A Comparison of Methods for Determining the Age Distribution of Star Clusters: Application to the Large Magellanic Cloud

Abstract

The age distribution of star clusters in nearby galaxies plays a crucial role in evaluating the lifetimes and disruption mechanisms of the clusters. Two very different results have been found recently for the age distribution chi(t) of clusters in the Large Magellanic Cloud (LMC). We found that chi(t) can be described approximately by a power law chi(t) propto t^{gamma}, with gamma -0.8, by counting clusters in the mass-age plane, i.e., by constructing chi(t) directly from mass-limited samples. Gieles & Bastian inferred a value of gamma~, based on the slope of the relation between the maximum mass of clusters in equal intervals of log t, hereafter the M_max method, an indirect technique that requires additional assumptions about the upper end of the mass function. However, our own analysis shows that the M_max method gives a result consistent with our direct counting method for clusters in the LMC, namely chi(t) propto t^-0.8 for t<10^9 yr. The reason for the apparent discrepancy is that our analysis includes many massive (M>1.5x10^3 M_sol), recently formed (t<10^7 yr) clusters, which are known to exist in the LMC, whereas Gieles & Bastian are missing such clusters. We compile recent results from the literature showing that the age distribution of young star clusters in more than a dozen galaxies, including dwarf and giant galaxies, isolated and interacting galaxies, irregular and spiral galaxies, has a similar declining shape. We interpret this approximately "universal" shape as due primarily to the progressive disruption of star clusters over their first ~few x 10^8 yr, starting soon after formation, and discuss some observational and physical implications of this early disruption for stellar populations in galaxies.