1The appearance of glass in the 3^rd millennium B.C. is the culmination of a long evolution of glassy materials such as glazes, faience and Egyptian blue. Glass vessels were first produced in the 17^th c. B.C.1

2There is a turning point in the history of the craft: the development of the glass blowing technique during the 1^st c. B.C. This breakthrough now divides the history of glass between pre-Roman and Roman productions. The emergence of glass blowing modified many aspects of the production: the recipes started to be different, the facilities had to be adapted, new equipment was created and prices changed. Glass blowing started as a simple technological advance, but it is still affecting our lives, socially and economically.2

3Glass is a very special material; common and cheap nowadays, but rare andexpensive in remote antiquity. Glass objects may artistically and culturally resemble their counterparts made in other materials, but their technology remains different. Stone, pottery, or metal craft industries remained essentially the same at least until the 19^th century. However, glass working has changed almost totally since the Classical period.

4Most of the artistic tendencies in ancient glass working reflect technological changes. As scientific researches on ancient glass evolved considerably within the 19^th and 20^th centuries, various new theories were conceived. Most of themwere influenced, if not dictated, by the studies of other materials. Several of these theories suffered criticisms and were abandoned by glass researchers. Unfortunately, erroneous interpretations and terminology still haunt this field.

1. Primary and secondary productions: Glass was perhaps not made in Greece, but was certainly worked in Greece

5Initially, attention must be drawn to the major issue concerning the distinction between primary and secondary production.3

6Glass manufacture is based on a simple recipe. Sand and soda are fused in a one-phase process. The procedure involves crucibles (later tanks), high temperatures and a complicated savoir-faire.4 The result is what we call raw glass, which can be shipped where ever in the form of ingots or chunks (fig. 1).5 Scholars working on metallurgy are familiar with this idea of primary production and shipping of ingots. The same was most likely happening with glass, although our knowledge concerning trade in raw glass is limited.

Fig. 1: Glass ingot (the content of one crucible) and glass chunks for glassworking.

7This first stage of production is glass making. Pre-roman facilities for glass making are nearly non-existent. There have been suspicions for glass making activities in Egypt, but these apply only to the Bronze Age. There is no evidence for the later periods and definitely nothing concerning Greece. We are thus still in the dark, although we luckily have some cuneiform texts and Pliny the Elder to supply us with information. Cuneiform texts from King Assurbanipal’s (668-627 B.C.) library provide us with precious evidences concerning early glass making technologies dating of the 2^nd millennium. They inform us on glass making procedures, furnaces, recipes, and fuel.6 On his side, Pliny the Elder devotes chapters to the “Old Method” of glass making, describing essentially the same elements as the cuneiform documents.7

8Raw glass was either colored at the place of its manufacture or in local workshops of secondary production. There are however one or two cases of excavated sites where the colorizing process might have taken place, thus establishing an intermediate production between the primary and the secondary one. Again, there is nothing beyond the Bronze Age and nothing concerning directly Greece.8

9The second stage of production is glass working, i.e. the manufacture of glass items. Glass ingots were broken into chunks and heated until they got softenough to be used for the fabrication of vessels, jewelry and inlays. This was done in local workshops, of which none have survived. Several attempts have been made to identify facilities, but unfortunately, some annoying problems remain: either we have workshop installations and no workshop remains, or we have workshop remains and miscellaneous objects scattered from the main installations. To date, only one Greek pre-Hellenistic workshop can be identified with a relative certainty: Olympia in Classical Greece.9

10Glass is a versatile material. As the craft evolved with the emergence of glass blowing, tens of techniques for manufacturing glass objects have been perfected. Some of the early techniques ceased, but some other continued and are still in practice nowadays. Pre-Roman techniques are:10

• molding, like lost wax casting, mold pressing, rotary pressing, sagging;

• building around a core, like core-forming and rod-forming;

• manipulating with hand-held tools;

• a simple firing of glass chunks in the furnace.11

11The study of pre-Roman techniques presents some methodological problems. Their identification and comprehension still represent a work in progress. At the beginning of the studies on glass, it seemed easy, as some of the pre-Roman techniques were close to the ones used for other materials. Scholars interpreted these techniques based on their knowledge of ancient crafts and present day practices.

12Some of the early interpretations made on glass working were actually drawn from the field of sculpture. As carving has been practiced for millennia on wood, bone and stone, many scholars thought that glass could have been carved like marble. This hypothesis was elaborated in order to explain the early massive and simple glass forms. Experiments were carried out, but unsuccessfully, as scholars were not able to reproduce the ancient glass recipes. Today, we can of course carve the modern stress-free glasses. However, this was not possible with the early recipes. Due to internal stresses, the glass would break with the first blow. Early glasses could be carefully drilled or abraded, but not carved. Actually, this became possible at a small scale only after the Roman period.

13The idea that the ancient Greeks did not work with glass is an old one. It is partially because glass working did not survive in Greece as a traditional craft. Another reason could be that glass finds were not numerous enough when archaeological researches began. Mycenaean beads and other ornaments were known, but at the time of their discovery scholars believed that those were not made of proper glass, but of glass paste. Glass finds were scarce, as the early excavations in Greece focused on large scale sites of religious character like temples and sanctuaries and not cemeteries where glass finds are usually unearthed. The Athenian necropoleis did yield few finds, but Athens probably never acted as a glass working center in pre-Roman times. On the other hand, Roman glass was ubiquitous: thus the scarcity of glass finds of the pre-Roman period seemed even greater. All this created the impression that the material was not worked in Greece. Consequently, the finds of pre-Roman date were considered as imported.

2. ‘The claim for Phoenicia’: Colored glass in Greece is not Phoenician, but Greek

14Archaic and Classical core-formed vessels and small objects of colored glass were tentatively attributed to Phoenician centers. The creation of the “Phoenician myth” was mostly based on a chapter of Pliny the Elder who narrates a story of accidental sand and natron fusion into glass on the Phoenician shore.12 Gustavus Eisen, in his essay on glass dated of 1927, referred to the Phoenician myth as “The claim for Phoenicia”, while at the same time he was expressing some doubts: “As we know, until the present day no such early glass has been discovered in Phoenicia, certainly none which can be much older than the 9^th century B.C.”.13

15This matter was sensibly solved in 1940 by Poul Fossing who wrote: “To Froehner (La verrerie antique, 1879) must be awarded the honor of having proved the untenability of the earlier assumption that these glass objects are Phoenician. In place of that assumption, he advances the opinion that not only the ancient glass vessels of the New Empire, but also the later vases, more Greek in their character, are all Egyptian. But because of the very fact that the shape of these vases is Greek, Froehner considers that they must date from the time after Alexander the Great, when Greek culture dominated Egypt. That this opinion is wrong, and that there are types bearing traces of Greek influence in finds that are more than two centuries earlier than Alexander’s time, will be shown in the following…”.14

16Unfortunately, the term Phoenician was then already established as a strong conventional name. Sixty-seven years later, many scholars are still using it. In modern glass researches, we are now using the terms “core-formed” for the objects and “Mediterranean” for the production.15

17The core-forming technique was the most popular for producing small bottles. It remained in use for fifteen centuries until the development of glass blowing. The basis of the technique implies the formation of a core around a metal rod; then the core was covered with glass. Over the years, several theories arose on how glass could be applied to the core, most of them biased by our knowledge of more recent practices of glass working. The oldest theory was that the core was dipped in a crucible of hot glass in the furnace,16 but we now know that in pre-Roman times, it was impossible to keep the fire going continuously and at the temperature needed to have the glass completely melted. Furthermore, this theory would imply that every small-scale factory needed a large-scale furnace for their productions.

18Then, the theory was modified: the core was wrapped with trails of glass, while the rod was turned by the artisan.17 This new hypothesis tried to explain the formation of the vessel body, but not the decoration. However, the fabric of the examined Greek vessels does not show traces of trails. All the theories and modern archaeological experimentations were extremely useful because they helped us to understand not just the exact procedures in use but also the apparatus employed by the craftsmen in the early workshops.

19The typical ancient glass kilns were thought erroneously to have been similar to the Medieval or Renaissance ones: a domed furnace in two or three storeys. Below, was the firing-chamber. The second or middle floor was equipped with one or several openings through which the artisans could reach the crucibles inside. From there, they would draw the molten glass. The annealing of the glass pieces (slow cooling process to prevent stress in the glass) was done in the upper compartment. All this may present some similarities with the Roman glass blowing workshops, but does not even remotely resemble the pre-Roman workshops. It is another example of projecting later knowledge to earlier periods as there was probably no domed furnace with openings in the pre-Roman times.

3. The Macedonian, Rhodian, and local core-forming workshops: Greek colored glass is not always Rhodian

20Several decades of research combined with the careful examination of the core-formed objects found in archaeological excavations helped us to get a general idea of what a glass workshop looked like. In addition to the still surviving primitive workshops used in undeveloped countries, the Hellenistic workshop remains uncovered on the island of Rhodes helped us immensely to get a picture of the early production centers. The “Rhodian” discovery is old and very important. It was initially published by Gladys Weinberg, but its exhaustive publication is currently being prepared. Although the finds date to the Hellenistic period, it is a treasure of information for the application of the core-forming technique in pre-Roman Greece.

21The find consists of debris from a workshop dating of the end of the 3^rd and the beginning of the 2^nd c. B.C. The debris had been used as a fill for the floor of a Rhodian house. We thus do not know with certainty where the workshop was located although we could suggest that it could have been in the vicinity. This workshop was mainly producing beads, gaming counters and vessels. Approximately 11,000 beads were found in the fill. The workshop remains also include glass chunks, glass rods, ceramic and bone tools, and small quantities of bronze and lead.18 The workshop operated with raw glass that was probably bought in chunks. Those were heated until they became soft enough to manipulate or, most probably, were crushed into a coarse powder. A source of heat was obviously needed. This could have been either a strong open fire or a semi-closed construction like a built brazier. In the same find, ceramic trays covered with melted colored glass are of extreme interest. On the surface of one tray is stuck one more minutetray with a second color of glass. The trays are so shallow that they could not have been some sort of crucible: they are color trays, something like pallets holding colored glass for the decoration of finished objects.

22Already in the 1960s Frederic Schuler examined the possibility that core-formed glass was made by applying crushed glass to the core.19 It is today supported by many scholars and practiced by Dudley Giberson, a designer of modern glass kilns and a glass artist.20 This hypothesis also finds some endorsement with the aspect of the finished objects. Crushed glass requires heating only for a short period of time and not at a very high temperature. The hot air moves easily around each minute fragment and melts the whole of its surface. Thus, the finished objects are obtained using minimal energy, labor and time. However, one disadvantage of this technique is that air is frequently trapped among the fragments, forming thousands of small bubbles within the glass fabric. These bubbles render the glass translucent and not transparent and can become thousands of starting points of corrosion, resulting from the powdery deteriorated glass. However, we should definitely pay attention to them and consider their presence valuable as they could be proof of the use of crushed glass (the bubbles could be also witness to the low temperature of the melt).

23For core-formed glasses, the erroneous term “glass paste” has been frequently used. That was a rather inferior preparation of glass powder used in the 19^th century when the major excavations started. For the excavators, the core-formed glasses appeared similar to the innovation of their time; thus was created the most widespread modern misunderstanding about ancient glass still in use today.

24The use of crushed glass led Giberson to a reconstruction of the core-forming kiln. He based his reconstitution on a Bronze Age kiln from Mesopotamian Nuzi, and also on the traditional bead-making kiln shapes from Africa. He proposed a domed construction with a top opening and a small opening near the base which helps air flow. Using such a kiln, he recreated the stages of core-forming using crushed raw glass and revived the extinct technique of core forming (fig. 2).21

Fig. 2: The core-forming furnace and technique, as proposed by Giberson.

25To date, two Greek core-forming productions have been identified for the Archaic and Classical periods. The Rhodian production was studied mainly by Donald Harden. It comprises most of the known shapes of the 6^th and 5^th c. BC and fewer of the 4^th c. The finds, mainly from the early Italian excavations in Rhodes, have been discussed by various scholars, but a thorough study dedicated to the Rhodian material does not exist yet (fig. 3).22

Fig. 3: Camiros, Rhodes, core-formed oenochoe, 5^th c. B.C.

26The Macedonian production is currently being studied by the author. An outline of the production cannot be given yet, but the material is very promising. Most of it comes from well-dated unlooted graves of the 6^th to the 4^th c. B.C. (fig. 4). Hands of different artisans can be discerned. This will certainly provide us with more information concerning the chaîne opératoire of the workshops.23

Fig. 4: Pydna, Macedonia, core-formed vessels, 5^th c. B.C.

27Initial comparisons between the Macedonian core-formed finds and the Rhodian ones show that, to a considerable extent, they belong to different productions, probably both local. This was somehow predictable, taking into consideration ancient politics and the geographical distance.

28Additionally, everywhere in Greece excavations reveal finds, which are difficult to attribute either to the Rhodian or the Macedonian production. Every excavator who tried to identify an exact parallel for his or her finds through the published typologies faced this problem. Although today the two above-mentioned workshops seem to be two of the most important ones, several smaller local workshops are perhaps still to be identified.

29Nevertheless, the establishment of glass workshops might not be like the pottery one, with a workshop operating almost in every village. For artisans familiar with pyrotechnology, core-forming glass would not be that difficult to practice, provided they had acess to suppliers of raw glass. Yet, above all, the biggest reason would be the need for manufacturing these objects. Religious, ceremonial and funerary purposes, as well as fashion and storage needs might have provided the motivations for the introduction of a glass workshop in various areas of Greece.

4. The Persian myth: Colorless glass in Greece is not Persian

30In contrast with the colored glasses, the colorless glass vessels, jewelry and inlays are rare finds. They constitute a very special group of luxury products throughout the pre-Roman period. “Colorless” basically means without added colorants, which means that only the natural greenish color of the substance remains. We also suspect that the intended addition of antimony in the batch aimed to either the decolorizing or the refinement of the glass, or both. Transparency was desirable in an attempt to imitate rock crystal, which was then considered the most precious material on Earth.

31The term hyalinon (made of glass) is encountered in a famous passage of Aristophanes’ Acharnians (74) which goes as follow: ξενιζόμενοι δέ πρὸς βίαν έπίνομεν ἐξ ὑαλίνων ἐκπωμάτων καί χρυσίδων ἂκρατον οἶνον ἡδύν. Greek ambassadors sent to the Persian court were actually forced to drink wine from glass vessels. This passage has been extensively used as an argument for the existence of colorless glass in the 5^th c. B.C. However, the archaeological finds do not support such an assumption as almost all the known vessels coming from sealed contexts are dating of the 4^th c. B.C. Additionally, research of the literary sources reveals that the term hyalinon was used in Classical antiquity to designate all transparent substances. These include glass, glaze, crystal, gems, lakes and amber. Bright planets, eye tunics and even urine were also described as glass-like substances.24

32Here rises another methodological problem associated with the use of the literary sources. It is not safe to base our interpretations on pre-Roman glass making techniques and terminology only on Roman sources: Pausanias (2.27.3), for example, describes a painting made by Pausias on the Epidaurian Tholos. A woman, the personification of Drunkenness was depicted drinking from a glass phiale so clear that one could see her face through the glass: γέγραπται δὲ ἐνταῦθα καὶ Μέθη, Παυσίου καὶ τοῦτο ἔργον, ἐξ' ὑαλίνης φιάλης πίνουσα· ἲδοις δέ κἂν έν τῇ γραφῇ φιάλην τε ὑάλου καὶ δι’ αὐτῆς γυναικὸς πρόσωπον. This fragment is thought to be a proof of the use of clear glass in Classical Greece. Yet, regardless of the terminology problem, we can consider Pausanias very accurate. In his time, hyalos really meant glass, as the material existed, was easily recognizable, and widely used. This does not mean however that the painting he saw in Epidaurus represented glass. At the time the painting was completed, before the middle of the 4^th c. B.C., a clear vessel called hyalinon was most probably a rock-crystal vessel.25

33It is even more dangerous to base our interpretations on texts of later scholiasts. They are usually unfamiliar with the technological details of the early periods they are commenting on, and they tend to explain the technological vocabulary according to the one in existence in their own technological conjuncture. Generally speaking, we tend to project very easily onto early antiquity knowledge gained from the study of Roman practices or, still worse, of present-day traditional techniques.

34The passage of Aristophanes was thought to be the best proof that glass cups existed in the Persian court in the 5^th c. B.C. In 1940, Paul Fossing based his study of a glass phiale from the Artemision in Ephesus on it. The vessel was found in a stratum dated before 330 B.C., but the excavator Hogarth was obviously so impressed by it that he published it as being “of fine fabric, worthy of the 5^th c.”26 In absence of any glass parallels, Fossing compared the decoration, but not the shape, with Persian metal vessels. So, the myth of the Achaemenid glass imports was created. It was enhanced a few years later when fragments of several vessels were found in the Persepolis treasury.27 They were also predating 330 B.C. Yet, several scholars concluded that they were dated to the 5^th c. B.C. Some based their arguments, and still do so, on the appearance of Achaemenid vessels in the hands of the tribute bearers on the Apadana staircase. This way, step by step, and on insecure ground, the myth of the Persian glass was established. One after the other, catalogue introductions and general books repeated and strengthened the myth. Needless to say, no exact parallels to any surviving glasses appear on the Persian reliefs. Non-specialists never had the chance to read what the excavators and specialists wrote. Erich Schmidt who published the Persepolis finds concluded that “the place of origin [of the glass phiale] must be sought in one of the western lowland provinces, for we have no clues for the manufacture of glass vessels on the Iranian plateau during this period”.

35And later, Shinji Fukai, in his treatise on Persian glass, thought that “the place where these vessels were actually made cannot be determined, but I suspect that it was the Mesopotamian region, not the Iranian highlands”.28

36We must look in the western satrapies for the beginning of the 4^th c. B.C. production of colorless glass. Several bowls had been found in Asia Minor at the beginning of the previous century, but these were mostly out of context. Later, other finds coming from controlled excavations were revealed.29 So far, the earliest securely dated colorless glass vessels of the classical period are those from the Mausoleum at Halikarnassos. They can be considered Achaemenid in shape and decoration. We can be sure that they were grave goods in the burial of Maussollos who died in 353/2 B.C.30

5. The Macedonian, Carian-Rhodian and regional workshops for colorless glass: Ancient glass cannot be carved like stone or cast like metals

37Up to now, the numerous colorless-glass finds that have been uncovered in Greece have been attributed to two production centers: the Macedonian and the Rhodian ones.

38The Rhodian finds of the Classical period are mainly phialae and calyx cups. Some have been attributed by Pavlos Triantaphyllidis to a Rhodian-Carian workshop, as Rhodes was under Carian rule at the time of the production. Some pieces are of special interest as they are from types not found outside Rhodes. They are thought to belong to an early local production influenced by Achaemenid workshops (fig. 5).31

Fig. 5: Rhodes, colorless glass vessels, 4^th c. B.C.

39The Macedonian finds appear in a variety of shapes that are not only restricted to vessels. There is a clear preference for drinking cups, but not for phialae (fig. 6). Other finds include jewelry, like rings and glass gems, and game counters. Hundreds of inlays made for wooden couches were also found32.

Fig. 6: Macedonia, colorless glass vessels, 4^th c. B.C.

40Isolated glass finds, non attributable to the Rhodian or the Macedonian productions, are also known from Abdera in Thrace,33 Lousoi in Arcadia34 and Lake Trichonitis in Western Greece.35 It is most likely that more workshops will be identified in the future, although not as many as we expect for the core-formed glasses.

41The process of manufacture of these objects still has in metalworking to be explained. The lost-wax technique is a very old one, still practiced today. One needs a wax model encased in clay or plaster. The wax is melted away and the void is filled with molten metal. It was thought this could have been done also with molten glass. Research revealed that it was impossible for the early glass-workers to reach temperatures high enough to pour their glass as a thin fluid. Only material with a low viscosity can fill all the recesses in the mold, as it is the case with metals. The best that a 1^st millennium B.C. glass-worker could have done would have been to work with a viscous glass sharing textural similarities with thick honey and this could not be poured from a crucible. They actually did not use crucibles and worked with chunks of already made glass which they heated until those became soft enough to manipulate.

42Yet it is obvious that most glass objects were molded. Schuler concluded that the molds were filled with crushed glass and then slowly heated to the point of melting.36 If the mold was heated long enough, then the resulting object would have been perfectly molded and also free of bubbles. So, when we refer to “cast glass”, we mean not poured, but molded glass. This type of glass “casting”/molding was probably used in Archaic and early Classical antiquity. However, following the experiments of Rosemarie Lierke, it is proposed that in the Late Classical period, a new technique evolved: the “glass pottery” also known as “rotary pressing” (fig. 7). This new technique involves the use of a mold secured on a turning wheel. The viscous glass was pressed in the mold with a plunger and as the wheel was turning, the hot glass was evenly distributed in the mold. This “glass pottery theory”, is still controversial, but several Macedonian vessels provide evidence for the use of the pressing method, either in its rotary or in its stationary version.37

Fig. 7: The rotary pressing technique, as proposed by Lierke.

43Both the “casting” and the “pressing” methods require the creation of molds. Glass technicians suggest that for the manufacture of vessels, a plaster mold would have been preferable because it would create a smoother glass surface and would have been easier to destroy in order to get the finished object. Open molds, like those for inlays, could also be made of clay or wood. We will probably never find the remains of wooden or plaster molds in Greece, but, at least, we are extremely lucky to have the clay molds that were uncovered at Olympia.

44In fact, the only surviving workshop remains for colorless glass manufacture in Classical Greece were found at Olympia. It is an extremely important find, but even if it was securely identified, would be still difficult to interpret. Glass working wasters were found within the complex of buildings identified by the excavators with the workshop of Pheidias. The stratum containing the debris was initially dated later than the Pheidias’ time, in the late 5^th c. and later, in the 4^th c. B.C.38 The workshop remains are unique and thus extremely valuable. The finds show how glass was worked, but unfortunately, they cannot be associated with any installations. Some nearby kilns were certainly used for bronze casting, but no glass kiln remains have so far been brought to light. The glass finds and their molds are stylistically dated to the first half of the 4^th c. B.C. in comparison to the later Macedonian finds.39

45There are several kinds of remains at Olympia (fig. 8):

• large scale inlays, which belonged to an anthemion-type decoration, palmette leaves and other elements. These are easy to identify, as they are comparable to the small-scale Macedonian inlays. Only a few inlays survived, but on the other hand, there are numerous clay molds;

• enigmatic glass folds, part of a garment. These are very problematic because they are inconsistent with the technological advances made prior to the 4^th c. B.C.;

• bone and iron tools, surplus material of ivory, horn, bronze, lead strips, obsidian, plaster, stone, and various colorants;

• glass chunks which were probably the raw glass in stock to be formed into objects.40 Among the latter group of objects, a single fragment with a sharp corner is valuable. This fragment shows that the chunks were crushed from a rectangular block. The crucible in which it was made was probably something like a small rectangular tank.41

Fig. 8: Olympia, workshop remains: glass, clay molds, bone, lead, bronze.

6. Multi and collaborating workshops

46At Olympia, the presence of so many miscellaneous objects means that glass was only one of many crafts worked there. We must imagine a large multi-workshop complex with skilled artisans working on different materials. Such workshops could have been set up or even built at the occasion of an important project, like the erection of a chryselephantine statue. Carpenters would have been needed for the inner construction, ivory benders for the revetments, sculptors and painters for the decoration, metalworkers for the gold, silver and bronze, potters for the clay molds, and ironsmiths for tools making. Such a project would have implied a large-scale operation directed by a master mind like Pheidias or another great artist.

47Similar or slightly different circumstances might have occurred for the Macedonian funerary couches. Those are composed of wood and leather cords fastened with bronze nails. They are decorated with ivory, glass, stucco, silver and gold foil, and are additionally painted (fig. 9). As we have manufacture evidence for almost a hundred of these couches, we cannot consider their production as a single project. It is more realistic to think about a complex containing several specialized and collaborating workshops. In a kingdom like Macedonia, we can imagine the existence of a “palace workshop”, although the products were certainly not only used in the palace or for palatial purposes. We can also suggest that a competent artist would have been coordinating the worshops. A mastermind is obviously behind the couches’ production and who could afford to have an in-house great artist but the king?

Fig. 9: Macedonia, 4^th c. BC, couch glass inlays.

48The production of colorless glass vessels probably differs. One of the widespread misconceptions is that the glass vessels imitate metal ones. It has often been suggested that moulds for glass were taken from metal vessels. Research proves this wrong; there are no exact parallels, while there is an evident discrepancy in dates. The typology in both evolves in the same artistic environment but independently, probably because they had different prototypes. Additionally, they had different clients, sizes of production and technical restrictions. Some interactions could be detected in situations where craftsmen could have shared use of pyrotechnology, but this was most likely the case with potters.

49Pottery and glass productions share common roots. Glass technology is considered an offspring of the very early ceramic technology. However in the 4^th c. B.C., this was reversed.42 The technique of rotary pressing was developed and, one century later, the plaster moulds were substituted for reusable ceramic ones in order to produce ceramic moldmade bowls. It is possible that the “in between experiments” were carried out with faience, but it is essentially the advanced glass technology which inspired this great Hellenistic ceramic production.43 The use of clay to produce glass molds is attested within the Olympia finds, but there is more than that. The firing of glasses in the molds was probably done in pottery kilns,44 ceramic wheels were obviously used for turning the molds in the rotary pressing, and a variety of ceramic tools and vessels facilitated the working of the glass craftsman.

50Finally, an unexpected testimony on the possible interactions occurring between workshops is offered to us by the plainest finds. Gaming counters start being produced in the 4^th c. B.C. and continued their life through the Hellenistic period until the Roman times, when they were also used as glass gems. Their irregular shape shows that they were not molded. They were actually made by a simple firing of the leftovers of other glass working productions.45 The fact that the gaming sets were composed, in the 4^th c. B.C., of two or three different colors obviously means that the artisans had access to leftovers of both the core-forming and the molding techniques. This could be seen as a proof of the actual spatial proximity of workshops and the coexistence of the two major pre-Roman productions.