This book will be required reading for anyone interested in the Greek or Roman family. Sabine R. Hübner exploits papyrological evidence which is of such a quality as to provide us with insights into family life unparalleled for any historical period earlier than the Renaissance. To supplement this material, she makes extensive use of anthropology. The benefit of such an approach is that it focuses research on the micro-dynamics of family structures. In particular, she considers the tensions inherent within families, focusing on those between generations and between women (wives and mothers-in-law). Given that family structures must emerge from individual decisions, this focus on micro-dynamics seems to me to be the way forward in family studies. H. overplays the innovative nature of her approach; nevertheless, there is little Classical scholarship that engages so thoroughly and with such sophistication with anthropological studies of the family.

Anthropologists tend to draw primarily on carefully collected qualitative data supported by quantitative analysis, but both these datasets are problematic for Roman Egypt. The quantity of the anecdotal material is slight and often the context is problematic so that moving from the anecdote to a generalization is fraught with difficulties. It becomes tempting to ‘cherry-pick’ stories that fit the model. The model suggests that new wives entering patriarchal households were in a difficult situation, and having argued previously along similar lines, I am sympathetic to the proposition. H. suggests that relations between wives and mothers-in-law were difficult and offers many anthropological parallels, but finds only a single letter, P.Petaus 29, to support the proposition. Similarly, she argues that adult women within a household were in competition, but can offer no documentary support. It is possible that prevalent social phenomena are hidden from view, but we have a considerable number of letters and petitions where one might expect the reporting of such tensions.

The core evidence for the analysis comes from the census returns. These are far from straightforward documents and H. might have considered more critically their value both as ‘data’ for the demographic regime and as representations of family structures. For the former, she relies completely on Bagnall and Frier's analysis (The Demography of Roman Egypt (1994)), without considering possible problems with their model. She also never seriously considers how the census process itself might affect the patterns of family that appear in the documentation. Each declaration offers a snapshot of residence, but does not record the processes by which that particular group came into being. The census declaration was an official and governmental text and the legalistic context may have influenced decisions on the representation of the census group. For example, we see many households centred on brothers in the documentation. But if one brother (or his descendants) separated their portion of the family property, they might submit their own return(s). It is not clear how this would alter living arrangements. Adult cousins might co-reside in a house, but not appear on the same return. It is also unclear how a census or residential group might relate to the more amorphous group of ‘family’.

The issue is particularly acute when we consider the status of women. Very often in the census records, women look as if they have been parachuted into a family and to be without prior close connection (and thus vulnerable), but in a society that practised close-kin marriage (however defined), in nearly all cases there must have been a pre-existing and close tie between a girl's natal and conjugal families. On marriage, a girl might relocate within the same house or from the house next door. In small communities, she was unlikely to be far from her natal family. Sometimes one has to look outside the particular document (or type of document) to understand social dynamics.

Indeed, in the aforementioned P. Petaus 29, we know about the fraught family circumstances precisely because the wife wrote to her father who wrote to the husband's brother about the behaviour of the mother. Her father may have been away (the absence generating the letter), but the wife was not alone. The social bonds that led to her marriage did not consequently dissolve. It is a mistake to see the ‘census household’ as an isolated social structure with clearly defined boundaries that separated its members from ‘neighbourhood’ or ‘community’.

The anthropological method with its focus on individual decision-making is somewhat at odds with the tendency to homogenize the ‘Roman family in Egypt’ into a single model, drawing for confirmatory material on Ptolemaic and late antique documents. Partly, H. struggles to adapt her approach to the extreme demographic régime adopted. Yet, it should follow from a micro-demographic approach that decision-making would vary according to individual circumstances. Family structures would adjust accordingly. In spite of what one would assume to have been a similar mortality régime, in Ptolemaic Egypt Egyptian and Greek families adopted differing household strategies. In the Roman period, urban and rural families also tended to be differently structured. If decisions were heavily influenced by economic circumstances, then the emergence of wealthier strata of society in the Roman period might have led to differential family strategies. Similarly, cultural factors, such as Roman law or Christianity, may also have had some effect. Finally, there is at least indicative evidence of regional variations in patterns of family formation. It seems to me that instead of worrying too much about structure, we should be considering the factors that lead to the adoption of particular familial and household strategies.

Such thoughts are to detract only marginally from this study. H.'s use of anthropology raises the standard for family studies in antiquity and challenges more positivistic approaches. She shows the benefits of an anthropologically sophisticated approach and offers propositions that should change methodologies and our understanding of the Roman family.