It is over 50 years since the Winchester Excavation Committee's 11 long seasons of excavations (1961–1971) concluded, and archaeological methodologies in and out of the field have advanced substantially since then. The first results of the Winchester programme, as far as the Roman period is concerned, followed in 1979 with the part publication of the late Roman cemetery at Lankhills; the critically important human osteology followed almost 30 years later in 2017 (Clarke Reference Clarke1979; Stuckert Reference Stuckert2017). Clarke provides a valuable reconsideration of the cemetery as an Appendix to volume 2. There is no explanation as to why it has taken so long to bring this work to publication, but it is clear from the acknowledgement to the De Laszlo Foundation and the list of individual donors and grant funders that these sources and the determined resolve of a team of volunteers with a small team of salaried researchers have been crucial to bringing the work to fruition. Co-author Francis Morris has taken a leading role in the project since 2014.

The two volumes are very well cross-referenced with the first reporting the excavations and the second reporting the finds, including some outstanding pieces found before 1961. Volume 1 has an introductory section detailing excavation and post-excavation methods, which is followed by general discussions on the contribution of early discoveries and the 1961–1971 excavations to deepen our understanding of the geology, prehistory, Roman and post-Roman and early Anglo-Saxon periods of Winchester and surrounding area. The meat of the volume follows with the reporting of the individual excavations themed by ‘On the Defences’, ‘Inside the Walls’ and ‘Outside the Walls’. Succinctly describing the excavations, the reports are of a very high standard and are supported by beautiful line drawings of plans and sections and excellent photography.

Given that much of what is reported here was found beneath the medieval and modern city, it is not surprising that the information is fragmentary about some of the sites: for example, the traces of a public building, presumed to be the Roman forum, beneath the medieval minsters at the heart of the town; Roman townhouses beneath medieval Wolvesey Palace; a Romano-Celtic temple beneath complex medieval and post-medieval properties in Lower Brook Street. Perhaps the most significant and stimulating discovery contributing to our knowledge of the early Roman town is that of the Neronian—early Flavian fort-style gate structure and associated defensive Rampart I at Southgate. Too late to belong to the initial Conquest period, one wonders whether this is a post-Boudican military reoccupation or a defence of a newly established community, possibly one of those awarded to the client king Togidubnus. We do not have the full circuit but the reports describe it as substantial, part of it followed by the later Roman defences on the western, northern and southern sides while the military-style ditches found in Lower Brook Street might give a clue to the location of the eastern side. Even if, at 29ha, the enclosed area was only around half that of the later town, a large settlement worthy of such protection is implied—but what was the nature and purpose of that settlement?

The Roman town overlies a large part of the Middle-to-Late Iron Age defended enclosure known as Oram's Arbour but there is no evidence of occupation of the very late Iron Age, the late first century BC/early first century AD, as there is at nearby Chichester and Silchester. The excavations deliver some evidence for a Conquest-period fort but whatever business that stimulated would have ceased on departure of the regiment, probably by the late AD 40s. With at least two roads dating to before the AD 70s and at least three others likely to be of a similar date, Winchester was clearly an important communication hub and a facility; a mansio, to control and to support the cursus publicus would have been essential. Was it the stimulus provided by the need to support a mansio that attracted people and created a town in that location, or was it in combination with an initiative of the client king to provide an administrative centre for that part of his kingdom which, after his death, became the civitas of the Belgae?

With regularity evident in much Roman town planning there is a temptation to extrapolate from sparse evidence of street sightings and here regular blocks of insulae and associated streets are confidently plotted to fill the whole of the walled area (cf. Illus 2.13 with Illus 4.5 where observations of streets are plotted). Yet the clearly awkward relationship between street SA and the hypothesised street running straight between the north and south gates and the street discovered in Lower Brook Street dividing the as-yet-unconfirmed insula XXIV suggests that arrangements might not have been as regular or extensive as projected. Similarly, the presence of five iron bands—four from the Cathedral Green excavation—of the type used to join wooden water pipes together are not sufficient to claim that between the second and fourth centuries AD the town was served by a gravity-fed water supply. It seems more likely that there was some localised distribution from cisterns of the type found at Gresham Street in Roman London (Blair et al. Reference Blair, Spain, Swift, Taylor and Goodburn2006); and the wells discovered in five insulae across the town were clearly important for the water supply.

Volume 2 focuses on individually recorded finds, reported by a multiplicity of individual authors; categories of bulk finds such as the pottery (other than decorated and stamped terra sigillata, amphorae and imported late Roman fine wares) and the animal bone will be published later in separate volumes. Nevertheless, it is extremely valuable to have all this material together in a single volume, offering the potential for synthesis and comparative analysis with other Winchester and urban assemblages. There is no mention of a discard policy and no reference to ceramic building material, and the number of querns recorded—only two—is puzzlingly small. Like the querns (p.754 & p.1038), the hones are also reported twice (pp.752–3 & pp.1035–7). There is also duplication in the reporting of the building stone, including the pieces of ornamental marble most of which are from post-Roman contexts, and probably including pieces that arrived in the medieval period.

These minor quibbles do not detract from the overall achievement of the publication of the Roman evidence from the 1961–1971 Winchester excavations. We salute and warmly thank all those who have brought this about.