Young open clusters are our laboratories for studying high-mass star formation and evolution. Unfortunately, the information that they provide is difficult to interpret, and sometimes contradictory. In this contribution, I present a few examples of the uncertainties that we face when confronting observations with theoretical models and our own assumptions.

Recent searches for supernova II-P progenitors in external galaxies have led to the identification of red objects with magnitudes and colours indicative of red supergiants, in most cases implying quite low luminosities, and hence masses. Current results would suggest that all explosions come from objects less massive than ~ 18 M⊙ and that some explosions come from stars with masses well below 10 M⊙. Stellar models, on the other hand, can only produce explosions from objects more massive than > 9 M⊙. What does our knowledge of local red supergiants tell us about the expected properties of such objects? We present the results of complementary large spectroscopic surveys and detailed studies of open clusters that point to substantially different observational characteristics depending on metallicity.

In the past few years, several clusters containing large numbers of red supergiants have been discovered. These clusters are amongst the most massive young clusters known in the Milky Way, with stellar masses reaching a few 104M⊙. They have provided us, for the first time, with large homogeneous samples of red supergiants of a given age. These large populations make them, despite heavy extinction along their sightlines, powerful laboratories to understand the evolutionary status of red supergiants. While some of the clusters, such as the eponymous RSGC1, are so obscured that their members are only observable in the near-IR, some of them are easily accessible, allowing for an excellent characterisation of cluster and stellar properties. The information gleaned so far from these clusters gives strong support to the idea that late-M type supergiants represent a separate class, characterised by very heavy mass loss. It also shows that the spectral-type distribution of red supergiants in the Milky Way is very strongly peaked towards M1, while suggesting a correlation between spectral type and evolutionary stage.