Over recent years the question of ancient hoards, in particular of precious metal, coins, plate or jewellery, has been the subject of numerous considerations (notably S. Gelichi and C. La Rocca (eds), Tesori. Forme di accumulazione della richezza nell'alto medioevo (secoli V–XI) (Rome, 2004)) in order to try to grasp the characteristics of a complex phenomenon that relates to multiple aspects of society in whatever period is under consideration: the economy, social organisation, the possible role of the images … The difficulties encountered by researchers when addressing these problems are illustrated by the ambivalence, indeed the ambiguity in many languages of the term ‘trésor/hoard’. Richard Hobbs has thus chosen, very judiciously, to take as his subject here ‘deposits of precious metal’, which defines the topic perfectly. On the other hand, one could question the descriptor ‘late Roman’ when applied to the period covered here, five centuries, from a.d. 200 to 700. There could be discussion over whether the third century should be included in Late Antiquity; others will challenge whether the sixth century still belongs to that same world. But from the first page H. effectively corrects his title by stating that it also covers the early Byzantine period, something I would feel is a better definition. It may certainly be felt that these are just questions of nomenclature, but they do have their importance for the topic of this study. All the same, the important thing is that H. wanted to study an extended period, as stated by the book's sub-title. One cannot but approve of his choice.

From the very beginning the reader is impressed by the amount of work that this volume represents: a very substantial bibliography (16 pages), and above all a major catalogue (‘database’) that gathers together in an Appendix (152–274, that is nearly half the volume) 2,045 deposits of precious metal grouped into 22 periods according to the principles on which H. bases his study — and we shall come back to these — to which are added six other groups which cover, if I have understood correctly, deposits whose dating has not apparently been determined with sufficient precision (a.d. 200–300, 300–400…). The deposits are placed in alphabetical order within each of these groupings, and an index enables one to find them with ease. Each entry is very rich in information, rigorously arranged according to a number of headings: H. presents these at the start of the Appendix (152) in a table which is indispensable reading before one tackles the catalogue. There is no point in discussing these at length here: the majority of them are sensible, though two are novel — the value of the metal expressed as an equivalent in weight of gold (‘Equivalent Gold Weight’), one of the fundamental tools for H.'s analysis, and a geographical classification by region, also an important element of the analysis, presented as fig. 3. The catalogue's approach is not simple, but it is perfectly usable by those who make use of it systematically and it indeed contains an extraordinary amount of information, though some of it does seem, it is true, somewhat superfluous: in particular, was it absolutely necessary within the frame of this work to present systematically the inscriptions on the objects (which formed part of the decoration)?

So this affords a substantial and very solid base (probably the most complete possible) for H.'s work: an examination of deposits of precious metal of Late Antiquity from within and without the frontiers of the Empire, though this last point for me poses a basic problem. Is it in fact legitimate to consider these two phenomena together as though, in relation to the question studied here, they obeyed the same rules? I would say here and now that this approach does not seem to me valid.

H. starts from the reasonable observation that these deposits, in particular hoards which contained silver plate, have rarely (if ever) been considered in the round, that is to say that the coins which were sometimes associated with them have as a general rule been left out of account, or necessarily employed as more or less precise indicators of date, or at best, when they were present in large quantity, treated separately as in the cases of the Kaiseraugst or Hoxne treasures. It is this that allows H. to underline from the very beginning (1) that (and this may come as a surprise to some) no researcher has ever dealt with these deposits in a satisfactory manner.

We shall pass over the pages dealing with the history of research on plate and coin (2–4), so summary as to be almost a caricature. By contrast, ch. 2 lays out clearly the principles on which the database was constituted and gives explicit justifications for the choices made by H.: the reasons for which, for example, bronze vessels were not considered, even though, one may imagine, they did have some definite value in the eyes of those who concealed them (so it might be asked why in the Vienne hoard, no. 1972 in the database, objects of iron and bronze were carefully deposited along with the silverware); though their Equivalent Gold Weight was, it is true, modest. It should be noted that metals in combination are going to complicate matters further: the inscriptions on the Paternus dish show that scrupulous account was kept of the different qualities of gold and of silver; the same was probably true for bronze, so often plated with silver.

But it is also in the same chapter that the basis for the chronological periodisation is put forward, that of the 22 coin periods borrowed, with some modifications for the latest ones, from that created for coins by Reece in 1972 and based on political events and monetary developments. This is a fundamental choice, one which would have merited more sustained justification: to what extent is it legitimate to apply to deposits consisting exclusively or for the most part of plate criteria established for the study of coin? Moreover, that is the reason that caused H. to create, over and above the 22 periods, the broader groupings in which to place the deposits which cannot, according to his scheme, be precisely dated. In fact there are matters for discussion here also, since many of these latter hoards can in fact be placed in a narrower date-range. They cannot be as precisely dated as a coin hoard, but that in turn raises another essential point; should a piece of silverware or a statuette be treated on a similar basis to coins? My response to this question is again in the negative. It is true that a stylistic dating must always be treated with caution, but not insignificant advances have been made in this area. On the whole I remain very hesitant about any simple equivalence between precious-metal objects and groups of coins and to their placing in periods that are sometimes quite narrow. Just to take a single example from among many possible ones: the silver reliquary from Ain Zirara in Algeria (catalogue no. 1872). In strict terms, there is no reason either stylistic or from the excavation to place it in the period a.d. 610–40. Stylistic analysis, the only criterion possible, clearly does not allow it to be dated to the reign of Heraclius.

The geographical criteria do not call for any particular remark; they make good sense. On the other hand, the final part of the chapter (17–25) is essential: it concerns the criteria for fixing the equivalences in terms of value of the different elements of the deposits by period. Proceeding from the weight and fineness of the object, H. thus establishes the equivalent in weight of gold (‘Equivalent Gold Weight’) of all these components, which allows him to make comparisons over the long term. This exercise is undoubtedly one open to debate, but well worth the trouble of undertaking. To guarantee its basis, H. surveys those texts that afford evidence on the intrinsic value of gold and of silver. They are well known and have been studied before, but this review, very clear and detailed, is most certainly very useful. The result is that H. adopts a ratio of gold to silver of 1:15, one which he regards as valid for the entire period under consideration.

The application of these principles does entail some complexity, but the method to be followed is explained without any confusion and one is happy to follow H. in his application of it throughout his study. I have myself insisted on the metal value of silver objects too often to be able to do otherwise. One must remain aware, however, that the texts show without the least ambiguity that the value in metal, which allows the establishment of an ‘Equivalent Gold Weight’, was not the only one the ancients took into account. Was the aesthetic quality of the objects or their value as works of art taken into account at the moment of choosing the pieces to be deposited? We do not know, but this possibility, even if theoretical, should cause us to nuance, even if only mentally, the preponderance given to the value in metal. So there are many elements that cannot be quantified: this is the problem of luxury. H. also and rightly declines to take into consideration ‘face value’, debatable enough for coins, but which moreover has hardly any meaning for the objects themselves.

The third and longest chapter (26–98) which constitutes the heart of the study reviews the deposits in chronological order and gives a detailed analysis for each period. A series of very clear maps shows the deposits by period and classified by their contents. One's only regret is that it is not that easy to distinguish between the symbols used to show the contents of the hoards, but this is a problem with all maps reduced for publication. And anyway the maps are very eloquent, showing particular situations at specific times: one just has for example to compare those for Periods 2 to 5 (a.d. 222–37, 238–59, 260–77, 275–95) (figs 5–8): the images are spectacular, because for Britain and Gaul between Periods 2 and 4 one passes from a near blank to a striking density before returning in Period 5 to a thinning-out for Gaul while the density remains high in Britain. At the same time one sees for Period 3 (a.d. 238–59) a sharp but short-lived upsurge along the Danube, especially on the lower Danube. And one notes also that the later one gets in time the more the maps covering the Mediterranean (and beyond) become empty: this is indeed very revealing in itself, but also poses a certain number of questions about the phenomena constituting hoarding and about the legitimacy of the division into periods that are sometimes very short (26 years for Period 20 – fig. 25; 16 for Period 21 – fig. 26). One does wonder what meaning the estimate of number of deposits per year for Period 17 for example (1) can have.

These reservations put forward (and in my eyes they are serious ones), the detail and the character of H.'s systematic analyses lead to some very telling conclusions. Some are reassuring, since they confirm observations already arrived at earlier: the sharp upsurge in deposits between a.d. 260 and 275 in Gaul does not come as a surprise, any more than does the particular place of Britain in the deposition of coin or plate. But this is really the first time that the patterns of distribution have been laid out over the long term. Moreover, the study does not stop there: the make-up of the deposits and their size, two crucial considerations, are also taken into account, revealing other developments over time in the geographical areas under consideration. Noteworthy, for example, are H.'s very interesting remarks (100 and table 72) concerning Africa: the small number of deposits needs to be set against their size and the ‘Equivalent Gold Weight’ of these finds. Therefore what is laid out here is a very complete picture of the phenomenon of deposits of precious metal. Ideas which hitherto had been simply impressionistic are here confirmed in a rigorous fashion. In addition H. revisits his conclusions in the fourth chapter (99–119) which he explicitly situates in the longue durée (‘Long range patterns with reference to the whole database’). In these pages the accent is put on regional aspects through time in order to characterise first the Western then the Eastern parts of the Empire, then the areas outside the frontiers. I have already stated above how unsatisfactory I find H.'s decision to treat these ‘non-Roman’ geographical areas on the same footing as the others. It has to be said that these pages, albeit illustrated by numerous histograms, are rather disappointing, contributing little that is new.

A last, fifth, chapter (120–34) constitutes in a way a more theoretical reflection on the problem of deposits of precious metals (‘Interpreting patterns of precious metal deposition’). Starting with a specific example, that of the Patching hoard, its entry in the database is examined in detail as a case-study of all that can be derived from that database (though by this point in the study the reader will have no doubt as to its excellent quality and its great worth); H. sets out to demonstrate everything that can be gained from the analysis of the points around which his study is articulated: the calculation in ‘Equivalent Gold Weight’, the placing within precise chronological sequences, the internal composition of the deposits. The last point is self-explanatory; on the chronological divisions I have expressed my reservations above — what is justified for coin is less so for objects; as for the first point it relates to a fundamental question for hoards of silver plate. These are indeed luxury objects, but as H. himself perfectly justifiably remarks, it is difficult to know what is represented in terms of wealth and of social standing by a hoard such as that from Mildenhall: scholars have probably often tended to overestimate them. A more objective form of calculation such as ‘Equivalent Gold Weight’ is without doubt a means of getting closer to the reality.

In order to forestall possible criticisms of his approach, H. returns to a topic needing debate: the question of the representativity of the finds, taking the example of Gaul, rich in hoards of silverware in the third century, largely lacking them in the fourth. He quite rightly rejects the idea that this is a consequence of recovery which fortuitously missed those of the end of antiquity. This is clearly unlikely, even if it is true that to an extent it depends on specific circumstances, in particular land-use and -management (on this topic let us remember, since H. raises the question of the ‘Portable Antiquities Scheme’ (126), that the position taken on this question in France, recently restated by the National Council on Archaeological Research, and other European countries is radically different. We do not share H.'s positive judgement on this scheme). But with equal justification H. has reservations over another interpretation of hoards (or rather of their absence), seductive though it may seem, that would have them reflect pretty faithfully the economic situation. The link often proposed in a rather mechanistic manner between precious-metal deposition (coins or objects) and ‘troubles’ likewise cannot be the one and only interpretation (129); H. takes the example of the sack of Rome by Alaric in a.d. 410 (and note the conclusions of the recent conference at Rome on this subject which were precisely that there is little evidence in the archaeology for this dramatic event). It should also be underlined that the troubles that led to the concealment of precious goods could be of very variable nature, doubtless often entirely unknown to us. It is, however, no less the case that this remains a major cause for hiding such deposits: some time ago Painter commented very revealingly on the behaviour, for another period to be sure, of Samuel Pepys faced with the threat of a Dutch invasion. H. also surveys other interpretations, such as changes in the fineness of the coinage, or particular regional traditions of hoarding, showing clearly that no one single explanation will do.

It is true that none of these proposals is novel, but rarely have they been so carefully and systematically laid out together, in a treatment that seeks to bring out all their implications and to balance one against the other. It would have been useful also to consider the very term ‘deposit’ which this book does not do sufficiently: the items registered in the database are far from being identical, so there is nothing in common, for instance, between the Kaiseraugst treasure and the Ain Zirara reliquary, and the presence of the latter is surprising since it is not a ‘deposit’ in the sense used here.

The amount of information that H.'s study presents was noted right at the beginning. It can sometimes be surprising for the manner in which it is presented, such as considering the worth of coins essentially as objects. Sometimes on the arid side, the exposition is of a remarkable clarity, even if one does not share all the trains of argument or the conclusions. If I have properly understood H.'s intention, it was not to resolve all the questions posed by the hoards. Many aspects are deliberately left to one side. But, when all is said and done, this study has one outstanding quality: it obliges the reader to weigh the author's arguments in detail. It is a work which, making available a mass of evidence rarely brought together, makes one think — and that does not happen that often. That is where H. wanted to guide the reader, as he says in a conclusion (134) of great modesty but one which cannot wholly conceal the justified ambition and the richness of this publication.