sources of supply. Similarly, a few fragmentary texts from Bogazköy record information about glass recipes, revealing the Hittites' interest in glass-making, although it has so far proved impossible to identify any piece of glass as an example of Hittite manufacture.
The first glass vessel found in Asia Minor dates no earlier than the late eighth century BC. It is the fine, colourless bowl (phiale mesomphalos) from tumulus P at Gordion, the capital of the Phrygian kingdom in central Anatolia. The vessel, decorated with a pattern of thirty-two radial petals, is the earliest surviving complete example of a glass vessel with cut decoration. In shape and decoration it mirrors exactly the metal bowls found in contemporary Assyrian and Phoenician contexts. It should, therefore, be regarded as an import, probably a gift to a local prince from the Assyrian royal court.
By the middle of the sixth century BC hard evidence is available for the existence of glass production in Asia Minor. It comes from a small workshop at Sardis attached to some Lydian domestic buildings that were probably destroyed during the capture of the city by the Achaemenid king, Cyrus the Great, between 547 and 542 BC. Sardis is, indeed, a fitting place to find the earliest, trace of local glass-working in Turkey, for it is there that we also have evidence for production towards the end of antiquity, in the early Byzantine period.
During the intervening centuries the glass industry must have flourished in Asia Minor just as it did elsewhere in the ancient world. Little trace, however, remains apart from the products themselves. A small number of kiln sites have been identified; in addition to Sardis, there is evidence for glass-making at Porsuk Höyük and Anemurium. Until archaeology produces a clearer picture of the physical remains of the ancient glass industry, one will have to depend largely on typological studies and analyses of distribution patterns to identify regional variations. So, it has been suggested that core-formed vessels may have been made in Turkey. Efforts have also been made to attribute certain types of Roman glass to centres in Asia Minor. The Erimtan Collection provides valuable additional information about the glass-ware that was commonly used there during the Roman period.
The majority of the pieces described in this catalogue are presumed to be from tombs. They well reflect the diversity of ancient burial goods. As well as the numerous plain bottles-both containers for liquid offerings to the dead and small vessels that were themselves taken gifts-there are some fine tablewares. The different types may represent the various social and economic classes to which the deceased belonged rather than the changing use of glass, although it is obvious that styles and tastes altered considerably with the passage of time. No information is available about the discovery of any of the vessels. Provenances, too, are unknown. It is, however, thought that a sizeable proportion of the collection comes from the area to the north and north-east of Ankara. The main centres from which the glass was collected are Çankiri (ancient Gangra-Germanicopolis), Sungurlu, Merzifon, Havza, Suluova, Amasya (ancient Amasea) and Niksar (ancient Neocaesarea). In addition, the odd vessel is attributed to Konya-Karaman, Adana and, further east, to Van, but no great reliance can be put on these claims.
There is, nevertheless, no reason to doubt that all were found somewhere in Turkey, and so some of the vessels can confidently be regarded as ancient imports. The late Hellenistic anphoriskos was probably imported. Either from the Syro-Palestinian region or from Cyprus. The cast ribbed bowls were also probably made elsewhere and than traded or otherwise conveyed to Asia Minor. The small mould-blown bottle is likely to be another import from the Syro-Palestinian coast. On the other hand, the square botles decorated with geometric designs on the bottom must be regarded as local, eastern products despite their close similarity to types found in the West. They coincide with other mould-blown bottles that bear inscriptions in Greek on the bottom. A number of these inscribed bottles are known in Turkish collections.
In addition to glass vessels, the Erimtan Collection contains a rich selection of miscellaneous glass objects. Spiral rods are found throughout the Roman world. The rods are usually between 20 and 30 centimetres long and terminate at one end in a flat disk, a spoon-like motif or a bird. Despite these variations, the rods all served the same purpose; namely, they were used for stirring up and dipping into cosmetics. Fragments found at Sardis have been securely dated to the early imperial period; one example came from a level below a second or third century AD stratum. The glass bracelets and rings, however, are more difficult to date since examples can range from the late antique to the Islamic period. It is to be regretted to that we do not know whether any of these objects were found in association with the glass vessels. It is, however, difficult to believe that the bracelets were recovered from a completely different set of tombs, and so they may well belong to the same time-scale as some of the later Roman vessels. The typology of Maud Spaer has been adopted for the bracelets, despite the obvious differences and variations that exist between the Palestinian examples and those found in Asia Minor.
Finally, there are a small number of post-Roman vessels in the Erimtan Collection. While the first piece is an example of a well-known Islamic type, found right across the Near and Middle East, the other two vessels are more difficult to place. The thick-walled, heavy glass bottle may have been used by a mediaeval alchemist.