210 TANIA PANAgOU
the other hand, even in the case of cities that did produce stamped amphoras,
these cities did not always stamp amphoras at a constant rate. The percentage
of stamped amphoras out of the total number of amphoras therefore varied
not only from city to city but also from one period to another in the same
city or from one workshop to another workshop in the same community
(Empereur 1982 ; Finkielsztejn 1995 : 282–3; Lund 1995 : 298; Lund 1999 : 188).^2
As a result, mere stamp counts do not offer us solid ground for the evaluation
of the extent of ancient trade. The study of shipwrecks illustrates the problem.
Although most of the cargos found in shipwrecks consist of amphoras, the
number of stamped amphoras represents an extremely small percentage of the
total number of amphoras found in these shipwrecks (Gibbins 2001 : 297–304).
This varying frequency of stamping applied to amphoras would cease to be
so confusing, if we had the answers to the still open question about the exact
meaning and purposes of the stamps (civic, fiscal purpose, certification of
capacity, organization in the place of production, etc.; see Garlan 2000 : 152–72;
Panagou 2010 : 431–58).
Another important obstacle to calculating the amount of trade conducted
by each city is presented by the varying capacities of the various amphora
types. For example, a Thasian stamped amphora may well correspond to a vol-
ume of 6.6 liters, while a Coan up to almost seven times more, that is, up to 44
liters.^3 One cannot therefore simply count the number of stamped amphoras
produced by one city and compare it with the number of stamped amphoras
from another city to get an idea about the comparative volume of trade con-
ducted by each city. Even if the number of stamped amphoras from one city
was the same as that from another city, we cannot conclude that they exported
a roughly similar amount because the capacities of the amphoras in one city
may have been much greater than those of the other city.
Furthermore, we are often uncertain about the attributions of stamps to
exporting cities. The most complicated case seems to be that of Cos. It has
long been taken for granted that every double handle originates from the
island of Cos. But as the study of stamped amphoras progresses, it has become
clear that Miletus, Panormos (near Miletus) and Colophon also produced
such types (Jöhrens 2009 : 213–19, 230; Mommsen, von Haugwitz, and Jöhrens
2010 : esp. 50). We may therefore need to reduce the number of double handles
attributed to Cos and in turn reassess our views about the volume of exports
from Cos. We should also remember that many stamps and amphora types
have not been identified yet (ca. 2,000 unidentified stamp types are recorded
in V. Grace’s files). On the other hand, it is doubtful that from the unattributed
stamps found so far any major production center remains unidentified.
Another factor that makes it difficult to interpret amphora stamps as evi-
dence for ancient trade is the secondary use of amphoras. Unbroken jars were