Abstract We reply here to the questions raised by Wu et al. concerning the results published by Mahdaoui et al. (2015). This paper describes experiments in which aqueous solutions containing ReO 4 − and OsCl 6 2− were brought in contact with natural oils at various ranges of concentration, time and temperature. The main observation is that the transfer of Re and Os to oils is very efficient under all experimental conditions. Wu et al. argue that thermodynamic equilibrium was not achieved in these experiments as apparent partition coefficients are inconsistent. They conclude that the experiments were flawed by possible leaking of reactors and that the conclusions by Mahdaoui et al. (2015) were not justified. In the following reply we explain that Mahdaoui et al. (2015) never claimed that thermodynamic equilibrium was achieved. Any calculations or considerations in this context are therefore meaningless. We recall the objectives of our publication, which were to experimentally test the behavior of ReO 4 − and OsCl 6 − (two plausible chemical forms of Re and Os in deep aquifers of petroleum systems) in aqueous solution-oil systems. To our knowledge these are the first experiments of their kind. The parameters that could influence the precision and reproducibility of our results were discussed in detail in Mahdaoui et al. (2015). The essential point is that all 60 of the experiments provide evidence of substantial transfer of Re and Os from water to oil. In contrast to what was mistakenly understood by Wu et al., the paper does not challenge the use of Re–Os to date geological events affecting petroleum. Instead, by providing a mechanism that might allow Os isotopic homogenization on a basin-wide scale, a critical step missing from most current models, it offers a possible explanation of how Re–Os geochronology in oils could potentially work. More generally, our study suggests that transfer of Re and Os from waters to oil may be an important phenomenon that should not be overlooked.