The Romano-Celtic temple complex in the Bois Du Flavier, near Mouzon in the French Ardennes, must be one of the most frequently cited but worst published temple sites in the field of Roman provincial religion. Caumont's monograph focuses on the site's most celebrated aspect — the nearly 1,000 miniature votive weapons found there. Information about the site and these finds had hitherto only been available in a series of short regional archaeology reports, published in the back of Gallia, as well as a few slightly longer articles in obscure local periodicals (for a summary, see P. Kiernan, Miniature Votive Offerings in the North-West Provinces of the Roman Empire (2009), 47–63). In addition to the published accounts of the site, C. has also enjoyed access to unpublished preliminary reports and the full collection of finds. C.'s use of broader comparanda and literature is less impressive. His bibliography, oddly placed on pages 389–400, seldom goes beyond the year 2000 and has nothing after 2004. One surprising lacuna is the lack of reference to comparable miniature weapons from the not-so-distant sanctuary of Mars at Blicquy (Hainaut, Belgium). Nonetheless, C. has produced the new comprehensive work both for the site and the votive offerings found there on which he concentrates.

C.'s introduction places the site within a broader regional archaeological setting in both the late La Tène and Roman periods (21–46). Its location on a ridge separating the territories of the Treveri and Remi clearly places it among other temples constructed in liminal situations. C. then turns (46–66) to the patchy history of the excavations conducted between 1966 and 1992, which explored an estimated 60 per cent of the site. The main structures uncovered include three rectangular cellae surrounded by an elliptical pathway, a square cella flanked by circular pathways on either side, and a temenos wall that once enclosed these buildings. Neither inscriptions nor figural representations have been recovered to allow the deities worshipped at Mouzon to be identified, though C. sees Hercules or Mars as plausible recipients of the weapon dedications. The coins, not dealt with by C., range from the first century b.c. to the fourth century a.d. The evolution of the site has been divided into five main phases. The earliest material belongs to the final phases of the Iron Age, no earlier than 70 b.c. The miniature and real weapons seem to have been deposited in the second and third phases, between 20/10 b.c. and a.d. 50/100. At this point, the site underwent a major rebuild that sealed most of them beneath a floor in the westernmost of the three rectangular cellae. C. illustrates the excavated areas and phases in a series of useful plans (figs 10–19). He deserves considerable credit for reconstructing a coherent picture from what is clearly a confusing corpus of excavation reports of varying detail and quality.

The introduction is followed by an intensely detailed study (69–388) of the most exciting finds from the site: the 916 miniature iron weapons and 47 fragments of real weapons and various pieces of horse-gear. The miniatures can be divided into 432 swords, 428 shields, 28 axe-heads and 83 projectile weapons, mostly spear- and arrowheads. C. also documents a small number of fragments of life-sized armour and shield fragments, most interestingly some scraps of chainmail and small pieces of lorica segmentata. Each find is carefully illustrated by C.'s superb line drawings and described in detail in an extensive catalogue. This is supplemented by an attempt to create typological groups, comparisons to life-sized military equipment, a short metallurgical report (by P. Merlusso), and descriptions of the method of manufacture. As an artefact study, the reviewer offers only one criticism: no inventory numbers are provided and it is nowhere clearly stated where the finds or unpublished site reports are now housed. Though minor, this detail is critical for future generations.

The final two chapters, ‘Problématiques’ (405–25) and ‘Conclusions’ (425–46), are certainly the most interesting parts of the book. C. believes that both the miniature and real military equipment at Mouzon was deposited by individual auxiliary soldiers of the Roman army in numerous single votive acts. The fragments of undeniably Roman armour recorded by C. seem to support this idea, but it is unclear if they are from the same phase and context as the miniatures. C. interprets these scraps as pars pro toto offerings, possibly left by veteran soldiers. C. sees the deposition of weapons in the temple as analogous to the use of real weapons found in Augustan and early first-century a.d. burials, such as those at Wederath. This interpretation differs substantially from that of the reviewer and others: that the collection was a substitute for a communal offering of war booty of the sort that characterised votive deposition in Iron Age sanctuaries, a theory supported by the similarity in manufacture among miniatures, their concentration in a single area, and by the fact that many were intentionally bent prior to deposition. In either case, and as C. rightly observes, these objects clearly belong to a transitional phase in the very earliest period of Roman provincial religion and are an important addition to our understanding of the development of provincial society.