We have derived the mean proper motions and space velocities of 154 Galactic globular clusters and the velocity dispersion profiles of 141 globular clusters based on a combination of Gaia DR2 proper motions with ground-based line-of-sight velocities. Combining the velocity dispersion profiles derived here with new measurements of the internal mass functions allows us to model the internal kinematics of 144 clusters, more than 90% of the currently known Galactic globular cluster population. We also derive the initial cluster masses by calculating the cluster orbits backwards in time applying suitable recipes to account for mass loss and dynamical friction. We find a correlation between the stellar mass function of a globular cluster and the amount of mass lost from the cluster, pointing to dynamical evolution as one of the mechanisms shaping the mass function of stars in clusters. The mass functions also show strong evidence that globular clusters started with a bottom-light initial mass function. Our simulations show that the currently surviving globular cluster population has lost about 80% of its mass since the time of formation. If globular clusters started from a log-normal mass function, we estimate that the Milky Way contained about 500 globular clusters initially, with a combined mass of about 2.5 · 10 8 M ⊙ . For a power-law initial mass function, the initial mass in globular clusters could have been a factor of three higher.

We present the chemical analysis of 49 giant stars of the globular cluster NGC 2419, using medium resolution spectra collected with the multi-object spectrograph DEIMOS@Keck. Previous analysis of this cluster revealed a large dispersion in the line strength of the infrared Ca II triplet, suggesting an intrinsic star-to-star scatter in its Fe or Ca content. From our analysis, we assess that all the investigated stars share the same [Fe/H], [Ca/Fe] and [Ti/Fe] abundance ratios, while a large spread in Mg and K abundances is detected. The distribution of [Mg/Fe] is bimodal, with ∼40 per cent of the observed targets having subsolar [Mg/Fe], down to [Mg/Fe] ∼ −1 dex, a level of Mg deficiency never observed before in globular clusters. It is found that the large dispersion in Mg abundances is likely the main origin of the observed dispersion of the Ca II triplet lines strengths (that can be erroneously interpreted in terms of Fe or Ca abundance scatter) because Mg plays a relevant role in the atmosphere of giant stars as an electron donor. A strong depletion in the Mg abundance leads to an increase of the line strength of the Ca II triplet, due to the variation in the electronic pressure, at a constant Fe and Ca abundance. Finally, we detect an anti-correlation between Mg and K abundances, not easily explainable within the framework of the current nucleosynthesis models.

We used deep observations collected with Advanced Camera for Surveys (ACS) at Hubble Space Telescope (HST) to derive the fraction of binary systems in a sample of 13 low‐density Galactic globular clusters. By analysing the colour distribution of main‐sequence stars we derived the minimum fraction of binary systems required to reproduce the observed colour–magnitude diagram morphologies. We found that all the analysed globular clusters contain a minimum binary fraction larger than 6 per cent within the core radius. The estimated global fractions of binary systems range from 10 to 50 per cent depending on the cluster. A dependence of the relative fraction of binary systems on the cluster age has been detected, suggesting that the binary disruption process within the cluster core is active and can significantly reduce the binary content in time.

No abstract