The stereotype of slave-run latifundia being turned into serf-worked estates is no longer credible as a model of the transition from antiquity to the middle ages, but Chris Wickham’s anomalous characterisation of the Roman Empire as ‘feudal’ is scarcely a viable alternative to that. If a fully-articulated feudal economy only emerged in the later middle ages, what do we make of the preceding centuries? By postulating a ‘general dominance of tenant production’ throughout the period covered by his book, Wickham fails to offer any basis for a closer characterisation of the post-Roman rural labour-force and exaggerates the degree of control that peasants enjoyed in the late Empire and post-Roman world. A substantial part of the rural labour-force of the sixth to eighth centuries comprised groups who, like Rosamond Faith’s inland-workers in Anglo-Saxon England, were more proletarian than peasant-like. The paper suggests the likely ways in which that situation reflected Roman traditions of direct management and the subordination of labour, and outlines what a Marxist theory of the so-called colonate might look like. After discussing Wickham’s handling of the colonate and slavery, and looking briefly at the nature of estates and the fate of the Roman aristocracy, I conclude by criticising the way Wickham uses the category of ‘mode of production’.