This volume is a province-wide study of a single artefact type, the seal-box, which is also cross-referenced to published continental assemblages. The history of previous work charts an early antiquarian recognition that study of these objects may prove fruitful, and Andrews rightly gives credit to James Tongue's undergraduate dissertation that was the first attempt to look at the Romano-British assemblage as a whole. Published in 2004 in Lucerna, the Roman Finds Group's Newsletter, Tongue's catalogue of 316 seal-boxes was formed from an extensive literature search. A. has 871 boxes, many of which are from museum collections and the PAS database. He takes a multi-variant approach to typology, combining seal-box shape with design and basal hole pattern to define stylistic groups that are then assessed for distribution and date and compared to the forms and designs found on the Continent. Three area case studies — London, Norfolk and the northern frontier — look more closely at context and attempt to tease out the social distribution and possible range of uses of the seal-boxes from their geographical relationships with a variety of site types and other finds, such as the association of seal-boxes and inkwells in London and seal-boxes and the military in the North. The chapter drawing the data together to establish a narrative for chronology and place of manufacture provides a high point of the work that will be a standard resort for future researchers.

Next A. tackles the questions of how seal-boxes were used and what they were used for, essentially two facets of a single topic. He initially takes a practical approach, examining the material remains of wax and fibres and from them devising a method of threading cord that fits the three-hole pattern found on the majority of boxes. From his experiments, from the archaeological evidence, such as the seal-box found with the Snettisham jeweller's hoard, and from their social distribution he concludes that seal-boxes were used for sealing pouches and bags containing money and other valuables rather than wooden writing-tablets. He links the halt in seal-box use to the decline in intrinsic value of the imperial currency and the dislocation of trade in the second part of the third century.

He then considers the significance of the designs used on seal-box lids. The apotropaic use of phalli, lunulae and ivy leaves and the connection of some designs with particular deities leads to the intriguing idea that even a plain box with a mirror-like tinned surface was designed to turn away a look of evil intent. His summing-up revisits his themes and briefly discusses the flaws inherent in archaeological data.

There is no doubt that A. has explored his subject in depth and provided a thorough and thoughtful piece of work. He has challenged accepted theories and sometimes found new evidence to sustain them or new arguments to refute them. I could wish he had attempted to depict social distribution as thoroughly as he did form and decoration, rather than admit defeat when faced with rural distribution in Norfolk, military distribution in the North and links to trade and literacy in London. Increasingly it is clear that functions of objects were as mutable as personal identity, and there is a need to acknowledge that use was contingent upon context; there may not be one answer, but several, to the questions of who used certain objects and how they used them. The seal-box lid found in the Flavian burial of a young woman at Winchester is a fine example. There were styli and penknives in the grave, but there was also a small collection of crepundia. The box lid may have belonged with either group of objects, but as it was fitted with a rivet in the shape of a hare, an animal occasionally used as a funerary motif, it may even have been a protective grave gift. The flaws in archaeological data can be more than compensated for by the rich plurality of meaning bound up in a single object.