Those who work on Romano-British epigraphy may count themselves fortunate. RIB I–III constitutes a published corpus of material that finds few comparisons in the other provinces of the Roman Empire, where the increasingly outdated volumes of CIL and IG still hold sway, supplemented to very varying degrees. RIB I published all material on stone up to 1955; RIB II performed the same task for instrumentum domesticum up to 1986; and now, in RIB III, the work is further updated with 550 inscriptions on stone ‘found or notified’ between 1 January 1955 and 31 December 2006. We await RIB IV, in which will appear curse tablets (on lead) and wooden writing-tablets. Although the editing of RIB III has been undertaken by Roger Tomlin, the work on these texts has been a continuous process of collection and publication since the 1950s, by Richard Wright, co-editor with Collingwood of RIB I, Mark Hassall and Roger Tomlin, who first joined and then succeeded Wright in his task; as Tomlin observes in his preface, ‘Their respective contributions cannot really be quantified’, and the triple credit on the cover seems wholly appropriate.

RIB III is an unashamed continuation of Volume I (although the numbering starts afresh from 3001 to signify Volume III; 2506–3000 remain unused). There are some innovations: diplomatic transcriptions now accompany the edited texts and translations; almost every entry includes a photograph as well as a drawing; and every stone includes a six-figure Ordnance Survey grid reference for its find-spot. This last is particularly welcome, although WGS (World Geodetic System) co-ordinates might have been more valuable to an international readership — and usable in e.g. Google Earth. Likewise, a concordance with AE, not just JRS and Britannia, might have been of immediate use to the wider epigraphic community. For those who do not have easy access to the complete set of OS maps for the UK, note the OS's free online ‘getamap’ service at: http://www.getamap.ordnancesurveyleisure.co.uk/; as well as the co-ordinate converter and Google Earth OS grid overlay available at http://www.nearby.org.uk/.

The question of mapping, however, also illustrates the more reactionary side to RIB III. In his review of RIB I (JRS 1966), Eric Birley bemoaned the lack of either a conspectus operis or a distribution map. The only gesture towards such a map anywhere in RIB remains the reproduction in RIB I add., 756 of Birley's own attempt to plot the geographical order of the inscriptions in RIB I (cross-referenced here on p. 1; cf. B. Jones and D. Mattingly, Atlas of Roman Britain (1990), fig. 5.10 for a distribution map for RIB I). In RIB I, the original editors delighted in noting Emil Huebner's misplacing (in CIL VII of 1873) of the Mendips in Derbyshire, and of Denbighshire in Scotland, but a chance to be more constructive has been missed.

The lack of any sort of conspectus remains a handicap: material from the line of Hadrian's Wall begins on p. 270, with a brief introduction noting a change in the organisation compared to RIB I; but one will find this only by turning the pages. Moreover, this and the very sparse introduction (covering working methods) is also a missed opportunity: the uniting of all the material published piecemeal in JRS and Britannia over the last 50 years (plus six previously unpublished: RIB 3019, 3030, 3109, 3178, 3370 and 3550*) offers a chance to take stock. The number of lapidary inscriptions from Britain appears to have increased by c. 25 per cent in 50 years (cf. J. Edmondson in A.E. Cooley (ed.), Becoming Roman, Writing Latin? (2002), 44–5 on both rates of discovery and the misleading impressions created by provincial-wide totals). I have not done so, but one would at least like to check the statement accompanying Mattingly and Jones' distribution map for RIB I (above) that post-1954 discoveries ‘would not affect the overall pattern greatly’. The rate of discovery, or at least of reporting, to judge by the concordance with JRS/Britannia, looks remarkably steady other than a marked slowdown 1993–1999 (a delayed consequence perhaps both of the recession at the start of the 1990s and the contemporary changes in policy concerning building development and archaeology?). It would be interesting to consider the pattern of new material against, for example, the concise observations of J.C. Mann based upon RIB I (‘Epigraphic consciousness’, JRS (1985), 204–6), since RIB I continues to provide the foundation for most subsequent discussion. Mann, for instance, noted the preponderance of military tombstones from the line of the Wall, suggesting a lack of take-up in the epigraphic habit in the region; total numbers are small, but the additonal nine included here (RIB 3290, 3364, 3365, 3366, 3398, 3400, 3445, 3472, 3473) weaken that assessment slightly: only three are clearly for soldiers, one is for a family, two for women, one perhaps for a child, the other two fragmentary (Mann noted 20 military against 15 clearly civilian, and 28 uncertain).

One must not, of course, complain about RIB being something it is not, and we have in our hands a wonderful research tool, promptly produced in the face of an increasingly unsupportive research environment, for which Tomlin in particular deserves our immense gratitude. But this volume is not as helpful as it might be. The question is perhaps best considered in the deliberately blunt terms of whether RIB III is much more than a very well-edited compilation in a single volume of all the material published in the last 50 years. Given the absence of conspectus, maps and indices, anyone wishing to use this material for the purposes of synthesis will have to work very hard; and as we shall see shortly, related concerns apply at the level of individual texts. The independent work in progress to generate an XML-based digital version of RIB is therefore a welcome development.

Some misgivings might also be raised from a methodological point of view, such as occasional inconsistency in presentation, and the deliberately tacit approach adopted in relation to previous scholarship. As an example of the former, both the recording of letter-heights and consistent indication of drawing scale are firmly eschewed in the Introduction. It is a little difficult to see why, when the emphasis is so unashamedly epigraphic (‘the prime concern of RIB III is the inscribed text’, 2). The suggestion that CSIR provides letter-heights is beside the point, quite apart from the fact that not every volume of CSIR in fact does so; neither does every stone appear in CSIR, nor is a concordance to CSIR included. Letter-heights are provided in some cases, and the inconsistency looks odder when letter-heights are invoked as the principal argument for the disaggregation of various of the monumental fragments from Southwark (nos 3016ff., especially 3019).

What I have called the ‘tacit approach’ to previous scholarship can best be illustrated through a couple of examples. RIB 3398 from Carvoran on Hadrian's Wall was previously published in JRS 1965 and CSIR 1.6. In the description of the relief that accompanies the text, RIB observes that ‘shortness of the tunic suggests a male figure although the person commemorated is female’ (this is Wright in JRS), ‘but the high neckband and outer garments are not typically masculine’ (this is Coulston and Phillips in CSIR). The only references are the listing of previous publications at the head of the lemma. This example is harmless, for sure, but does it count as good practice? Where this becomes irritating, or worse, is in the effort it requires of the interested reader, and in the difficulty it creates of assessing the interpretations presented. RIB 3025 is a small fragment (two letters) of a fine marble inscription. Without discussion, the RIB entry concludes ‘probably no later than c. 175 and funerary’. Working back through the original Britannia notice and the excavation report, it remains wholly unclear what the basis for the dating might be (not in any case a common feature in RIB it seems, where dates are often avoided, even when previous publications offered them — the previous example, 3398, is dated to the third century in CSIR, but no date is offered in RIB; notwithstanding the difficulty of dating epigraphic texts that seems to look a gift horse in the mouth). The proposal that the fragment is funerary appears to be a silent inference from the statement in Britannia that the fragment was found in secondary deposition in a pit not far from two burial monuments; but that report itself omitted the additional detail of the original excavation report that the fragment was found much closer to a building to which it could have been affixed, and the excavators reasonably wrote only of a monumental inscription. The problem with the tacit approach is that all statements are presented as being of the same status, yet some are derivative, some are inference, and some appear to be the editors' own judgement. Working out which is which is made unnecessarily difficult, and uncertainty will therefore remain as to what the basis is for any particular assertion.

None of this should be allowed seriously to detract from what is a hugely welcome addition to the materials available for the study of Roman Britain. The inclusion of translations already in RIB I was a significant step towards increased accessibility; the inclusion of photographs as well as drawings throughout this volume is a welcome further step (contrast the continued austerity of CIL). The production values as a whole are high, the cost is relatively low. If every province were so well served, the study of the Roman world would look very different.