Bishopstone in East Sussex has been well known within Anglo-Saxon archaeology for some time, as a result of the excavation in the 1960s and 1970s of a settlement and cemetery, dating from as early as the fifth and sixth century, on Rookery Hill (Bell Reference Bell1977). Gabor Thomas's fieldwork and researches over the past decade, published here with creditable promptitude, have shown it to be an exceptionally informative site for the whole of the Anglo-Saxon period. East of Rookery Hill, only half a kilometre across a dry valley, is St Andrew's Church, containing substantial late Anglo-Saxon fabric. Charters, and a curious mid-eleventh-century hagiographical text from Flanders, shed further insights. Now there is also excavated evidence of settlement features and deposits in an area north of that church: some two dozen structures, seventy-eight Anglo-Saxon pits and a length of ditch. The buildings include a range of forms of timber architecture: mostly post-in-trench but two with only earth-fast posts and two aisled buildings combining these forms, and an already famous cellared tower. The evidence of economic life retrieved and analysed, in the form of faunal and botanic remains and pottery in particular, is unusually strong. The ‘small finds’ from the site are not copious, but a hoard of iron artefacts apparently deliberately deposited for some ritual purpose rather than for merely practical storage in the cellar below the tower is of considerable interest.

Thomas describes an understandable concatenation of chance and design that led him to this project: coming to the area originally with quite different objectives in view, its potential within a sophisticated research framework for Anglo-Saxon archaeology was recognized. The question of whether a general Middle Anglo-Saxon settlement ‘shift’ can be posited, and a growing emphasis on the need to assess the particular circumstances and histories of individual locales or pays, are carefully discussed at the beginning and end of the report. Among the phenomena for which Bishopstone now adds significant new evidence are the shrinkage in extent of original church burial grounds of the early Middle Ages (Zadora-Rio Reference Zadora-Rio2003) and, quite differently, a detailed exposition of waste-disposal practices and sorting from the pits (although I confess to smiling at the earnestness of the declaration that ‘most of the pits analysed from Bishopstone displayed complex biographies marked by sharp transitions in deposit formation processes and depositionary pathways’). While the detail of the subsistence evidence is valuable, it contains few surprises. One such, however, is the evidence for the continuing cultivation of spelt, from a deposit in pit 2363. Attention should also be drawn to the identification of a possible feast deposit in another pit, and to the apparently ritual burial of a cat, a dog and a sheep in a layer sealing the cache of ironwork in the cellar under the tower.

There is little to quibble with in the contents of this report. In the historical discussion, the use of the accusative in the phrase adquisierat in hereditatem propriam in the charter of ad 825 ought to mean that Offa had added the property of Beddington to his own heritable patrimony, not that he had inherited it himself (pp 17–18). In places the ordering of the information is a little odd: for instance, the iron hoard is discussed initially as a special deposit (pp 102–6) but its contents are considered and illustrated by category through the general review of iron artefacts (pp 114–27). The presentation of the radiocarbon dates and a Bayesian model is inevitably dense, but crucial detail has been omitted in respect of the dating of Structure K. This catches attention because that is a well-constructed and well-preserved post-in-trench ‘hall’, apparently with a similar predecessor (Structure J), and sited prominently both at the top of a slope and at the very edge of the original burial zone. The discussion on page 202 quotes an estimate of its date of construction as cal ad 700–860 (95% probability) or cal ad 730–820 (68% probability), but there is no event building in Figure 8.2 as is stated, and a careful review of the text indicates that the date must be extrapolated from the radiocarbon date of a pig-bone in pit 2876, which cuts the wall-trench of this building. The estimated date-range quoted for the construction of Structure K is consistent with the posterior density estimate for that dated sample in Table 8.3, but this is not explained. Bayesian modelling processes are undoubtedly complex and relatively difficult to understand, but we must constantly battle against the impression that the whole method is baffling jiggery-pokery that should just be handed over to experts to do the work like technicians or mechanics and provide the answers for us.

In the end, Thomas suggests that, with the undoubtedly earlier than expected radiocarbon dates of both the settlement deposits and the cemetery, and Mark Gardiner's suggestion that a building of the seventh or eighth century may be identifiable on Rookery Hill (Gardiner Reference Gardiner2003), there is no longer a missing Middle Anglo-Saxon phase between the two settlements at Bishopstone, and he rethinks the possible nature of settlement re-disposition accordingly. He may well be correct, although it seems to me that the gap has been significantly reduced rather than conclusively bridged. The elephant in the room is the regrettably unpublished cemetery evidence from Rookery Hill. The concluding discussion also finally notes the vulnerability of a site like Bishopstone to Viking raids, and might further have drawn attention to the escapades of Wulfnoth cild, father of Earl Godwine, in this area in the later years of Æthelræd's reign. But, a few such detailed comments apart, this has been an exemplary project, brought to an exemplary fulfilment in this report, and Gabor Thomas is to be congratulated on what he has achieved.