This erudite and densely argued study of the legal texts from the Roman world which deal with sales of wine is not an easy read, but it more than repays the effort. The subject, how in legal theory and practice the Romans and their subjects dealt with risk in large-scale purchases of wine, is a prime case-study of ‘transaction costs’ in the Roman Empire, of particular value because wine was a ubiquitous, relatively expensive and unstable good which has therefore left us substantial documentary and juristic evidence. Jakab is also a standard-bearer for the ‘real life’ approach to studying Roman law, and wine sales allow her to collate and compare the juristic sources with actual contracts and related documents from Roman and Byzantine Egypt, and also, looking back to the Republic, the specimen contract in Cato's De Agricultura. Her principal conclusion is that, despite the general Roman legal principle of periculum emptoris, both Roman legal writings and the contracts from Egypt developed similar clauses to limit, or at least clarify, the exposure of the purchaser to the particular risks, principally to do with quality and measurement, inherent in the practice of bulk-buying wine as must at the vintage for collection or delivery some months later, well after vinification.

J.'s detailed analysis of so many difficult texts, sometimes pressing them until they squeak, inevitably invites occasional disagreement. She does not directly address the broader question of whether the common contractual elements she finds reflect a trend to empire-wide legal norms or are inevitably similar responses to common practical problems. J. focuses on sale in advance (Lieferungskauf) because that is the best attested type in the legal writings and contracts, and sometimes slips into claiming it was the main type actually used. However, she notes that direct sales of ready wine may not have required contracts or caused juristic debate, and I suspect that most bulk dealing in wine to supply major cities and the army will have involved stocks of ready wine. Furthermore, fifty-two of the fifty-seven usable contracts from Egypt are of the fifth to seventh centuries, and only one before the third. Matched against total numbers of published texts per century, this type of contract looks to have been a predominantly Byzantine usage in Egypt. There is also the curious fact that forty-four of the contracts do not specify the price, which Bagnall has suggested was to mask usury while J. suggests that there may have been a separate receipt. To progress we need to put these contracts in their socio-economic context. First, by noting that these wine sales belong to a Byzantine boom in contracts of sale in advance of all kinds of goods, and also of labour contracts for vineyards (see P.Heid. V). Second, by looking at the positions of the contracting parties. P.Oxy. LXXVII 5123 (a.d. 555), just published, reveals a sale in advance of wine to Apion II, the great landowner at Oxyrhynchus, by one of his ‘registered tenants’. Were these sales in advance a way in which small producers maintained their cash-flow and some economic independence or a means for the rich of keeping or gaining control, and were they an alternative or an adjunct to labour contracts? J.'s study has opened up fascinating issues for further investigation and debate.