A number of writers have been attracted to the idea that some of the puzzling features of quantum mechanics might be manifestations of 'reverse' or 'retro' causality, at a level underlying that of the usual quantum description. The main motivation for this view stems from EPR/Bell phenomena, where it offers two virtues. First, as was noted by its earliest proponent, 1 it has the potential to provide a timelike decomposition of the nonlocal correlations revealed in EPR cases -i.e., as we would now put it, 2 for the violation of the Bell inequalities in the quantum world. Second, Bell's derivation of his famous inequality depends explicitly on the assumption that hidden states do not depend on future measurement settings -so that its violation simply invites a retrocausal explanation, at least from the point of view of anyone who has already been bitten by the retrocausal bug. Most people working in the foundations of quantum mechanics remain resolutely unbitten, however. It is common for the retrocausal option to be ignored altogether, or, as in this rather careful recent survey article, relegated to the footnotes with other unmemorabilia:To be scrupulous, there are perhaps four other ways [i.e., other than nonlocality] that the correlations in [an EPR-Bohm] experiment could be explained away. (1) One could simply 'refuse to consider the correlations mysterious'. (2) One could deny that the experimenters have free will to choose the settings of their measurement devices at random, as required for a statistically * This note is based on a talk given at workshops at