Feedback from supernovae is often invoked as an important process in limiting star formation, removing gas from galaxies and hence as a determining process in galaxy formation. Here we report on numerical simulations investigating the interaction between supernova explosions and the natal molecular cloud. We also consider the cases with and without previous feedback from the high-mass star in the form of ionising radiation and stellar winds. The supernova is able to find weak points in the cloud and create channels through which it can escape, leaving much of the well shielded cloud largely unaffected. This effect is increased when the channels are pre-existing due to the effects of previous stellar feedback. The expanding supernova deposits its energy in the gas that is in these exposed channels, and hence sweeps up less mass when feedback has already occurred, resulting in faster outflows with less radiative losses. The full impact of the supernova explosion is then able to impact the larger scale of the galaxy in which it abides. We conclude that supernova explosions only have moderate effects on their dense natal environments but that with pre-existing feedback, the energetic effects of the supernova are able to escape and affect the wider scale medium of the galaxy.

Molecular oxygen has been the subject of many observational searches as chemical models predicted it to be a reservoir of oxygen. Although it has been detected in two regions of the interstellar medium, its rarity is a challenge for astrochemical models. In this paper, we have combined the physical conditions computed with smoothed particle hydrodynamics (SPH) simulations with our full gas-grain chemical model Nautilus, to study the predicted O2 abundance in interstellar material forming cold cores. We thus follow the chemical evolution of gas and ices in parcels of material from the diffuse interstellar conditions to the cold dense cores. Most of our predicted O2 abundances are below 1e-8 (with respect to the total proton density) and the predicted column densities in simulated cold cores is at maximum a few 1e14 cm-2, in agreement with the non detection limits. This low O2 abundance can be explained by the fact that, in a large fraction of the interstellar material, the atomic oxygen is depleted onto the grain surface (and hydrogenated to form H2O) before O2 can be formed in the gas-phase and protected from UV photo-dissociations. We could achieve this result only because we took into account the full history of the evolution of the physical conditions from the diffuse medium to the cold cores.

This paper investigates the gravitational trapping of HII regions predicted by steady-state analysis using radiation hydrodynamical simulations. We present idealised spherically symmetric radiation hydrodynamical simulations of the early evolution of HII regions including the gravity of the central source. As with analytic steady state solutions of spherically symmetric ionised Bondi accretion flows, we find gravitationally trapped HII regions with accretion through the ionisation front onto the source. We found that, for a constant ionising luminosity, fluctuations in the ionisation front are unstable. This instability only occurs in this spherically symmetric accretion geometry. In the context of massive star formation, the ionising luminosity increases with time as the source accretes mass. The maximum radius of the recurring HII region increases on the accretion timescale until it reaches the sonic radius, where the infall velocity equals the sound speed of the ionised gas, after which it enters a pressure-driven expansion phase. This expansion prevents accretion of gas through the ionisation front, the accretion rate onto the star decreases to zero, and it stops growing from accretion. Because of the time required for any significant change in stellar mass and luminosity through accretion our simulations keep both mass and luminosity constant and follow the evolution from trapped to expanding in a piecewise manner. Implications of this evolution of HII regions include a continuation of accretion of material onto forming stars for a period after the star starts to emit ionising radiation, and an extension of the lifetime of ultracompact HII regions.

We investigate the general stability of 1D spherically symmetric ionized Bondi accretion onto a massive object in the specific context of accretion onto a young stellar object. We first derive a new analytic expression for a steady state two temperature solution that predicts the existence of compact and hypercompact HII regions. We then show that this solution is only marginally stable if ionization is treated self-consistently. This leads to a recurring collapse of the HII region over time. We derive a semi-analytic model to explain this instability, and test it using spatially converged 1D radiation hydrodynamical simulations. We discuss the implications of the 1D instability on 3D radiation hydrodynamics simulations of supersonic accreting flows.

We revisit an alternate explanation for the turbulent nature of molecular clouds - namely, that velocity dispersions matching classical predictions of driven turbulence can be generated by the passage of clumpy material through a shock. While previous work suggested this mechanism can reproduce the observed Larson relation between velocity dispersion and size scale ($\sigma \propto L^{\Gamma}$ with $\Gamma \approx 0.5$), the effects of self-gravity and magnetic fields were not considered. We run a series of smoothed particle magnetohydrodynamics experiments, passing clumpy gas through a shock in the presence of a combination of self-gravity and magnetic fields. We find powerlaw relations between $\sigma$ and $L$ throughout, with indices ranging from $\Gamma=0.3-1.2$. These results are relatively insensitive to the strength and geometry of magnetic fields, provided that the shock is relatively strong. $\Gamma$ is strongly sensitive to the angle between the gas' bulk velocity and the shock front, and the shock strength (compared to the gravitational boundness of the pre-shock gas). If the origin of the $\sigma-L$ relation is in clumpy shocks, deviations from the standard Larson relation constrain the strength and behaviour of shocks in spiral galaxies.

We develop a semi-analytic model to investigate how accretion onto wide low-mass binary stars can result in a close high-mass binary system. The key ingredient is to allow mass accretion while limiting the gain in angular momentum. We envision this process as being regulated by an external magnetic field during infall. Molecular clouds are made to collapse spherically with material either accreting onto the stars or settling in a disk. Our aim is to determine what initial conditions are needed for the resulting binary to be both massive and close. Whether material accretes, and what happens to the binary separation as a result, depends on the relative size of its specific angular momentum, compared to the specific angular momentum of the binary. When we add a magnetic field we are introducing a torque to the system which is capable of stripping the molecular cloud of some of its angular momentum, and consequently easing the formation of high-mass binaries. Our results suggest that clouds in excess of 1000 M$_\odot$ and radii of 0.5 pc or larger, can easily form binary systems with masses in excess of 25 M$_\odot$ and separations of order 10 R$_\odot$ with magnetic fields of order 100 {\mu}G (mass-to-flux ratios of order 5).

We demonstrate numerical techniques for automatic identification of individual spiral arms in hydrodynamic simulations of astrophysical discs. Building on our earlier work, which used tensor classification to identify regions that were "spiral-like", we can now obtain fits to spirals for individual arm elements. We show this process can even detect spirals in relatively flocculent spiral patterns, but the resulting fits to logarithmic "grand-design" spirals are less robust. Our methods not only permit the estimation of pitch angles, but also direct measurements of the spiral arm width and pattern speed. In principle, our techniques will allow the tracking of material as it passes through an arm. Our demonstration uses smoothed particle hydrodynamics simulations, but we stress that the method is suitable for any finite-element hydrodynamics system. We anticipate our techniques will be essential to studies of star formation in disc galaxies, and attempts to find the origin of recently observed spiral structure in protostellar discs.

Aims. We study the effect of large scale dynamics on the molecular composition of the dense interstellar medium during the transition between diffuse to dense clouds. Methods. We followed the formation of dense clouds (on sub-parsec scales) through the dynamics of the interstellar medium at galac- tic scales. We used results from smoothed particle hydrodynamics (SPH) simulations from which we extracted physical parameters that are used as inputs for our full gas-grain chemical model. In these simulations, the evolution of the interstellar matter is followed for ~50 Myr. The warm low-density interstellar medium gas flows into spiral arms where orbit crowding produces the shock formation of dense clouds, which are held together temporarily by the external pressure. Results. We show that depending on the physical history of each SPH particle, the molecular composition of the modeled dense clouds presents a high dispersion in the computed abundances even if the local physical properties are similar. We find that carbon chains are the most affected species and show that these differences are directly connected to differences in (1) the electronic fraction, (2) the C/O ratio, and (3) the local physical conditions. We argue that differences in the dynamical evolution of the gas that formed dense clouds could account for the molecular diversity observed between and within these clouds. Conclusions. This study shows the importance of past physical conditions in establishing the chemical composition of the dense medium.

High-mass stars are commonly found in stellar clusters promoting the idea that their formation occurs due to the physical processes linked with a young stellar cluster. It has recently been reported that isolated high-mass stars are present in the Large Magellanic Cloud. Due to their low velocities it has been argued that these are high-mass stars which formed without a surrounding stellar cluster. In this paper we present an alternative explanation for the origin of these stars in which they formed in a cluster environment but are subsequently dispersed into the field as their natal cluster is tidally disrupted in a merger with a higher-mass cluster. They escape the merged cluster with relatively low velocities typical of the cluster interaction and thus of the larger scale velocity dispersion, similarly to the observed stars. $N$-body simulations of cluster mergers predict a sizeable population of low velocity ($\le$ 20 km s$^{-1}$), high-mass stars at distances of > 20 pc from the cluster. High-mass clusters in which gas poor mergers are frequent would be expected to commonly have halos of young stars, including high-mass stars, that were actually formed in a cluster environment.

We present high-resolution smoothed particle hydrodynamics simulations of a region of gas flowing in a spiral arm and identify dense gas clouds to investigate their kinematics with respect to a Milky Way model. We find that, on average, the gas in the arms can have a net radial streaming motion of $v_R \approx -9 \,\mathrm{km/s}$ and rotate $\approx 6 \,\mathrm{km/s}$ slower than the circular velocity. This translates to average peculiar motions towards the Galaxy centre and opposite to Galactic rotation. These results may be sensitive to the assumed spiral arm perturbation, which is $\approx 3\%$ of the disc potential in our model. We compare the actual distance and the kinematic estimate and we find that streaming motions introduce systematic offsets of $\approx 1$ kpc. We find that the distance error can be as large as $\pm 2$ kpc and the recovered cloud positions have distributions that can extend significantly into the inter-arm regions. We conclude that this poses a difficulty in tracing spiral arm structure in molecular cloud surveys.

Voronoi grids have been successfully used to represent density structures of gas in astronomical hydrodynamics simulations. While some codes are explicitly built around using a Voronoi grid, others, such as Smoothed Particle Hydrodynamics (SPH), use particle-based representations and can benefit from constructing a Voronoi grid for post-processing their output. So far, calculating the density of each Voronoi cell from SPH data has been done numerically, which is both slow and potentially inaccurate. This paper proposes an alternative analytic method, which is fast and accurate. We derive an expression for the integral of a cubic spline kernel over the volume of a Voronoi cell and link it to the density of the cell. Mass conservation is ensured rigorously by the procedure. The method can be applied more broadly to integrate a spherically symmetric polynomial function over the volume of a random polyhedron.

We investigate the triggering of star formation and the formation of stellar clusters in molecular clouds that form as the ISM passes through spiral shocks. The spiral shock compresses gas into $\sim$100 pc long main star formation ridge, where clusters forming every 5-10 pc along the merger ridge. We use a gravitational potential based cluster finding algorithm, which extracts individual clusters, calculates their physical properties and traces cluster evolution over multiple time steps. Final cluster masses at the end of simulation range between 1000 and 30000 M$_{\odot}$ with their characteristic half-mass radii between 0.1 pc and 2 pc. These clusters form by gathering material from 10-20 pc size scales. Clusters also show a mass - specific angular momentum relation, where more massive clusters have larger specific angular momentum due to the larger size scales, and hence angular momentum from which they gather their mass. The evolution shows that more massive clusters experiences hierarchical merging process, which increases stellar age spreads up to 2-3 Myr. Less massive clusters appear to grow by gathering nearby recently formed sinks, while more massive clusters with their large global gravitational potentials are increasing their mass growth from gas accretion.

We present 1.05 mm ALMA observations of the deeply embedded high-mass protocluster G11.92-0.61, designed to search for low-mass cores within the accretion reservoir of the massive protostars. Our ALMA mosaic, which covers an extent of ~0.7 pc at sub-arcsecond (~1400 au) resolution, reveals a rich population of 16 new millimetre continuum sources surrounding the three previously-known millimetre cores. Most of the new sources are located in the outer reaches of the accretion reservoir: the median projected separation from the central, massive (proto)star MM1 is ~0.17 pc. The derived physical properties of the new millimetre continuum sources are consistent with those of low-mass prestellar and protostellar cores in nearby star-forming regions: the median mass, radius, and density of the new sources are 1.3 Msun, 1600 au, and n(H2)~10^7 cm^-3. At least three of the low-mass cores in G11.92-0.61 drive molecular outflows, traced by high-velocity 12CO(3-2) (observed with the SMA) and/or by H2CO and CH3OH emission (observed with ALMA). This finding, combined with the known outflow/accretion activity of MM1, indicates that high- and low-mass stars are forming (accreting) simultaneously within this protocluster. Our ALMA results are consistent with the predictions of competitive-accretion-type models in which high-mass stars form along with their surrounding clusters.

We investigate how star formation efficiency can be significantly decreased by the removal of a molecular cloud's envelope by feedback from an external source. Feedback from star formation has difficulties halting the process in dense gas but can easily remove the less dense and warmer envelopes where star formation does not occur. However, the envelopes can play an important role keeping their host clouds bound by deepening the gravitational potential and providing a constraining pressure boundary. We use numerical simulations to show that removal of the cloud envelopes results in all cases in a fall in the star formation efficiency (SFE). At 1.38 free-fall times our 4 pc cloud simulation experienced a drop in the SFE from 16 to six percent, while our 5 pc cloud fell from 27 to 16 per cent. At the same time, our 3 pc cloud (the least bound) fell from an SFE of 5.67 per cent to zero when the envelope was lost. The star formation efficiency per free-fall time varied from zero to $\approx$ 0.25 according to $\alpha$, defined to be the ratio of the kinetic plus thermal to gravitational energy, and irrespective of the absolute star forming mass available. Furthermore the fall in SFE associated with the loss of the envelope is found to even occur at later times. We conclude that the SFE will always fall should a star forming cloud lose its envelope due to stellar feedback, with less bound clouds suffering the greatest decrease.

We investigate the role of magnetic fields in the fragmentation of self-gravitating discs using 3D global ideal magnetohydrodynamic simulations performed with the "phantom" smoothed particle hydrodynamics code. For initially toroidal fields, we find two regimes. In the first, where the cooling time is greater than five times the dynamical time, magnetic fields reduce spiral density wave amplitudes, which in turn suppresses fragmentation. This is the case even if the magnetic pressure is only a tenth of the thermal pressure. The second regime occurs when the cooling time is sufficiently short that magnetic fields cannot halt fragmentation. We find that magnetised discs produce more massive fragments, due to both the additional pressure exerted by the magnetic field, and the additional angular momentum transport induced by Maxwell stresses. The fragments are confined to a narrower range of initial semimajor axes than those in unmagnetised discs. The orbital eccentricity and inclination distributions of unmagnetised and magnetised disc fragments are similar. Our results suggest the fragmentation boundary could be at cooling times a factor of two lower than predicted by purely hydrodynamical models.

We investigate the triggering of star formation in clouds that form in Galactic scale flows as the ISM passes through spiral shocks. We use the Lagrangian nature of SPH simulations to trace how the star forming gas is gathered into self-gravitating cores that collapse to form stars. Large scale flows that arise due to Galactic dynamics create shocks of order 30 km/s that compress the gas and form dense clouds $(n> $several $\times 10^2$ cm$^{-3}$) in which self-gravity becomes relevant. These large-scale flows are necessary for creating the dense physical conditions for gravitational collapse and star formation. Local gravitational collapse requires densities in excess of $n>10^3$ cm$^{-3}$ which occur on size scales of $\approx 1$ pc for low-mass star forming regions ($M<100 M_{\odot}$), and up to sizes approaching 10 pc for higher-mass regions ($M>10^3 M_{\odot}$). Star formation in the 250 pc region lasts throughout the 5 Myr timescale of the simulation with a star formation rate of $\approx 10^{-1} M_{\odot}$ yr$^{-1}$ kpc$^{-2}$. In the absence of feedback, the efficiency of the star formation per free-fall time varies from our assumed 100 % at our sink accretion radius to values of $< 10^{-3}$ at low densities.

As hydrodynamic simulations increase in scale and resolution, identifying structures with non-trivial geometries or regions of general interest becomes increasingly challenging. There is a growing need for algorithms that identify a variety of different features in a simulation without requiring a "by-eye" search. We present tensor classification as such a technique for smoothed particle hydrodynamics (SPH). These methods have already been used to great effect in N-Body cosmological simulations, which require smoothing defined as an input free parameter. We show that tensor classification successfully identifies a wide range of structures in SPH density fields using its native smoothing, removing a free parameter from the analysis and preventing the need for tesselation of the density field, as required by some classification algorithms. As examples, we show that tensor classification using the tidal tensor and the velocity shear tensor successfully identifies filaments, shells and sheet structures in giant molecular cloud simulations, as well as spiral arms in discs. The relationship between structures identified using different tensors illustrates how different forces compete and co-operate to produce the observed density field. We therefore advocate the use of multiple tensors to classify structure in SPH simulations, to shed light on the interplay of multiple physical processes.

We examine the combined effects of winds and photoionizing radiation from O--type stars on embedded stellar clusters formed in model turbulent molecular clouds covering a range of masses and radii. We find that feedback is able to increase the quantities of dense gas present, but decreases the rate and efficiency of the conversion of gas to stars relative to control simulations in which feedback is absent. Star formation in these calculations often proceeds at a rate substantially slower than the freefall rate in the dense gas. This decoupling is due to the weakening of, and expulsion of gas from, the deepest parts of the clouds' potential wells where most of the star formation occurs in the control simulations. This results in large fractions of the stellar populations in the feedback simulation becoming dissociated from dense gas. However, where star formation \emph{does} occur in both control and feedback simulations, it does so in dense gas, so the correlation between star formation activity and dense gas is preserved. The overall dynamical effects of feedback on the \emph{clusters} are minimal, with only small fraction of stars becoming unbound, despite large quantities of gas being expelled from some clouds. This owes to the settling of the stars into virialised and stellar--dominated configurations before the onset of feedback. By contrast, the effects of feedback on the observable properties of the clusters -- their U--, B-- and V--band magnitudes -- are strong and sudden. The timescales on which the clusters become visible and unobscured are short compared with the timescales which the clouds are actually destroyed.

Aims. We investigate the effects of ionising photons on accretion and stellar mass growth in a young star forming region, using a Monte Carlo radiation transfer code coupled to a smoothed particle hydrodynamics (SPH) simulation. Methods. We introduce the framework with which we correct stellar cluster masses for the effects of photoionising (PI) feedback and compare to the results of a full ionisation hydrodynamics code. Results. We present results of our simulations of star formation in the spiral arm of a disk galaxy, including the effects of photoionising radiation from high mass stars. We find that PI feedback reduces the total mass accreted onto stellar clusters by approximately 23 per cent over the course of the simulation and reduces the number of high mass clusters, as well as the maximum mass attained by a stellar cluster. Mean star formation rates (SFRs) drop from 0.042 solar masses per year in our control run to 0.032 solar masses per year after the inclusion of PI feedback with a final instantaneous SFR reduction of 62 per cent. The overall cluster mass distribution appears to be affected little by PI feedback. Conclusions. We compare our results to the observed extra-galactic Schmidt-Kennicutt relation and the observed properties of local star forming regions in the Milky Way and find that internal photoionising (PI) feedback is unlikely to reduce star formation rates by more than a factor of approximately 2 and thus may play only a minor role in regulating star formation.

Turbulence is ubiquitous in the interstellar medium (ISM) of the Milky Way and other spiral galaxies. The energy source for this turbulence has been much debated with many possible origins proposed. The universality of turbulence, its reported large-scale driving, and that it occurs also in starless molecular clouds, challenges models invoking any stellar source. A more general process is needed to explain the observations. In this work we study the role of galactic spiral arms. This is accomplished by means of three-dimensional hydrodynamical simulations which follow the dynamical evolution of interstellar diffuse clouds (100cm-3) interacting with the gravitational potential field of the spiral pattern. We find that the tidal effects of the arm's potential on the cloud result in internal vorticity, fragmentation and hydrodynamical instabilities. The triggered turbulence result in large-scale driving, on sizes of the ISM inhomogeneities, i.e. as large as 100pc, and efficiencies in converting potential energy into turbulence in the range 10 to 25 percent per arm crossing. This efficiency is much higher than those found in previous models. The statistics of the turbulence in our simulations are strikingly similar to the observed power spectrum and Larson scaling relations of molecular clouds and the general ISM. The dependency found from different models indicate that the ISM turbulence is mainly related to local spiral arm properties, such as its mass density and width. This correlation seems in agreement with recent high angular resolution observations of spiral galaxies, e.g. M51 and M33.

Context: Star formation efficiency (SFE) theories are currently based on statistical distributions of turbulent cloud structures and a simple model of star formation from cores. They remain poorly tested, especially at the highest densities. Aims: We investigate the effects of gas density on the SFE through measurements of the core formation efficiency (CFE). With a total mass of $\sim2\times10^4$ M$_\odot$, the W43-MM1 ridge is one of the most convincing candidate precursor of starburst clusters and thus one of the best place to investigate star formation. Methods: We used high-angular resolution maps obtained at 3 mm and 1 mm within W43-MM1 with the IRAM Plateau de Bure Interferometer to reveal a cluster of 11 massive dense cores (MDCs), and, one of the most massive protostellar cores known. An Herschel column density image provided the mass distribution of the cloud gas. We then measured the 'instantaneous' CFE and estimated the SFE and the star formation rate (SFR) within subregions of the W43-MM1 ridge. Results: The high SFE found in the ridge ($\sim$6% enclosed in $\sim$8 pc$^3$) confirms its ability to form a starburst cluster. There is however a clear lack of dense cores in the northern part of the ridge, which may be currently assembling. The CFE and the SFE are observed to increase with volume gas density while the SFR steeply decreases with the virial parameter, $\alpha_{vir}$. Statistical models of the SFR may well describe the outskirts of the W43-MM1 ridge but struggle to reproduce its inner part, which corresponds to measurements at low $\alpha_{vir}$. It may be that ridges do not follow the log-normal density distribution, Larson relations, and stationary conditions forced in the statistical SFR models.

We model the combined effects of photoionization and momentum--driven winds from O--stars on molecular clouds spanning a parameter space of initial conditions. The dynamical effects of the winds are very modest. However, in the lower--mass clouds, they influence the morphologies of the HII regions by creating 10pc--scale central cavities.\\ The inhomogeneous structures of the model GMCs make them highly permeable to photons, ionized gas and supernova ejecta, and the leaking of ionized gas in particular strongly affects their evolution, reducing the effectiveness of feedback. Nevertheless, feedback is able to expel large fractions of the mass of the lower escape--velocity clouds. Its impact on star formation is more modest, decreasing final star formation efficiencies by 10--20$\%$, and the rate of change of the star formation efficiency per freefall time by about one third. However, the clouds still form stars substantially faster than observed GMCs.

The innermost parsec around Sgr A* has been found to play host to two discs or streamers of O and W-R stars. They are misaligned by an angle approaching 90 degrees. That the stars are approximately coeval indicates that they formed in the same event rather than independently. We have performed SPH simulations of the infall of a single prolate cloud towards a massive black hole. As the cloud is disrupted, the large spread in angular momentum can, if conditions allow, lead to the creation of misaligned gas discs. In turn, stars may form within those discs. We are now investigating the origins of these clouds in the Galactic Centre (GC) region.

We analyse a hydrodynamical simulation of star formation. Sink particles in the simulations which represent stars show episodic growth, which is presumably accretion from a core that can be regularly replenished in response to the fluctuating conditions in the local environment. The accretion rates follow $\dot{m} \propto m^{2/3}$, as expected from accretion in a gas-dominated potential, but with substantial variations over-laid on this. The growth times follow an exponential distribution which is tapered at long times due to the finite length of the simulation. The initial collapse masses have an approximately lognormal distribution with already an onset of a power-law at large masses. The sink particle mass function can be reproduced with a non-linear stochastic process, with fluctuating accretion rates $\propto m^{2/3}$, a distribution of seed masses and a distribution of growth times. All three factors contribute equally to the form of the final sink mass function. We find that the upper power law tail of the IMF is unrelated to Bondi-Hoyle accretion.

We examine the effect of momentum-driven OB-star stellar winds on a parameter space of simulated turbulent Giant Molecular Clouds using SPH hydrodynamical simulations. By comparison with identical simulations in which ionizing radiation was included instead of winds, we show that momentum-driven winds are considerably less effective in disrupting their host clouds than are HII regions. The wind bubbles produced are smaller and generally smoother than the corresponding ionization-driven bubbles. Winds are roughly as effective in destroying the very dense gas in which the O-stars are embedded, and thus shutting down the main regions of star-forming activity in the model clouds. However, their influence falls off rapidly with distance from the sources, so they are not as good at sweeping up dense gas and triggering star formation further out in the clouds. As a result, their effect on the star formation rate and efficiency is generally more negative than that of ionization, if they exert any effect at all.

Observations of atomic or molecular lines can provide important information about the physical state of star forming regions. In order to investigate the line profiles from dynamical collapsing massive star forming regions (MSFRs), we model the emission from hydrodynamic simulations of a collapsing cloud in the absence of outflows. By performing radiative transfer calculations, we compute the optically thick HCO+ and optically thin N2H+ line profiles from two collapsing regions at different epochs. Due to large-scale collapse, the MSFRs have large velocity gradients, reaching up to 20 km/s/pc across the central core. The optically thin lines typically contain multiple velocity components resulting from the superposition of numerous density peaks along the line-of-sight. The optically thick lines are only marginally shifted to the blue side of the optically thin line profiles, and frequently do not have a central depression in their profiles due to self-absorption. As the regions evolve the lines become brighter and the optically thick lines become broader. The lower order HCO+ (1-0) transitions are better indicators of collapse than the higher order (4-3) transitions. We also investigate how the beam sizes affect profile shapes. Smaller beams lead to brighter and narrower lines that are more skewed to the blue in HCO+ relative to the true core velocity, but show multiple components in N2H+. High resolution observations (e.g. with ALMA) can test these predictions and provide insights into the nature of MSFRs.

We follow the near radial infall of a prolate cloud onto a 4 x 10^6 Msun supermassive black hole in the Galactic Centre using smoothed particle hydrodynamics (SPH). We show that a prolate cloud oriented perpendicular to its orbital plane naturally produces a spread in angular momenta in the gas which can translate into misaligned discs as is seen in the young stars orbiting Sagittarius A*. A turbulent or otherwise highly structured cloud is necessary to avoid cancelling too much angular momentum through shocks at closest approach. Our standard model of a 2 x 10^4 Msun gas cloud brought about the formation of a disc within 0.3 pc from the black hole and a larger, misaligned streamer at 0.5 pc. A total of 1.5 x 10^4 Msun of gas formed these structures. Our exploration of the simulation parameter space showed that when star formation occurred, it resulted in top-heavy IMFs with stars on eccentric orbits with semi-major axes 0.02 to 0.3 pc and inclinations following the gas discs and streamers. We suggest that the single event of an infalling prolate cloud can explain the occurrence of multiple misaligned discs of young stars.

We present the fourth in a series of papers detailing our SPH study of the effects of ionizing feedback from O--type stars on turbulent star forming clouds. Here, we study the effects of photoionization on a series of initially partially unbound clouds with masses ranging from $10^{4}$--$10^{6}$M$_{\odot}$ and initial sizes from 2.5-45pc. We find that ionizing feedback profoundly affects the structure of the gas in most of our model clouds, creating large and often well-cleared bubble structures and pillars. However, changes in the structures of the embedded clusters produced are much weaker and not well correlated to the evolution of the gas. We find that in all cases, star formation efficiencies and rates are reduced by feedback and numbers of objects increased, relative to control simulations. We find that local triggered star formation does occur and that there is a good correlation between triggered objects and pillars or bubble walls, but that triggered objects are often spatially-mixed with those formed spontaneously. Some triggered objects acquire large enough masses to become ionizing sources themselves, lending support to the concept of propagating star formation. We find scant evidence for spatial age gradients in most simulations, and where we do see them, they are not a good indicator of triggering, as they apply equally to spontaneously-formed objects as triggered ones. Overall, we conclude that inferring the global or local effects of feedback on stellar populations from observing a system at a single epoch is very problematic.

Massive stars are known to have a high multiplicity, with examples of higher order multiples among the nearest and best studied objects. In this paper we study hierarchical multiple systems (an inner binary as a component of a wider binary) of massive stars in a clustered environment, in which a system with a size of 100--1000 au will undergo many close encounters during the short lifetime of a massive star. Using two types of N-body experiment we determine the post-formation collision probabilities of these massive hierarchies. We find that, depending on the specifics of the environment, the hierarchy, and the amount of time that is allowed to pass, tens of percent of hierarchies will experience a collision, typically between the two stars of the inner binary. In addition to collisions, clusters hosting a hierarchical massive system produce high velocity runaways at an enhanced rate. The primordial multiplicity specifics of massive stars appear to play a key role in the generation of these relatively small number events in cluster simulations, complicating their use as diagnostics of a cluster's history.

Understanding star formation is problematic as it originates in the large scale dynamics of a galaxy but occurs on the small scale of an individual star forming event. This paper presents the first numerical simulations to resolve the star formation process on sub-parsec scales, whilst also following the dynamics of the interstellar medium (ISM) on galactic scales. In these models, the warm low density ISM gas flows into the spiral arms where orbit crowding produces the shock formation of dense clouds, held together temporarily by their external pressure. Cooling allows the gas to be compressed to sufficiently high densities that local regions collapse under their own gravity and form stars. The star formation rates follow a Schmidt-Kennicutt \Sigma_{SFR} ~ \Sigma_{gas}^{1.4} type relation with the local surface density of gas while following a linear relation with the cold and dense gas. Cooling is the primary driver of star formation and the star formation rates as it determines the amount of cold gas available for gravitational collapse. The star formation rates found in the simulations are offset to higher values relative to the extragalactic values, implying a constant reduction, such as from feedback or magnetic fields, is likely to be required. Intriguingly, it appears that a spiral or other convergent shock and the accompanying thermal instability can explain how star formation is triggered, generate the physical conditions of molecular clouds and explain why star formation rates are tightly correlated to the gas properties of galaxies.

We have carried out a survey of the NGC 2068 region in the Orion B molecular cloud using HARP on the JCMT, in the 13CO and C18O (J = 3-2) and H13CO+ (J = 4-3) lines. We used 13CO to map the outflows in the region, and matched them with previously defined SCUBA cores. We decomposed the C18O and H13CO+ into Gaussian clumps, finding 26 and 8 clumps respectively. The average deconvolved radii of these clumps is 6200 +/- 2000 AU and 3600 +/- 900 AU for C18O and H13CO+ respectively. We have also calculated virial and gas masses for these clumps, and hence determined how bound they are. We find that the C18O clumps are more bound than the H13CO+ clumps (average gas mass to virial mass ratio of 4.9 compared to 1.4). We measure clump internal velocity dispersions of 0.28 +/- 0.02 kms-1 and 0.27 +/- 0.04 kms-1 for C18O and H13CO+ respectively, although the H13CO+ values are heavily weighted by a majority of the clumps being protostellar, and hence having intrinsically greater linewidths. We suggest that the starless clumps correspond to local turbulence minima, and we find that our clumps are consistent with formation by gravoturbulent fragmentation. We also calculate inter-clump velocity dispersions of 0.39 +/- 0.05 kms-1 and 0.28 +/- 0.08 kms-1 for C18O and H13CO+ respectively. The velocity dispersions (both internal and external) for our clumps match results from numerical simulations of decaying turbulence in a molecular cloud. However, there is still insufficient evidence to conclusively determine the type of turbulence and timescale of star formation, due to the small size of our sample.

We extend our previous SPH parameter study of the effects of photoionization from O-stars on star-forming clouds to include initially unbound clouds. We generate a set of model clouds in the mass range $10^{4}-10^{6}$M$_{\odot}$ with initial virial ratios $E_{\rm kin}/E_{\rm pot}$=2.3, allow them to form stars, and study the impact of the photoionizing radiation produced by the massive stars. We find that, on the 3Myr timescale before supernovae are expected to begin detonating, the fractions of mass expelled by ionizing feedback is a very strong function of the cloud escape velocities. High-mass clouds are largely unaffected dynamically, while lower-mass clouds have large fractions of their gas reserves expelled on this timescale. However, the fractions of stellar mass unbound are modest and significant portions of the unbound stars are so only because the clouds themselves are initially partially unbound. We find that ionization is much more able to create well-cleared bubbles in the unbound clouds, owing to their intrinsic expansion, but that the presence of such bubbles does not necessarily indicate that a given cloud has been strongly influenced by feedback. We also find, in common with the bound clouds from our earlier work, that many of the systems simulated here are highly porous to photons and supernova ejecta, and that most of them will likely survive their first supernova explosions.

We present a detailed study of star formation occurring in bound star--forming clouds under the influence of internal ionizing feedback from massive stars across a spectrum of cloud properties. We infer which objects are triggered by comparing our feedback simulations with control simulations in which no feedback was present. We find feedback always results in a lower star--formation efficiency and usually but not always results in a larger number of stars or clusters. Cluster mass functions are not strongly affected by feedback, but stellar mass functions are biased towards lower masses. Ionization also affects the geometrical distribution of stars in ways that are robust against projection effects, but may make the stellar associations more or less subclustered depending on the background cloud environment. We observe a prominent pillar in one simulation which is the remains of an accretion flow feeding the central ionizing cluster of its host cloud and suggest that this may be a general formation mechanism for pillars such as those observed in M16. We find that the association of stars with structures in the gas such as shells or pillars is a good but by no means foolproof indication that those stars have been triggered and we conclude overall that it is very difficult to deduce which objects have been induced to form and which formed spontaneously simply from observing the system at a single time.

We discuss three different ways in which stellar feedback may alter the outcome of star cluster formation: triggering or suppressing star formation, and redistributing the stellar population in space. We use detailed Smoothed Particle Hydrodynamics (SPH) simulations of HII regions in turbulent molecular clouds to show that all three of these may happen in the same system, making inferences about the effects of feedback problematic.

Infant mortality brought about by the expulsion of a star cluster's natal gas is widely invoked to explain cluster statistics at different ages. While a well studied problem, most recent studies of gas expulsion's effect on a cluster have focused on massive clusters, with stellar counts of order $10^4$. Here we argue that the evolutionary timescales associated with the compact low-mass clusters typical of the median cluster in the Solar neighborhood are short enough that significant dynamical evolution can take place over the ages usually associated with gas expulsion. To test this we perform {\it N}-body simulations of the dynamics of a very young star forming region, with initial conditions drawn from a large-scale hydrodynamic simulation of gravitational collapse and fragmentation. The subclusters we analyse, with populations of a few hundred stars, have high local star formation efficiencies and are roughly virialised even after the gas is removed. Over 10 Myr they expand to a similar degree as would be expected from gas expulsion if they were initially gas-rich, but the expansion is purely due to the internal stellar dynamics of the young clusters. The expansion is such that the stellar densities at 2 Myr match those of YSOs in the Solar neighborhood. We argue that at the low-mass end of the cluster mass spectrum, a deficit of clusters at 10s of Myr does not necessarily imply gas expulsion as a disruption mechanism.

We present an SPH parameter study of the dynamical effect of photoionization from O--type stars on star--forming clouds of a range of masses and sizes during the time window before supernovae explode. Our model clouds all have the same degree of turbulent support initially, the ratio of turbulent kinetic energy to gravitational potential energy being set to $E_{\rm kin}/|E_{\rm pot}|$=0.7. We allow the clouds to form stars and study the dynamical effects of the ionizing radiation from the massive stars or clusters born within them. We find that dense filamentary structures and accretion flows limit the quantities of gas that can be ionized, particularly in the higher density clusters. More importantly, the higher escape velocities in our more massive (10$^{6}$M$_{\odot}$) clouds prevent the HII regions from sweeping up and expelling significant quantities of gas, so that the most massive clouds are largely dynamically unaffected by ionizing feedback. However, feedback has a profound effect on the lower--density 10$^{4}$ and 10$^{5}$M$_{\odot}$ clouds in our study, creating vast evacuated bubbles and expelling tens of percent of the neutral gas in the 3Myr timescale before the first supernovae are expected to detonate, resulting in clouds highly porous to both photons and supernova ejecta.

We present the first numerical simulations that self-consistently follow the formation of dense molecular clouds in colliding flows. Our calculations include a time-dependent model for the H2 and CO chemistry that runs alongside a detailed treatment of the dominant heating and cooling processes in the ISM. We adopt initial conditions characteristic of the warm neutral medium and study two different flow velocities - a slow flow with v = 6.8 km/s and a fast flow with v = 13.6 km/s. The clouds formed by the collision of these flows form stars, with star formation beginning after 16 Myr in the case of the slower flow, but after only 4.4 Myr in the case of the faster flow. In both flows, the formation of CO-dominated regions occurs only around 2 Myr before the onset of star formation. Prior to this, the clouds produce very little emission in the J = 1 -> 0 transition line of CO, and would probably not be identified as molecular clouds in observational surveys. In contrast, our models show that H2-dominated regions can form much earlier, with the timing depending on the details of the flow. In the case of the slow flow, small pockets of gas become fully molecular around 10 Myr before star formation begins, while in the fast flow, the first H2-dominated regions occur around 3 Myr before the first prestellar cores form. Our results are consistent with models of molecular cloud formation in which the clouds are dominated by "dark" molecular gas for a considerable proportion of their assembly history.

We report on Smoothed Particle Hydrodynamics (SPH) simulations of the impact on a turbulent $\sim2\times10^{3}$ M$_{\odot}$ star--forming molecular cloud of irradiation by an external source of ionizing photons. We find that the ionizing radiation has a significant effect on the gas morphology, but a less important role in triggering stars. The rate and morphology of star formation are largely governed by the structure in the gas generated by the turbulent velocity field, and feedback has no discernible effect on the stellar initial mass function. Although many young stars are to be found in dense gas located near an ionization front, most of these objects also form when feedback is absent. Ionization has a stronger effect in diffuse regions of the cloud by sweeping up low--density gas that would not otherwise form stars into gravitationally--unstable clumps. However, even in these regions, dynamical interactions between the stars rapidly erase the correlations between their positions and velocities and that of the ionization front.

The fraction of star formation that results in bound star clusters is influenced by the density spectrum in which stars are formed and by the response of the stellar structure to gas expulsion. We analyse hydrodynamical simulations of turbulent fragmentation in star-forming regions to assess the dynamical properties of the resulting population of stars and (sub)clusters. Stellar subclusters are identified using a minimum spanning tree algorithm. When considering only the gravitational potential of the stars and ignoring the gas, we find that the identified subclusters are close to virial equilibrium (the typical virial ratio Q_vir~0.59, where virial equilibrium would be Q_vir~0.5). This virial state is a consequence of the low gas fractions within the subclusters, caused by the accretion of gas onto the stars and the accretion-induced shrinkage of the subclusters. Because the subclusters are gas-poor, up to a length scale of 0.1-0.2 pc at the end of the simulation, they are only weakly affected by gas expulsion. The fraction of subclusters that reaches the high density required to evolve to a gas-poor state increases with the density of the star-forming region. We extend this argument to star cluster scales, and suggest that the absence of gas indicates that the early disruption of star clusters due to gas expulsion (infant mortality) plays a smaller role than anticipated, and is potentially restricted to star-forming regions with low ambient gas densities. We propose that in dense star-forming regions, the tidal shocking of young star clusters by the surrounding gas clouds could be responsible for the early disruption. This `cruel cradle effect' would work in addition to disruption by gas expulsion. We suggest possible methods to quantify the relative contributions of both mechanisms.

We use Smoothed Particle Hydrodynamics to simulate the formation of a massive (10^6Msun) stellar cluster system formed from the gravitational collapse of a turbulent molecular cloud. We investigate the hierarchical clustering properties of our model system and we study the influence of the photoionizing radiation produced by the system's multiple O-type stars on the evolution of the protocluster. We find that dense gas near the ionizing sources prevents the radiation from eroding the filaments in which most of the star formation occurs and that instead, ionized gas fills pre-existing voids and bubbles originally created by the turbulent velocity field.

We present the first detailed classification of the structures of Class 0 cores in a high resolution simulation of a giant molecular cloud. The simulated cloud contains 10^4 solar masses and produces over 350 cores which allows for meaningful statistics. Cores are classified into three types according to how much they depart from spherical symmetry. We find that three quarters of the cores are better described as irregular filaments than as spheres. Recent Herschel results have shown that cores are formed within a network of filaments, which we find has had a significant impact on the resulting core geometries. We show that the column densities and ram pressure seen by the protostar are not uniform and generally peak along the axes of the filament. The angular momentum vector of the material in the cores varies both in magnitude and direction, which will cause the rotation vector of the central source to fluctuate during the collapse of the core. In the case of the more massive stars, accretion from the environment outside the original core volume is even more important than that from the core itself. This additional gas is primarily accreted onto the cores along the dense filaments in which the cores are embedded, and the sections of the surfaces of the cores which do not coincide with a filament have very little additional material passing through them. The assumption of spherical symmetry cannot be applied to the majority of collapsing cores, and is never a good description of how stars accrete gas from outside the original core radius. This has ramifications for our understanding of collapsing cores, in particular their line profiles, the effect of radiation upon them and their ability to fragment.

Super Star Clusters (Mecl > 10^5 Msol) are the largest stellar nurseries in our local Universe, containing hundreds of thousands to millions of young stars within a few light years. Many of these systems are found in external galaxies, especially in pairs of interacting galaxies, and in some dwarf galaxies, but relatively few in disk galaxies like our own Milky Way. We show that a possible explanation for this difference is the presence of shear in normal spiral galaxies which impedes the formation of the very large and dense super star clusters but prefers the formation of loose OB associations possibly with a less massive cluster at the center. In contrast, in interacting galaxies and in dwarf galaxies, regions can collapse without having a large-scale sense of rotation. This lack of rotational support allows the giant clouds of gas and stars to concentrate into a single, dense and gravitationally bound system.

We investigate the formation of both clustered and distributed populations of young stars in a single molecular cloud. We present a numerical simulation of a 10,000 solar mass elongated, turbulent, molecular cloud and the formation of over 2500 stars. The stars form both in stellar clusters and in a distributed mode which is determined by the local gravitational binding of the cloud. A density gradient along the major axis of the cloud produces bound regions that form stellar clusters and unbound regions that form a more distributed population. The initial mass function also depends on the local gravitational binding of the cloud with bound regions forming full IMFs whereas in the unbound, distributed regions the stellar masses cluster around the local Jeans mass and lack both the high-mass and the low-mass stars. The overall efficiency of star formation is ~ 15 % in the cloud when the calculation is terminated, but varies from less than 1 % in the the regions of distributed star formation to ~ 40 % in regions containing large stellar clusters. Considering that large scale surveys are likely to catch clouds at all evolutionary stages, estimates of the (time-averaged) star formation efficiency for the giant molecular cloud reported here is only ~ 4 %. This would lead to the erroneous conclusion of 'slow' star formation when in fact it is occurring on a dynamical timescale.

In this contribution we review our recent numerical work discussing the essential role of the local cluster environment in assembling massive stars. First we show that massive stars are formed from low mass pre-stellar cores and become massive due to accretion. Proto-stars that benefit from this accretion are those situated at the centre of a cluster's potential well, which is the focal point of the contraction of the cluster gas. Given that most of the mass which makes up a massive star in this model comes from the cluster environment rather than the core, it is important to model the molecular cloud environment accurately. Preliminary results of a simulation which accurately treats the chemistry and time-dependent thermodynamics of a molecular cloud show quantitatively similar star formation to previous models, but allow a true comparison to be made between simulation and observations. This method can also be applied to cases with varying metallicities allowing star formation in primordial gas to be studied. In general, these numerical studies of clustered star formation yield IMFs which are compatible with the Salpeter mass function. The only possible exception to this is in low density unbound regions of molecular clouds which lack very low and high mass stars.

With the use of N-body calculations the amount and properties of escaping stars from low-N (N = 100 and 1000) young embedded star clusters prior to gas expulsion are studied over the first 5 Myr of their existence. Besides the number of stars also different initial radii and binary populations are examined as well as virialised and collapsing clusters. It is found that these clusters can loose substantial amounts (up to 20%) of stars within 5 Myr with considerable velocities up to more than 100 km/s. Even with their mean velocities between 2 and 8 km/s these stars will still be travelling between 2 and 30 pc during the 5 Myr. Therefore can large amounts of distributed stars in star-forming regions not necessarily be counted as evidence for the isolated formation of stars.

We undertake a systematic analysis of the early (< 0.5 Myr) evolution of clustering and the stellar initial mass function in turbulent fragmentation simulations. These large scale simulations for the first time offer the opportunity for a statistical analysis of IMF variations and correlations between stellar properties and cluster richness. The typical evolutionary scenario involves star formation in small-n clusters which then progressively merge; the first stars to form are seeds of massive stars and achieve a headstart in mass acquisition. These massive seeds end up in the cores of clusters and a large fraction of new stars of lower mass is formed in the outer parts of the clusters. The resulting clusters are therefore mass segregated at an age of 0.5 Myr, although the signature of mass segregation is weakened during mergers. We find that the resulting IMF has a smaller exponent (alpha=1.8-2.2) than the Salpeter value (alpha=2.35). The IMFs in subclusters are truncated at masses only somewhat larger than the most massive stars (which depends on the richness of the cluster) and an universal upper mass limit of 150 Msun is ruled out. We also find that the simulations show signs of the IGIMF effect proposed by Weidner & Kroupa, where the frequency of massive stars is suppressed in the integrated IMF compared to the IMF in individual clusters. We identify clusters through the use of a minimum spanning tree algorithm which allows easy comparison between observational survey data and the predictions of turbulent fragmentation models. In particular we present quantitative predictions regarding properties such as cluster morphology, degree of mass segregation, upper slope of the IMF and the relation between cluster richness and maximum stellar mass. [abridged]

We present a thorough literature study of the most-massive star, m_max, in several young star clusters in order to assess whether or not star clusters are populated from the stellar initial mass function (IMF) by random sampling over the mass range 0.01 < m < 150 M_sol without being constrained by the cluster mass, M_ecl. The data reveal a partition of the sample into lowest mass objects (M_ecl < 10^2 M_sol), moderate mass clusters (10^2 M_sol < M_ecl < 10^3 M_sol) and rich clusters above 10^3 M_sol. Additionally, there is a plateau of a constant maximal star mass (m_max ~ 25 M_sol) for clusters with masses between 10^3 M_sol and 4 10^3 M_sol. Statistical tests of this data set reveal that the hypothesis of random sampling from the IMF between 0.01 and 150 M_sol is highly unlikely for star clusters more massive than 10^2 M_sol with a probability of p ~ 2 10^-7 for the objects with M_ecl between 10^2 M_sol and 10^3 M_sol and p ~ 3 10^-9 for the more massive star clusters. Also, the spread of m_max values at a given M_ecl is smaller than expected from random sampling. We suggest that the basic physical process able to explain this dependence of stellar inventory of a star cluster on its mass may be the interplay between stellar feedback and the binding energy of the cluster-forming molecular cloud core. Given these results, it would follow that an integrated galactic initial mass function (IGIMF) sampled from such clusters would automatically be steeper in comparison to the IMF within individual star clusters.

We show that massive stars and stellar clusters are formed simultaneously, the global evolution of the forming cluster is what allows the central stars to become massive. We predict that massive star forming clumps, such as those observed in Motte et al. 2007, contract and grow in mass leading to the formation of massive stars. This occurs as mass is continually channeled from large radii onto the central proto-stars, which can become massive through accretion. Using SPH simulations of massive star forming clumps in a Giant Molecular Cloud, we show that clumps are initially diffuse and filamentary, and become more concentrated as they collapse. Simulated interferometry observations of our data provide an explanation as to why young massive star forming regions show more substructure than older ones. The most massive stars in our model are found within the most bound cluster. Most of the mass accreted by the massive stars was originally distributed throughout the clump at low densities, and was later funneled to the star due to global in-fall. Even with radiative feedback no massive pre-stellar cores are formed. The original cores are of intermediate mass and gain their additional mass in the proto-stellar stage. We also find that cores which form low mass stars exist within the volume from which the high mass stars accrete, but are largely unaffected by this process.

The nearest site of massive star formation in Orion is dominated by the Trapezium subsystem, with its four OB stars and numerous companions. The question of how these stars came to be in such close proximity has implications for our understanding of massive star formation and early cluster evolution. A promising route toward rapid mass segregation was proposed by McMillan et al. (2007), who showed that the merger product of faster-evolving sub clusters can inherit their apparent dynamical age from their progenitors. In this paper we briefly consider this process at a size and time scale more suited for local and perhaps more typical star formation, with stellar numbers from the hundreds to thousands. We find that for reasonable ages and cluster sizes, the merger of sub-clusters can indeed lead to compact configurations of the most massive stars, a signal seen both in Nature and in large-scale hydrodynamic simulations of star formation from collapsing molecular clouds, and that sub-virial initial conditions can make an un-merged cluster display a similar type of mass segregation. Additionally, we discuss a variation of the minimum spanning tree mass-segregation technique introduced by Allison et al. (2009).

The observed slope at the high-mass end of the initial mass function (IMF) displays a remarkable universality in a wide variety of physical environments. We predict that competitive accretion, the ongoing accretion of gas from a common reservoir by a collection of protostellar cores, can provide a natural explanation for such a universal slope in star forming regions with metallicities roughly greater than 1e-5 the solar value. In our discussion, we point out that competitive accretion will occur whenever a gaseous region has multiple Jeans masses of material and contains large-scale motions that are controlled by the gravitational potential well. We describe how and when these conditions can be reached during the chemical enrichment of the Universe, showing that they can occur for a wide range of metallicities and environmental conditions. We also discuss the ability of other physical processes to limit the effects of further accretion onto protostellar cores. Current theoretical and numerical studies show that competitive accretion is robust against disrupting effects - such as feedback from young stars, supersonic turbulence and magnetic fields - in all but the most extreme cases.

Mass segregation is observed in many star clusters, including several that are less than a few Myr old. Timescale arguments are frequently used to argue that these clusters must be displaying primordial segregation, because they are too young to be dynamically relaxed. Looking at this argument from the other side, the youth of these clusters and the limited time available to mix spatially distinct populations of stars can provide constraints on the amount of initial segregation that is consistent with current observations. We present n-body experiments testing this idea, and discuss the implications of our results for theories of star formation. For system ages less than a few crossing times, we show that star formation scenarios predicting general primordial mass segregation are inconsistent with observed segregation levels.

We present an analysis of star-forming gas cores in an SPH simulation of a Giant Molecular Cloud. We identify cores using their deep potential wells. This yields a smoother distribution with clearer boundaries than density. Additionally, this gives an indication of future collapse, as bound potential cores (p-cores) represent the earliest stages of fragmentation in molecular clouds. We find that the mass function of the p-cores resembles the stellar IMF and the observed clump mass function, although p-core masses (~0.7 Msol) are smaller than typical density clumps. The bound p-cores are generally subsonic, have internal substructure, and are only quasi-spherical. We see no evidence of massive bound cores supported by turbulence. We trace the evolution of the p-cores forward in time, and investigate the connection between the original p-core mass and the stellar mass that formed from it. We find that there is a poor correlation, with considerable scatter suggesting accretion onto the core is dependent on more factors than just the initial core mass. During the accretion process the p-cores accrete from beyond the region first bound, highlighting the importance of the core environment to its subsequent evolution.

The presence of young massive stars orbiting on eccentric rings within a few tenths of a parsec of the supermassive black hole in the Galactic centre is challenging for theories of star formation. The high tidal shear from the black hole should tear apart the molecular clouds that form stars elsewhere in the Galaxy, while transporting the stars to the Galactic centre also appears unlikely during their stellar lifetimes. We present numerical simulations of the infall of a giant molecular cloud that interacts with the black hole. The transfer of energy during closest approach allows part of the cloud to become bound to the black hole, forming an eccentric disc that quickly fragments to form stars. Compressional heating due to the black hole raises the temperature of the gas to 100-1000K, ensuring that the fragmentation produces relatively high stellar masses. These stars retain the eccentricity of the disc and, for a sufficiently massive initial cloud, produce an extremely top-heavy distribution of stellar masses. This potentially repetitive process can therefore explain the presence of multiple eccentric rings of young stars in the presence of a supermassive black hole.

Using an SPH simulation of a star-forming region in a molecular cloud, we show that the emergence of a clump mass function (CMF) resembling the stellar initial mass function (IMF) is a ubiquitous feature of molecular cloud structure, but caution against its over-interpretation. We employ three different techniques to extract the clumps in this study. In the first two, we interpolate the SPH particle data to 2 and 3 dimensional grids before performing the clump-find, using position-position (PP) and position-position-velocity (PPV) information respectively. In the last technique, the clump-finding is performed on the SPH data directly, making use of the full 3 dimensional position information. Although the CMF is typically similar to that observed in regions of nearby star formation, the individual clumps and their masses are found to be unreliable since they depend strongly on the parameters and the method of the clump-finding. In particular we find that the resolution and orientation of the data make a significant difference to the resulting properties of the identified clumps in the PP and PPV cases. We conclude that making comparisons between a CMF and the stellar IMF should be done with caution, since the definition of a clump boundary, and hence the number of clumps and their properties, are arbitrary in the extraction method. This is especially true if molecular clouds are truly scale free.

We present SPH simulations of protoclusters including the effects of winds from massive stars. Using a particle-injection method, we investigate the effect of structure close to the wind sources and the short-timescale influence of winds on protoclusters. Structures such as disks and gaseous filaments have a strong collimating effect. By a different technique of injecting momentum from point sources, we compare the large-scale, long-term effects of isotropic and intrinsically-collimated winds and find them to be similar. Both types of wind dramatically slow the global star formation process, but the timescale on which they expel significant mass from the cluster is rather long (approaching 10 freefall times). Clusters may then experience rapid star formation early in their lifetimes, before switching to a mode where gas is gradually expelled, while star formation proceeds much more slowly. This complicates conclusions regarding slow star formation derived from measuring the star-formation efficiency per freefall time. Estimates of the efficacy of winds in dispersing clusters derived simply from the total wind momentum flux may not be very reliable. This is due to material being expelled from deep within stellar potential wells, often to velocities well in excess of the cluster escape velocity, and also to the loss of momentum flux through holes in the gas distribution. Winds have little effect on the accretion--driven stellar IMF except at the very high-mass end, where they serve to prevent some of the most massive objects accreting. We also find that the morphology of the gas, the rapid motions of the wind sources and the action of accretion flows prevent the formation of bubble-like structures. This may make it difficult to discern the influence of winds on very young clusters.

We investigate the formation of brown dwarfs and very low-mass stars through the gravitational fragmentation of infalling gas into stellar clusters. The gravitational potential of a forming stellar cluster provides the focus that attracts gas from the surrounding molecular cloud. Structures present in the gas grow, forming filaments flowing into the cluster centre. These filaments attain high gas densities due to the combination of the cluster potential and local self-gravity. The resultant Jeans masses are low, allowing the formation of very low-mass fragments. The tidal shear and high velocity dispersion present in the cluster preclude any subsequent accretion thus resulting in the formation of brown dwarfs or very low-mass stars. Ejections are not required as the brown dwarfs enter the cluster with high relative velocities, suggesting that their disc and binary properties should be similar to that of low-mass stars. This mechanism requires the presence of a strong gravitational potential due to the stellar cluster implying that brown dwarf formation should be more frequent in stellar clusters than in distributed populations of young stars. Brown dwarfs formed in isolation would require another formation mechanism such as due to turbulent fragmentation.

We investigate how the dynamical state of a turbulently supported, 1000 solar mass, molecular cloud affects the properties of the cluster it forms, focusing our discussion on the star formation efficiency (SFE) and the initial mass function (IMF). A variety of initial energy states are examined in this paper, ranging from clouds with PE = 0.1 KE to clouds with PE = 10 KE, and for both isothermal and piece-wise polytropic equations of state (similar to that suggested by Larson). It is found that arbitrary star formation efficiencies are possible, with strongly unbound clouds yielding very low star formation efficiencies. We suggest that the low star formation efficiency in the Maddelena cloud may be a consequence of the relatively unbound state of its internal structure. It is also found that competitive accretion results in the observed IMF when the clouds have initial energy states of PE >= KE. We show that under such conditions the shape of the IMF is independent of time in the calculations. This demonstrates that the global accretion process can be terminated at any stage in the cluster's evolution, while still yielding a distribution of stellar masses that is consistent with the observed IMF. As the clouds become progressively more unbound, competitive accretion is less important and the protostellar mass function flattens. These results predict that molecular clouds should be permeated with a distributed population of stars that follow a flatter than Salpeter IMF.

We describe simulations of the response of a gaseous disc to an active spiral potential. The potential is derived from an N-body calculation and leads to a multi-armed time-evolving pattern. The gas forms long spiral arms typical of grand design galaxies, although the spiral pattern is asymmetric. The primary difference from a grand-design spiral galaxy, which has a consistent 2/4-armed pattern, is that instead of passing through the spiral arms, gas generally falls into a developing potential minimum and is released only when the local minimum dissolves. In this case, the densest gas is coincident with the spiral potential, rather than offset as in the grand-design spirals. We would there fore expect no offset between the spiral shock and star formation, and no obvious co-rotation radius. Spurs which occur in grand-design spirals when large clumps are sheared off leaving the spiral arms, are rare in the active, time-evolving spiral reported here. Instead, large branches are formed from spiral arms when the underlying spiral potential is dissolving due to the N-body dynamics. We find that the molecular cloud mass spectrum for the active potential is similar to that for clouds in grand design calculations, depending primarily on the ambient pressure rather than the nature of the potential. The largest molecular clouds occur when spiral arms collide, rather than by agglomeration within a spiral arm.

The formation of massive stars is currently an unsolved problems in astrophysics. Understanding the formation of massive stars is essential because they dominate the luminous, kinematic, and chemical output of stars. Furthermore, their feedback is likely to play a dominant role in the evolution of molecular clouds and any subsequent star formation therein. Although significant progress has been made observationally and theoretically, we still do not have a consensus as to how massive stars form. There are two contending models to explain the formation of massive stars, Core Accretion and Competitive Accretion. They differ primarily in how and when the mass that ultimately makes up the massive star is gathered. In the core accretion model, the mass is gathered in a prestellar stage due to the overlying pressure of a stellar cluster or a massive pre-cluster cloud clump. In contrast, competitive accretion envisions that the mass is gathered during the star formation process itself, being funneled to the centre of a stellar cluster by the gravitational potential of the stellar cluster. Although these differences may not appear overly significant, they involve significant differences in terms of the physical processes involved. Furthermore, the differences also have important implications in terms of the evolutionary phases of massive star formation, and ultimately that of stellar clusters and star formation on larger scales. Here we review the dominant models, and discuss prospects for developing a better understanding of massive star formation in the future.

This paper describes a James Clerk Maxwell Telescope (JCMT) legacy survey that has been awarded roughly 500 hrs of observing time to be carried out from 2007 to 2009. In this survey we will map with SCUBA-2 (Submillimetre Common User Bolometer Array 2) almost all of the well-known low-mass and intermediate-mass star-forming regions within 0.5 kpc that are accessible from the JCMT. Most of these locations are associated with the Gould Belt. From these observations we will produce a flux-limited snapshot of star formation near the Sun, providing a legacy of images, as well as point-source and extended-source catalogues, over almost 700 square degrees of sky. The resulting images will yield the first catalogue of prestellar and protostellar sources selected by submillimetre continuum emission, and should increase the number of known sources by more than an order of magnitude. We will also obtain CO maps with the array receiver HARP (Heterodyne Array Receiver Programme), in three CO isotopologues, of a large typical sample of prestellar and protostellar sources. We will then map the brightest hundred sources with the SCUBA-2 polarimeter (POL-2), producing the first statistically significant set of polarization maps in the submillimetre. The images and source catalogues will be a powerful reference set for astronomers, providing a detailed legacy archive for future telescopes, including ALMA, Herschel and JWST.

We present ongoing hydrodynamic and MHD simulations of molecular cloud formation in spiral galaxies. The hydrodynamic results show the formation of molecular gas clouds where spiral shocks compress atomic gas to high densities. The spiral shocks also produce structure in the spiral arms, provided the gas is cold (< 1000 K). When both hot and cold components of the ISM are modeled, this structure is enhanced. Properties such as the clump mass spectra and spatial distribution will be compared from clouds identified in these simulations. In particular the multiphase simulations predict the presence of much more interarm molecular gas than when a single phase is assumed. We also discuss very recent results from galactic-scale MHD calculations. From observational comparisons of the magnetic and thermal pressure, magnetic fields are expected to be a major factor in explaining the dynamics of the ISM, from kpc scales to those of star formation. We describe the difference in structure of the spiral arms, and the evolution of the global magnetic field for a range of field strengths.

Recent studies of dense clumps/cores in a number of regions of low-mass star formation have shown that the mass distribution of these clumps closely resembles the initial mass function (IMF) of field stars. One possible interpretation of these observations is that we are witnessing the fragmentation of the clouds into the IMF, and the observed clumps are bound pre-stellar cores. In this paper, we highlight a potential difficulty in this interpretation, namely that clumps of varying mass are likely to have systematically varying lifetimes. This timescale problem can effectively destroy the similarity bewteen the clump and stellar mass functions, such that a stellar-like clump mass function (CMF) results in a much steeper stellar IMF. We also discuss some ways in which this problem may be avoided.

We extend recent numerical results (Dobbs et. al. 2006) on molecular cloud formation in spiral galaxies by including a multi-phase medium. The addition of a hot phase of gas enhances the structure in the cold gas, and significantly increases the fraction of molecular hydrogen that is formed when the cold gas passes through a spiral shock. The difference in structure is reflected in the mass power spectrum of the molecular clouds, which is steeper for the multi-phase calculations. The increase in molecular gas occurs as the addition of a hot phase leads to higher densities in the cold gas. In particular, cold gas is confined in clumps between the spiral arms and retains a higher molecular fraction. Unlike the single phase results, molecular clouds are present in the inter-arm regions for the multi-phase medium. However the density of the inter-arm molecular hydrogen is generally below that which can be reliably determined from CO measurements. We therefore predict that for a multi-phase medium, there will be low density clouds containing cold atomic and molecular hydrogen, which are potentially entering the spiral arms.

We conduct Smoothed Particle Hydrodynamics simulations of the `collect and collapse' scenario (Elmegreen and Lada, 1977) for star formation triggered by an expanding HII region. We simulate the evolution of a spherical uniform molecular cloud with an ionising source at its centre. The gas in the cloud is self-gravitating, although the cloud is prevented from globally collapsing. We find that the shell driven by the HII region fragments to form numerous self--gravitating objects. We repeat our calculations at four numerical resolutions to ensure that they are converged. We compare our results to the analytical model of Whitworth et al, 1994 and show that our simulations and the predictions of Whitworth et al are in good agreement in the sense that the shell fragments at the time and radius predicted by Whitworth et al to within 20% and 25% respectively. Most of the fragments produced in our two highest resolution calculations are approximately half the mass of those predicted, but this conclusion is robust against both numerical resolution and the presence of random noise (local fluctuations in density of a factor of $\sim2$) in the initial gas distribution. We conclude that such noise has little impact on the fragmentation process.

We present numerical simulations of the passage of clumpy gas through a galactic spiral shock, the subsequent formation of giant molecular clouds (GMCs) and the triggering of star formation. The spiral shock forms dense clouds while dissipating kinetic energy, producing regions that are locally gravitationally bound and collapse to form stars. In addition to triggering the star formation process, the clumpy gas passing through the shock naturally generates the observed velocity dispersion size relation of molecular clouds. In this scenario, the internal motions of GMCs need not be turbulent in nature. The coupling of the clouds' internal kinematics to their externally triggered formation removes the need for the clouds to be self-gravitating. Globally unbound molecular clouds provides a simple explanation of the low efficiency of star formation. While dense regions in the shock become bound and collapse to form stars, the majority of the gas disperses as it leaves the spiral arm.

We present an alternative explanation for the nature of turbulence in molecular clouds. Often associated with classical models of turbulence, we instead interpret the observed gas dynamics as random motions, induced when clumpy gas is subject to a shock. From simulations of shocks, we show that a supersonic velocity dispersion occurs in the shocked gas provided the initial distribution of gas is sufficiently non-uniform. We investigate the velocity size-scale relation $\sigma \propto r^{\alpha}$ for simulations of clumpy and fractal gas, and show that clumpy shocks can produce realistic velocity size-scale relations with mean $\alpha \thicksim 0.5$. For a fractal distribution, with a fractal dimension of 2.2 similar to what is observed in the ISM, we find $\sigma \propto r^{0.4}$. The form of the velocity size-scale relation can be understood as due to mass loading, i.e. the post-shock velocity of the gas is determined by the amount of mass encountered as the gas enters the shock. We support this hypothesis with analytical calculations of the velocity dispersion relation for different initial distributions. A prediction of this model is that the line-of sight velocity dispersion should depend on the angle at which the shocked gas is viewed.

We present Smoothed Particle Hydrodynamics (SPH) simulations of molecular cloud formation in spiral galaxies. These simulations model the response of a non-self-gravitating gaseous disk to a galactic potential. The spiral shock induces high densities in the gas, and considerable structure in the spiral arms, which we identify as molecular clouds. We regard the formation of these structures as due to the dynamics of clumpy shocks, which perturb the flow of gas through the spiral arms. In addition, the spiral shocks induce a large velocity dispersion in the spiral arms, comparable with the magnitude of the velocity dispersion observed in molecular clouds. We estimate the formation of molecular hydrogen, by post-processing our results and assuming the gas is isothermal. Provided the gas is cold ($T\le100$ K), the gas is compressed sufficiently in the spiral shock for molecular hydrogen formation to occur in the dense spiral arm clumps. These molecular clouds are largely confined to the spiral arms, since most molecular gas is photodissociated to atomic hydrogen upon leaving the arms.

An unsolved issue in the standard core accretion model for gaseous planet formation is how kilometre-sized planetesimals form from, initially, micron-sized dust grains. Solid growth beyond metre sizes can be difficult both because the sticking efficiency becomes very small, and because these particles should rapidly migrate into the central star. We consider here how metre-sized particles evolve in self-gravitating accretion discs using simulations in which the gravitational influence of the solid particles is also included. Metre-sized particles become strongly concentrated in the spiral structures present in the disc and, if the solid to gas density ratio is sufficiently high, can fragment due to their own self-gravity to form planetesimals directly. This result suggests that planetesimal formation may occur very early in the star formation process while discs are still massive enough to be self-gravitating. The dependence of this process on the surface density of the solids is also consistent with the observation that extrasolar planets are preferentially found around high metallicity stars.

Competitive accretion, a process to explain the origin of the IMF, occurs when stars in a common gravitational potential accrete from a distributed gaseous component. We show that concerns recently raised on the efficiency of competitive accretion are incorrect as they use globally averaged properties which are inappropriate for the detailed physics of a forming stellar cluster. A full treatment requires a realistic treatment of the cluster potential, the distribution of turbulent velocities and gas densities. Accreting gas does not travel at the global virial velocity of the system due to the velocity-sizescale relation inherent in turbulent gas and due to the lower velocity dispersion of small-N clusters in which much of the accretion occurs. Stars located in the gas-rich centres of such systems initially accrete from low relative velocity gas attaining larger masses before needing to accrete the higher velocity gas. Stars not in the centres of such potentials, or that enter the cluster later when the velocity dispersion is higher, do not accrete significantly and thus retain their low-masses. In competitive accretion, most stars do not continue to accrete significantly such that their masses are set from the fragmentation process. It is the few stars which continue to accrete that become higher-mass stars. Competitive accretion is therefore likely to be responsible for the formation of higher-mass stars and can explain the mass distribution, mass segregation and binary frequency of these stars. Global kinematics of competitive accretion models include large-scale mass infall, with mean inflow velocities of order 0.5 km/s at scales of 0.5 pc, but infall signatures are likely to be confused by the large tangential velocities and the velocity dispersion present.

In this paper, we examine whether clumpy, colliding, flows could be responsible for the clump mass functions that have been observed in several regions of embedded star formation, which have been shown to be described by a Salpeter type slope. The flows presented here, which comprise a population of initially identical clumps and the calculations are performed with and without the inclusion of self-gravity. When the shock region is at its densest, we find that the clump mass spectrum is always well modelled by a Salpeter type slope. This is true regardless of whether the self-gravity is included in the simulations or not. In the non-self-gravitating simulations, this slope is retained at lower Mach numbers (Mach 5 and 10) as the simulations progress past the densest phase. In the simulations which include self-gravity, we find that low Mach number runs yield a flatter mass function after the densest phase. This is simply a result of increased coagulation due to gravitational collapse of the flows. In the high Mach number runs (Mach 20) the Salpeter slope is always lost. The self gravitating calculations also show that the sub-group of gravitationally bound clumps in which star formation occurs, always contain the most massive clumps in the population. The mass function of these bound star forming clumps is not at all similar to the Salpeter type mass function observed for stars in the field. We conclude that the clump mass function may not only have nothing to do with gravity, but also nothing to do with the star formation process and the resulting mass distribution of stars. This raises doubt over the claims that the clump mass function is the origin of the stellar IMF, for regions such as rho Oph, Serpens and the Orion B cloud.

We review recent advances in our understanding of the origin of the initial mass function (IMF). We emphasize the use of numerical simulations to investigate how each physical process involved in star formation affects the resulting IMF. We stress that it is insufficient to just reproduce the IMF, but that any successful model needs to account for the many observed properties of star forming regions including clustering, mass segregation and binarity. Fragmentation involving the interplay of gravity, turbulence, and thermal effects is probably responsible for setting the characteristic stellar mass. Low-mass stars and brown dwarfs can form through the fragmentation of dense filaments and disks, possibly followed by early ejection from these dense environments which truncates their growth in mass. Higher-mass stars and the Salpeter-like slope of the IMF are most likely formed through continued accretion in a clustered environment. The effects of feedback and magnetic fields on the origin of the IMF are still largely unclear. Lastly, we discuss a number of outstanding problems that need to be addressed in order to develop a complete theory for the origin of the IMF.

We use numerical simulations of the fragmentation of a 1000 solar mass molecular cloud and the formation of a stellar cluster to study how the initial conditions for star formation affect the resulting initial mass function (IMF). In particular, we are interested in the relation between the thermal Jeans mass in a cloud and the knee of the initial mass function, i.e. the mass separating the region with a flat IMF slope from that typified by a steeper, Salpeter-like, slope. In three isothermal simulations with M_J=1 solar mass, M_J=2 solar masses and M_J=5 solar masses, the number of stars formed, at comparable dynamical times, scales roughly with the number of initial Jeans masses in the cloud. The mean stellar mass also increases (though less than linearly) with the initial Jeans mass in the cloud. It is found that the IMF in each case displays a prominent knee, located roughly at the mass scale of the initial Jeans mass. Thus clouds with higher initial Jeans masses produce IMFs which are shallow to higher masses. This implies that a universal IMF requires a physical mechanism that sets the Jeans mass to be near 1 solar mass. Simulations including a barotropic equation of state as suggested by Larson, with cooling at low densities followed by gentle heating at higher densities, are able to produce realistic IMFs with the knee located at approximately 1 solar mass, even with an initial M_J=5 solar masses. We therefore suggest that the observed universality of the IMF in the local Universe does not require any fine tuning of the initial conditions in star forming clouds but is instead imprinted by details of the cooling physics of the collapsing gas.

We present Smoothed Particle Hydrodynamics (SPH) simulations of the response of gas discs to a spiral potential. These simulations show that the commonly observed spurs and feathering in spiral galaxies can be understood as being due to structures present in the spiral arms that are sheared by the divergent orbits in a spiral potential. Thus, dense molecular cloud-like structures generate the perpendicular spurs as they leave the spiral arms. Subsequent feathering occurs as spurs are further sheared into weaker parallel structures as they approach the next spiral passage. Self-gravity of the gas is not included in these simulations, stressing that these features are purely due to the hydrodynamics in spiral shocks. Instead, a necessary condition for this mechanism to work is that the gas need be relatively cold (1000 K or less) in order that the shock is sufficient to generate structure in the spiral arms, and such structure is not subsequently smoothed by the gas pressure.

We present numerical simulations of the passage of clumpy gas through a galactic spiral shock and the subsequent formation of giant molecular clouds (GMCs). The spiral shock forms dense clouds while dissipating kinetic energy, producing regions that are locally gravitationally bound and collapse to form stars. The effect of the clumpiness of gas as it passes through the shock is to generate chaotic internal motions in the gas. The kinematics of these motions are found to agree with the observed velocity-dispersion/size relation found in star-forming regions. In contrast to the standard picture where continuously driven turbulence generates the density inhomogeneities in star-forming clouds, we find here that it is the clumpiness of the interstellar gas that produces the chaotic motions as it passes through the spiral shock and initiates the star formation process. The velocity dispersion can be understood as being due to the random mass loading of clumps as they converge in the spiral shock. In this model there is no need for any internal or external continuous driving mechanism for the 'turbulence'. The coupling of the clouds' internal kinematics to their externally triggered formation removes the need for the clouds to be self-gravitating. Indeed, while clearly some parts of the clouds are self-gravitating and able to form stars, most of the molecular material remains gravitationally unbound. This can provide a simple explanation for the low efficiency of star formation.

We present a model for the formation of high-mass close binary systems in the context of forming massive stars through gas accretion in the centres of stellar clusters. A low-mass wide binary evolves under mass accretion towards a high-mass close binary, attaining system masses of order 30-50 solar masses at separations of order 1 AU. The resulting high frequency of binary systems with two massive components is in agreement with observations. These systems are typically highly eccentric and may evolve to have periastron separations less than their stellar radii. Mergers of these binary systems are therefore likely and can lead to the formation of the most massive stars, circumventing the problem of radiation pressure stopping the accretion. The stellar density required to induce binary mergers is one million stars per cubic parsec, or approximately one percent that required for direct stellar collisions.

We investigate the formation of star clusters in an unbound GMC, where the supporting kinetic energy is twice as large as the cloud's self-gravity. This cloud manages to form a series of star clusters and disperse, all within roughly 2 crossing times (10 Myr), supporting recent claims that star formation is a rapid process. Simple assumptions about the nature of the star formation occurring in the clusters allows us to place an estimate for the star formation efficiency at about 5 to 10 %, consistent with observations. We also propose that unbound clouds can act as a mechanism for forming OB associations. The clusters that form in the cloud behave as OB subgroups. These clusters are naturally expanding from one another due to unbound nature of the flows that create them. The properties of the cloud we present here are are consistent with those of classic OB associations.

We present numerical investigations into the formation of massive stars from centrally condensed turbulent cores. The results of five hydrodynamical simulations are described, following the collapse of the core, fragmentation and the formation of small clusters of protostars. We generate two different initial turbulent velocity fields corresponding to power-law spectra with power varying as k^{-4} and k^{-3.5}, and apply two different initial core radii. Calculations are included for both completely isothermal collapse, and a non-isothermal equation of state above a critical density 10^{-14}gcm^{-3}. Our calculations reveal the preference of fragmentation over monolithic star formation in turbulent cores. Fragmentation was prevalent in all the isothermal cases. Although disc fragmentation was largely suppressed in the non-isothermal runs due to the small dynamic range between the initial density and the critical density, our results show that some fragmentation still persisted. This is inconsistent with previous suggestions that turbulent cores result in the formation of a single massive star. We conclude that turbulence cannot be measured as an isotropic pressure term.

The formation of massive stars in close binary systems is complicated due to their high radiation pressure, the crowded environment and the expected minimum separation for fragmentation being many times greater than the orbital separation. I discuss how massive star formation can be understood as being due to competitive accretion in stellar clusters. Massive binary systems are then formed due to accretion onto wider low-mass systems. The combination of accretion and dynamical interactions with other stars in the cluster then result in producing a close binary system. Tidal and 3-body captures in dense stellar cores can also play a role in forming massive binary systems while stellar mergers of such close binary systems due to interactions may play an important role in overcoming the radiation pressure.

Competitive accretion occurs when stars in a cluster accrete from a shared reservoir of gas. The competition arises due to the relative attraction of stars as a function of their mass and location in the cluster. The low relative motions of the stars and gas in young, gas dominated clusters results in a tidal limit to the accretion whereas in the stellar dominated cluster cores, the high relative velocities results in Bondi-Hoyle accretion. The combination of these two accretion processes produces a two power-law IMF with $\gamma \approx -1.5$, for low-mass stars which accrue their mass in the gas dominated regime, and a steeper, $\gamma\approx -2.5$, IMF for higher-mass stars that form in the core of a cluster. Simulations of the fragmentation and formation of a stellar cluster show that the final stellar masses, and IMF, are due to competitive accretion. Competitive accretion also naturally results in a mass segregated cluster and in a direct correlation between the richness of a cluster and the mass of the most massive star therein. The {\sl knee} where the IMF slope changes occurs near the Jeans mass of the system.

We present the first ever hydrodynamic calculations of star cluster formation that incorporate the effect of feedback from ionising radiation. In our simulations, the ionising source forms in the cluster core at the intersection of several dense filaments of inflowing gas. We show that these filaments collimate ionised outflows and suggest such an environmental origin for at least some observed outflows in regions of massive star formation. Our simulations show both positive feedback (i.e. promotion of star formation in neutral gas compressed by expanding HII regions) and negative feedback (i.e. suppression of the accretion flow in to the central regions). We show that the volume filling factor of ionised gas is very different in our simulations than would result from the case where the central source interacted with an azimuthally smoothed gas density distribution. As expected, gas density is the key parameter in determining whether clusters are unbound by photoionising radiation. Nevertheless, we find - on account of the acceleration of a small fraction of the gas to high velocities in the outflows - that the deposition in the gas of an energy that exceeds the binding energy of the cluster is not a sufficient criterion for unbinding the bulk of the cluster mass.

We investigate the dependence of stellar properties on the mean thermal Jeans mass in molecular clouds. We compare the results from the two largest hydrodynamical simulations of star formation to resolve the fragmentation process down to the opacity limit, the first of which was reported by Bate, Bonnell & Bromm. The initial conditions of the two calculations are identical except for the radii of the clouds, which are chosen so that the mean densities and mean thermal Jeans masses of the clouds differ by factors of nine and three, respectively. We find that the denser cloud, with the lower mean thermal Jeans mass, produces a higher proportion of brown dwarfs and has a lower characteristic (median) mass of the stars and brown dwarfs. This dependence of the initial mass function (IMF) on the density of the cloud may explain the observation that the Taurus star-forming region appears to be deficient in brown dwarfs when compared with the Orion Trapezium cluster. The new calculation also produces wide binaries (separations >20 AU), one of which is a wide binary brown dwarf system. Based on the hydrodynamical calculations, we develop a simple accretion/ejection model for the origin of the IMF. In the model, all stars and brown dwarfs begin with the same mass (set by the opacity limit for fragmentation) and grow in mass until their accretion is terminated stochastically by their ejection from the cloud through dynamically interactions. The model predicts that the main variation of the IMF in different star-forming environments should be in the location of the peak (due to variations in the mean thermal Jeans mass of the cloud) and in the substellar regime. However, the slope of the IMF at high-masses may depend on the dispersion in the accretion rates of protostars.

In this paper we consider the evolution of small planetesimals in marginally stable, self-gravitating protoplanetary discs. The drag force between the disc gas and the embedded planetesimals generally causes the planetesimals to drift inwards through the disc at a rate that depends on the particle size. In a marginally stable, self-gravitating disc, however, the planetesimals are significantly influenced by the non-axisymmetric spiral structures resulting from the growth of the gravitational instability. For small particles, that are strongly coupled to the disc gas, and for large particles, that have essentially decoupled from the disc gas, the effect is not particularly significant. Intermediate sized particles, which would generally have the largest radial drift rates, do, however, become significantly concentrated at the peaks of the spiral arms. These high density regions may persist for, of order, an orbital period and may attain densities comparable to that of the disc gas. Although at the end of the simulation only ~ 25 % of the planetesimal particles lie in regions of enhanced density, during the course of the simulation at least 75 % of the planetesimal particles have at some stage been in a such a region. We find that the concentration of particles in the spiral arms results in an increased collision rate, an effect that could significantly accelerate planetesimal growth. The density enhancements may also be sufficient for the growth of planetesimals through direct gravitational collapse. The interaction between small planetesimals and self-gravitating spiral structures may therefore play an important role in the formation of large planetesimals that will ultimately coagulate to form terrestrial planets or the cores of gas/ice giant planets.

We investigate the physical processes which lead to the formation of massive stars. Using a numerical simulation of the formation of a stellar cluster from a turbulent molecular cloud, we evaluate the relevant contributions of fragmentation and competitive accretion in determining the masses of the more massive stars. We find no correlation between the final mass of a massive star, and the mass of the clump from which it forms. Instead, we find that the bulk of the mass of massive stars comes from subsequent competitive accretion in a clustered environment. In fact, the majority of this mass infalls onto a pre-existing stellar cluster. Furthermore, the mass of the most massive star in a system increases as the system grows in numbers of stars and in total mass. This arises as the infalling gas is accompanied by newly formed stars, resulting in a larger cluster around a more massive star. High-mass stars gain mass as they gain companions, implying a direct causal relationship between the cluster formation process, and the formation of higher-mass stars therein.

We present the results of a numerical simulation in which star formation proceeds from an initially unbound molecular cloud core. The turbulent motions, which dominate the dynamics, dissipate in shocks leaving a quiescent region which becomes gravitationally bound and collapses to form a small multiple system. Meanwhile, the bulk of the cloud escapes due to its initial supersonic velocities. In this simulation, the process naturally results in a star formation efficiency of 50%. The mass involved in star formation depends on the gas fraction that dissipates sufficient kinetic energy in shocks. Thus, clouds with larger turbulent motions will result in lower star formation efficiencies. This implies that globally unbound, and therefore transient giant molecular clouds (GMCs), can account for the low efficiency of star formation observed in our Galaxy without recourse to magnetic fields or feedback processes. Observations of the dynamic stability in molecular regions suggest that GMCs may not be self-gravitating, supporting the ideas presented in this letter.

Self-gravitating protostellar discs are unstable to fragmentation if the gas can cool on a time scale that is short compared to the orbital period. We use a combination of hydrodynamic simulations and N-body orbit integrations to study the long term evolution of a fragmenting disc with an initial mass ratio to the star of M_disc/M_star = 0.1. For a disc which is initially unstable across a range of radii, a combination of collapse and subsequent accretion yields substellar objects with a spectrum of masses extending (for a Solar mass star) up to ~0.01 M_sun. Subsequent gravitational evolution ejects most of the lower mass objects within a few million years, leaving a small number of very massive planets or brown dwarfs in eccentric orbits at moderately small radii. Based on these results, systems such as HD 168443 -- in which the companions are close to or beyond the deuterium burning limit -- appear to be the best candidates to have formed via gravitational instability. If massive substellar companions originate from disc fragmentation, while lower-mass planetary companions originate from core accretion, the metallicity distribution of stars which host massive substellar companions at radii of ~1 au should differ from that of stars with lower mass planetary companions.

It has already been shown, using a local model, that accretion discs with cooling times t_cool <= 3 Omega^-1 fragment into gravitationally bound objects, while those with cooling times t_cool > 3 Omega^-1 evolve into a quasi-steady state. We present results of three-dimensional simulations that test if the local result still holds globally. We find that the fragmentation boundary is close to that determined using the local model, but that fragmentation may occur for longer cooling times when the disc is more massive or when the mass is distributed in such a way as to make a particular region of the disc more susceptible to the growth of the gravitational instability. These results have significant implications for the formation of of gaseous planets in protoplanetary discs and also for the redistribution of angular momentum which could be driven by the presence of relatively massive, bound objects within the disc.

We present an analytical method for constraining the substellar initial mass function in globular clusters, based on the observed frequency of transit events. Globular clusters typically have very high stellar densities where close encounters are relatively common, and thus tidal capture can occur to form close binary systems. Encounters between main sequence stars and lower-mass objects can result in tidal capture if the mass ratio is > 0.01. If brown dwarfs exist in significant numbers, they too will be found in close binaries, and some fraction of their number should be revealed as they transit their stellar companions. We calculate the rate of tidal capture of brown dwarfs in both segregated and unsegregated clusters, and find that the tidal capture is more likely to occur over an initial relaxation time before equipartition occurs. The lack of any such transits in recent HST monitoring of 47 Tuc implies an upper limit on the frequency of brown dwarfs (< 15 % relative to stars) which is significantly below that measured in the galactic field and young clusters.

Recent surveys of star forming regions have shown that most stars, and probably all massive stars, are born in dense stellar clusters. The mechanism by which a molecular cloud fragments to form several hundred to thousands of individual stars has remained elusive. Here, we use a numerical simulation to follow the fragmentation of a turbulent molecular cloud and the subsequent formation and early evolution of a stellar cluster containing more than 400 stars. We show that the stellar cluster forms through the hierarchical fragmentation of a turbulent molecular cloud. This leads to the formation of many small subclusters which interact and merge to form the final stellar cluster. The hierarchical nature of the cluster formation has serious implications in terms of the properties of the new-born stars. The higher number-density of stars in subclusters, compared to a more uniform distribution arising from a monolithic formation, results in closer and more frequent dynamical interactions. Such close interactions can truncate circumstellar discs, harden existing binaries, and potentially liberate a population of planets. We estimate that at least one-third of all stars, and most massive stars, suffer such disruptive interactions.

We investigate the dynamical effects of rapid gas expulsion from the core of a young stellar cluster. The aims of this study are to determine 1) whether a mass-segregated core survives the gas expulsion and 2) the probable location of any massive stars that have escaped from the core. Feedback from massive stars is expected to remove the gas from the core of the cluster first, as that is where most massive stars are located. We find that gas expulsion has little effect on the core for a core star formation efficiency, of greater than 50%. For lower values of star formation efficiency down to 20%, a reduced core survives containing the majority of the massive stars while some of them are dispersed into the rest of the cluster. In fact we find that ejected stars migrate from radial to tangential orbits due to stellar encounters once they leave the core. Thus, the location of massive stars outside of the core does not exclude their forming in the dense cluster core. Few massive stars are expected to remain in the core for a star formation efficiency lower than 20%.

We present results from the first hydrodynamical star formation calculation to demonstrate that close binary stellar systems (separations $\lsim 10$ AU) need not be formed directly by fragmentation. Instead, a high frequency of close binaries can be produced through a combination of dynamical interactions in unstable multiple systems and the orbital decay of initially wider binaries. Orbital decay may occur due to gas accretion and/or the interaction of a binary with its circumbinary disc. These three mechanisms avoid the problems associated with the fragmentation of optically-thick gas to form close systems directly. They also result in a preference for close binaries to have roughly equal-mass components because dynamical exchange interactions and the accretion of gas with high specific angular momentum drive mass ratios towards unity. Furthermore, due to the importance of dynamical interactions, we find that stars with greater masses ought to have a higher frequency of close companions, and that many close binaries ought to have wide companions. These properties are in good agreement with the results of observational surveys.

We present results from the largest numerical simulation of star formation to resolve the fragmentation process down to the opacity limit. The simulation follows the collapse and fragmentation of a large-scale turbulent molecular cloud to form a stellar cluster and, simultaneously, the formation of circumstellar discs and binary stars. This large range of scales enables us to predict a wide variety of stellar properties for comparison with observations. The calculation clearly demonstrates that star formation is a highly-dynamic and chaotic process. Star-disc encounters form binaries and truncate discs. Stellar encounters disrupt bound multiple systems. The cloud produces roughly equal numbers of stars and brown dwarfs, with masses down to the opacity limit for fragmentation (~5 Jupiter masses). The initial mass function is consistent with a Salpeter slope (Gamma=-1.35) above 0.5 Msun, a roughly flat distribution (Gamma=0) in the range 0.006-0.5 Msun, and a sharp cutoff below ~0.005 Msun. This is consistent with recent observational surveys. The brown dwarfs form by the dynamical ejection of low-mass fragments from dynamically unstable multiple systems before the fragments have been able to accrete to stellar masses. Close binary systems (with separations <10 AU) are not formed by fragmentation in situ. Rather, they are produced by hardening of initially wider multiple systems through a combination of dynamical encounters, gas accretion, and/or the interaction with circumbinary and circumtriple discs. Finally, we find that the majority of circumstellar discs have radii less than 20 AU due to truncation in dynamical encounters. This is consistent with observations of the Orion Trapezium Cluster and implies that most stars and brown dwarfs do not form large planetary systems.

Using a local model Gammie (2001) has shown that accretion discs with cooling times t_cool <= 3/Omega fragment into gravitationally bound objects, while those with cooling times t_cool > 3/Omega evolve into a quasi-steady state. We use three-dimensional smoothed particle hydrodynamic simulations of protoplanetary accretion discs to test if the local results hold globally. We find that for disc masses appropriate for T Tauri discs, the fragmentation boundary still occurs at a cooling time close to t_cool = 3/Omega. For more massive discs, which are likely to be present at an earlier stage of the star formation process, fragmentation occurs for longer cooling times, but still within a factor of two of that predicted using a local model. These results have implications not only for planet formation in protoplanetary discs and star formation in AGN discs, but also for the redistribution of angular momentum which could be driven by the presence of relatively massive objects within the accretion disc.

The orbital elements of extreme mass ratio binaries will be modified by interactions with surrounding circumstellar disks. For brown dwarf companions to Solar-type stars the resulting orbital migration is sufficient to drive short period systems to merger, creating a brown dwarf desert at small separations. We highlight the similarities and the differences between the migration of brown dwarfs and massive extrasolar planets, and discuss how observations can test a migration model for the brown dwarf desert.

We use high resolution numerical simulations to study whether gravitational instabilities within circumstellar discs can produce astrometrically detectable motion of the central star. For discs with masses of M_disc = 0.1 M_star, which are permanantly stable against fragmentation, we find that the magnitude of the astrometric signal depends upon the efficiency of disc cooling. Short cooling times produce prominent filamentary spiral structures in the disc, and lead to stellar motions that are potentially observable with future high precision astrometric experiments. For a disc that is marginally unstable within radii of \~10 au, we estimate astrometric displacements of 10-100 microarcsec on decade timescales for a star at a distance of 100 pc. The predicted displacement is suppressed by a factor of several in more stable discs in which the cooling time exceeds the local dynamical time by an order of magnitude. We find that the largest contribution comes from material in the outer regions of the disc and hence, in the most pessimistic scenario, the stellar motions caused by the disc could confuse astrometric searches for low mass planets orbiting at large radii. They are, however, unlikely to present any complications in searches for embedded planets orbiting at small radii, relative to the disc size, or Jupiter mass planets or greater orbiting at large radii.

We present results from the first hydrodynamical star formation calculation to demonstrate that brown dwarfs are a natural and frequent product of the collapse and fragmentation of a turbulent molecular cloud. The brown dwarfs form via the fragmentation of dense molecular gas in unstable multiple systems and are ejected from the dense gas before they have been able to accrete to stellar masses. Thus, they can be viewed as `failed stars'. Approximately three quarters of the brown dwarfs form in gravitationally-unstable circumstellar discs while the remainder form in collapsing filaments of molecular gas. These formation mechanisms are very efficient, producing roughly the same number of brown dwarfs as stars, in agreement with recent observations. However, because close dynamical interactions are involved in their formation, we find a very low frequency of binary brown dwarf systems ($\lsim 5$%) and that those binary brown dwarf systems that do exist must be close $\lsim 10$ AU. Similarly, we find that young brown dwarfs with large circumstellar discs (radii $\gsim 10$ AU) are rare ($\approx 5$%).

We show that the dearth of brown dwarfs in short-period orbits around Solar-mass stars - the brown dwarf desert - can be understood as a consequence of inward migration within an evolving protoplanetary disc. Brown dwarf secondaries forming at the same time as the primary star have masses which are comparable to the initial mass of the protoplanetary disc. Subsequent disc evolution leads to inward migration, and destruction of the brown dwarf, via merger with the star. This is in contrast with massive planets, which avoid this fate by forming at a later epoch when the disc is close to being dispersed. Within this model, a brown dwarf desert arises because the mass at the hydrogen burning limit is coincidentally comparable to the initial disc mass for a Solar mass star. Brown dwarfs should be found in close binaries around very low mass stars, around other brown dwarfs, and around Solar-type stars during the earliest phases of star formation.

In this review, I present the case for how massive stars may form through stellar collisions. This mechanism requires very high stellar densities, up to 4 orders of magnitude higher than are observed in the cores of dense young clusters. In this model, the required stellar densities arise due to gas accretion onto stars in the cluster core, including the precursers of the massive stars. This forces the core to contract until the stellar densities are sufficiently high for collisions to occur. Gas accretion is also likely to play a major role in determining the distribution of stellar masses in the cluster as well as the observed mass segregation. One of the main advantages of this mechanism is that it explicitly relies on the cluster environment in order to produce the massive stars. It is thus in a position to explain the relation between clustered and massive star formation which is not an obvious outcome of an isolated accretion mechanism. A recent numerical simulation supports this model as the cluster core increases its density by $10^5$ during gas accretion. Approximately 15 stellar collisions occur (with $R_{coll}=1$ au) in the cluster core, making a significant contribution to the mass of the most massive star.

We present time series spectroscopy of the Classical T Tauri star VZ Cha. We follow spectral variations at intermediate resolution over five successive nights, or approximately two rotation periods. We see profile features which persist on timescales longer than the expected infall time from the inner disc, and we see expected evidence of rotational variations in the lines, but we also note that rotation alone cannot produce all the observed variability and some other mechanism must be invoked. The behaviour of H-alpha is observed to be markedly different from that of the other lines. In particular, the evidence of rotational effects is lacking at H-alpha, and the activity in the red and blue wings of the line is not significantly correlated, in contrast to the other Balmer lines.

We present a simple physical mechanism that can account for the observed stellar mass spectrum for masses $\ms \simgreat 0.5 \solm$. The model depends solely on the competitive accretion that occurs in stellar clusters where each star's accretion rate depends on the local gas density and the square of the accretion radius. In a stellar cluster, there are two different regimes depending on whether the gas or the stars dominate the gravitational potential. When the cluster is dominated by cold gas, the accretion radius is given by a tidal-lobe radius. This occurs as the cluster collapses towards a $\rho\propto R^{-2}$ distribution. Accretion in this regime results in a mass spectrum with an asymptotic limit of $\gamma=-3/2$ (where Salpeter is $\gamma=-2.35$). Once the stars dominate the potential and are virialised, which occurs first in the cluster core, the accretion radius is the Bondi-Hoyle radius. The resultant mass spectrum has an asymptotic limit of $\gamma=-2$ with slightly steeper slopes ($\gamma\approx-2.5$) if the stars are already mass-segr egated. Simulations of accretion onto clusters containing 1000 stars show that as expected, the low-mass stars accumulate the majority of their masses during the gas dominated phase whereas the high-mass stars accumulate the majority of their massed during the stellar dominated phase. This results in a mass spectrum with a relatively shallow $\gamma\approx 3/2$ power-law for low-mass stars and a steeper, power-law for high-mass stars $ -2.5\simless\gamma\le -2$. This competitive accretion model also results in a mass segregated cluster.

We investigate the physics of gas accretion in young stellar clusters. Accretion in clusters is a dynamic phenomenon as both the stars and the gas respond to the same gravitational potential. Accretion rates are highly non-uniform with stars nearer the centre of the cluster, where gas densities are higher, accreting more than others. This competitive accretion naturally results in both initial mass segregation and a spectrum of stellar masses. Accretion in gas-dominated clusters is well modelled using a tidal-lobe radius instead of the commonly used Bondi-Hoyle accretion radius. This works as both the stellar and gas velocities are under the influence of the same gravitational potential and are thus comparable. The low relative velocity that results means that the tidal radius is smaller than the Bondi-Hoyle radius in these systems. In contrast, when the stars dominate the potential and are virialised, the Bondi-Hoyle radius is smaller than the tidal radius and thus Bondi-Hoyle accretion is a better fit to the accretion rates.

We have simulated encounters between planetary systems and single stars in various clustered environments. This allows us to estimate the fraction of systems liberated, the velocity distribution of the liberated planets, and the separation and eccentricity distributions of the surviving bound systems. Our results indicate that, for an initial distribution of orbits that is flat in log space and extends out to 50AU, 50% of the available planets can be liberated in a globular cluster, 25% in an open cluster, and less than 10% in a young cluster. These fractions are reduced to 25%, 12% and 2% if the initial population extends only to 20AU. Furthermore, these free-floating planets can be retained for longer than a crossing time only in a massive globular cluster. It is therefore difficult to see how planets, which by definition form in a disc around a young star, could be subsequently liberated to form a significant population of free floating substellar objects in a cluster.

We investigate how the formation and evolution of extrasolar planetary systems can be affected by stellar encounters that occur in the crowded conditions of a stellar cluster. Using plausible estimates of cluster evolution, we show how planet formation may be supressed in globular clusters while planets wider than 0.1 AU that do form in such environments can be ejected from their stellar system. Less crowded systems such as open clusters have a much reduced effect on any planetary system. Planet formation is unaffected in open clusters and only the wider planetary systems will be disrupted during the cluster's lifetime. The potential for free-floating planets in these environments is also discussed.

We consider the dynamics of a protostellar disc in a binary system where the disc is misaligned with the orbital plane of the binary, with the aim of determining the observational consequences for such systems. The disc wobbles with a period approximately equal to half the binary's orbital period and precesses on a longer timescale. We determine the characteristic timescale for realignment of the disc with the orbital plane due to dissipation. If the dissipation is determined by a simple isotropic viscosity then we find, in line with previous studies, that the alignment timescale is of order the viscous evolution timescale. However, for typical protostellar disc parameters, if the disc tilt exceeds the opening angle of the disc, then tidally induced shearing within the disc is transonic. In general, hydrodynamic instabilities associated with the internally driven shear result in extra dissipation which is expected to drastically reduce the alignment timescale. For large disc tilts the alignment timescale is then comparable to the precession timescale, while for smaller tilt angles $\delta$, the alignment timescale varies as $(\sin \delta)^{-1}$. We discuss the consequences of the wobbling, precession and rapid realignment for observations of protostellar jets and the implications for binary star formation mechanisms.

We present SCUBA observations of the protomultiple system NGC1333/IRAS4 at 450um and 850um. The 850um map shows significant extended emission which is most probably a remnant of the initial cloud core. At 450um, the component 4A is seen to have an elongated shape suggestive of a disk. Also we confirm that in addition to the 4A and 4B system, there exists another component 4C, which appears to lie out of the plane of the system and of the extended emission. Deconvolution of the beam reveals a binary companion to IRAS4B. Simple considerations of binary dynamics suggest that this triple 4A-4BI-4BII system is unstable and will probably not survive in its current form. Thus IRAS4 provides evidence that systems can evolve from higher to lower multiplicity as they move towards the main sequence. We construct a map of spectral index from the two wavelengths, and comment on the implications of this for dust evolution and temperature differences across the map. There is evidence that in the region of component 4A the dust has evolved, probably by coagulating into larger or more complex grains. Furthermore, there is evidence from the spectral index maps that dust from this object is being entrained in its associated outflow.

Observations have revealed that most stars are born in clusters. As these clusters typically contain more mass in gas than in stars, accretion can play an important role in determining the final stellar masses. Numerical simulations of gas accretion in stellar clusters have found that the stars compete for the available reservoir of gas. The accretion rates are highly nonuniform and are determined primarily by each star's position in the cluster. Stars in the centre accrete more gas, resulting in initial mass segregation. This competitive accretion naturally results in a mass spectrum and is potentially the dominant mechanism for producing the initial mass function. Furthermore, accretion on to the core of a cluster forces it to shrink, which may result in formation of massive stars through collisions.

Observations have revealed that most stars are born in clusters. These systems, containing from tens to thousands of stars and typically significant mass in gas in the youngest systems, evolve due to a combination of stellar and star-gas interactions. Simulations of pure stellar systems are used to investigate possible initial configurations including ellipticity, substructure and mass segregation. Simulations of gas-rich clusters investigate the effects of accretion on the cluster dynamics and on the individual masses that result in a stellar mass spectrum. Further stellar interactions, including binary destruction and eventually cluster dissolution are also discussed.

The position of pre-main-sequence or protostars in the Hertzsprung--Russell diagram is often used to determine their mass and age by comparison with pre-main-sequence evolution tracks. On the assumption that the stellar models are accurate, we demonstrate that, if the metallicity is known, the mass obtained is a good estimate. However, the age determination can be very misleading because it is significantly (generally different by a factor of two to five) dependent on the accretion rate and, for ages less than about one million years, the initial state of the star. We present a number of accreting protostellar tracks that can be used to determine age if the initial conditions can be determined and the underlying accretion rate has been constant in the past. Because of the balance established between the Kelvin-Helmholtz, contraction timescale and the accretion timescale a pre-main-sequence star remembers its accretion history. Knowledge of the current accretion rate, together with an H--R-diagram position gives information about the rate of accretion in the past but does not necessarily improve any age estimate. We do not claim that ages obtained by comparison with these particular accreting tracks are likely to be any more reliable than those from comparisons with non-accreting tracks. Instead we stress the unreliability of any such comparisons and use the disparities between various tracks to estimate the likely errors in age and mass estimates. We also show how a set of coeval accreting objects do not appear coeval when compared with non-accreting tracks. Instead accreting pre-main-sequence stars of around a solar mass are likely to appear older than those of either smaller or larger mass.

We review recent work that investigates the formation of stellar clusters, ranging in scale from globular clusters through open clusters to the small scale aggregates of stars observed in T associations. In all cases, recent advances in understanding have been achieved through the use of state of the art stellar dynamical and gas dynamical calculations, combined with the possibility of intercomparison with an increasingly large dataset on young clusters. Among the subjects that are highlighted are the frequency of cluster-mode star formation, the possible relationship between cluster density and the highest stellar mass, subclustering and the dynamical interactions that occur in compact aggregates, such as binary star formation. We also consider how the spectrum of stellar masses may be shaped by the process of competitive accretion in dense environments and how cluster properties, such as mass segregation and cluster morphology, can be used in conjunction with numerical simulations to investigate the initial conditions for cluster formation. Lastly, we contrast bottom-up and top-down scenarios for cluster formation and discuss their applicability to the formation of clusters on a range of scales.

We present a method of determining lower limits on the masses of pre-main-sequence (PMS) stars and so constraining the PMS evolutionary tracks. This method uses the red-shifted absorption feature observed in some emission-line profiles of T Tauri stars indicative of infall. The maximum velocity of the accreting material measures the potential energy at the stellar surface, which, combined with an observational determination of the stellar radius, yields the stellar mass. This estimate is a lower limit owing to uncertainties in the geometry and projection effects. Using available data, we show that the computed lower limits can be larger than the masses derived from PMS evolutionary tracks for M < 0.5 M_\odot. Our analysis also supports the notion that accretion streams do not impact near the stellar poles but probably hit the stellar surface at moderate latitudes.

We present a model for the formation of massive ($M > 10 M_\odot$) stars through accretion-induced collisions in the cores of embedded dense stellar clusters. This model circumvents the problem of accreting onto a star whose luminosity is sufficient to reverse the infall of gas. Instead, the central core of the cluster accretes from the surrounding gas, thereby decreasing its radius until collisions between individual components become significant. These components are, in general, intermediate-mass stars that have formed through accretion onto low-mass protostars. Once a sufficiently massive star has formed to expel the remaining gas, the cluster expands in accordance with this loss of mass, halting further collisions. This process implies a critical stellar density for the formation of massive stars, and a high rate of binaries formed by tidal capture.

This chapter reviews the dynamical processes in young stellar clusters. The accretion of gas by individual stars affects the dynamics of the cluster, and the masses of the stars. Dynamical mass segregation cannot explain the degree of mass segregation observed in clusters such as the Trapezium Cluster in Orion, implying that the location of the massive stars is an indication of where they formed. This can, however, be explained by the competitive accretion model. If most, if not all, stars form in binary systems, then the high proportion of pre-main sequence binary systems can evolve in clusters to the smaller proportion of Galactic field stars through binary-binary and binary-single interactions. These models make definite predictions for the distribution of binary properties. The early evolution of a cluster sensitively depends on the primordial binary star proportion. Close encounters between stars with circum-stellar discs have drastic effects on the discs and on the stellar orbits. The discs are truncated at radii comparable to the encounter peri-astron, limiting their lifetimes and affecting their potential for planet formation. If the gas represents a significant fraction of the total cluster mass, its removal on dynamical time-scales can unbind the cluster allowing the stars to escape and populate the field. (shortened)

A method for following fragmentation simulations further in time using smoothed particle hydrodynamics (SPH) is presented. In a normal SPH simulation of the collapse and fragmentation of a molecular cloud, high-density regions of gas that form protostars are represented by many particles with small separations. These high-density regions require small time steps, limiting the time for which the simulation can be followed. Thus, the end result of the fragmentation can never be definitively ascertained, and comparisons between cloud fragmentation calculations and the observed characteristics of stellar systems cannot be made. In this paper, each high-density region is replaced by a single, non-gaseous particle, with appropriate boundary conditions, which contains all the mass in the region and accretes any infalling mass. This enables the evolution of the cloud and the resulting protostars to be followed for many orbits or until most of the original cloud mass has been accreted.

We examine the implications of a beat frequency modulated model of T Tauri accretion. In particular we show that measurements of the variability of accretion generated lines can be used in conjunction with existing photometry to obtain a measurement of the underlying photospheric and disc flux. This provides an independent way of checking spectral energy distribution modelling. In addition, we show how spectroscopy of T Tauri stars can reveal the inclination angle between the magnetic axis and the plane of the disc.

A general model of magnetically controlled accretion onto T Tauri stars is presented. In this model the magnetic field is oriented arbitrarily in relation to the star's rotation axis. The resultant interplay between the magnetic field and accretion disc causes a variable accretion rate. The dominant timescale of this variability is the beat frequency between the stellar rotation frequency and the orbital frequency at the magnetosphere boundary. This model is analogous to that developed to explain quasi-periodic oscillations in low-mass X-ray binaries.

A viable solution to the origin of close binary systems, unaccounted for in recent theories, is presented. Fragmentation, occurring at the end of the secondary collapse phase (during which molecular hydrogen is dissociating), can form binary systems with separations less than 1 au. Two fragmentation modes are found to occur after the collapse is halted. The first consists of the fragmentation of a protostellar disc due to rotational instabilities in a protostellar core, involving both an $m=1$ and an $m=2$ mode. This fragmentation mechanism is found to be insensitive to the initial density distribution: it can occur in both centrally condensed and uniform initial conditions. The second fragmentation mode involves the formation of a rapidly rotating core at the end of the collapse phase which is unstable to the axisymmetric perturbations. This core bounces into a ring which quickly fragments into several components. The binary systems thus formed contain less than 1 per cent of a solar mass and therefore will need to accrete most of their final mass if they are to form a binary star system. Their orbital properties will thus be determined by the properties of the accreted matter.