K. M., Battersby, C., Rathborne, J., Contreras, Grisdale, K., Agertz, O., Renaud, F., Romeo, A. B., Jan. 2018. Physi- Y., Longmore, S., Walker, D., Dawson, J., Lopez, J. A. P., Dec. cal properties and scaling relations of molecular clouds: the effect 2015. High-mass star-forming cloud G0.38+0.04 in the Galactic of stellar feedback. ArXiv e-prints. center dust ridge contains H2CO and SiO masers. A&A584, L7. Grisdale, K., Agertz, O., Romeo, A. B., Renaud, F., Read, J. I., Apr. Glatt, K., Grebel, E. K., Sabbi, E., Gallagher, III, J. S., Nota, 2017. The impact of stellar feedback on the density and velocity A., Sirianni, M., Clementini, G., Tosi, M., Harbeck, D., Koch, structure of the interstellar medium. MNRAS466, 1093–1110. A., Kayser, A., Da Costa, G., Oct. 2008. Age Determination of Gualandris, A., Merritt, D., May 2008. Ejection of Supermassive Six Intermediate-Age Small Magellanic Cloud Star Clusters with Black Holes from Galaxy Cores. ApJ678, 780–797. HST/ACS. AJ136, 1703–1727. Guillard, N., Emsellem, E., Renaud, F., Oct. 2016. New insights on Glover, S. C. O., Clark, P. C., Jan. 2014. Molecular cooling in the the formation of nuclear star clusters. MNRAS461, 3620–3629. diffuse interstellar medium. MNRAS437, 9–20. Habibi, M., Stolte, A., Brandner, W., Hußmann, B., Motohara, K., Glover, S. C. O., Clark, P. C., Mar. 2016. Is atomic carbon a good Aug. 2013. The Arches cluster out to its tidal radius: dynamical tracer of molecular gas in metal-poor galaxies? MNRAS456, mass segregation and the effect of the extinction law on the stellar 3596–3609. mass function. A&A556, A26. Gnedin, O. Y., Hernquist, L., Ostriker, J. P., Mar. 1999a. Tidal Shock- Has¸egan, M., Jordan,´ A., Cotˆ e,´ P., Djorgovski, S. G., McLaugh- ing by Extended Mass Distributions. ApJ514, 109–118. lin, D. E., Blakeslee, J. P., Mei, S., West, M. J., Peng, E. W., Gnedin, O. Y., Lee, H. M., Ostriker, J. P., Sep. 1999b. Effects of Tidal Ferrarese, L., Milosavljevic,´ M., Tonry, J. L., Merritt, D., Jul. Shocks on the Evolution of Globular Clusters. ApJ522, 935–949. 2005. The ACS Virgo Cluster Survey. VII. Resolving the Connec- Gnedin, O. Y., Ostriker, J. P., Jan. 1997. Destruction of the Galactic tion between Globular Clusters and Ultracompact Dwarf Galaxies. Globular Cluster System. ApJ474, 223. ApJ627, 203–223. Gnedin, O. Y., Ostriker, J. P., Mar. 1999. On the Self-consistent Re- Hancock, M., Smith, B. J., Struck, C., Giroux, M. L., Hurlock, S., sponse of Stellar Systems to Gravitational Shocks. ApJ513, 626– Jun. 2009. Candidate Tidal Dwarf Galaxies in Arp 305: Lessons on 637. Dwarf Detachment and Globular Cluster Formation. AJ137, 4643– Gnedin, O. Y., Ostriker, J. P., Tremaine, S., Apr. 2014. Co-evolution 4654. of Galactic Nuclei and Globular Cluster Systems. ApJ785, 71. Harris, W. E., 1991. Globular cluster systems in galaxies beyond the Goerdt, T., Moore, B., Read, J. I., Stadel, J., Dec. 2010. Core Cre- Local Group. ARA&A29, 543–579. ation in Galaxies and Halos Via Sinking Massive Objects. ApJ725, Harris, W. E., Oct. 1996. A Catalog of Parameters for Globular Clus- 1707–1716. ters in the Milky Way. AJ112, 1487. Goerdt, T., Moore, B., Read, J. I., Stadel, J., Zemp, M., May 2006. Harris, W. E., Blakeslee, J. P., Harris, G. L. H., Feb. 2017. Galactic Does the Fornax dwarf spheroidal have a central cusp or core? Dark Matter Halos and Globular Cluster Populations. III. Exten- MNRAS368, 1073–1077. sion to Extreme Environments. ApJ836, 67. Goldman, S. R., van Loon, J. T., Zijlstra, A. A., Green, J. A., Wood, Harris, W. E., Harris, G. L. H., Alessi, M., Aug. 2013. A Catalog of P. R., Nanni, A., Imai, H., Whitelock, P. A., Matsuura, M., Groe- Globular Cluster Systems: What Determines the Size of a Galaxy’s newegen, M. A. T., Gomez,´ J. F., Feb. 2017. The wind speeds, Globular Cluster Population? ApJ772, 82. dust content, and mass-loss rates of evolved AGB and RSG stars at Hartmann, M., Debattista, V. P., Seth, A., Cappellari, M., Quinn, T. R., varying metallicity. MNRAS465, 403–433. Dec. 2011. Constraining the role of star cluster mergers in nuclear Goodman, J., Feb. 1987. On gravothermal oscillations. ApJ313, 576– cluster formation: simulations confront integral-field data. MN- 595. RAS418, 2697–2714. Goodwin, S. P., Apr. 1997. The initial conditions of young globular Hashimoto, Y., Funato, Y., Makino, J., Jan. 2003. To Circularize clusters in the Large Magellanic Cloud. MNRAS286, 669–680. or Not To Circularize?-Orbital Evolution of Satellite Galaxies. Goodwin, S. P., Kroupa, P., Aug. 2005. Limits on the primordial stel- ApJ582, 196–201. lar multiplicity. A&A439, 565–569. Hassan, S., Dave,´ R., Mitra, S., Finlator, K., Ciardi, B., Santos, M. G., Graham, A. W., May 2012. Extending the Mbh-σ diagram with dense Jan. 2018. Constraining the contribution of active galactic nuclei to nuclear star clusters. MNRAS422, 1586–1591. reionization. MNRAS473, 227–240. Grand, R. J. J., Kawata, D., Cropper, M., Apr. 2012. The dynamics of Hayli, A., Jul. 1970. Numerical Experiments on the Escape from Non stars around spiral arms. MNRAS421, 1529–1538. Isolated Clusters. A&A7, 17. Gratier, P., Braine, J., Rodriguez-Fernandez, N. J., Schuster, K. F., Heggie, D., Hut, P., Feb. 2003. The Gravitational Million-Body Prob- Kramer, C., Corbelli, E., Combes, F., Brouillet, N., van der Werf, lem: A Multidisciplinary Approach to Star Cluster Dynamics, by P. P., Rollig,¨ M., Jun. 2012. Giant molecular clouds in the Local Douglas Heggie and Piet Hut. Cambridge University Press, 2003, Group galaxy M 33. A&A542, A108. 372 pp. Gratton, R. G., Bonifacio, P., Bragaglia, A., Carretta, E., Castel- Heggie, D. C., 1973. Regularization Using a Time-Transformation lani, V., Centurion, M., Chieffi, A., Claudi, R., Clementini, G., Only. In: Tapley, B. D., Szebehely, V. (Eds.), Recent Advances in D’Antona, F., Desidera, S., Franc¸ois, P., Grundahl, F., Lucatello, Dynamical Astronomy. Vol. 39 of Astrophysics and Space Science S., Molaro, P., Pasquini, L., Sneden, C., Spite, F., Straniero, O., Library. p. 34. Apr. 2001. The O-Na and Mg-Al anticorrelations in turn-off and Heggie, D. C., Dec. 1975. Binary evolution in stellar dynamics. MN- early subgiants in globular clusters. A&A369, 87–98. RAS173, 729–787. Gratton, R. G., Carretta, E., Bragaglia, A., Feb. 2012. Multiple pop- Heggie, D. C., Dec. 2014. Towards an N-body model for the globular

51 cluster M4. MNRAS445, 3435–3443. Tanaka, M., Ishigaki, M. N., Akiyama, M., Arimoto, N., Garmilla, Heggie, D. C., Giersz, M., Apr. 2014. MOCCA code for star cluster J. A., Lupton, R. H., Strauss, M. A., Furusawa, H., Miyazaki, S., simulations - III. Stellar-mass black holes in the globular cluster Murayama, H., Nishizawa, A. J., Takada, M., Usuda, T., Wang, M22. MNRAS439, 2459–2467. S.-Y., Nov. 2016. A New Milky Way Satellite Discovered in the Helmi, A., White, S. D. M., de Zeeuw, P. T., Zhao, H., Nov. 1999. Subaru/Hyper Suprime-Cam Survey. ApJ832, 21. Debris streams in the solar neighbourhood as relicts from the for- Hong, J., Kim, E., Lee, H. M., Spurzem, R., Apr. 2013. Comparative mation of the Milky Way. Nature402, 53–55. study between N-body and Fokker-Planck simulations for rotat- Henault-Brunet,´ V., Gieles, M., Evans, C. J., Sana, H., Bastian, N., ing star clusters - II. Two-component models. MNRAS430, 2960– Ma´ız Apellaniz,´ J., Taylor, W. D., Markova, N., Bressert, E., de 2972. Koter, A., van Loon, J. T., Sep. 2012. The VLT-FLAMES Taran- Hopkins, P. F., Keres,ˇ D., Onorbe,˜ J., Faucher-Giguere,` C.-A., tula Survey. VI. Evidence for rotation of the young massive cluster Quataert, E., Murray, N., Bullock, J. S., Nov. 2014. Galaxies R136. A&A545, L1. on FIRE (Feedback In Realistic Environments): stellar feedback Hennebelle, P., Falgarone, E., Nov. 2012. Turbulent molecular clouds. explains cosmologically inefficient star formation. MNRAS445, A&A Rev.20, 55. 581–603. Henon,´ M., Feb. 1961. Sur l’evolution´ dynamique des amas globu- Hopkins, P. F., Murray, N., Quataert, E., Thompson, T. A., Jan. 2010. laires. Annales d’Astrophysique 24, 369. A maximum stellar surface density in dense stellar systems. MN- Henon,´ M., Feb. 1965. Sur l’evolution´ dynamique des amas globu- RAS401, L19–L23. laires. II. Amas isol and eacute. Annales d’Astrophysique 28, 62– Hopkins, P. F., Quataert, E., Sep. 2010. How do massive black holes +. get their gas? MNRAS407, 1529–1564. Henon,´ M., Jun. 1969. Rates of Escape from Isolated Clusters with an Hopkins, P. F., Quataert, E., Murray, N., Apr. 2012. The structure Arbitrary Mass Distribution. A&A2, 151. of the interstellar medium of star-forming galaxies. MNRAS421, Henon,´ M., Nov. 1970. Numerical exploration of the restricted prob- 3488–3521. lem. VI. Hill’s case: Non-periodic orbits. A&A9, 24–36. Hou, L. G., Han, J. L., Sep. 2014. The observed spiral structure of the Hernandez, X., Gilmore, G., Jun. 1998. Dynamical friction in dwarf Milky Way. A&A569, A125. galaxies. MNRAS297, 517–525. Howard, C. S., Pudritz, R. E., Harris, B. E., Klessen, R. S., Oct. 2017. Hernandez, X., Jimenez,´ M. A., Allen, C., Feb. 2013. Flattened ve- Simulating the UV Escape Fractions from Molecular Cloud Popu- locity dispersion profiles in globular clusters: Newtonian tides or lations in Star-forming Dwarf and Spiral Galaxies. ArXiv e-prints. modified gravity? MNRAS428, 3196–3205. Hudson, M. J., Robison, B., Jul. 2017. The correlation between the Hernquist, L., Aug. 1989. Tidal triggering of starbursts and nuclear sizes of globular cluster systems and their host dark matter haloes. activity in galaxies. Nature340, 687–691. ArXiv e-prints. Herrera, C. N., Boulanger, F., Nesvadba, N. P. H., Oct. 2011. From Hughes, A., Meidt, S. E., Schinnerer, E., Colombo, D., Pety, J., Leroy, large scale gas compression to cluster formation in the Antennae A. K., Dobbs, C. L., Garc´ıa-Burillo, S., Thompson, T. A., Dumas, overlap region. A&A534, A138. G., Schuster, K. F., Kramer, C., Dec. 2013. Probability Distribution Herrera, C. N., Boulanger, F., Nesvadba, N. P. H., Falgarone, E., Feb. Functions of 12CO(J = 1 - 0) Brightness and Integrated Intensity 2012. ALMA CO and VLT/SINFONI H2 observations of the An- in M51: The PAWS View. ApJ779, 44. tennae overlap region: mass and energy dissipation. A&A538, L9. Hunter, D. A., Elmegreen, B. G., Dupuy, T. J., Mortonson, M., Oct. Heyer, M., Krawczyk, C., Duval, J., Jackson, J. M., Jul. 2009. Re- 2003. Cluster Mass Functions in the Large and Small Magellanic Examining Larson’s Scaling Relationships in Galactic Molecular Clouds: Fading and Size-of-Sample Effects. AJ126, 1836–1848. Clouds. ApJ699, 1092–1103. Hut, P., Makino, J., Jan. 1999. Astrophysics on the GRAPE Family of Heyer, M. H., Terebey, S., Jul. 1998. The Anatomy of the Perseus Special-Purpose Computers. Science 283, 501. Spiral Arm: 12CO and IRAS Imaging Observations of the W3- Hut, P., McMillan, S., Goodman, J., Mateo, M., Phinney, E. S., Pryor, W4-W5 Cloud Complex. ApJ502, 265–277. C., Richer, H. B., Verbunt, F., Weinberg, M., Nov. 1992. Binaries Hills, J. G., Feb. 1980. The effect of mass loss on the dynamical evo- in globular clusters. PASP104, 981–1034. lution of a stellar system - Analytic approximations. ApJ235, 986– Hwang, N., Lee, M. G., Lee, J. C., Park, W.-K., Park, H. S., Kim, 991. S. C., Park, J.-H., Sep. 2011. Extended Star Clusters in the Remote Hirano, S., Yoshida, N., Sakurai, Y., Fujii, M. S., Nov. 2017. Forma- Halo of the Intriguing Dwarf Galaxy NGC 6822. ApJ738, 58. tion of the first star cluster and massive star binaries by fragmenta- Ibata, R., Irwin, M., Lewis, G., Ferguson, A. M. N., Tanvir, N., Jul. tion of filamentary primordial gas clouds. ArXiv e-prints. 2001. A giant stream of metal-rich stars in the halo of the galaxy Hirota, A., Kuno, N., Sato, N., Nakanishi, H., Tosaki, T., Sorai, K., M31. Nature412, 49–52. Aug. 2011. Giant Molecular Clouds in the Spiral Arm of IC 342. Ibata, R., Nipoti, C., Sollima, A., Bellazzini, M., Chapman, S. C., Da- ApJ737, 40. lessandro, E., Feb. 2013a. Do globular clusters possess dark matter Hoekstra, H., Hsieh, B. C., Yee, H. K. C., Lin, H., Gladders, M. D., haloes? A case study in NGC 2419. MNRAS428, 3648–3659. Dec. 2005. Virial Masses and the Baryon Fraction in Galaxies. Ibata, R. A., Lewis, G. F., Conn, A. R., Irwin, M. J., McConnachie, ApJ635, 73–85. A. W., Chapman, S. C., Collins, M. L., Fardal, M., et al., Jan. Holtzman, J. A., Faber, S. M., Shaya, E. J., Lauer, T. R., Groth, J., 2013b. A vast, thin plane of corotating dwarf galaxies orbiting the Hunter, D. A., Baum, W. A., Ewald, S. P., Hester, J. J., Light, Andromeda galaxy. Nature493, 62–65. R. M., Lynds, C. R., O’Neil, E. J., Westphal, J. A., Mar. 1992. Plan- Ibata, R. A., Lewis, G. F., Irwin, M. J., Quinn, T., Jun. 2002. Uncov- etary Camera observations of NGC 1275 - Discovery of a central ering cold dark matter halo substructure with tidal streams. MN- population of compact massive blue star clusters. AJ103, 691–702. RAS332, 915–920. Holtzman, J. A., Watson, A. M., Mould, J. R., Gallagher, III, J. S., Ibata, R. A., Lewis, G. F., Martin, N. F., Mar. 2016. Feeling the Pull: Ballester, G. E., Burrows, C. J., Clarke, J. T., Crisp, D., Evans, a Study of Natural Galactic Accelerometers. I. Photometry of the R. W., Griffiths, R. E., Hester, J. J., Hoessel, J. G., Scowen, P. A., Delicate Stellar Stream of the Palomar 5 Globular Cluster. ApJ819, Stapelfeldt, K. R., Trauger, J. T., Westphal, J. A., Aug. 1996. Star 1. Clusters in Interacting and Cooling Flow Galaxies. AJ112, 416. Ibata, R. A., Lewis, G. F., Thomas, G., Martin, N. F., Chapman, Homma, D., Chiba, M., Okamoto, S., Komiyama, Y., Tanaka, M., S., Jun. 2017. Feeling the Pull: A Study of Natural Galactic

52 Accelerometers. II. Kinematics and Mass of the Delicate Stellar dwarf galaxy globular clusters. MNRAS444, 2377–2395. Stream of the Palomar 5 Globular Cluster. ApJ842, 120. Kauffmann, J., Pillai, T., Zhang, Q., Menten, K. M., Goldsmith, P. F., Inagaki, S., 1986. Post-collapse evolution of globular clusters with Lu, X., Guzman,´ A. E., Jul. 2017. The Galactic Center Molecular finite number of stars in the core. PASJ38, 853–863. Cloud Survey. I. A steep linewidth-size relation and suppression of Inoue, S., Aug. 2009. The test for suppressed dynamical friction in a star formation. A&A603, A89. constant density core of dwarf galaxies. MNRAS397, 709–716. Kaufman, M., Grupe, D., Elmegreen, B. G., Elmegreen, D. M., Inoue, S., Dekel, A., Mandelker, N., Ceverino, D., Bournaud, F., Pri- Struck, C., Brinks, E., Nov. 2012. NGC 2207/IC 2163: A Graz- mack, J., Feb. 2016. Non-linear violent disc instability with high ing Encounter with Large-scale Shocks. AJ144, 156. Toomre’s Q in high-redshift clumpy disc galaxies. MNRAS456, Keel, W. C., Kennicutt, Jr., R. C., Hummel, E., van der Hulst, J. M., 2052–2069. May 1985. The effects of interactions on spiral galaxies. I - Nuclear Inoue, T., Fukui, Y., Sep. 2013. Formation of Massive Molecular activity and star formation. AJ90, 708–730. Cloud Cores by Cloud-Cloud Collision. ApJ774, L31. Keenan, D. W., Innanen, K. A., Apr. 1975. Numerical investigation of Irwin, J. A., Jul. 1994. Arcs and bridges in the interacting galaxies galactic tidal effects on spherical stellar systems. AJ80, 290–302. NGC 5775/NGC 5774. ApJ429, 618–633. Kenney, J. D. P., Lord, S. D., Nov. 1991. Orbit crowding of molecular Izumi, N., Kobayashi, N., Yasui, C., Tokunaga, A. T., Saito, M., gas at a bar-spiral arm transition zone in M83. ApJ381, 118–129. Hamano, S., Nov. 2014. Discovery of Star Formation in the Ex- Kennicutt, R. C., Evans, N. J., Sep. 2012. Star Formation in the Milky treme Outer Galaxy Possibly Induced by a High-velocity Cloud Way and Nearby Galaxies. ARA&A50, 531–608. Impact. ApJ795, 66. Kennicutt, Jr., R. C., 1998. Star Formation in Galaxies Along the Hub- Jeans, J. H., 1902. The Stability of a Spherical Nebula. Philosophical ble Sequence. ARA&A36, 189–232. Transactions of the Royal Society of London Series A 199, 1–53. Kessel-Deynet, O., Burkert, A., Jan. 2003. Radiation-driven implo- Jog, C. J., Jul. 2013. Jeans instability criterion modified by external sion of molecular cloud cores. MNRAS338, 545–554. tidal field. MNRAS434, L56–L60. Kim, J.-h., Ma, X., Grudic,´ M. Y., Hopkins, P. F., Hayward, C. C., Jog, C. J., Jun. 2014. Effective Q Criterion for Disk Stability in an Wetzel, A., Faucher-Giguere,` C.-A., Keres,ˇ D., Garrison-Kimmel, External Potential. AJ147, 132. S., Murray, N., Mar. 2018. Formation of globular cluster candi- Jog, C. J., Solomon, P. M., Mar. 1992. A triggering mechanism for dates in merging proto-galaxies at high redshift: a view from the enhanced star formation in colliding galaxies. ApJ387, 152–161. FIRE cosmological simulations. MNRAS474, 4232–4244. Johnson, L. C., Seth, A. C., Dalcanton, J. J., Beerman, L. C., Foues- Kim, W.-T., Kim, Y., Kim, J.-G., Jul. 2014. Nature of the Wiggle neau, M., Weisz, D. R., Bell, T. A., Dolphin, A. E., Sandstrom, Instability of Galactic Spiral Shocks. ApJ789, 68. K., Williams, B. F., Apr. 2017. Panchromatic Hubble Andromeda Kim, W.-T., Ostriker, E. C., Jul. 2006. Formation of Spiral-Arm Spurs Treasury. XVIII. The High-mass Truncation of the Star Cluster and Bound Clouds in Vertically Stratified Galactic Gas Disks. Mass Function. ApJ839, 78. ApJ646, 213–231. Johnston, K. V., Mar. 1998. A Prescription for Building the Milky Kimm, T., Cen, R., Rosdahl, J., Yi, S. K., May 2016. Formation of Way’s Halo from Disrupted Satellites. ApJ495, 297–308. Globular Clusters in Atomic-cooling Halos Via Rapid Gas Con- Johnston, K. V., Spergel, D. N., Haydn, C., May 2002. How Lumpy densation and Fragmentation during the Epoch of Reionization. Is the Milky Way’s Dark Matter Halo? ApJ570, 656–664. ApJ823, 52. Jordan,´ A., McLaughlin, D. E., Cotˆ e,´ P., Ferrarese, L., Peng, E. W., Kimmig, B., Seth, A., Ivans, I. I., Strader, J., Caldwell, N., Anderton, Mei, S., Villegas, D., Merritt, D., Tonry, J. L., West, M. J., Jul. T., Gregersen, D., Feb. 2015. Measuring Consistent Masses for 25 2007. The ACS Virgo Cluster Survey. XII. The Luminosity Func- Milky Way Globular Clusters. AJ149, 53. tion of Globular Clusters in Early-Type Galaxies. ApJS171, 101– King, I., Apr. 1958. The escape of stars from clusters 11. A simple 145. theory of the evolution of an isolated cluster. AJ63, 114–+. Jordan,´ A., Peng, E. W., Blakeslee, J. P., Cotˆ e,´ P., Eyheramendy, S., King, I., Oct. 1962. The structure of star clusters. I. an empirical den- Ferrarese, L., Mei, S., Tonry, J. L., West, M. J., Jan. 2009. The sity law. AJ67, 471. ACS Virgo Cluster Survey XVI. Selection Procedure and Catalogs King, I. R., Feb. 1966. The structure of star clusters. III. Some simple of Globular Cluster Candidates. ApJS180, 54–66. dynamical models. AJ71, 64. Joung, M. K. R., Mac Low, M.-M., Dec. 2006. Turbulent Structure of King, I. R., Sep. 1981. The Dynamics of Globular Clusters. QJRAS22, a Stratified Supernova-driven Interstellar Medium. ApJ653, 1266– 227. 1279. Knierman, K. A., Gallagher, S. C., Charlton, J. C., Hunsberger, S. D., Just, A., Berczik, P., Petrov, M. I., Ernst, A., Jan. 2009. Quantitative Whitmore, B., Kundu, A., Hibbard, J. E., Zaritsky, D., Sep. 2003. analysis of clumps in the tidal tails of star clusters. MNRAS392, From Globular Clusters to Tidal Dwarfs: Structure Formation in 969–981. the Tidal Tails of Merging Galaxies. AJ126, 1227–1244. Just, A., Khan, F. M., Berczik, P., Ernst, A., Spurzem, R., Feb. 2011. Koch, A., Cotˆ e,´ P., McWilliam, A., Nov. 2009. All quiet in the outer Dynamical friction of massive objects in galactic centres. MN- halo: chemical abundances in the globular cluster Pal 3. A&A506, RAS411, 653–674. 729–743. Just, A., Penarrubia,˜ J., Mar. 2005. Large scale inhomogeneity and Koch, A., Grebel, E. K., Odenkirchen, M., Mart´ınez-Delgado, D., local dynamical friction. A&A431, 861–877. Caldwell, J. A. R., Nov. 2004. Mass Segregation in the Globular Kartha, S. S., Forbes, D. A., Alabi, A. B., Brodie, J. P., Romanowsky, Cluster Palomar 5 and its Tidal Tails. AJ128, 2274–2287. A. J., Strader, J., Spitler, L. R., Jennings, Z. G., Roediger, J. C., Koda, J., Scoville, N., Sawada, T., La Vigne, M. A., Vogel, S. N., May 2016. The SLUGGS survey*: exploring the globular cluster Potts, A. E., Carpenter, J. M., Corder, S. A., Wright, M. C. H., systems of the Leo II group and their global relationships. MN- White, S. M., Zauderer, B. A., Patience, J., Sargent, A. I., Bock, RAS458, 105–126. D. C. J., Hawkins, D., Hodges, M., Kemball, A., Lamb, J. W., Katz, H., Ricotti, M., Jul. 2013. Two epochs of globular cluster forma- Plambeck, R. L., Pound, M. W., Scott, S. L., Teuben, P., Woody, tion from deep field luminosity functions: implications for reion- D. P., Aug. 2009. Dynamically Driven Evolution of the Interstellar ization and the Milky Way satellites. MNRAS432, 3250–3261. Medium in M51. ApJ700, L132–L136. Katz, H., Ricotti, M., Nov. 2014. Clues on the missing sources of Koenig, X. P., Leisawitz, D. T., Benford, D. J., Rebull, L. M., Padgett, reionization from self-consistent modelling of Milky Way and D. L., Assef, R. J., Jan. 2012. Wide-field Infrared Survey Explorer

53 Observations of the Evolution of Massive Star-forming Regions. D. W., Kroupa, P., Santiago, B. X., Apr. 2015. Globular Cluster ApJ744, 130. Streams as Galactic High-Precision Scales the Poster Child Palo- Konstantopoulos, I. S., Smith, L. J., Adamo, A., Silva-Villa, E., Gal- mar 5. ApJ803, 80. lagher, J. S., Bastian, N., Ryon, J. E., Westmoquette, M. S., Zack- Kupper,¨ A. H. W., Kroupa, P., Baumgardt, H., Heggie, D. C., Oct. risson, E., Larsen, S. S., Weisz, D. R., Charlton, J. C., May 2013. 2010a. Peculiarities in velocity dispersion and surface density pro- The Snapshot Hubble U-band Cluster Survey (SHUCS). I. Survey files of star clusters. MNRAS407, 2241–2260. Description and First Application to the Mixed Star Cluster Popu- Kupper,¨ A. H. W., Kroupa, P., Baumgardt, H., Heggie, D. C., Jan. lation of NGC 4041. AJ145, 137. 2010b. Tidal tails of star clusters. MNRAS401, 105–120. Koposov, S. E., Rix, H.-W., Hogg, D. W., Mar. 2010. Constraining the Kupper,¨ A. H. W., MacLeod, A., Heggie, D. C., Jul. 2008. On the Milky Way Potential with a Six-Dimensional Phase-Space Map of structure of tidal tails. MNRAS387, 1248–1252. the GD-1 Stellar Stream. ApJ712, 260–273. Kuruwita, R. L., Federrath, C., Ireland, M., Sep. 2017. Binary star Kraan-Korteweg, R. C., Lahav, O., 2000. The Universe behind the formation and the outflows from their discs. MNRAS470, 1626– Milky Way. A&A Rev.10, 211–261. 1641. Kraft, R. P., Jun. 1994. Abundance differences among globular-cluster Kustaanheimo, P., Stiefel, E., 1965. J. Reine Angew. Math 218, 204. giants: Primordial versus evolutionary scenarios. PASP106, 553– Lada, C. J., Forbrich, J., Lombardi, M., Alves, J. F., Feb. 2012. Star 565. Formation Rates in Molecular Clouds and the Nature of the Extra- Kraft, R. P., Sneden, C., Smith, G. H., Shetrone, M. D., Langer, G. E., galactic Scaling Relations. ApJ745, 190. Pilachowski, C. A., Jan. 1997. Proton Capture Chains in Globu- Lada, C. J., Lada, E. A., 2003. Embedded Clusters in Molecular lar Cluster Stars.II.Oxygen, Sodium, Magnesium, and Aluminum Clouds. ARA&A41, 57–115. Abundances in M13 Giants Brighter Than the Horizontal Branch. Lamers, H. J. G. L. M., Baumgardt, H., Gieles, M., Nov. 2010. Mass- AJ113, 279. loss rates and the mass evolution of star clusters. MNRAS409, Kraljic, K., Bournaud, F., Martig, M., Sep. 2012. The Two-phase For- 305–328. mation History of Spiral Galaxies Traced by the Cosmic Evolution Lane, R. R., Kiss, L. L., Lewis, G. F., Ibata, R. A., Siebert, A., Bed- of the Bar Fraction. ApJ757, 60. ding, T. R., Szekely,´ P., Balog, Z., Szabo,´ G. M., Aug. 2010. Halo Krause, M., Charbonnel, C., Decressin, T., Meynet, G., Prantzos, globular clusters observed with AAOmega: dark matter content, N., Diehl, R., Oct. 2012. Superbubble dynamics in globular clus- metallicity and tidal heating. MNRAS406, 2732–2742. ter infancy. I. How do globular clusters first lose their cold gas? Lane, R. R., Kupper,¨ A. H. W., Heggie, D. C., Jul. 2012. The tidal A&A546, L5. tails of 47 Tucanae. MNRAS423, 2845–2853. Krauss, L. M., Chaboyer, B., Jan. 2003. Age Estimates of Globular Lardo, C., Pancino, E., Bellazzini, M., Bragaglia, A., Donati, P., Clusters in the Milky Way: Constraints on Cosmology. Science Gilmore, G., Randich, S., Feltzing, S., Jeffries, R. D., Vallenari, 299, 65–70. A., Alfaro, E. J., Allende Prieto, C., Flaccomio, E., Koposov, S. E., Kravtsov, A. V., Gnedin, O. Y., Apr. 2005. Formation of Globular Recio-Blanco, A., Bergemann, M., Carraro, G., Costado, M. T., Clusters in Hierarchical Cosmology. ApJ623, 650–665. Damiani, F., Hourihane, A., Jofre,´ P., de Laverny, P., Marconi, G., Kritsuk, A. G., Norman, M. L., Padoan, P., Wagner, R., Aug. 2007. Masseron, T., Morbidelli, L., Sacco, G. G., Worley, C. C., Jan. The Statistics of Supersonic Isothermal Turbulence. ApJ665, 416– 2015. The Gaia-ESO Survey: Kinematics of seven Galactic globu- 431. lar clusters. A&A573, A115. Kroupa, P., Oct. 1998. The dynamical evolution of stellar superclus- Larsen, S. S., Sep. 2002. The Luminosity Function of Star Clusters in ters. MNRAS300, 200–204. Spiral Galaxies. AJ124, 1393–1409. Kruijssen, J. M. D., Nov. 2012. On the fraction of star formation oc- Larsen, S. S., Mar. 2004. The structure and environment of young curring in bound stellar clusters. MNRAS426, 3008–3040. stellar clusters in spiral galaxies. A&A416, 537–553. Kruijssen, J. M. D., Dec. 2015. Globular clusters as the relics of regu- Larsen, S. S., Jan. 2006. Star formation in clusters. In: Livio, M., lar star formation in ‘normal’ high-redshift galaxies. MNRAS454, Casertano, S. (Eds.), Planets to Cosmology: Essential Science in 1658–1686. the Final Years of the Hubble Space Telescope. pp. 35–+. Kruijssen, J. M. D., Longmore, S. N., Elmegreen, B. G., Murray, N., Larsen, S. S., Feb. 2009. The mass function of young star clusters in Bally, J., Testi, L., Kennicutt, R. C., Jun. 2014. What controls star spiral galaxies. A&A494, 539–551. formation in the central 500 pc of the Galaxy? MNRAS440, 3370– Larsen, S. S., Brodie, J. P., Elmegreen, B. G., Efremov, Y. N., Hodge, 3391. P. W., Richtler, T., Aug. 2001. Structure and Mass of a Young Kruijssen, J. M. D., Pelupessy, F. I., Lamers, H. J. G. L. M., Porte- Globular Cluster in NGC 6946. ApJ556, 801–812. gies Zwart, S. F., Icke, V., Jun. 2011. Modelling the formation and Larsen, S. S., Brodie, J. P., Forbes, D. A., Strader, J., May 2014. evolution of star cluster populations in galaxy simulations. MN- Chemical composition and constraints on mass loss for globular RAS414, 1339–1364. clusters in dwarf galaxies: WLM and IKN. A&A565, A98. Krumholz, M. R., Jun. 2014. The big problems in star formation: The Larsen, S. S., Richtler, T., Feb. 2000. Young massive star clusters in star formation rate, stellar clustering, and the initial mass function. nearby spiral galaxies. III. Correlations between cluster popula- Phys. Rep.539, 49–134. tions and host galaxy properties. A&A354, 836–846. Krumholz, M. R., Dekel, A., McKee, C. F., Jan. 2012. A Universal, Larsen, S. S., Strader, J., Brodie, J. P., Aug. 2012. Constraints on mass Local Star Formation Law in Galactic Clouds, nearby Galaxies, loss and self-enrichment scenarios for the globular clusters of the High-redshift Disks, and Starbursts. ApJ745, 69. Fornax dSph. A&A544, L14. Krumholz, M. R., Tan, J. C., Jan. 2007. Slow Star Formation in Dense Larson, R. B., Mar. 1981. Turbulence and star formation in molecular Gas: Evidence and Implications. ApJ654, 304–315. clouds. MNRAS194, 809–826. Krumholz, M. R., Thompson, T. A., Dec. 2012. Direct Numerical Leaman, R., VandenBerg, D. A., Mendel, J. T., Nov. 2013. The bi- Simulation of Radiation Pressure-driven Turbulence and Winds in furcated age-metallicity relation of Milky Way globular clusters Star Clusters and Galactic Disks. ApJ760, 155. and its implications for the accretion history of the galaxy. MN- Kundic, T., Ostriker, J. P., Jan. 1995. Tidal-shock relaxation: A reex- RAS436, 122–135. amination of tidal shocks in stellar systems. ApJ438, 702–707. Lee, H. M., Ostriker, J. P., Nov. 1987. The evolution and final disinte- Kupper,¨ A. H. W., Balbinot, E., Bonaca, A., Johnston, K. V., Hogg, gration of spherical stellar systems in a steady galactic tidal field.

54 ApJ322, 123–132. van de Ven, G., Ferrarese, L., Jordan,´ A., Infante, L., Peng, E. W., Lee, Y.-N., Hennebelle, P., Jun. 2016. Formation of a protocluster: Jun. 2013. The complex nature of the nuclear star cluster in FCC A virialized structure from gravoturbulent collapse. II. A two- 277. MNRAS431, 3364–3372. dimensional analytical model for a rotating and accreting system. Mackey, A. D., Broby Nielsen, P., Ferguson, A. M. N., Richardson, A&A591, A31. J. C., Jul. 2008. Multiple Stellar Populations in Three Rich Large Leigh, N., Giersz, M., Webb, J. J., Hypki, A., De Marchi, G., Kroupa, Magellanic Cloud Star Clusters. ApJ681, L17. P., Sills, A., Dec. 2013. The state of globular clusters at birth: Mackey, A. D., Da Costa, G. S., Ferguson, A. M. N., Yong, D., Jan. emergence from the gas-embedded phase. MNRAS436, 3399– 2013. A VLT/FLAMES Study of the Peculiar Intermediate-age 3412. Large Magellanic Cloud Star Cluster NGC 1846. I. Kinematics. Leigh, N., Umbreit, S., Sills, A., Knigge, C., de Marchi, G., Glebbeek, ApJ762, 65. E., Sarajedini, A., May 2012. Quantifying the universality of the Mackey, A. D., Gilmore, G. F., Dec. 2004. Comparing the properties stellar initial mass function in old star clusters. MNRAS422, 1592– of local globular cluster systems: implications for the formation of 1600. the Galactic halo. MNRAS355, 504–534. Leigh, N. W. C., Georgiev, I. Y., Boker,¨ T., Knigge, C., den Brok, Mackey, A. D., van den Bergh, S., Jun. 2005. The properties of Galac- M., Jul. 2015. Nuclear star cluster formation in energy-space. MN- tic globular cluster subsystems. MNRAS360, 631–645. RAS451, 859–869. Madrid, J. P., Hurley, J. R., Martig, M., Apr. 2014. The Impact of Leitet, E., Bergvall, N., Hayes, M., Linne,´ S., Zackrisson, E., May Galaxy Geometry and Mass Evolution on the Survival of Star Clus- 2013. Escape of Lyman continuum radiation from local galax- ters. ApJ784, 95. ies. Detection of leakage from the young starburst Tol 1247-232. Madrid, J. P., Leigh, N. W. C., Hurley, J. R., Giersz, M., Sep. 2017. A&A553, A106. Mass evaporation rate of globular clusters in a strong tidal field. Leroy, A. K., Bolatto, A. D., Ostriker, E. C., Rosolowsky, E., Walter, MNRAS470, 1729–1737. F., Warren, S. R., Donovan Meyer, J., Hodge, J., Meier, D. S., Maji, M., Zhu, Q., Li, Y., Charlton, J., Hernquist, L., Knebe, A., Aug. Ott, J., Sandstrom, K., Schruba, A., Veilleux, S., Zwaan, M., Mar. 2017. The Formation and Evolution of Star Clusters in Interacting 2015. ALMA Reveals the Molecular Medium Fueling the Nearest Galaxies. ApJ844, 108. Nuclear Starburst. ApJ801, 25. Makino, J., Nov. 1996. Postcollapse Evolution of Globular Clusters. Li, H., Gnedin, O. Y., Nov. 2014. Modeling the Formation of Globular ApJ471, 796. Cluster Systems in the Virgo Cluster. ApJ796, 10. Makino, J., Fukushige, T., Koga, M., Namura, K., Dec. 2003. Li, H., Gnedin, O. Y., Gnedin, N. Y., Meng, X., Semenov, V. A., GRAPE-6: Massively-Parallel Special-Purpose Computer for As- Kravtsov, A. V., Jan. 2017. Star Cluster Formation in Cosmological trophysical Particle Simulations. PASJ55, 1163–1187. Simulations. I. Properties of Young Clusters. ApJ834, 69. Makino, J., Taiji, M., Apr. 1998. Scientific Simulations with Special- Lin, C. C., Shu, F. H., Feb. 1966. On the Spiral Structure of Disk Purpose Computers–the GRAPE Systems. Galaxies, II. Outline of a Theory of Density Waves. Proceedings Mandelbaum, R., Seljak, U., Cool, R. J., Blanton, M., Hirata, C. M., of the National Academy of Science 55, 229–234. Brinkmann, J., Oct. 2006. Density profiles of galaxy groups and Linden, S. T., Evans, A. S., Rich, J., Larson, K. L., Armus, L., D´ıaz- clusters from SDSS galaxy-galaxy weak lensing. MNRAS372, Santos, T., Privon, G. C., Howell, J., Inami, H., Kim, D.-C., Chien, 758–776. L.-H., Vavilkin, T., Mazzarella, J. M., Modica, F., Surace, J. A., Mandushev, G., Staneva, A., Spasova, N., Dec. 1991. Dynamical Manning, S., Abdullah, A., Blake, A., Yarber, A., Lambert, T., Jul. masses for Galactic globular clusters. A&A252, 94–99. 2017. Massive Star Cluster Formation and Destruction in Lumi- Mao, S., Schneider, P., Apr. 1998. Evidence for substructure in lens nous Infrared Galaxies in GOALS. ApJ843, 91. galaxies? MNRAS295, 587. Loeb, A., Barkana, R., 2001. The Reionization of the Universe by the Mapelli, M., May 2017. Rotation in young massive star clusters. MN- First Stars and Quasars. ARA&A39, 19–66. RAS467, 3255–3267. Lofthouse, E. K., Kaviraj, S., Conselice, C. J., Mortlock, A., Hartley, Maraston, C., Bastian, N., Saglia, R. P., Kissler-Patig, M., Schweizer, W., Mar. 2017. Major mergers are not significant drivers of star F., Goudfrooij, P., Mar. 2004. The dynamical mass of the young formation or morphological transformation around the epoch of cluster W3 in NGC 7252. Heavy-weight globular cluster or ultra peak cosmic star formation. MNRAS465, 2895–2900. compact dwarf galaxy? A&A416, 467–473. Longmore, S. N., Bally, J., Testi, L., Purcell, C. R., Walsh, A. J., Martig, M., Bournaud, F., Teyssier, R., Dekel, A., Dec. 2009. Mor- Bressert, E., Pestalozzi, M., Molinari, S., et al., Feb. 2013. Varia- phological Quenching of Star Formation: Making Early-Type tions in the Galactic star formation rate and density thresholds for Galaxies Red. ApJ707, 250–267. star formation. MNRAS429, 987–1000. Martinez-Medina, L. A., Pichardo, B., Moreno, E., Peimbert, A., Ve- Loren, R. B., Oct. 1976. Colliding clouds and star formation in NGC lazquez, H., Jan. 2016. On the Origin of High-altitude Open Clus- 1333. ApJ209, 466. ters in the Milky Way. ApJ817, L3. Lotz, J. M., Telford, R., Ferguson, H. C., Miller, B. W., Stiavelli, M., Martinez-Medina, L. A., Pichardo, B., Peimbert, A., Moreno, E., Jan. Mack, J., May 2001. Dynamical Friction in DE Globular Cluster 2017. On the Survival of High-altitude Open Clusters within the Systems. ApJ552, 572–581. Milky Way Galaxy Tides. ApJ834, 58. Lutz, D., May 1991. NGC 3597 - Formation of an elliptical via merg- Mashchenko, S., Sills, A., Jan. 2005. Globular Clusters with Dark ing? A&A245, 31–40. Matter Halos. II. Evolution in a Tidal Field. ApJ619, 258–269. Lutzgendorf,¨ N., Baumgardt, H., Kruijssen, J. M. D., Oct. 2013. N- Mastrobuono-Battisti, A., Di Matteo, P., Montuori, M., Haywood, M., body simulations of globular clusters in tidal fields: Effects of Oct. 2012. Clumpy streams in a smooth dark halo: the case of intermediate-mass black holes. A&A558, A117. Palomar 5. A&A546, L7. Lynden-Bell, D., Eggleton, P. P., May 1980. On the consequences of Mastrobuono-Battisti, A., Perets, H. B., Loeb, A., Nov. 2014. Ef- the gravothermal catastrophe. MNRAS191, 483–498. fects of Intermediate Mass Black Holes on Nuclear Star Clusters. Lynden-Bell, D., Wood, R., 1968. The gravo-thermal catastrophe in ApJ796, 40. isothermal spheres and the onset of red-giant structure for stellar Matsui, H., Saitoh, T. R., Makino, J., Wada, K., Tomisaka, K., systems. MNRAS138, 495–+. Kokubo, E., Daisaka, H., Okamoto, T., Yoshida, N., Feb. 2012. Lyubenova, M., van den Bosch, R. C. E., Cotˆ e,´ P., Kuntschner, H., Origin of Multiple Nuclei in Ultraluminous Infrared Galaxies.

55 ApJ746, 26. Globular Clusters. III. The Double Subgiant Branch of NGC 1851. McConnachie, A. W., Jul. 2012. The Observed Properties of Dwarf ApJ673, 241–250. Galaxies in and around the Local Group. AJ144, 4. Milosavljevic,´ M., Apr. 2004. On the Origin of Nuclear Star Clusters McKee, C. F., Ostriker, E. C., Sep. 2007. Theory of Star Formation. in Late-Type Spiral Galaxies. ApJ605, L13–L16. ARA&A45, 565–687. Minchev, I., Famaey, B., Oct. 2010. A New Mechanism for Ra- McLaughlin, D. E., Harris, W. E., Hanes, D. A., Feb. 1994. The spatial dial Migration in Galactic Disks: Spiral-Bar Resonance Overlap. structure of the M87 globular cluster system. ApJ422, 486–507. ApJ722, 112–121. McMillan, S., Portegies Zwart, S., van Elteren, A., Whitehead, Mineo, S., Rappaport, S., Levine, A., Pooley, D., Steinhorn, B., A., Jul. 2012. Simulations of Dense Stellar Systems with the Homan, J., Dec. 2014. A Comprehensive X-Ray and Multiwave- AMUSE Software Toolkit. In: Capuzzo-Dolcetta, R., Limongi, length Study of the Colliding Galaxy Pair NGC 2207/IC 2163. M., Tornambe,` A. (Eds.), Advances in Computational Astro- ApJ797, 91. physics: Methods, Tools, and Outcome. Vol. 453 of Astronomical Misgeld, I., Hilker, M., Jul. 2011. Families of dynamically hot stellar Society of the Pacific Conference Series. p. 129. systems over 10 orders of magnitude in mass. MNRAS414, 3699– Meidt, S. E., Schinnerer, E., Garc´ıa-Burillo, S., Hughes, A., Colombo, 3710. D., Pety, J., Dobbs, C. L., Schuster, K. F., Kramer, C., Leroy, Miville-Deschenes,ˆ M.-A., Murray, N., Lee, E. J., Jan. 2017. Physical A. K., Dumas, G., Thompson, T. A., Dec. 2013. Gas Kinemat- Properties of Molecular Clouds for the Entire Milky Way Disk. ics on Giant Molecular Cloud Scales in M51 with PAWS: Cloud ApJ834, 57. Stabilization through Dynamical Pressure. ApJ779, 45. Mo, H., van den Bosch, F. C., White, S., May 2010. Galaxy Formation Mengel, S., Lehnert, M. D., Thatte, N. A., Vacca, W. D., Whitmore, and Evolution. B., Chandar, R., Oct. 2008. Young star clusters in interacting galax- Mo, H. J., Mao, S., White, S. D. M., Apr. 1998. The formation of ies - NGC 1487 and NGC 4038/4039. A&A489, 1091–1105. galactic discs. MNRAS295, 319–336. Mesz´ aros,´ S., Martell, S. L., Shetrone, M., Lucatello, S., Troup, N. W., Molinari, S., Bally, J., Noriega-Crespo, A., Compiegne,` M., Bernard, Bovy, J., Cunha, K., Garc´ıa-Hernandez,´ D. A., Overbeek, J. C., J. P., Paradis, D., Martin, P., Testi, L., Barlow, M., Moore, T., Allende Prieto, C., Beers, T. C., Frinchaboy, P. M., Garc´ıa Perez,´ Plume, R., Swinyard, B., Zavagno, A., Calzoletti, L., Di Giorgio, A. E., Hearty, F. R., Holtzman, J., Majewski, S. R., Nidever, D. L., A. M., Elia, D., Faustini, F., Natoli, P., Pestalozzi, M., Pezzuto, S., Schiavon, R. P., Schneider, D. P., Sobeck, J. S., Smith, V. V., Piacentini, F., Polenta, G., Polychroni, D., Schisano, E., Traficante, Zamora, O., Zasowski, G., May 2015. Exploring Anticorrelations A., Veneziani, M., Battersby, C., Burton, M., Carey, S., Fukui, Y., and Light Element Variations in Northern Globular Clusters Ob- Li, J. Z., Lord, S. D., Morgan, L., Motte, F., Schuller, F., Stringfel- served by the APOGEE Survey. AJ149, 153. low, G. S., Tan, J. C., Thompson, M. A., Ward-Thompson, D., Michie, R. W., 1963. The dynamics of spherical stellar systems, IV. White, G., Umana, G., Jul. 2011. A 100 pc Elliptical and Twisted MNRAS126, 499–+. Ring of Cold and Dense Molecular Clouds Revealed by Herschel Michie, R. W., Bodenheimer, P. H., 1963. The dynamics of spherical Around the Galactic Center. ApJ735, L33. stellar systems, II. MNRAS126, 269. Moore, B., Ghigna, S., Governato, F., Lake, G., Quinn, T., Stadel, Mieske, S., Frank, M. J., Baumgardt, H., Lutzgendorf,¨ N., Neumayer, J., Tozzi, P., Oct. 1999. Dark Matter Substructure within Galactic N., Hilker, M., Oct. 2013. On central black holes in ultra-compact Halos. ApJ524, L19–L22. dwarf galaxies. A&A558, A14. Moore, T. J. T., Urquhart, J. S., Morgan, L. K., Thompson, M. A., Oct. Mieske, S., Hilker, M., Infante, L., Mar. 2002. Ultra compact objects 2012. The effect of spiral arms on star formation in the Galaxy. in the Fornax cluster of galaxies: Globular clusters or dwarf galax- MNRAS426, 701–707. ies? A&A383, 823–837. Moreno, E., Pichardo, B., Velazquez,´ H., Oct. 2014. Tidal Radii and Mieske, S., Hilker, M., Infante, L., Jordan,´ A., May 2006. Spec- Destruction Rates of Globular Clusters in the Milky Way due to troscopic Metallicities for Fornax Ultracompact Dwarf Galax- Bulge-Bar and Disk Shocking. ApJ793, 110. ies, Globular Clusters, and Nucleated Dwarf Elliptical Galaxies. Moreno, J., Torrey, P., Ellison, S. L., Patton, D. R., Bluck, A. F. L., AJ131, 2442–2451. Bansal, G., Hernquist, L., Apr. 2015. Mapping galaxy encounters Mieske, S., Hilker, M., Jordan,´ A., Infante, L., Kissler-Patig, M., Re- in numerical simulations: the spatial extent of induced star forma- jkuba, M., Richtler, T., Cotˆ e,´ P., Baumgardt, H., West, M. J., Fer- tion. MNRAS448, 1107–1117. rarese, L., Peng, E. W., Sep. 2008. The nature of UCDs: Internal Morris, M., 1989. On Heating, Ionization, and Star Formation in the dynamics from an expanded sample and homogeneous database. Galactic Center Region. In: Morris, M. (Ed.), The Center of the A&A487, 921–935. Galaxy. Vol. 136 of IAU Symposium. p. 171. Mieske, S., Hilker, M., Misgeld, I., Jan. 2012. The specific frequen- Motte, F., Nguyenˆ Luong, Q., Schneider, N., Heitsch, F., Glover, S., cies of ultra-compact dwarf galaxies. A&A537, A3. Carlhoff, P., Hill, T., Bontemps, S., Schilke, P., Louvet, F., Henne- Miholics, M., Webb, J. J., Sills, A., Dec. 2014. The size of star clusters mann, M., Didelon, P., Beuther, H., Nov. 2014. The formation of accreted by the Milky Way. MNRAS445, 2872–2877. the W43 complex: constraining its atomic-to-molecular transition Miholics, M., Webb, J. J., Sills, A., Feb. 2016. The dynamical evolu- and searching for colliding clouds. A&A571, A32. tion of accreted star clusters in the Milky Way. MNRAS456, 240– Mucciarelli, A., Dalessandro, E., Ferraro, F. R., Origlia, L., Lanzoni, 247. B., Sep. 2014. No Evidence of Chemical Anomalies in the Bi- Mikkola, S., May 2008. A Brief History of Regularisation. In: Ves- modal Turnoff Cluster NGC 1806 in the Large Magellanic Cloud. perini, E., Giersz, M., Sills, A. (Eds.), Dynamical Evolution of ApJ793, L6. Dense Stellar Systems. Vol. 246 of IAU Symposium. pp. 218–227. Mullan, B., Konstantopoulos, I. S., Kepley, A. A., Lee, K. H., Charl- Milgrom, M., Jul. 1983. A modification of the Newtonian dynamics ton, J. C., Knierman, K., Bastian, N., Chandar, R., et al., Apr. 2011. as a possible alternative to the hidden mass hypothesis. ApJ270, Star Clusters in the Tidal Tails of Interacting Galaxies: Cluster 365–370. Populations Across a Variety of Tail Environments. ApJ731, 93. Milone, A. P., Bedin, L. R., Piotto, G., Anderson, J., King, I. R., Muratov, A. L., Gnedin, O. Y., Aug. 2010. Modeling the Metallicity Sarajedini, A., Dotter, A., Chaboyer, B., Mar´ın-Franch, A., Ma- Distribution of Globular Clusters. ApJ718, 1266–1288. jewski, S., Aparicio, A., Hempel, M., Paust, N. E. Q., Reid, I. N., Murray, N., Quataert, E., Thompson, T. A., Jan. 2010. The Disrup- Rosenberg, A., Siegel, M., Jan. 2008. The ACS Survey of Galactic tion of Giant Molecular Clouds by Radiation Pressure and the Ef-

56 ficiency of Star Formation in Galaxies. ApJ709, 191–209. Giguere,` C.-A., Murray, N., Feb. 2017. Giant clumps in the FIRE Naab, T., Ostriker, J. P., Aug. 2017. Theoretical Challenges in Galaxy simulations: a case study of a massive high-redshift galaxy. MN- Formation. ARA&A55, 59–109. RAS465, 952–969. Najarro, F., Figer, D. F., Hillier, D. J., Kudritzki, R. P., Apr. 2004. Ostriker, J. P., Spitzer, L. J., Chevalier, R. A., Sep. 1972. On the Evo- Metallicity in the Galactic Center: The Arches Cluster. ApJ611, lution of Globular Clusters. ApJ176, L51+. L105–L108. O’Sullivan, E., Ponman, T. J., Apr. 2004. XMM-Newton and Chandra Nakamura, F., Li, Z.-Y., Jun. 2007. Protostellar Turbulence Driven by observations of three X-ray-faint early-type galaxies. MNRAS349, Collimated Outflows. ApJ662, 395–412. 535–546. Naoz, S., Narayan, R., Aug. 2014. Globular Clusters and Dark Satel- Pancino, E., Romano, D., Tang, B., Tautvaisienˇ e,˙ G., Casey, A. R., lite Galaxies through the Stream Velocity. ApJ791, L8. Gruyters, P., Geisler, D., San Roman, I., Randich, S., Alfaro, E. J., Neumayer, N., Walcher, C. J., 2012. Are Nuclear Star Clusters the Bragaglia, A., Flaccomio, E., Korn, A. J., Recio-Blanco, A., Smil- Precursors of Massive Black Holes? Advances in Astronomy janic, R., Carraro, G., Bayo, A., Costado, M. T., Damiani, F., Jofre,´ 2012, 709038. P., Lardo, C., de Laverny, P., Monaco, L., Morbidelli, L., Sbordone, Newberg, H. J., Yanny, B., Rockosi, C., Grebel, E. K., Rix, H.-W., L., Sousa, S. G., Villanova, S., May 2017. The Gaia-ESO Survey. Brinkmann, J., Csabai, I., Hennessy, G., Hindsley, R. B., Ibata, R., Mg-Al anti-correlation in iDR4 globular clusters. A&A601, A112. Ivezic,´ Z., Lamb, D., Nash, E. T., Odenkirchen, M., Rave, H. A., Parker, R. J., Goodwin, S. P., Wright, N. J., Meyer, M. R., Quanz, Schneider, D. P., Smith, J. A., Stolte, A., York, D. G., Apr. 2002. S. P., Jun. 2016. Mass segregation in star clusters is not energy The Ghost of Sagittarius and Lumps in the Halo of the Milky Way. equipartition. MNRAS459, L119–L123. ApJ569, 245–274. Parmentier, G., Fritze, U., Jan. 2009. When Efficient Star Formation Ngan, W. H. W., Carlberg, R. G., Jun. 2014. Using Gaps in N-body Drives Cluster Formation. ApJ690, 1112–1118. Tidal Streams to Probe Missing Satellites. ApJ788, 181. Pasquali, A., Bik, A., Zibetti, S., Ageorges, N., Seifert, W., Brandner, Nguyen, N. K., Pettitt, A. R., Tasker, E. J., Okamoto, T., Dec. 2017. W., Rix, H.-W., Jutte,¨ M., Knierim, V., Buschkamp, P., Feiz, C., The Impact of Galactic Disc Environment on Star-Forming Clouds. Gemperlein, H., Germeroth, A., Hofmann, R., Laun, W., Lederer, ArXiv e-prints. R., Lehmitz, M., Lenzen, R., Mall, U., Mandel, H., Muller,¨ P., Nguyen-Luong, Q., Anderson, L. D., Motte, F., Kim, K.-T., Schilke, Naranjo, V., Polsterer, K., Quirrenbach, A., Schaffner,¨ L., Storz, P., Carlhoff, P., Beuther, H., Schneider, N., Didelon, P., Kramer, C., Weiser, P., Apr. 2011. Infrared Narrowband Tomography of C., Louvet, F., Nony, T., Bihr, S., Rugel, M., Soler, J., Wang, Y., the Local Starburst NGC 1569 with the Large Binocular Tele- Bronfman, L., Simon, R., Menten, K. M., Wyrowski, F., Walmsley, scope/LUCIFER. AJ141, 132. C. M., Aug. 2017. Large-scale Map of Millimeter-wavelength Hy- Patton, D. R., Torrey, P., Ellison, S. L., Mendel, J. T., Scudder, J. M., drogen Radio Recombination Lines around a Young Massive Star Jun. 2013. Galaxy pairs in the Sloan Digital Sky Survey - VI. The Cluster. ApJ844, L25. orbital extent of enhanced star formation in interacting galaxies. Nitadori, K., Aarseth, S. J., Jul. 2012. Accelerating NBODY6 with MNRAS433, L59–L63. graphics processing units. MNRAS424, 545–552. Penarrubia,˜ J., Benson, A. J., Walker, M. G., Gilmore, G., Mc- Noguchi, M., Sep. 1988. Gas dynamics in interacting disc galaxies - Connachie, A. W., Mayer, L., Aug. 2010. The impact of dark mat- Fuelling of nuclei by induced bars. A&A203, 259–272. ter cusps and cores on the satellite galaxy population around spiral Norris, M. A., Kannappan, S. J., Jun. 2011. The ubiquity and dual galaxies. MNRAS406, 1290–1305. nature of ultra-compact dwarfs. MNRAS414, 739–758. Penarrubia,˜ J., Varri, A. L., Breen, P. G., Ferguson, A. M. N., Sanchez-´ Norris, M. A., Kannappan, S. J., Forbes, D. A., Romanowsky, A. J., Janssen, R., Oct. 2017. Stellar envelopes of globular clusters em- Brodie, J. P., Faifer, F. R., Huxor, A., Maraston, C., Moffett, A. J., bedded in dark mini-haloes. MNRAS471, L31–L35. Penny, S. J., Pota, V., Smith-Castelli, A., Strader, J., Bradley, D., Pearson, S., Kupper,¨ A. H. W., Johnston, K. V., Price-Whelan, A. M., Eckert, K. D., Fohring, D., McBride, J., Stark, D. V., Vaduvescu, Jan. 2015. Tidal Stream Morphology as an Indicator of Dark Mat- O., Sep. 2014. The AIMSS Project - I. Bridging the star cluster- ter Halo Geometry: The Case of Palomar 5. ApJ799, 28. galaxy divide. MNRAS443, 1151–1172. Pearson, S., Price-Whelan, A. M., Johnston, K. V., Mar. 2017. Gaps Odenkirchen, M., Brosche, P., Geffert, M., Tucholke, H.-J., Nov. in Globular Cluster Streams: Pal 5 and the Galactic Bar. ArXiv 1997. Globular cluster orbits based on Hipparcos proper motions. e-prints. New A2, 477–499. Peebles, P. J. E., Dicke, R. H., Dec. 1968. Origin of the Globular Star Odenkirchen, M., Grebel, E. K., Dehnen, W., Rix, H., Yanny, Clusters. ApJ154, 891. B., Newberg, H. J., Rockosi, C. M., Mart´ınez-Delgado, D., Pelupessy, F. I., van Elteren, A., de Vries, N., McMillan, S. L. W., Brinkmann, J., Pier, J. R., Nov. 2003. The Extended Tails of Palo- Drost, N., Portegies Zwart, S. F., Sep. 2013. The Astrophysical mar 5: A 10deg Arc of Globular Cluster Tidal Debris. AJ126, Multipurpose Software Environment. A&A557, A84. 2385–2407. Peng, E. W., Jordan,´ A., Cotˆ e,´ P., Blakeslee, J. P., Ferrarese, L., Mei, Odenkirchen, M., Grebel, E. K., Rockosi, C. M., Dehnen, W., Ibata, S., West, M. J., Merritt, D., Milosavljevic,´ M., Tonry, J. L., Mar. R., Rix, H.-W., Stolte, A., Wolf, C., Anderson, Jr., J. E., Bahcall, 2006. The ACS Virgo Cluster Survey. IX. The Color Distributions N. A., Brinkmann, J., Csabai, I., Hennessy, G., Hindsley, R. B., of Globular Cluster Systems in Early-Type Galaxies. ApJ639, 95– Ivezic,´ Z.,ˇ Lupton, R. H., Munn, J. A., Pier, J. R., Stoughton, C., 119. York, D. G., Feb. 2001. Detection of Massive Tidal Tails around the Perret, V., Renaud, F., Epinat, B., Amram, P., Bournaud, F., Con- Globular Cluster Palomar 5 with Sloan Digital Sky Survey Com- tini, T., Teyssier, R., Lambert, J.-C., Feb. 2014. Evolution of the missioning Data. ApJ548, L165–L169. mass, size, and star formation rate in high redshift merging galax- Offner, S. S. R., Klein, R. I., McKee, C. F., Krumholz, M. R., Sep. ies. MIRAGE - A new sample of simulations with detailed stellar 2009. The Effects of Radiative Transfer on Low-Mass Star Forma- feedback. A&A562, A1. tion. ApJ703, 131–149. Petts, J. A., Gualandris, A., Read, J. I., Dec. 2015. A semi-analytic dy- Oka, T., Geballe, T. R., Goto, M., Usuda, T., McCall, B. J., Oct. namical friction model that reproduces core stalling. MNRAS454, 2005. Hot and Diffuse Clouds near the Galactic Center Probed by 3778–3791. + Metastable H 31,. ApJ632, 882–893. Petts, J. A., Read, J. I., Gualandris, A., Nov. 2016. A semi-analytic Oklopciˇ c,´ A., Hopkins, P. F., Feldmann, R., Keres,ˇ D., Faucher- dynamical friction model for cored galaxies. MNRAS463, 858–

57 869. Star cluster survival and compressive tides in Antennae-like merg- Peuten, M., Zocchi, A., Gieles, M., Gualandris, A., Henault-Brunet,´ ers. MNRAS391, L98–L102. V., Nov. 2016. A stellar-mass black hole population in the globular Renaud, F., Boily, C. M., Naab, T., Theis, C., Nov. 2009. Fully Com- cluster NGC 6101? MNRAS462, 2333–2342. pressive Tides in Galaxy Mergers. ApJ706, 67–82. Pfeffer, J., Baumgardt, H., Aug. 2013. Ultra-compact dwarf galaxy Renaud, F., Bournaud, F., Duc, P.-A., Jan. 2015a. A parsec-resolution formation by tidal stripping of nucleated dwarf galaxies. MN- simulation of the Antennae galaxies: formation of star clusters dur- RAS433, 1997–2005. ing the merger. MNRAS446, 2038–2054. Pfeffer, J., Griffen, B. F., Baumgardt, H., Hilker, M., Nov. 2014. Con- Renaud, F., Bournaud, F., Emsellem, E., Agertz, O., Athanassoula, tribution of stripped nuclear clusters to globular cluster and ultra- E., Combes, F., Elmegreen, B., Kraljic, K., Motte, F., Teyssier, R., compact dwarf galaxy populations. MNRAS444, 3670–3683. Dec. 2015b. Environmental regulation of cloud and star formation Pflamm-Altenburg, J., Kroupa, P., Jul. 2009. Recurrent gas accre- in galactic bars. MNRAS454, 3299–3310. tion by massive star clusters, multiple stellar populations and mass Renaud, F., Bournaud, F., Emsellem, E., Elmegreen, B., Teyssier, thresholds for spheroidal stellar systems. MNRAS397, 488–494. R., Mar. 2014a. Beads on a String and Spurs in Galactic Disks. Piatti, A. E., Cole, A., Sep. 2017. Observational hints of a real age In: Seigar, M. S., Treuthardt, P. (Eds.), Structure and Dynamics spread in the young LMC star cluster NGC 1971. MNRAS470, of Disk Galaxies. Vol. 480 of Astronomical Society of the Pacific L77–L81. Conference Series. p. 247. Pichardo, B., Martos, M., Moreno, E., Jul. 2004. Models for the Grav- Renaud, F., Bournaud, F., Emsellem, E., Elmegreen, B., Teyssier, itational Field of the Galactic Bar: An Application to Stellar Or- R., Alves, J., Chapon, D., Combes, F., et al., Dec. 2013. A sub- bits in the Galactic Plane and Orbits of Some Globular Clusters. parsec resolution simulation of the Milky Way: global structure of ApJ609, 144–165. the interstellar medium and properties of molecular clouds. MN- Piotto, G., Bedin, L. R., Anderson, J., King, I. R., Cassisi, S., Milone, RAS436, 1836–1851. A. P., Villanova, S., Pietrinferni, A., Renzini, A., May 2007. A Renaud, F., Bournaud, F., Kraljic, K., Duc, P.-A., Jul. 2014b. Star- Triple Main Sequence in the Globular Cluster NGC 2808. ApJ661, bursts triggered by intergalactic tides andinterstellar compressive L53–L56. turbulence. MNRAS442, L33–L37. Piskunov, A. E., Kharchenko, N. V., Roser,¨ S., Schilbach, E., Scholz, Renaud, F., Gieles, M., Apr. 2013. The role of galaxy mergers on the R.-D., Jan. 2006. Revisiting the population of Galactic open clus- evolution of star clusters. MNRAS431, L83–L87. ters. A&A445, 545–565. Renaud, F., Gieles, M., Apr. 2015a. A flexible method to evolve col- Ploeckinger, S., Oct. 2015. Tides or dark matter sub-haloes: Which lisional systems and their tidal debris in external potentials. MN- ones are more attractive? MNRAS452, 3742–3751. RAS448, 3416–3422. Popa, C., Naoz, S., Marinacci, F., Vogelsberger, M., Aug. 2016. Gas- Renaud, F., Gieles, M., May 2015b. The effect of secular galactic rich and gas-poor structures through the stream velocity effect. growth on the evolution of star clusters. MNRAS449, 2734–2740. MNRAS460, 1625–1639. Renaud, F., Gieles, M., Boily, C. M., Dec. 2011. Evolution of star Portegies Zwart, S. F., Hut, P., Makino, J., McMillan, S., Sep. 1998. clusters in arbitrary tidal fields. MNRAS418, 759–769. On the dissolution of evolving star clusters. A&A337, 363–371. Renaud, F., Kraljic, K., Bournaud, F., Nov. 2012. Star Formation Portegies Zwart, S. F., McMillan, S. L. W., Gieles, M., Sep. 2010. Laws and Thresholds from Interstellar Medium Structure and Tur- Young Massive Star Clusters. ARA&A48, 431–493. bulence. ApJ760, L16. Pota, V., Forbes, D. A., Romanowsky, A. J., Brodie, J. P., Spitler, Rey-Raposo, R., Dobbs, C., Agertz, O., Alig, C., Jan. 2017. The L. R., Strader, J., Foster, C., Arnold, J. A., Benson, A., Blom, C., roles of stellar feedback and galactic environment in star-forming Hargis, J. R., Rhode, K. L., Usher, C., Jan. 2013. The SLUGGS molecular clouds. MNRAS464, 3536–3551. Survey: kinematics for over 2500 globular clusters in 12 early-type Rey-Raposo, R., Dobbs, C., Duarte-Cabral, A., Jan. 2015. Are tur- galaxies. MNRAS428, 389–420. bulent spheres suitable initial conditions for star-forming clouds? Press, W. H., Schechter, P., Feb. 1974. Formation of Galaxies and MNRAS446, L46–L50. Clusters of Galaxies by Self-Similar Gravitational Condensation. Ricci, F., Marchesi, S., Shankar, F., La Franca, F., Civano, F., Feb. ApJ187, 425–438. 2017. Constraining the UV emissivity of AGN throughout cosmic Prieto, J. L., Gnedin, O. Y., Dec. 2008. Dynamical Evolution of Glob- time via X-ray surveys. MNRAS465, 1915–1925. ular Clusters in Hierarchical Cosmology. ApJ689, 919–935. Rich, R. M., Ryde, N., Thorsbro, B., Fritz, T. K., Schultheis, M., Rafikov, R. R., May 2001. The local axisymmetric instability criterion Origlia, L., Jonsson,¨ H., Dec. 2017. Detailed Abundances for the in a thin, rotating, multicomponent disc. MNRAS323, 445–452. Old Population near the Galactic Center. I. Metallicity Distribution Raghavan, D., McAlister, H. A., Henry, T. J., Latham, D. W., Marcy, of the Nuclear Star Cluster. AJ154, 239. G. W., Mason, B. D., Gies, D. R., White, R. J., ten Brummelaar, Ricotti, M., Parry, O. H., Gnedin, N. Y., Nov. 2016. A Common Ori- T. A., Sep. 2010. A Survey of Stellar Families: Multiplicity of gin for Globular Clusters and Ultra-faint Dwarfs in Simulations of Solar-type Stars. ApJS190, 1–42. the First Galaxies. ApJ831, 204. Rahner, D., Pellegrini, E. W., Glover, S. C. O., Klessen, R. S., Oct. Rieder, S., Ishiyama, T., Langelaan, P., Makino, J., McMillan, 2017. Winds and radiation in unison: a new semi-analytic feedback S. L. W., Portegies Zwart, S., Dec. 2013. Evolution of star clus- model for cloud dissolution. MNRAS470, 4453–4472. ters in a cosmological tidal field. MNRAS436, 3695–3706. Rahner, D., Pellegrini, E. W., Glover, S. C. O., Klessen, R. S., Jan. Robertson, B. E., Ellis, R. S., Furlanetto, S. R., Dunlop, J. S., Apr. 2018. Forming clusters within clusters: how 30 Doradus recol- 2015. Cosmic Reionization and Early Star-forming Galaxies: A lapsed and gave birth again. MNRAS473, L11–L15. Joint Analysis of New Constraints from Planck and the Hubble Read, J. I., Agertz, O., Collins, M. L. M., Jul. 2016. Dark matter cores Space Telescope. ApJ802, L19. all the way down. MNRAS459, 2573–2590. Roche, N., Humphrey, A., Gomes, J. M., Papaderos, P., Lagos, P., Read, J. I., Wilkinson, M. I., Evans, N. W., Gilmore, G., Kleyna, J. T., Sanchez,´ S. F., Nov. 2015. CALIFA spectroscopy of the interacting Feb. 2006. The tidal stripping of satellites. MNRAS366, 429–437. galaxy NGC 5394 (Arp 84): starbursts, enhanced [N II]6584 and Renaud, F., Agertz, O., Gieles, M., Mar. 2017. The origin of the Milky signs of outflows and shocks. MNRAS453, 2349–2363. Way globular clusters. MNRAS465, 3622–3636. Roman-Duval, J., Jackson, J. M., Heyer, M., Rathborne, J., Simon, R., Renaud, F., Boily, C. M., Fleck, J.-J., Naab, T., Theis, C., Nov. 2008. Nov. 2010. Physical Properties and Galactic Distribution of Molec-

58 ular Clouds Identified in the Galactic Ring Survey. ApJ723, 492– Scarpa, R., Marconi, G., Gilmozzi, R., Carraro, G., Jan. 2007. Using 507. globular clusters to test gravity in the weak acceleration regime: Romanowsky, A. J., Strader, J., Spitler, L. R., Johnson, R., Brodie, NGC 7099. A&A462, L9–L12. J. P., Forbes, D. A., Ponman, T., Jun. 2009. Mapping The Dark Side Scheepmaker, R. A., Lamers, H. J. G. L. M., Anders, P., Larsen, S. S., with DEIMOS: Globular Clusters, X-Ray Gas, and Dark Matter in Jan. 2009. The spatial distribution of star and cluster formation in the NGC 1407 Group. AJ137, 4956–4987. M 51. A&A494, 81–93. Romeo, A. B., Falstad, N., Aug. 2013. A simple and accurate ap- Schinnerer, E., Meidt, S. E., Colombo, D., Chandar, R., Dobbs, C. L., proximation for the Q stability parameter in multicomponent and Garc´ıa-Burillo, S., Hughes, A., Leroy, A. K., Pety, J., Querejeta, realistically thick discs. MNRAS433, 1389–1397. M., Kramer, C., Schuster, K. F., Feb. 2017. The PdBI Arcsecond Rosolowsky, E., Jan. 2007. Giant Molecular Clouds in M31. I. Molec- Whirlpool Survey (PAWS): The Role of Spiral Arms in Cloud and ular Cloud Properties. ApJ654, 240–251. Star Formation. ApJ836, 62. Rosolowsky, E., Keto, E., Matsushita, S., Willner, S. P., Jun. 2007. Schinnerer, E., Meidt, S. E., Pety, J., Hughes, A., Colombo, D., High-Resolution Molecular Gas Maps of M33. ApJ661, 830–844. Garc´ıa-Burillo, S., Schuster, K. F., Dumas, G., Dobbs, C. L., Ross, D. J., Mennim, A., Heggie, D. C., Feb. 1997. Escape from a Leroy, A. K., Kramer, C., Thompson, T. A., Regan, M. W., tidally limited star cluster. MNRAS284, 811–814. Dec. 2013. The PdBI Arcsecond Whirlpool Survey (PAWS). I. A Rossa, J., van der Marel, R. P., Boker,¨ T., Gerssen, J., Ho, L. C., Cloud-scale/Multi-wavelength View of the Interstellar Medium in Rix, H.-W., Shields, J. C., Walcher, C.-J., Sep. 2006. Hubble Space a Grand-design Spiral Galaxy. ApJ779, 42. Telescope STIS Spectra of Nuclear Star Clusters in Spiral Galax- Schweizer, F., Jan. 1982. Colliding and merging galaxies. I - Evidence ies: Dependence of Age and Mass on Hubble Type. AJ132, 1074– for the recent merging of two disk galaxies in NGC 7252. ApJ252, 1099. 455–460. Rossi, L. J., Hurley, J. R., Dec. 2015a. Estimating the impact of the Schweizer, F., 1987. Star formation in colliding and merging galaxies. Galactic bar on the evolution of Galactic star clusters from N-body In: Faber, S. M. (Ed.), Nearly Normal Galaxies. From the Planck simulations. MNRAS454, 1453–1467. Time to the Present. pp. 18–25. Rossi, L. J., Hurley, J. R., Feb. 2015b. Reconstructing the initial mass Scott, N., Graham, A. W., Feb. 2013. Updated Mass Scaling Relations function of disc-bulge Galactic globular clusters from N-body sim- for Nuclear Star Clusters and a Comparison to Supermassive Black ulations. MNRAS446, 3389–3403. Holes. ApJ763, 76. Roskar,ˇ R., Debattista, V. P., Quinn, T. R., Wadsley, J., Nov. 2012. Scudder, J. M., Ellison, S. L., Torrey, P., Patton, D. R., Mendel, J. T., Radial migration in disc galaxies - I. Transient spiral structure and Oct. 2012. Galaxy pairs in the Sloan Digital Sky Survey - V. Trac- dynamics. MNRAS426, 2089–2106. ing changes in star formation rate and metallicity out to separations Ruchti, G. R., Read, J. I., Feltzing, S., Pipino, A., Bensby, T., Oct. of 80 kpc. MNRAS426, 549–565. 2014. The hunt for the Milky Way’s accreted disc. MNRAS444, Semenov, V. A., Kravtsov, A. V., Gnedin, N. Y., Aug. 2017. The Phys- 515–526. ical Origin of Long Gas Depletion Times in Galaxies. ApJ845, Ruchti, G. R., Read, J. I., Feltzing, S., Serenelli, A. M., McMil- 133. lan, P., Lind, K., Bensby, T., Bergemann, M., et al., Jul. 2015. Sesar, B., Bovy, J., Bernard, E. J., Caldwell, N., Cohen, J. G., Foues- The Gaia-ESO Survey: a quiescent Milky Way with no significant neau, M., Johnson, C. I., Ness, M., Ferguson, A. M. N., Mar- dark/stellar accreted disc. MNRAS450, 2874–2887. tin, N. F., Price-Whelan, A. M., Rix, H.-W., Schlafly, E. F., Bur- Ryde, N., Schultheis, M., Jan. 2015. Chemical abundances of M gi- gett, W. S., Chambers, K. C., Flewelling, H., Hodapp, K. W., ants in the Galactic centre: A single metal-rich population with low Kaiser, N., Magnier, E. A., Platais, I., Tonry, J. L., Waters, C., [α/Fe]. A&A573, A14. Wyse, R. F. G., Aug. 2015. The Nature and Orbit of the Ophiuchus Sabbi, E., Lennon, D. J., Gieles, M., de Mink, S. E., Walborn, N. R., Stream. ApJ809, 59. Anderson, J., Bellini, A., Panagia, N., van der Marel, R., Ma´ız Seth, A. C., Cappellari, M., Neumayer, N., Caldwell, N., Bastian, N., Apellaniz,´ J., Aug. 2012. A Double Cluster at the Core of 30 Do- Olsen, K., Blum, R. D., Debattista, V. P., McDermid, R., Puzia, T., radus. ApJ754, L37. Stephens, A., May 2010. The NGC 404 Nucleus: Star Cluster and Sackett, P. D., Aug. 1999. The Shape of Dark Matter Halos. In: Mer- Possible Intermediate-mass Black Hole. ApJ714, 713–731. ritt, D. R., Valluri, M., Sellwood, J. A. (Eds.), Galaxy Dynamics Seth, A. C., Dalcanton, J. J., Hodge, P. W., Debattista, V. P., Dec. - A Rutgers Symposium. Vol. 182 of Astronomical Society of the 2006. Clues to Nuclear Star Cluster Formation from Edge-on Spi- Pacific Conference Series. rals. AJ132, 2539–2555. Sanchez,´ S. F., Rosales-Ortega, F. F., Iglesias-Paramo,´ J., Molla,´ Seth, A. C., van den Bosch, R., Mieske, S., Baumgardt, H., Brok, M., Barrera-Ballesteros, J., Marino, R. A., Perez,´ E., Sanchez-´ M. D., Strader, J., Neumayer, N., Chilingarian, I., Hilker, M., Mc- Blazquez, P., et al., Mar. 2014. A characteristic oxygen abundance Dermid, R., Spitler, L., Brodie, J., Frank, M. J., Walsh, J. L., Sep. gradient in galaxy disks unveiled with CALIFA. A&A563, A49. 2014. A supermassive black hole in an ultra-compact dwarf galaxy. Sanchez-Salcedo,´ F. J., Reyes-Iturbide, J., Hernandez, X., Aug. 2006. Nature513, 398–400. An extensive study of dynamical friction in dwarf galaxies: the Shanks, T., Metcalfe, N., Chehade, B., Findlay, J. R., Irwin, M. J., role of stars, dark matter, halo profiles and MOND. MNRAS370, Gonzalez-Solares, E., Lewis, J. R., Yoldas, A. K., Mann, R. G., 1829–1840. Read, M. A., Sutorius, E. T. W., Voutsinas, S., Aug. 2015. The Sanders, J. L., Bovy, J., Erkal, D., Apr. 2016. Dynamics of stream- VLT Survey Telescope ATLAS. MNRAS451, 4238–4252. subhalo interactions. MNRAS457, 3817–3835. Shapiro, K. L., Genzel, R., Forster¨ Schreiber, N. M., Mar. 2010. Star- Sandford, E., Kupper,¨ A. H. W., Johnston, K. V., Diemand, J., Sep. forming galaxies at z ˜ 2 and the formation of the metal-rich glob- 2017. Quantifying tidal stream disruption in a simulated Milky ular cluster population. MNRAS403, L36–L40. Way. MNRAS470, 522–538. Shara, M. M., Hurley, J. R., Jul. 2006. Dynamical Effects Dominate Sbordone, L., Bonifacio, P., Marconi, G., Buonanno, R., 2004. Chem- the Evolution of Cataclysmic Variables in Dense Star Clusters. ical abundances in Terzan 7. Mem. Soc. Astron. Italiana75, 396. ApJ646, 464–473. Scarpa, R., Falomo, R., Nov. 2010. Testing Newtonian gravity in the Sheth, K., Regan, M., Hinz, J. L., Gil de Paz, A., Menendez-´ low acceleration regime with globular clusters: the case of ω Cen- Delmestre, K., Munoz-Mateos,˜ J.-C., Seibert, M., Kim, T., et al., tauri revisited. A&A523, A43. Dec. 2010. The Spitzer Survey of Stellar Structure in Galaxies

59 (S4G). PASP122, 1397. B., Fang, M., Jun. 2017. Molecular Clouds in the Extreme Outer Shetty, R., Ostriker, E. C., Aug. 2006. Global Modeling of Spur For- Galaxy between l = 34.deg75 to 45.deg25. ApJS230, 17. mation in Spiral Galaxies. ApJ647, 997–1017. Swinbank, A. M., Papadopoulos, P. P., Cox, P., Krips, M., Ivison, Shima, K., Tasker, E. J., Habe, A., May 2017. Does feedback help R. J., Smail, I., Thomson, A. P., Neri, R., Richard, J., Ebeling, or hinder star formation? The effect of photoionization on star H., Nov. 2011. The Interstellar Medium in Distant Star-forming formation in giant molecular clouds. MNRAS467, 512–523. Galaxies: Turbulent Pressure, Fragmentation, and Cloud Scaling Sippel, A. C., Hurley, J. R., Mar. 2013. Multiple stellar-mass black Relations in a Dense Gas Disk at z = 2.3. ApJ742, 11. holes in globular clusters: theoretical confirmation. MNRAS430, Tacconi, L. J., Genzel, R., Neri, R., Cox, P., Cooper, M. C., Shapiro, L30–L34. K., Bolatto, A., Bouche,´ N., et al., Feb. 2010. High molecular Smilgys, R., Bonnell, I. A., Dec. 2017. Formation of stellar clusters. gas fractions in normal massive star-forming galaxies in the young MNRAS472, 4982–4991. Universe. Nature463, 781–784. Smith, B. J., Soria, R., Struck, C., Giroux, M. L., Swartz, D. A., Tacconi, L. J., Neri, R., Genzel, R., Combes, F., Bolatto, A., Cooper, Yukita, M., Mar. 2014. Extra-nuclear Starbursts: Young Luminous M. C., Wuyts, S., Bournaud, F., Burkert, A., Comerford, J., Cox, Hinge Clumps in Interacting Galaxies. AJ147, 60. P., Davis, M., Forster¨ Schreiber, N. M., Garc´ıa-Burillo, S., Gracia- Smith, R., Goodwin, S., Fellhauer, M., Assmann, P., Jan. 2013. Infant Carpio, J., Lutz, D., Naab, T., Newman, S., Omont, A., Saintonge, mortality in the hierarchical merging scenario: dependence on gas A., Shapiro Griffin, K., Shapley, A., Sternberg, A., Weiner, B., expulsion time-scales. MNRAS428, 1303–1311. May 2013. Phibss: Molecular Gas Content and Scaling Relations Sofue, Y., Rubin, V., 2001. Rotation Curves of Spiral Galaxies. in z ˜ 1-3 Massive, Main-sequence Star-forming Galaxies. ApJ768, ARA&A39, 137–174. 74. Soifer, B. T., Rowan-Robinson, M., Houck, J. R., de Jong, T., Neuge- Takahashi, K., Portegies Zwart, S. F., Jun. 2000. The Evolution of bauer, G., Aumann, H. H., Beichman, C. A., Boggess, N., Clegg, Globular Clusters in the Galaxy. ApJ535, 759–775. P. E., Emerson, J. P., Gillett, F. C., Habing, H. J., Hauser, M. G., Takahira, K., Tasker, E. J., Habe, A., Sep. 2014. Do Cloud-Cloud Col- Low, F. J., Miley, G., Young, E., Mar. 1984. Infrared galaxies in lisions Trigger High-mass Star Formation? I. Small Cloud Colli- the IRAS minisurvey. ApJ278, L71–L74. sions. ApJ792, 63. Sollima, A., Bellazzini, M., Smart, R. L., Correnti, M., Pancino, E., Tan, J. C., Jun. 2000. Star Formation Rates in Disk Galaxies and Cir- Ferraro, F. R., Romano, D., Jul. 2009. The non-peculiar velocity cumnuclear Starbursts from Cloud Collisions. ApJ536, 173–184. dispersion profile of the stellar system ω Centauri. MNRAS396, Tanikawa, A., Fukushige, T., Oct. 2010. Mass-Loss Timescale of Star 2183–2193. Clusters in an External Tidal Field. II. Effect of Mass Profile of Solomon, P. M., Rivolo, A. R., Barrett, J., Yahil, A., Aug. 1987. Mass, Parent Galaxy. PASJ62, 1215–1230. luminosity, and line width relations of Galactic molecular clouds. Tasker, E. J., Tan, J. C., Jul. 2009. Star Formation in Disk Galaxies. I. ApJ319, 730–741. Formation and Evolution of Giant Molecular Clouds via Gravita- Sormani, M. C., Binney, J., Magorrian, J., May 2015. Gas flow in tional Instability and Cloud Collisions. ApJ700, 358–375. barred potentials. MNRAS449, 2421–2435. Tasker, E. J., Wadsley, J., Pudritz, R., Mar. 2015. Star Formation in Spitler, L. R., Forbes, D. A., Jan. 2009. A new method for estimat- Disk Galaxies. III. Does Stellar Feedback Result in Cloud Death? ing dark matter halo masses using globular cluster systems. MN- ApJ801, 33. RAS392, L1–L5. Terlevich, E., 1980. N-Body Simulations of Open Clusters. In: Spitzer, L., 1987. Dynamical evolution of globular clusters. Princeton, Hesser, J. E. (Ed.), Star Clusters. Vol. 85 of IAU Symposium. p. NJ, Princeton University Press, 1987, 191 p. 165. Spitzer, L. J., Jan. 1958. Distribution of Galactic Clusters. ApJ127, Terlevich, E., Jan. 1987. Evolution of n-body open clusters. MN- 17. RAS224, 193–225. Spitzer, L. J., Chevalier, R. A., Jul. 1973. Random Gravitational En- The Dark Energy Survey Collaboration, Oct. 2005. The Dark Energy counters and the Evolution of Spherical Systems. V. Gravitational Survey. ArXiv Astrophysics e-prints. Shocks. ApJ183, 565–582. Theis, C., Jul. 2002. On a Formation Scenario of Star Clusters. Spurzem, R., Sep. 1999. Direct N-body Simulations. Journal of Com- Ap&SS281, 97–100. putational and Applied Mathematics 109, 407–432. Thomas, G. F., Famaey, B., Ibata, R., Lughausen,¨ F., Kroupa, P., Jul. Spurzem, R., Aarseth, S. J., Sep. 1996. Direct collisional simulation 2017a. Stellar streams as gravitational experiments. I. The case of of 10000 particles past core collapse. MNRAS282. Sagittarius. A&A603, A65. Stiavelli, M., Piotto, G., Capaccioli, M., Ortolani, S., 1991. Disk- Thomas, G. F., Famaey, B., Ibata, R., Renaud, F., Martin, N. F., shocking and the mass function of globular clusters. In: Janes, K. Kroupa, P., Sep. 2017b. Stellar streams as gravitational experi- (Ed.), The Formation and Evolution of Star Clusters. Vol. 13 of As- ments II. Asymmetric tails of globular cluster streams. ArXiv e- tronomical Society of the Pacific Conference Series. pp. 449–451. prints. Stolte, A., Ghez, A. M., Morris, M., Lu, J. R., Brandner, W., Thomas, G. F., Ibata, R., Famaey, B., Martin, N. F., Lewis, G. F., Aug. Matthews, K., Mar. 2008. The Proper Motion of the Arches Cluster 2016. Exploring the reality of density substructures in the Palomar with Keck Laser-Guide Star Adaptive Optics. ApJ675, 1278–1292. 5 stellar stream. MNRAS460, 2711–2719. Strader, J., Brodie, J. P., Cenarro, A. J., Beasley, M. A., Forbes, Tiongco, M. A., Vesperini, E., Varri, A. L., Sep. 2016. Kinematical D. A., Oct. 2005. Extragalactic Globular Clusters: Old Spectro- evolution of tidally limited star clusters: the role of retrograde stel- scopic Ages and New Views on Their Formation. AJ130, 1315– lar orbits. MNRAS461, 402–411. 1323. Tonini, C., Jan. 2013. The Metallicity Bimodality of Globular Cluster Strader, J., Caldwell, N., Seth, A. C., Jul. 2011. Star Clusters in M31. Systems: A Test of Galaxy Assembly and of the Evolution of the V. Internal Dynamical Trends: Some Troublesome, Some Reassur- Galaxy Mass-Metallicity Relation. ApJ762, 39. ing. AJ142, 8. Toomre, A., May 1964. On the gravitational stability of a disk of stars. Strader, J., Chomiuk, L., Maccarone, T. J., Miller-Jones, J. C. A., ApJ139, 1217–1238. Seth, A. C., Oct. 2012. Two stellar-mass black holes in the globular Toomre, A., Toomre, J., Dec. 1972. Galactic Bridges and Tails. cluster M22. Nature490, 71–73. ApJ178, 623–666. Sun, Y., Su, Y., Zhang, S.-B., Xu, Y., Chen, X.-P., Yang, J., Jiang, Z.- Trancho, G., Konstantopoulos, I. S., Bastian, N., Fedotov, K., Gal-

60 lagher, S., Mullan, B., Charlton, J. C., Apr. 2012. Gemini Spec- Vesperini, E., Varri, A. L., McMillan, S. L. W., Zepf, S. E., Sep. 2014. troscopic Survey of Young Star Clusters in Merging/Interacting Kinematical fingerprints of star cluster early dynamical evolution. Galaxies. IV. Stephan’s Quintet. ApJ748, 102. MNRAS443, L79–L83. Tremaine, S., Weinberg, M. D., Jul. 1984. A kinematic method for Villanova, S., Geisler, D., Carraro, G., Moni Bidin, C., Munoz,˜ C., measuring the pattern speed of barred galaxies. ApJ282, L5–L7. Dec. 2013. Ruprecht 106: The First Single Population Globular Tremaine, S. D., Ostriker, J. P., Spitzer, Jr., L., Mar. 1975. The forma- Cluster? ApJ778, 186. tion of the nuclei of galaxies. I - M31. ApJ196, 407–411. Vinko,´ J., Sarneczky,´ K., Balog, Z., Immler, S., Sugerman, B. E. K., Trenti, M., Padoan, P., Jimenez, R., Aug. 2015. The Relative and Ab- Brown, P. J., Misselt, K., Szabo,´ G. M., Csizmadia, S., Kun, M., solute Ages of Old Globular Clusters in the LCDM Framework. Klagyivik, P., Foley, R. J., Filippenko, A. V., Csak,´ B., Kiss, L. L., ApJ808, L35. Apr. 2009. The Young, Massive, Star Cluster Sandage-96 After the Trenti, M., van der Marel, R., Nov. 2013. No energy equipartition in Explosion of Supernova 2004dj in NGC 2403. ApJ695, 619–635. globular clusters. MNRAS435, 3272–3282. Voggel, K., Hilker, M., Richtler, T., Feb. 2016. Globular cluster clus- Tsatsi, A., Mastrobuono-Battisti, A., van de Ven, G., Perets, H. B., tering and tidal features around ultra-compact dwarf galaxies in the Bianchini, P., Neumayer, N., Jan. 2017. On the rotation of nuclear halo of NGC 1399. A&A586, A102. star clusters formed by cluster inspirals. MNRAS464, 3720–3727. von Hoerner, S., Mar. 1957. Internal structure of globular clusters. Ueda, J., Iono, D., Petitpas, G., Yun, M. S., Ho, P. T. P., Kawabe, ApJ125. R., Mao, R.-Q., Mart´ın, S., et al., Jan. 2012. Unveiling the Phys- von Hoerner, S., 1960. Die numerische Integration des n-Korper-¨ ical Properties and Kinematics of Molecular Gas in the Antennae Problemes fur¨ Sternhaufen. I. ZAp50. Galaxies (NGC 4038/9) through High-resolution CO (J = 3-2) Ob- von Hoerner, S., 1963. Die numerische Integration des n-Korper-¨ servations. ApJ745, 65. Problems fur¨ Sternhaufen, II. ZAp57. Valluri, M., May 1993. Compressive tidal heating of a disk galaxy in Wada, K., Koda, J., Mar. 2004. Instabilities of spiral shocks - I. Onset a rich cluster. ApJ408, 57–70. of wiggle instability and its mechanism. MNRAS349, 270–280. van de Ven, G., van den Bosch, R. C. E., Verolme, E. K., de Zeeuw, Walcher, C. J., Yates, R. M., Minchev, I., Chiappini, C., Bergemann, P. T., Jan. 2006. The dynamical distance and intrinsic structure of M., Bruzual, G., Charlot, S., Coelho, P. R. T., Gallazzi, A., Martig, the globular cluster ω Centauri. A&A445, 513–543. M., Oct. 2016. Self-similarity in the chemical evolution of galaxies van den Bergh, S., Jun. 1971. The Post-Eruptive Galaxy M 82. and the delay-time distribution of SNe Ia. A&A594, A61. A&A12, 474. Wang, L., Spurzem, R., Aarseth, S., Giersz, M., Askar, A., Berczik, P., van den Bergh, S., Apr. 2000. Updated Information on the Local Naab, T., Schadow, R., Kouwenhoven, M. B. N., May 2016. The Group. PASP112, 529–536. DRAGON simulations: globular cluster evolution with a million Vanzella, E., Calura, F., Meneghetti, M., Mercurio, A., Castellano, stars. MNRAS458, 1450–1465. M., Caminha, G. B., Balestra, I., Rosati, P., Tozzi, P., De Barros, Wang, Z., Fazio, G. G., Ashby, M. L. N., Huang, J.-S., Pahre, M. A., S., Grazian, A., D’Ercole, A., Ciotti, L., Caputi, K., Grillo, C., Smith, H. A., Willner, S. P., Forrest, W. J., et al., Sep. 2004. The Merlin, E., Pentericci, L., Fontana, A., Cristiani, S., Coe, D., Jun. Off-Nuclear Starbursts in NGC 4038/4039 (The Antennae Galax- 2017a. Paving the way for the JWST: witnessing globular cluster ies). ApJS154, 193–198. formation at z ¿ 3. MNRAS467, 4304–4321. Ward, R. L., Benincasa, S. M., Wadsley, J., Sills, A., Couchman, Vanzella, E., Castellano, M., Meneghetti, M., Mercurio, A., Caminha, H. M. P., Jan. 2016. The properties of bound and unbound molec- G. B., Cupani, G., Calura, F., Christensen, L., Merlin, E., Rosati, ular cloud populations formed in galactic disc simulations. MN- P., Gronke, M., Dijkstra, M., Mignoli, M., Gilli, R., De Barros, RAS455, 920–929. S., Caputi, K., Grillo, C., Balestra, I., Cristiani, S., Nonino, M., Watkins, L. L., van der Marel, R. P., Bellini, A., Anderson, J., Oct. Giallongo, E., Grazian, A., Pentericci, L., Fontana, A., Comastri, 2015. Hubble Space Telescope Proper Motion (HSTPROMO) Cat- A., Vignali, C., Zamorani, G., Brusa, M., Bergamini, P., Tozzi, alogs of Galactic Globular Clusters. III. Dynamical Distances and P., Jun. 2017b. Magnifying the Early Episodes of Star Formation: Mass-to-Light Ratios. ApJ812, 149. Super Star Clusters at Cosmological Distances. ApJ842, 47. Webb, J. J., Harris, W. E., Sills, A., Hurley, J. R., Feb. 2013. The Vanzella, E., de Barros, S., Vasei, K., Alavi, A., Giavalisco, M., Siana, Influence of Orbital Eccentricity on Tidal Radii of Star Clusters. B., Grazian, A., Hasinger, G., et al., Jul. 2016. Hubble Imaging of ApJ764, 124. the Ionizing Radiation from a Star-forming Galaxy at Z=3.2 with Webb, J. J., Leigh, N., Sills, A., Harris, W. E., Hurley, J. R., Aug. fesc¿50. ApJ825, 41. 2014a. The effect of orbital eccentricity on the dynamical evolution Varri, A. L., Bertin, G., Oct. 2009. Properties of Quasi-relaxed Stellar of star clusters. MNRAS442, 1569–1577. Systems in an External Tidal Field. ApJ703, 1911–1922. Webb, J. J., Patel, S. S., Vesperini, E., Feb. 2017. The Early Evolution Varri, A. L., Bertin, G., Apr. 2012. Self-consistent models of quasi- of Star Clusters in Compressive and Extensive Tidal Fields. ArXiv relaxed rotating stellar systems. A&A540, A94. e-prints. Vasei, K., Siana, B., Shapley, A. E., Quider, A. M., Alavi, A., Rafel- Webb, J. J., Sills, A., Harris, W. E., Hurley, J. R., Nov. 2014b. The ef- ski, M., Steidel, C. C., Pettini, M., Lewis, G. F., Nov. 2016. The fects of orbital inclination on the scale size and evolution of tidally Lyman Continuum Escape Fraction of the Cosmic Horseshoe: A filling star clusters. MNRAS445, 1048–1055. Test of Indirect Estimates. ApJ831, 38. Webb, J. J., Vesperini, E., Jan. 2017. On the link between energy Vasiliev, E., Nov. 2014. Evolution of binary supermassive black holes equipartition and radial variation in the stellar mass function of and the final-parsec problem. ArXiv e-prints. star clusters. MNRAS464, 1977–1983. Vegetti, S., Koopmans, L. V. E., Jan. 2009. Bayesian strong Wehner, E. H., Harris, W. E., Jun. 2006. From Supermassive Black gravitational-lens modelling on adaptive grids: objective detection Holes to Dwarf Elliptical Nuclei: A Mass Continuum. ApJ644, of mass substructure in Galaxies. MNRAS392, 945–963. L17–L20. Vesperini, E., Jan. 2010. Star cluster dynamics. Philosophical Trans- Weidner, C., Bonnell, I. A., Zinnecker, H., Dec. 2010. Super Star actions of the Royal Society of London Series A 368, 829–849. Clusters Versus OB Associations. ApJ724, 1503–1508. Vesperini, E., Heggie, D. C., Aug. 1997. On the effects of dynamical Weinberg, M. D., Jan. 1986. Orbital decay of satellite galaxies in evolution on the initial mass function of globular clusters. MN- spherical systems. ApJ300, 93–111. RAS289, 898–920. Weinberg, M. D., Oct. 1994. Adiabatic invariants in stellar dynamics.

61 1: Basic concepts. AJ108, 1398–1402. meijer, A., Shrotriya, T. S., May 2017. Delay-time distribution of Weisz, D. R., Boylan-Kolchin, M., Jul. 2017. Local Group ultra-faint core-collapse supernovae with late events resulting from binary in- dwarf galaxies in the reionization era. MNRAS469, L83–L88. teraction. A&A601, A29. White, S. D. M., Apr. 1977. Mass segregation and missing mass in the Zaritsky, D., McCabe, K., Aravena, M., Athanassoula, E., Bosma, Coma cluster. MNRAS179, 33–41. A., Comeron,´ S., Courtois, H. M., Elmegreen, B. G., Elmegreen, Whitmore, B., Schweizer, F., Leitherer, C., Borne, K., Robert, C., D. M., Erroz-Ferrer, S., Gadotti, D. A., Hinz, J. L., Ho, L. C., Oct. 1993. Hubble Space Telescope discovery of candidate young Holwerda, B., Kim, T., Knapen, J. H., Laine, J., Laurikainen, E., globular clusters in the merger remnant NGC 7252. AJ106, 1354– Munoz-Mateos,˜ J. C., Salo, H., Sheth, K., Feb. 2016. Globular 1370. Cluster Populations: Results Including S4G Late-type Galaxies. Whitmore, B. C., 2003. The formation of star clusters. In: A Decade ApJ818, 99. of Hubble Space Telescope Science. pp. 153–178. Zepf, S. E., Ashman, K. M., English, J., Freeman, K. C., Sharples, Whitmore, B. C., Chandar, R., Schweizer, F., Rothberg, B., Leitherer, R. M., Aug. 1999. The Formation and Evolution of Candidate C., Rieke, M., Rieke, G., Blair, W. P., Mengel, S., Alonso-Herrero, Young Globular Clusters in NGC 3256. AJ118, 752–764. A., Jul. 2010. The Antennae Galaxies (NGC 4038/4039) Revisited: Zhang, Q., Fall, S. M., Dec. 1999. The Mass Function of Young Star Advanced Camera for Surveys and NICMOS Observations of a Clusters in the “Antennae” Galaxies. ApJ527, L81–L84. Prototypical Merger. AJ140, 75–109. Zhong, S., Berczik, P., Spurzem, R., Sep. 2014. Super Massive Black Whitmore, B. C., Schweizer, F., Mar. 1995. Hubble space telescope Hole in Galactic Nuclei with Tidal Disruption of Stars. ApJ792, observations of young star clusters in NGC-4038/4039, ’the anten- 137. nae’ galaxies. AJ109, 960–980. Zinn, R., Jun. 1985. The globular cluster system of the galaxy. IV - Whitmore, B. C., Zhang, Q., Leitherer, C., Fall, S. M., Schweizer, F., The halo and disk subsystems. ApJ293, 424–444. Miller, B. W., Oct. 1999. The Luminosity Function of Young Star Zonoozi, A. H., Haghi, H., Kupper,¨ A. H. W., Baumgardt, H., Frank, Clusters in “the Antennae” Galaxies (NGC 4038-4039). AJ118, M. J., Kroupa, P., Jun. 2014. Direct N-body simulations of globular 1551–1576. clusters - II. Palomar 4. MNRAS440, 3172–3183. Wielen, R., 1985. Dynamics of open star clusters. In: Goodman, J., Zonoozi, A. H., Kupper,¨ A. H. W., Baumgardt, H., Haghi, H., Kroupa, Hut, P. (Eds.), IAU Symp. 113: Dynamics of Star Clusters. pp. P., Hilker, M., Mar. 2011. Direct N-body simulations of globular 449–460. clusters - I. Palomar 14. MNRAS411, 1989–2001. Wise, J. H., Abel, T., Aug. 2007. Resolving the Formation of Proto- Zuckerman, B., Song, I., Sep. 2004. Young Stars Near the Sun. galaxies. I. Virialization. ApJ665, 899–910. ARA&A42, 685–721. Woolley, R. V. D. R., 1954. A study of the equilibrium of globular clusters. MNRAS114, 191. Worthey, G., Nov. 1994. Comprehensive stellar population models and the disentanglement of age and metallicity effects. ApJS95, 107–149. Wu, X., Zhao, H., Famaey, B., Jun. 2010. Lopsidedness of cluster galaxies in modified gravity. J. Cosmology Astropart. Phys.6, 010. Wyse, R. F. G., 2001. The Merging History of the Milky Way Disk. In: Funes, J. G., Corsini, E. M. (Eds.), Galaxy Disks and Disk Galax- ies. Vol. 230 of Astronomical Society of the Pacific Conference Series. pp. 71–80. Yanny, B., Newberg, H. J., Grebel, E. K., Kent, S., Odenkirchen, M., Rockosi, C. M., Schlegel, D., Subbarao, M., Brinkmann, J., Fukugita, M., Ivezic, Z.,ˇ Lamb, D. Q., Schneider, D. P., York, D. G., May 2003. A Low-Latitude Halo Stream around the Milky Way. ApJ588, 824–841. Yoon, J. H., Johnston, K. V., Hogg, D. W., Apr. 2011. Clumpy Streams from Clumpy Halos: Detecting Missing Satellites with Cold Stellar Structures. ApJ731, 58. Yoon, S.-J., Yi, S. K., Lee, Y.-W., Feb. 2006. Explaining the Color Distributions of Globular Cluster Systems in Elliptical Galaxies. Science 311, 1129–1132. Yoshida, N., Springel, V., White, S. D. M., Tormen, G., Dec. 2000. Weakly Self-interacting Dark Matter and the Structure of Dark Ha- los. ApJ544, L87–L90. Yusef-Zadeh, F., Hewitt, J. W., Arendt, R. G., Whitney, B., Rieke, G., Wardle, M., Hinz, J. L., Stolovy, S., Lang, C. C., Burton, M. G., Ramirez, S., Sep. 2009. Star Formation in the Central 400 pc of the Milky Way: Evidence for a Population of Massive Young Stellar Objects. ApJ702, 178–225. Zanella, A., Daddi, E., Le Floc’h, E., Bournaud, F., Gobat, R., Valentino, F., Strazzullo, V., Cibinel, A., Onodera, M., Perret, V., Renaud, F., Vignali, C., May 2015. An extremely young massive clump forming by gravitational collapse in a primordial galaxy. Nature521, 54–56. Zapartas, E., de Mink, S. E., Izzard, R. G., Yoon, S.-C., Badenes, C., Gotberg,¨ Y., de Koter, A., Neijssel, C. J., Renzo, M., Schoote-

62