The substantial two-volume Poultry report describes demanding rescue excavations undertaken by the Museum of London opposite the Royal Exchange. Much has already been published in summary reports (as P. Rowsome's Heart of the City (2000)), and on individual finds (see R. Tomlin on a text describing the sale of a slave in Britannia 34 (2003), 41–52), but this definitive account presents a wealth of new material and is essential reading for all with a serious interest in the archaeology of Roman London. The first volume offers a dense description of the archaeological sequences supported by selective publication of key finds assemblages, embracing some smaller satellite sites in the environs. Impatient readers will skip to the synthetic accounts found in the second volume where key information is repeated alongside a stimulating discussion of a range of topics.

The importance of the site — beyond its scale and the resources dedicated to its study — derives from its location astride Roman London's main east–west road (decumanus maximus) on the west bank of the Walbrook stream, where deep water-logged deposits offered rich archaeological pickings. Drains built here using timbers felled a.d. 47/48 defined the line of this principal road and provide the earliest fixed date in the archaeology of London. The site lay just beyond the western boundary of the earliest settlement (city or fort), and questions are left hanging as to whether the road was built in the act of London's foundation, or — as this reviewer thinks more likely — exited a site of slightly earlier origin.

A series of timber buildings was built alongside roads converging on the nearby Walbrook crossing. The remains of 88 such buildings, within an estimated 13 separate Roman properties, are described. Most were of early Roman date, although new houses were being built here down to the mid-fourth century. These structures illustrated imported and local vernacular traditions in earth-fast and timber-framed construction techniques, which showed continuity throughout the Roman period. Reconstructions prepared for the Museum of London's important ‘High Street Londinium’ exhibition (2000–2001) add detail to our understanding of building appearance, form and fabric.

Finds reports are enriched by material from a large rubbish dump, which included stable and granary waste, alongside the road a.d. 53–55 (264–72). Early assemblages included important continental imports, local products and finds with military associations: attesting to ‘a population that was highly “Romanised” from the start and dependent to a great degree on imported artefacts’ (439). Several buildings appear to have been associated with street-side commerce, although there was little evidence of industrial production, and included a tavern and a shop destroyed in the revolt of a.d. 60/61. The shop contained a stock of samian, spoons and spices, showing similarities with contemporary shop assemblages from Colchester and Verulamium. A Flavian bakery or cookhouse (Building 18/48) is also identified. This was set within an unusual brick construction with an outer portico, which was perhaps a public building. The importance of the area for milling/baking was also illustrated by a large assemblage of broken lava quernstones that had been re-used in a cobbled surface, and by the ubiquity of insect species associated with spoilt grain. Among other specialist chapters Hammerson's contribution on the Roman coins (not all of which could be cleaned and studied) is particularly valuable for its comparative discussion of London's larger coin assemblages (and in the frustrating absence of any overall synthesis of Roman coins from the City).

The authors present several significant observations on periods of rapid site-wide growth and contraction, which are likely to reflect on the wider circumstances of Roman London. A concentration of tree-ring dates suggests that a ‘great expansion’ was underway in a.d. 59–61, immediately prior to the Boudican revolt (440). There are conflicting signals about the speed with which recovery took place after the revolt: pottery assemblages suggest a hiatus of ten years, but dendro' dates point to the swift restoration of the public infrastructure and it is possible that some individual houses belonged to a phase of Neronian reconstruction (e.g. Building 21). Another significant phase of expansion took place after a.d. 73, with a cluster of tree-ring dates associated with new roads and drains c. a.d. 78. As elsewhere in London, the pace of building slowed considerably from the middle of the second century, with property amalgamations common in the post-Hadrianic period. Later Roman houses were more expensively decorated, with a greater use of stone-built construction, although this report shows that timber buildings were more common in the later city than is sometimes assumed.

Some of the most interesting results that are newly presented here concern the disuse of the roads and the drainage system in the second half of the fourth century. Occupation at Poultry continued into the last decades of the century but not, it would seem, beyond, and there are indications in the different character of the finds assemblages of this period that ‘abandonment, when it came, may have been abrupt’ (447). A decapitated human body found in a disused roadside culvert illustrates the changing nature of the urban landscape in the final quarter of the century. There is an emerging pattern, supported in particular by evidence from Roman Southwark, of late Roman burials encroaching into previously built-over parts of Roman London. It is tempting to suggest that London was being remodelled as a smaller settlement, with an urban core restricted to the area east of the Walbrook, following a pattern of contraction similar to that witnessed in some cities in Gaul. If so this was a short-lived concept, since the smaller site was not enclosed with walls in the Gallic fashion and excavations in more central areas fail to suggest that London remained recognisably urban beyond the end of the fourth century.

The importance of the Poultry site is such that the report on the PCA excavations at Tokenhouse Yard can only be given brief attention in this review. The excavations consisted of two small but deep trenches in the upper reaches of the Walbrook: a far more peripheral part of the Roman settlement than was examined at Poultry. Here a stream channel had been infilled early in the Roman period, followed by various episodes of land reclamation and drainage associated with buildings and gardens represented by short lengths of timber walls and fences. Some of the finds may have been associated with votive offerings. This clearly presented report is a useful addition to our knowledge of this part of London, without introducing new data that might require us to rethink current interpretive models.