‘‘Pottery can be represented as a sort of spy, or a symptom of a much more complex reality’’ (Pucci 1983: 106). Pottery is overwhelmingly the most abundant manufactured material surviving from antiquity. This gives it a peculiar importance in archaeological studies of the ancient economy, and one that is out of all proportion to its actual economic value in Roman society (Carandini 1983; K. Greene 1992; Peacock
1982) . With the help of improved characterization studies, the origins of many distinctive forms of pottery are now known and their distributions can be traced and their local occurrence assessed quantitatively (W. V. Harris 1993a). It is less certain what we can prove from distribution patterns alone, especially where it is suspected that the pottery has been traded ‘‘piggy-back’’ alongside archaeologically less-visible commodities.
The study of amphorae offers the greatest potential for understanding Roman trade, since the containers are representative of commerce in their contents (Amphores 1989; Peacock and Williams 1986). In many instances a particular commodity can be ascribed to a distinctive form of amphora (olive oil for the Dressel 20, wine for the Dressel 1, and so on). Studies of regional pottery supply patterns are also starting to demonstrate the local and long-range connections of a number of sites and how these fluctuated over time (Fulford 1989; P. Reynolds 1995). The quantities of amphorae known to have been shipped are impressive. Monte Testaccio, for instance, is conservatively estimated to represent c.60-80 million oil amphorae. Several large deposits of Italian wine amphorae dredged from the bed of the Rhone in the nineteenth century are reckoned to have totaled more than 100,000 vessels (Tchernia
1983) . Recent research has also emphasized a massive increase during the Roman period in production and trade in fish products, including the celebrated rotted fish sauces (Ben Lazreg et al. 1995; Curtis 1991a, b).
The overall volume of manufactures and their distribution patterns are extremely impressive. Above all, one is struck by evidence of large-scale activities, well above the subsistence needs of individuals or inward-looking communities. First-century ce kilns at La Graufesenque in southwest France could apparently accommodate up to 30,000 high-quality pots in a single firing and millions of these vessels were evidently distributed across the western empire during the operational period of the site (Peacock 1982: 114-28). A pottery workshop lease in Egypt specifies the annual production by the potter for the estate owner of 15,000 amphorae with a total capacity of c.100,000 liters of wine (Cockle 1981).
World History
