Debates about the population composition at Roman military bases have been vibrant for over two decades. Though the presence of women and children in military settlements is now accepted by most, there have been few attempts to reach beyond showing their presence alone, to scrutinizing the identity, origin or daily life of these non-combatants. To date, there have been few book-length works dedicated to this dialogue and no single-authored monograph. For this reason Penelope Allison's book, which investigates distribution patterns of artefacts in five German military bases from the first and second centuries a.d. in order to study the populations living inside Roman forts and the activities that occupied them, is much welcomed and will be important for researchers and students alike.

A. covers three important aspects of the debate. The first five chapters provide background on the history of the problem and her methodological approach: Introduction followed by ch. 2 on scholarship of women and children in the military sphere; ch. 3 provides a theoretical background for using artefact distribution to determine social practice; ch. 4 discusses site selection and data processing; ch. 5 explains how categorizations for each artefact type were chosen. Chs 6–10 provide the case studies of each fort with descriptions and introductions to the artefact assemblages which form the basis of the discussion (with appendices A–G presenting plans, tables and drawings). The data were compiled with GIS plotting of previously published fort excavations (‘legacy data’). It should be noted that some of the case studies were previously published (Allison, Arch.Dial. 13.1 (2006), 1–20; Allison, BRGK 87 (2006), 387–452), particularly on Ellingen, Vetera I and Oberstimm, but their compilation and expansion here in a single volume is of great value (data are added on Rottweil and Hesselbach). Chs 11–13 present overarching analyses of the material and conclusions about the rôle of women and children in military forts based on the evidence presented. Ch. 12 is the most vital chapter in this regard, for its discussion of the strong evidence for women's and children's presence inside some Roman forts and will become standard reading for anyone diving into this often confrontational debate.

A strength of the volume is its willingness to return to topics that have been dismissed by previous scholars. An example is the subject of officers' wives and families living inside the fort, whose presence has been accepted because they were allowed legal marriage (13–14), but with almost no further discussion of their social identity. A. points to these discrepancies in past research (2) and insists that we move beyond the mere acceptance of the presence of non-combatants and start to investigate social identity. Whether she fulfills this aim overall varies in each case study. Equally welcome is A.'s treatment of some of the military fort as domestic space (3). Though the military milieu requires one to remember the idiosyncrasies specific to this context, it does not necessitate a unique definition of space used by human groups. Also welcome is A.'s attempt to deconstruct past structuralist approaches to military space (for example, H. von Petrikovits, Die Innenbauten römischer Legionslager während der Prinzipatzeit (1975)), and to deny the common approach of categorizing military spaces and artefacts with a single label. A. approaches the artefact assemblages not immediately as items associated with the Roman army, but rather as the product of human production and consumption, and from this premise analyses their function and ability to reveal socio-spatial patterns.

While A.'s approach to mapping the social use of space through artefact deposition is especially welcome in Roman military studies, there are a few points to consider critically. One potential downfall is addressed by Allison early on (39; 51–4): can we expect artefact distribution patterns to reflect actual spaces of use and loss or are we merely quantifying the random distribution of demolition, levelling and rubbish accumulation and mistaking it for localized habitational debris? A. suggests we not throw the baby out with the bathwater. Certainly some accumulated finds represent rubbish disposal or levelling for subsequent building, while other artefacts, such as many of those used in this volume, slip through the cracks and remain there (51–2). Each context needs to be evaluated in its own right, and its formation processes understood in order to say something relevant about the meaning of perceived patterns. This is precisely what A. does in order to better understand the social use of space within Roman military forts. Another potential detractor is the ambiguity of the use of some artefacts and their attribution to a certain category of user. For example, when examining fasteners, A. concludes that ‘… it is not always certain whether they were part of combat dress, civilian dress, horse equipment or had other uses’ (69). Bronze needles and pins are by necessity categorized as ‘dress? cloth-working? toilet?’ and as ‘male? female?’ (100). This conservative approach is appreciated, and other artefacts indeed provide greater assurance, but at times the uncertainties seem too pronounced to formulate a comfortable conclusion about the identities of those who used the artefacts under scrutiny. In this regard, ch. 11 is most useful for boiling down the percentages and patterns in a concise discussion of artefact distribution throughout spaces in the forts investigated. A.'s conclusion here that we stop looking for uni-functional definitions of space is important and she urges that we consider military spaces as multi-purpose and dynamic (317).

Archaeology more often speaks in probabilities and percentages with varying levels of certainty. The author knows that ascriptions will be disputed, but argues that the overall patterns that emerge are meaningful and hopes that this evidence will encourage a dialogue about socio-spatial practices inside Roman forts (108). It is also hoped that work along these lines will encourage a new level of artefact recording and reporting that would allow further studies such as this one. A.'s conclusions that the presence of women and children beyond the officers' families was commonplace (336) and that they likely took part in activities that helped maintain the garrison (342) are important and now need to be considered in any future treatment of military space and the broader communities supporting Roman garrisons.