Though the earliest evidence for the use of fire is a subject of debate, it is clear that by the late Middle Paleolithic, Neandertals in southwest France were able to use fire. The archaeological record of fire use in this place and time is, however, quite patchy. While there are a growing number of sites with impressive evidence for fire use, there are also a much larger number of sites without such evidence. Based primarily on evidence from two recently excavated well-stratified Middle Paleolithic sites, we argue here that taphonomic issues, sampling bias, or site use are not sufficient explanations to account for the relative lack of evidence for fire. Given that modern huntergatherers use fire daily and in a wide variety of circumstances, the prolonged periods of Mousterian occupation without fires, even during some of the harshest conditions of the late Pleistocene, raises significant issues regarding the role of fire during these times. In our view, the evidence suggests that Western European Neandertals were not habitual fire users. One explanation advanced here is that at least some Neandertals, even in the late Middle Paleolithic, lacked the technological skill to make fire on demand, and thus relied on access to natural sources of fire. (Sandgathe et al. 2011a). So while the results provided by them make it clear that there is increasing evidence for fire through the Middle Paleolithic, it seems likely that this pattern is driven by the number of dated Middle Paleolithic contexts, which also increases through time. We therefore suggest that this taphonomic explanation for the pattern should be the null hypothesis. This would help reconcile the authors' observation that the evidence for fire use increases over time with the work of others who have demonstrated an apparent scarcity of fire evidence overall, which appears in only a very small percentage of occupation layers (Cohen-Ofri et al. 2006;Davies and Underdown 2006;Dibble et al. 2009;Gowlett 2006;James 1989;Perlès 1981;Roebroeks and Tuffreau 1999).Here we present evidence, based on recent excavations by the authors at two Mousterian sites in southwest France, Pech de l'Azé IV and Roc de Marsal, for the persistent scarcity of fire well after its first occurrence in the European Middle Paleolithic record. What these data strongly suggest is that while Neandertals occasionally used fire, there were also major periods of time when fires were either not present at these sites or were present only sporadically, even during periods of relatively cold conditions. As will be argued below, the near absence of fire evidence in some levels at these sites cannot be explained by taphonomic processes, excavation bias, or changes in site function, and furthermore, the length of time these sites were occupied without using fire to any significant degree is totally inconsistent with modern hunter-gatherer use of fire. This evidence challenges the assumption of immediate and widespread use of fire, and at least in the case of western European Neandertals, i...