4It begins with a detailed discussion of systems used for the expression of dates, including the issue of when the New Year began (according to Roche in the spring), names of months, eras used, changes under the Roman Empire, etc. This leads to the first of the dozens of useful tables incorporated into the book, “Tableau I”, listing the equivalent dates in the Babylonian and Gregorian sequence of months. Like so many aspects of this book, the list is sure to become a regular source used by other epigraphists for checking dates, since it often takes longer to interpret the date in an inscription than to decipher it!

5The dated inscriptions are arranged in four chapters: I. Nabatène; II. Hégra; III. Périphéries and IV. Arabie. Once the reader becomes accustomed to this division, there is no problem, though in fact “IV. Arabie” might usefully have been entitled “Arabie de la période romaine”, since, of course, Ḥegra too is in Arabia and the existing title of the chapter is ambiguous. As already indicated, however, the book is provided with such a profusion of lists and indexes that it is always very easy to find the inscription one is looking for. Indeed, it is encyclopaedic, with invaluable lists and indexes which facilitate the finding of particular inscriptions from a number of different starting-points.

8Immediately before this point in the same inscription Roche has read lwḥ (for the older reading of lwt, a known preposition) in the phrase lwḥ dwtrʾ, “le panneau de Celui du Lieu”. This new reading is then derived from lwḥʾ, “panneau”, otherwise unattested in Nabataean, and in this case appearing, for no clear reason, in the construct state: we might have expected lwḥʾ dy dwtrʾ. At the end of the preceding line we have a restoration of wdnh qrbtʾ, “et voici l’offrande”. Roche’s 2014 article read wdnh mqrbtʾ and it is arguable that the shorter reading is to be preferred because it would fit the lacuna better. But there are two problems here. First, the statement that “voici l’offrande” is “une formule fréquente” is a little odd. Certainly the formulaic “This is…” is common, but I cannot find any trace of either Nabataean qrbtʾ or mqrbtʾ, “offering”, in the sources available to me (or in closely related Aramaic sources). Even if the reading is correct, the demonstrative could have been feminine dʾ (though there is some evidence of loss of the dnh/dʾ distinction). This is a not big issue: a hypothetical restoration is… hypothetical. But more explanation would be desirable.

9Although all the inscriptions have been reread and some new readings are recorded, a full discussion of the readings of the earlier editions is not provided. This is, of course, understandable and inevitable, since full re-editions with epigraphic commentary would have been a gigantic task. This does mean, however, that anyone who consults the Roche inventory may be disappointed that there is sometimes no specific discussion, even of dates. While I was reading the book, a colleague asked me about a particular dated Nabataean inscription, wanting to know whether the year-number in the existing editions was clear and correct, whether the widely accepted restoration of the era was likely and whether the style of writing fitted the supposed date. He was unaware of Roche’s book, but would probably have turned to it if he had known about it. The inscription in question is Roche 122 (R122) which is normally dated through its reference to 147 ce — šnt XX + XX + II lh[prkyʾ]. Roche does not discuss the interpretation of the numerals, though they are not entirely clear, or the restoration of the word hprkyʾ (which is given by Roche as hprkyhyʾ, presumably a misprint). The dating to 147 ce is probably correct, but it could have been discussed a little more. One might add that both the script charts and the chart of number signs (Plates 2-4) are rather old, being photocopies of publications of 1885 and 1898: the numeral chart in particular could have been updated.

10Such a complex book does, of course, have typographical and similar errors here and there. A regular typographical error is the printing of {ʿ} as {ʾ} and vice versa (e.g. p. 40, 78 ʾšryn, p. 307 ʿdr), evidently mistakes which have arisen in typesetting. Also there are some inconsistencies in references to the Bibliography. For example, “M&S 1970” (p. 47) is nowhere explained, having been omitted from the list of Abbreviations. “M&S” refers to “Milik and Starcky” (not a British chain store!). However, these errors are trivial.

11While not adding many new Nabataean inscriptions to the database, this book fulfils the equally important (and much more laborious) task of presenting a major section of the accumulated corpus. It not only presents that material, but gives a commentary on the inscriptions concerned (and relevant sections of the papyri), improving readings and casting much new light on individual inscriptions. But this is not all: Marie-Jeanne Roche has focussed on “Le temps en Nabatène” and provides a thorough and impressive analysis of all the data related to this theme. She not only presents conclusions, but also provides the world of Nabataean Studies with what will become one of its most useful tools. This is a very significant contribution to the study of Nabataean history.