In a foreword, C. Young — who undertook a small excavation at Ickham in 1974 — states that ‘virtually all the hypotheses put forward by me after the excavation have now been disproved’ (xi). The full report confirms that Ickham had four Roman watermills, built entirely from timber, and evidence for metalworking abounds. In 1981 Young had raised the site's profile by proposing official or military status, and the production of corn and metalwork there for Saxon Shore forts (Collectanea Historica: Essays in Memory of Stuart Rigold). The late Roman buckles which prompted Young's suggestion are discussed very carefully (157–9), but no clear verdict is reached, partly because views about their official status have changed, but also because many were probably manufactured at Ickham for use elsewhere. The coexistence of mills and extensive metalworking, combined with the discovery of a large hammerhead (not illustrated at the time), had led R.J. Spain to suggest in 1984 that it might have come from a water-powered trip-hammer. This radical proposal has been widely repeated by historians of technology, and is perpetuated in Wikipedia's (2011) entries on trip-hammers and stamp-mills. Although convincing evidence now exists for cam-operated industrial machinery from other Roman sites, the connection between mills and metalworking at Ickham has not been pursued in the final report, and the hammerhead has been reinterpreted as a tool for dressing millstones (353–4).

In order to understand the chronological and economic context of Ickham's watermills it is essential to assess the quality of other evidence. This report concerns rescue work carried out during destruction by gravel extraction between 1972 and 1975, including a short professional excavation in 1974. The late Jim Bradshaw and local amateurs managed to salvage a rich range of artefacts, but the high standard of presentation achieved by the editors should not let us forget that it has been assembled from individual specialist reports written at varying dates. All readers should begin by reading the short introduction (1–13) and note that ‘There was no overall strategy for recording, nor indeed any clear vision of the post-excavation methodology to be employed or the analytical destiny of the site records’ (12). P. Bennett's stimulating concluding essay (321–45) on the wider significance of Ickham should be read next, for it consistently urges caution about using poor-quality data.

Wood survived well in Ickham's waterlogged conditions, and several timber-lined water channels and associated wooden structures were recorded. Their association with fragmentary millstones led Spain to identify post-built structures as watermills, and to bring them to wider attention in 1984 in History of Technology and Archaeologia Cantiana. Since the water channels and related features provided assemblages of datable artefacts, the mills and their associated structural features provide a more coherent set of data than the site's other excavated features and material. The first of Ickham's four watermills is now dated to the third century, rather than the second as originally proposed by Spain, although one early water channel could possibly have served an unlocated fifth mill (322–3). The long delay in publishing Ickham allowed further consideration of the structural remains, and comparisons with more recently excavated mills. Spain's convincing ground-plans and reconstruction drawings illustrate a detailed account of the working of these undershot mills, together with their water channels and tail races (42–62). A discussion of their wider significance (with A. Wilson, 62–8) notes the low quality of their woodworking in comparison with normal Romano-British carpentry, and its similarity to that of early medieval watermills in Ireland and Denmark: ‘the Irish evidence reflects a greater European tradition in mill construction which, on the basis of the Ickham evidence, almost certainly has its origins in the late Roman period’ (67).

Only half of Ickham's fascinating range of small finds (which make up a substantial report of just over 160 pages) were stratified, another third were roughly assignable to an area of the site, nearly 18 per cent have neither context nor location (303), and ‘weeding’ had taken place to reduce their bulk (339). A summary of the distribution and chronology of small finds (303–10) is rather undermined by the variable size and arbitrary numbering of site areas (10–12 and fig. 4). A traditional typological structure for the small finds report would have been preferable to thematic classification according to function, especially as much of the metalwork probably arrived on the site as scrap for recycling (e.g. buckles (159); pewter tableware (241); and even coins (341–2)).

Molinologists and other historians of technology would have welcomed a shorter monograph that presented the mills alongside the millstones and other related artefacts, which are widely dispersed through the report (the large hammerhead formerly associated with water-powered machinery is relegated to an inconspicuous appendix). The exemplary discussion of the millstones includes an instructive table illustrating reuse of fragments as quernstones, whetstones or bearings (288), but the bearing blocks for the waterwheels are (unaccountably) not illustrated alongside their discussion here (see Archaeologia Cantiana 100 (1984), 126, fig. 12).

Bennett's useful concluding essay demonstrates that despite the problematic way in which its finds were recovered, Ickham remains a fascinating site. It underlines the capacity of the economy, even in later Roman times, to support dynamic industrial and food-processing activities in a wealthy area of Kent (which contains urban and military centres, as well as villas and other rural settlements, all linked by roads and rivers). Resources devoted to high-quality excavation and publication of similar sites in Kent and elsewhere might clarify the circumstances in which it was advantageous to invest in the construction and maintenance of mechanical mills rather than to rely upon traditional hand-powered querns.