Introduction Three different types of transport amphorae are each characteristic of three different phases at Panayia Ematousa. The characteristic Cypriot amphorae with horizontal handles reaching above the rim of the amphora dominate the Archaic to Classical period (A1). The late Hellenistic to Early Roman period is dominated by the double barrel- handled amphorae (A24), and finally the Late Roman period, the 5 th to 7 th centuries AD are dominated by the Late Roman 1 amphorae (A6). This pattern suggests that although amphorae from many different places arrived at Panayia Ematousa, especially in the Hellenistic to Early Roman peri- ods, each period seems to have been dominated by one commercial interest. The Cypriot amphorae with horizontal handles and (LR1) amphorae were produced in Cyprus, and the fabric of a large group of Koan amphorae suggest a similar prove- nance (see A24). The small-scale survey in the area has revealed no kiln sites, 2 and the amphorae found at the site most likely reflect consumption rather than local production. Each period of occupation appears to have imported one major commodity. The import is regional from somewhere in Cyprus, and imports from outside the island were never numerous. The Cypriot amphorae with horizontal handles could have carried both wine and oil 3 and this is also the case with the Late Roman 1 amphorae, 4 but the Koan amphorae are linked to wine production. Viticulture could have been practiced locally as well as indicated by one of the current names for the area Panayia ton Ampelion, “our Lady of the vines. The laws of statistics are notoriously difficult to apply to ceramic studies and the methodology has been debated intensely. 5 The methodological problems are very discouraging, but how can you compare “many” to “a lot”? Some numbers are necessary in order to get a more precise view of the distribution of amphorae. 6 The question is what is the best way to quantify amphorae. To begin with they come in very different sizes. A Cypriot amphora with horizontal handles is more than twice as big as a LR1 amphora. An equal number of amphorae would produce very differ- ent amounts of individual sherds if we counted all identified fragments of these two types. Also the ribbed body fragments of the LR1 amphorae are easily recognisable while the body fragments of the Cypriot amphora with horizontal handles are diffi- cult to distinguish from local utility ceramics. Another issue is the relationship between size and amount of goods carried. Cypriot amphora with horizontal handles could carry a lot more than the equal number of LR1 amphorae. The logical thing is to count or even better weigh a certain type of diagnostic sherds only. At Panayia Ematousa we counted every single fragment of pottery, but we did not weigh them so in this case weight is not an option. Transport amphorae have only one base, but not all base-types are distinctive and well pre- served. LR1 amphorae have thin rounded bases and they are very rarely preserved. The rim is another possibility, but rims of the Cypriot amphora with horizontal handles are difficult to distinguish from utility ceramics. Left are the han- dles and here the problem is how easily they break. Handles of the double barrel handled amphorae such as Koan and Sub-Koan amphorae tend to 303 1 I thank Lone Wriedt Sørensen, Director of the excavations at Panayia Ematousa, for permitting and encouraging me to study and publish this material and providing me with the funding and John Lund, National Museum of Copenhagen, for his guidance and support. 2 Methenitis, PE I, 426. 3 Jacobsen 2002. 4 Empereur & Picon 1989, 242. 5 E.g. Orton 1993 6 Jacobsen 2004. Transport amphorae 1 Kristina Winther Jacobsen