c©**Copyright Year* * Astronomical Society of the Pacific Cluster mass dependent truncation of the upper IMF: evidence from observations and simulations

Abstract. We attempt to evaluate whether the integrated galactic IMF (IGIMF) is expected to be steeper than the IMF within individual clusters through direct evaluation of whether there is a systematic dependence of maximum stellar mass on cluster mass. We show that the result is sensitive to observational selection biases and requires an accurate knowledge of cluster ages, particularly in more populous clusters. At face value there is no compelling evidence for non-random selection of stellar masses in low mass clusters but there is arguably some evidence that the maximum stellar mass is anomalously low (compared with the expectations of random mass selection) in clusters containing more than several thousand stars. Whether or not this effect is then imprinted on the IGIMF then depends on the slope of the cluster mass function. We argue that a more economical approach to the problem would instead involve direct analysis of the upper IMF in clusters using statistical tests for truncation of the mass function. When such an approach is applied to data from hydrodynamic simulations we find evidence for truncated mass functions even in the case of simulations without feedback. 1