Archaeological and Anthropological Sciences

, Volume 8, Issue 1, pp 51–63

Millet, the late comer: on the tracks of Panicum miliaceum in prehistoric Greece


    • Aristotle University of Thessaloniki
Original Paper

DOI: 10.1007/s12520-013-0152-5

Cite this article as:
Valamoti, S.M. Archaeol Anthropol Sci (2016) 8: 51. doi:10.1007/s12520-013-0152-5


Archaebotanical evidence for Panicum miliaceum is reviewed for prehistoric Greece including published and unpublished recent finds, providing a basis for exploring the context of the appearance of millet in Greece, the timing of its introduction and cultivation, and its significance in terms of contacts, movement of people, and cultural identity as expressed through culinary practice and food consumption. To this end, the archaeobotanical record is examined together with human isotopic, archaeozoological, and artefactual evidence. Millet is introduced to the northern part of Greece sometime during the end of the 3rd millennium bc and established as a widely used crop during the Late Bronze Age. Isotopic evidence suggests that millet consumption during the Late Bronze Age was not widespread but confined to certain regions, settlements, or individuals. Millet is suggested to reach Greece from the north after its spread westwards from China through Central Asia and the steppes of Eurasia. The timing of the introduction of millet and the horse in northern Greece coincide; the possibility therefore that they are both introduced through contacts with horse breeding cultures cultivating millet in the north and/or northeast is raised. Intensified contact networks during the Bronze Age, linking prehistoric northern Greece to central Europe and the Pontic Steppes, would have opened the way to the introduction of millet, overland via river valleys leading to the Danube, or via maritime routes, linking the Black Sea to the north Aegean. Alternatively, millet could have been introduced by millet-consuming populations, moving southwards from the Eurasian steppes.


Panicum miliaceumHorse domesticationBronze Age GreeceCrop introductions


Neolithic agricultural communities appear in Greece early in the 7th millennium bc inhabiting settlements in the form of tells or flat extended settlements in a densely wooded landscape. Within the sites themselves, structures and artifacts provide ample evidence for cooking in indoor and outdoor hearths, storage, pottery production, stone tool making, weaving, jewelry and figurine making. (e.g. Demoule and Perlès 1993; Kotsakis 1999; Halstead 1999; Perlès 2001). The Bronze Age begins sometime during the second half of the 4th millennium bc ending around 1100 bc and is characterized by changes in settlement pattern, increased complexity, regional differentiation, and changes in the form and production of various artifacts, including the production of copper alloys, arsenic, and tin bronzes (Andreou 2010). During these 6,000 years of the Neolithic and the Bronze Age, a wide range of plant species appears to have been used by the inhabitants of Greece, as is indicated by archaeobotanical research systematically carried out during the last three decades. A variety of cereals, pulses, fruits, and nuts, found in pure, dense concentrations, either harvested from cultivated fields or the wild, were probably consumed as food during the Neolithic and the Bronze Age (Valamoti 2003, 2009; Megaloudi 2006). Cereal species of Neolithic Greece are einkorn, emmer, new glume wheat type, free threshing wheat, and barley, with the first three dominating the assemblages. These wheats need to be dehusked to remove the glumes from the grain, a necessary step in the preparation of these cereals for food. These cereal species of the Neolithic continue to be used during the Bronze Age. In the north of Greece, there is evidence that during the Early Bronze Age einkorn wheat was ground, probably after boiling, stored, and consumed throughout the year, suggesting an early form of bulgur or trachanas and that likewise, barley was ground and stored in a similar form (Valamoti 2002, 2011a; Valamoti et al. 2008). Ground cereals have also been identified at Late Bronze Age Akrotiri on the island of Santorini (Sarpaki 2001).

Thus, for more than 5,000 years, the glume wheats are the main cereals consumed by Neolithic and Bronze Age societies of Greece, corresponding to their “traditional staff of life.” The Bronze Age, however, witnesses the influx of new crop species, cereals, pulses and medicinal, psychoactive, and/or oil plants. These crops do not seem to arrive simultaneously in the form of a “package,” rather, they are introduced at different periods during the course of the Bronze Age and they seem to underline intensive contacts with the north and the east during this period (Valamoti 2007a). Opium poppy is probably introduced from central Europe where it is a common component of the Neolithic LBK archaeobotanical assemblages while Lallemantia sp., with a central Asian/Transcaucasian distribution, is probably introduced from the east (Jones and Valamoti 2005).

Among the cereal species that make their appearances during the Bronze Age are spelt wheat and broomcorn millet (Panicum miliaceum). In the sections that follow, the archaeobotanical record of broomcorn millet from prehistoric Greece will be examined in detail providing a basis for a discussion of the context of its introduction, cultivation, and use in the area.

Broomcorn millet in prehistoric Greece: old evidence and new finds

Archaeobotanical reports for a total of 85 sites provide a relatively good body of data, despite differences in sampling intensity and sample numbers retrieved from each site and a more systematic exploration in the north (Megaloudi 2006; Valamoti 2009). Millet finds from Greek Neolithic and Bronze Age sites are summarized in Table 1 and Fig. 1 in an attempt to evaluate their significance for each period and in order to understand the processes of the beginnings of the cultivation of this crop by prehistoric communities of Greece. A general observation concerning finds of broomcorn millet in Greece is that it is rare in the 5th millennium bc and becomes established as a crop during the Late Bronze Age (Hansen 1988; Kroll 1991; Valamoti and Jones 2003; Valamoti 2007, 2009; Zohary and Hopf 2000; Stika and Heiss 2013). Megaloudi (2006, Tables 5.3, 5.4, and 5.5, p. 36) reaches a similar conclusion although the relevant tables in the publication show a regular occurrence of millet during the Early and Middle Neolithic which does not seem to be the case. The earliest mention of broomcorn millet in the archaeobotanical record from Greece dates to the Early Neolithic (7th millennium bc) and has been found at Argissa in Thessaly. It is, however, a single seed, identified as cf. P. miliaceum (Hopf 1962a). P. miliaceum (cf.) is also reported by Kroll (1981, 1983) for the Middle Neolithic levels of Otzaki Magoula, also in Thessaly, without an indication of the number of seeds. Sporadic occurrences of broomcorn millet are reported from a few northern Greek sites, dated to the 5th millennium bc such as those found at Limenaria (no number provided) and Agios Ioannis Loukas (two seeds) on Thassos island in the north Aegean (Megaloudi 2006), a single seed from Dikili Tash (Valamoti 2004), and eight seeds from Mandalo1 (Valamoti and Jones 2003), all dated to the Late/Final Neolithic (4500–3500 bc approximately). All these Late/Final Neolithic sites have overlying Bronze Age occupation phases. These occasional finds of millet as single seeds that occur in certain site reports for Neolithic layers, may be intrusive, given their extremely limited presence among hundreds of seeds of cereals like the glume wheats and barley and more importantly as nearly all these sites have Bronze Age levels overlying the Neolithic ones. This picture of an absence of broomcorn millet as a crop in the Neolithic is confirmed by a recent addition of a large body of archaeobotanical data from Neolithic northern Greece over the last decade, from sites such as Apsalos (Valamoti 2006), Mavropigi (Valamoti 2011b), Kremasti Koiladas (Karathanou and Valamoti 2011), Kleitos (Stylianakou and Valamoti, in preparation), and Promachon/Topolnitsa (Valamoti 2007b). It would be safe therefore to conclude that broomcorn millet is not a crop during the Neolithic in Greece or a crop contaminant, or a field weed as has been previously suggested for this region (Hunt et al 2008) and for the Neolithic of central Europe (Kreuz et al. 2005; Kohler Schneider and Caneppele 2009). Therefore, the presence of broomcorn millet since the Early Neolithic in Greece (contra M. Jones 2004; see also Hunt et al 2008) seems very unlikely. The only rich concentration of broomcorn millet contained in a miniature pot is from Final Neolithic Olynthus (Mylonas 1929) but the context of its identification is uncertain.
Table 1

Presence of Panicum miliaceum seeds in prehistoric Greece


Total number of seeds


Number of samples with millet





1 cf



6500–5800 bc

Early Neolithic

Hopf (1962a)

Otzaki Magoula



5800–5300 bc

Middle Neolithic

Kroll (1981 1981, 1983)





4500–3500 bc

Final Neolithic

Valamoti and Jones (2003)

Dikili Tash




4500–3500 bc

Final Neolithic

Valamoti (2004)

Agios Ioannis Loukas




4500–3500 bc

Final Neolithic

Megaloudi (2006)






Seed concentration



4500–3500 bc

Final Neolithic

Mylonas (1929)

Skala Sotiros


367 rich concentration


2200–2000 bc

Early Bronze Age

Valamoti (in preparation)





2100–1900 bc

Early Bronze Age

Valamoti et al. (in preparation)







Renfrew (2003)

Agios Mamas




Middle Bronze Age

Becker and Kroll (2008)

Agios Mamas




Late Bronze Age

Becker and Kroll (2008)

Toumba Thessalonikis


Rich concentration


Late Bronze Age

Valamoti et al. (in preparation)



Rich concentration


1350 bc

Late Bronze Age

Jones et al. (1986)



559,112 rich concentration


Late Bronze Age

Kroll (1983)



Rich concentration


Post-1500 bc

Late Bronze Age

Valamoti et al. (in preparation)



251 rich concentration


1500–1200 bc

Late Bronze Age

Valamoti (2010)





Late Bronze Age

Kroll (1982)

Single seed occurrences identified as Panicum/Setaria are not included. More than 50 seeds in a single sample are reported as “rich concentration”
Fig. 1

Map of Greece with Neolithic and Bronze Age sites where P. miliaceum has been found. 1 Limenaria, 2 Agios Ioannis Loukas, 3 Skala Sotiros, 4 Dikili Tash, 5 Sitagroi, 6 Agios Mamas/Olynthus, 7 Toumba Thessalonikis, 8 Assiros, 9 Kastanas, 10 Archondiko, 11 Mandalo, 12 Angelochori, 13 Argissa, 14 Otzaki Magoula, and 15 Tiryns
Fig. 2

P. miliaceum seeds from Skala Sotiros (Thassos island), 2200–2000 bc

Equally scarce are finds of broomcorn millet seeds among Early Bronze Age contexts (approximately 3500–2000 bc). A single seed is reported from Sitagroi, a multi-period site in northern Greece (Renfrew 2003) and eight seeds from Kastanas. Finds of millet, dated to the end of the 3rd millennium bc and the beginning of the second, have recently been identified at the sites of Skala Sotiros, on the island of Thassos (Valamoti, in preparation), and Archondiko, a site located on the western part of the region of Macedonia, northern Greece (Valamoti et al. 2008, Valamoti et al., in preparation), reported here for the first time.2 The Early Bronze Age finds from Archondiko consist of 40 seeds. They have been found within different floor contexts, dated on the basis or pottery and stratigraphic sequence to the end of the 3rd millennium bc and the beginning of the 2nd millennium bc (2200–1900 bc; Papadopoulou et al. 2010). Skala Soriros, on the island of Thassos, has yielded numerous finds of millet in several samples (367 seeds in eight samples, the richest sample containing 122 broomcorn millet grains; Fig. 2), and the context has been C14 dated both from charcoal and seed finds at 2200–2030 bc.3 On present evidence, the finds from Skala Soriros are the richest ones belonging to the end of the 3rd millennium bc. Limited published evidence is available for the Middle Bronze Age (approximately 1900–1600): three millet seeds are encountered at Agios Mamas, near Olynthus (Becker and Kroll 2008). Thus the earliest so far finds of a rich concentration of seeds comes from Skala Soriros on the island of Thassos, dated to the end of the 3rd millennium or the early part of the 2nd millennium bc and in light of the seeds from Archondiko, of a similar date, this find probably does not stand on its own. It is, however, absent from Agios Athanasios (Moniaki 2009), an Early Bronze Age site inhabited between 2400 and 2100 bc (Mavroidi 2013) indicating that millet was perhaps not adopted by all communities in the end of the 3rd millennium bc.
Fig. 3

P. miliaceum from Archondiko (loose seeds broken off from a conglomeration. Late Bronze Age

The presence of broomcorn millet certainly takes off during the Late Bronze Age (approximately 1500–1200 bc) where it is encountered in dense, pure concentrations at several sites, in particular in the region of central Macedonia which is well investigated archaeobotanically. These sites are Archondiko, Kastanas, Assiros, Agios Mamas, and Angelochori. The finds from Archondiko originate from a context dated to the Late Bronze Age, as the underlying context has a C14 date of around 1500–1400 bc They consist of a conglomeration of millet seeds fused together and loose seeds most likely broken off the charred millet mass (Fig. 3). At Kastanas, thousands of seeds have been found in a pure, dense concentration dated to this period (Kroll 1983). At Assiros, clear evidence exists that millet was a stored crop together with glume wheats and free-threshing cereals in the large storerooms destroyed by fire around 1300 bc, preserving by charring the crops stored within (Jones et al. 1986). At Angelochori, the millet seeds are found in mixed deposits originating from hearths and pits, containing kitchen refuse as well as in a pure, rich concentration, represented by a mass of seeds fused together (Fig. 4) (Valamoti 2010). Unpublished data from Late Bronze Age Toumba Thessalonikis also indicate the presence of millet in a large proportion of the samples, their numbers ranging from a few seeds to 200 (Mangafa, unpublished data). Thus, millet occurs as a common find, occurring either regularly but in low numbers or in rich concentrations in samples belonging to the Late Bronze Age (Figs. 1 and 5).
Fig. 4

P. miliaceum from Angelochori (conglomeration of fused seeds). Late Bronze Age
Fig. 5

a Early (3500–1900 bc) and Middle Bronze Age (1900–1600 bc) finds of broomcorn millet from Greece. b Late Bronze Age (1600–1200 bc) finds of broomcorn millet from Greece
Fig. 6

Representation of a pyravnos pot from Archondiko (Late Bronze Age), after Deliopoulos (2007)

There appears to be a north-south gradient, with millet being more common and prominent in the north rather than in the south of Greece where there are no records from the Neolithic or the Bronze Age except for ten millet seeds identified from the Late Bronze Age Tiryns. Bronze Age sites with rich archaeobotanical finds from the south lacking millet include Lerna in the Peloponnese (Hopf 1961, 1962b) and Knossos and Myrtos on Crete (Jones 1984; Renfrew 1972, respectively). This picture is further confirmed by the preliminary examination of recently retrieved data from Middle Bronze Age Agia Paraskevi near Lamia (Gkotsinas et al. 2012, based on 45 samples) and further to the south from Late Bronze Age Mitrou in Atalandi (Karathanou et al., in preparation, based on 300 samples). Likewise, the preliminary examination of 170 samples from Agios Vassileios south of Sparta has not provided any indications for a millet concentration (Valamoti and Karathanou, in preparation). This geographic distribution of millet in prehistoric Greece is only partly a reflection of sampling and recovery. Admittedly, archaeobotanical research is much more intensive and systematic in the northern part of the country; therefore, the presence of crops is better documented for this area, introducing a bias for comparisons between the north and the south. However, several sites from the south have also been investigated archaeobotanically and broomcorn millet, if representing an important crop, would have been more prominent in the archaeobotanical record of the region. Broomcorn millet is known for its wide range of adaptation to different climates, soils and altitudes, having low water and nutrient requirements and is cultivated in temperate, mainly, as well as in subtropical regions, its short growing cycle allowing for flexibility in sowing time (e.g., Arber 1934; Langer and Hill 1981; Ahlgren 1949; Doggett 1989). Although some climatic differences of temperature and precipitation exist between different areas of Greece, with a more Mediterranean climate prevailing in the south and continental influences in the north, the latter is by no means better suited to millet cultivation compared with the south. This north-south gradient therefore in millet’s distribution in prehistoric Greece is unlikely to have resulted from climatic differences between different parts of the country. A more likely explanation would be a slightly later introduction of the crop from the north to the south.

The context of millet consumption in prehistoric Greece

Millet could have been used as food or fodder. The contexts of millet finds are not indicative of its use and at present no food remains like processed cereals, consisting of, or including, millet, are reported from Greece (cf. Valamoti 2011a). Millet as food could be used for gruels or ground into flour for various food preparations like those encountered today in India where dehulled millet is cooked as poridge or for the preparation of flat, unleavened bread or mixed with butter-milk and allowed to ferment and subsequently used for making porridge (e.g., Pushpamma 1989; Malleshi 1989). Millet was the traditional cereal for preparing mamalinga in Rumania, a thick porridge-like preparation (like italian polenta) but its cooking in this form nowadays is confined in some households and millet has been largely replaced by maize (Cârciumaru et al. 2005). Millet as food could have also been used for the production of a fermented drink involving lactic acid fermentation which improves organoleptic quality by fermentation, digestibility and nutritive value compared with the unfermented counterpart like the boza or Bragă prepared in various parts of the Ottoman Empire and Iran and still prepared in countries such as Bulgaria, Rumania, Ukraine, Turkey, etc. (e.g., Cârciumaru et al. 2005; Starodub 2008; Hancioğlu Ö and Karapinar 1997; Arici and Daglioglu 2002; Petrov et al. 1978).4 It is legitimate to wonder whether the introduction of millet to Bronze Age cuisines of northern Greece was accompanied by new cooking techniques and vessels used for the preparation of millet-based recipes. Wardle (1993) points out the introduction of a very common late bronze age cooking ware, known as “pyraunos,” encountered only in northern Greece at sites like Assiros, Kastanas, and recently Archondiko and other northern Greek sites (Deliopoulos 2007; Psaraki and Mavroeidi 2012 for a recent review), but not in the south, where the Mycenaean palaces flourished. This cooking pot is unknown in southern Greece but occurs from northern Greece to as far north as Hungary (Wardle 1993). Pyraunos is a coarse ware pot consisting of a ceramic container fixed to a cylindrical stand bearing a large opening to allow the introduction of burning fuel underneath, and openings across its body to allow air circulation and facilitate fuel burning (Fig. 6). Any possible association between the pot and the crop is at present only speculative and needs to be further explored geographically and temporally across regions where millet is introduced earlier than northern Greece, for example Central Europe, Hungary, Rumania, and the Eurasian steppes. The recent finds of broomcorn millet at Skala Sotiros and Archondiko dated to the end of the 3rd millennium bc appear to predate the introduction of pyraunos in the region and indicate that the two do not reach northern Greece as a “cooking kit” comprising both the pot and the ingredient.

Millet becomes a well established crop in the north during the course of the 2nd millennium bc as it is encountered in pure, dense concentrations, mainly stored, in houses (e.g., Kastanas) or central store-rooms (e.g., Assiros) at several sites. This observation based on the archaeobotanical record is further corroborated by isotopic analyses of human skeletal remains from northern Greece. A study of five sites in northern Greece spanning the Neolithic through to the Iron Age, suggests that during the Neolithic and Early/Middle Bronze Age, people based their diet on C3 plants (Triantafyllou 2001, 2013). The same is also observed for several Late Bronze Age sites. During the Late Bronze Age and in particular during the Iron Age, certain populations demonstrate an intake of C4 foodstuffs that could correspond either to the direct consumption of C4 plants or of animals fed on C4 plants (Triantafyllou 2001; 2013). The limited presence of millet in the archaeobotanical record of southern Greece discussed above is further supported by isotopic evidence from Middle Bronze Age Peloponnese, from human skeletal remains from cemeteries at Lerna, Aspis and Asine. During the Middle Helladic period (2100–1700 bc) in southern Greece, a C3 terrestrial-type diet prevailed and no C4 plants were consumed by humans (Triantaphyllou et al. 2008a, b; Invargson et al. 2009). For the Mycenaean Palaces, on the other hand, the few seeds from Tiryns indicate that millet was not unknown in the region, while isotopic analyses from Late Bronze Age human skeletons from the Peloponnese and central mainland Greece have shown that at Agia Triada and Almyri in the Peloponnese, some of the individuals were partly basing their diet on C4 plants (Petroutsa and Manolis 2010). Thus, unlike the scarce archaeobotanical record, the isotopic studies offer a more nuanced picture of millet intake in the Late Bronze Age south of Greece with certain individuals representing perhaps a few millet consumers settled in the south through mating or other networks. The isotopic analyses on human skeletal remains indicate that even in the Late Bronze Age, when millet occurs regularly among archaeobotanical assemblages, in the north at least, it does not constitute a dietary element of all sites or of all individuals living within a community.

The archaeobotanical and isotopic evidence combined suggest that millet, after its introduction sometime towards the 3rd millennium bc in the north of Greece, it became a systematically cultivated and consumed cereal crop, its consumption being more evident in northern Greece in the Late Bronze Age compared with the south, in particular as regards the archaeobotanical record. Despite the prominance of the crop in the archaeobotanical record of the north, the isotopic evidence indicates both for the north and the south of Late Bronze Age Greece that millet was not a main element of diets of the region but most likely a foodstuff characterising and differentiating certain individuals or communities.

In light of the isotopic indications discussed above, the status of the crop and of the people consuming it may have varied as well as the intake of millet between members of the same community as well as between different settlements. It is uncertain whether in the early years of its introduction broomcorn millet constituted a regular human food, an occasional famine food or animal fodder. Nor is it known whether its consumption during the Bronze Age was generalized or practiced by certain individuals according to status or origin. This raises very interesting questions regarding the status of this new-comer crop and the people or village communities cultivating and consuming it (for feeding to their animals), their origins and position in Late Bronze Age societies of south-eastern Europe. Millet consumption may have constituted a cultural signifier indicating origin from millet growing areas in the north or north east or access to contact networks with these regions. Brides from the north bringing seeds of millet, a crop of their ancestral homes (cf. Hastorf 1998), travelers trading goods or whole communities moving southwards could have brought millet initially to northern Greece and then further to the south. Differences in the status of the plant between northern Greece and the south are also a possibility, given the absence of this crop in the south prior to Mycenean times and its selective consumption in the Late Bronze Age. Reflections on status differences or regional preferences related to millet consumption are available from later periods and although not directly applicable to the Bronze Age past of Greece they provide some insight as to how different cereals can be used to underline status differences or different origin. As can be gleaned by ancient Greek written sources like Herodote, Demosthenes, Xenophon, and Sophocles millet was looked down upon by Athenians but cultivated in Sparta, Thrace and along the coasts of the Black Sea as well as in Cilicia in the East; in Roman times Pliny mentions that millet was the preferred cereal of the people inhabiting the regions of the Black Sea (Ammouretti 1986). Further isotopic analyses on Bronze Age human skeletal remains from Greece may shed light on such issues.

Along the millet “route”: the introduction of millet to Bronze Age Greece

Millet in its journey to the northern coast of the Aegean had to travel great distances from northern China through the Eurasian steppes before reaching the warm landscapes of northern Greece. Although the precise area of broomcorn millet domestication is a subject of controversy and different areas of norteasten China have been suggested (Hunt et al. 2011 for a review of the literature), increasing archaeobotanical and isotopic research has shown that millet, an initially gathered plant (on the basis of broomcorn millet phytoliths from Cishan in the Yellow River valley dated to the 8th millennium bc), was subsequently domesticated and became a major cereal in northeast China cultivated at Dadiwan, Xinglonggou and Yuezhuang in eastern Inner Mongolia and northeast China, in the 6th millenium bc (Liu et al. 2004, 2009; Yang et al. 2012; Hunt et al. 2011; Zhao 2005a, b; Weber and Fuller 2006; Lee et al. 2007). Millet cultivation is regarded generally as established in the Yellow River basin by 5500 bc (the Beixin, Cishan, Peiligang, and Dadiwan cultures), although no clear evidence for transitions from gathering to cultivation have been identified in this region (Lu 2002 quoted in Weber and Fuller 2006). The archaeobotanical evidence is further supported by isotopic studies of skeletal remains from northern China, according to which millet contributed significantly to the Neolithic diets of people in that area (e.g., Barton et al. 2009; Lanehart et al. 2011; Liu et al. 2012 for a recent review of the literature). Isotopic evidence from human skeletal remains from Chifeng in eastern Inner Mongolia suggests a substantial intake of C4 plants by the Early Neolithic (5800–5300) increasing during the course of the Neolithic (Liu et al. 2012). Therefore, millet formed a staple food plant in the Neolithic while by the Bronze Age the feeding of domestic animals on C4 plants, i.e., millet, suggest an agriculture heavily dependent on millet surplus production for providing both fodder and human food (Liu et al. 2012). Thus, both archaeobotanical and skeletal isotopic analyses demonstrate broomcorn millet consumption during the Neolithic in China, in particular in the drier central regions of the north part of the country.

Between northern China and southeastern Europe lie the Eurasian steppes and millet in all likelihood spread westwards, travelling across space and through time, being handed down from community to community until it reached the mountains and narrow river plains of northern Greece. Based on early finds of millet in Europe, unless they represent independent domestication episodes (Jones 2004), millet has been argued to have spread from the East to the west through cultural contacts between Central Asia and Europe, existing already from the 5th millennium bc. (Jones 2004; Jones et al. 2011; Hunt et al. 2011). According to recent genetic analysis of modern broomcorn millet landraces, two millet groups, a western and an eastern one have been identified and millet could have spread from northern China to Central Asia and from there could have followed two different trajectories, one northwards to Mongolia and Siberia and another westwards to Central Asia and the Eurasian Steppes (Hunt et al. 2011). Such a movement is consistent with archaeobotanical finds from Eastern Europe and the Balkan Peninsula. Further to the north and northeast of Greece millet appears in earlier archaeological contexts dating to the 5th millennium bc in eastern Germany, Slovakia (Baden culture), Poland (TRB culture), Ukraine (Tripolye culture), Rumania (Vinca culture), and Moldova (Nesbitt and Summers 1988; Zohary and Hopf 2000; Hunt et al 2008). In Slovakia, millet was already grown from the Baden period onward (Hajnalová 1986, 1989), and there are still earlier records of its cultivation from Poland (Wasylikowa et al. 1991), Hungary and Romania (Gyulai 2007; Monah 2007). Recent finds from two Austrian sites of the Jevišovice culture (3200–2800) of the Final Neolithic in the area indicate an early cultivation of broomcorn millet as a crop (Kohler Schneider and Caneppele 2009). Further to the south, in Bulgaria, the evidence as regards millet points to a situation similar to that of Greece: millet is encountered as a crop at Late Bronze Age sites dated to the period between 2200–1200 bc such as Kamenska Chuka, Nebet Tepe, Adata, Okrazna Bolnitza, and Goljama Detelina (Popova 2009, 2010). Early Neolithic finds of millet in Bulgaria seem to be intrusive (Marinova and Valamoti 2013): single finds of millet have been found at two early Neolithic sites, Drenkovo and Kapitan Dimitrievo, while millet is found at Late Chalcolithic contexts at the sites of Durankulak (Marinova 2006) and Provadia (Marinova 2008). Neither the south of Greece nor northeast Turkey have yielded millet remains prior to the Late Bronze Age. Millet occurs at Late Bronze Age Troad (Riehl 1999). As suggested by Riehl and Nesbitt (2003), millet first became established in Europe and subsequently in Anatolia and the Near East. The Ješovice finds in Austria are much earlier than those from southeastern Europe (Greece and Bulgaria) pointing to a north-south itinerary concerning millet’s introduction to Greece, supported, for the mainland, by both the archaeobotanical record and isotopic analyses on human skeletal remains from prehistoric Greece (see above), confirming earlier suggestions for a north to south adoption/introduction of broomcorn millet in Greece (Valamoti 2007a) and the Mediterranean in general (van Zeist 1980). The regions through which millet reached Greece are probably those areas in central and Eastern Europe where it occurs earlier than in the south. Compared with these central and eastern European finds, in Greece, large quantities of millet are encountered among Late Bronze Age deposits, the earlier finds perhaps corresponding to “exotica” introduced via contact networks, rather than a regular crop.

Archaeobotanical and artefactual evidence points to cultural contacts between northern Greece and regions further to the north and east, during the course of the 3rd millennium bc. Lallemantia, a plant with a transcaucasian/central Asian origin appears in northern Greece during this period (Jones and Valamoti 2005) linked perhaps to the introduction of tin bronzes from the East (Valamoti and Jones 2010). Spelt wheat and perhaps opium poppy indicate contacts with central Europe (Valamoti 2007a). Artefactual evidence suggests that the coastal areas of the north Aegean from the Thermaic Gulf to the west, to Skala Sotiros in the east had contacts with the Eastern Aegean, indicative of an active maritime network in the area. Contacts over large areas are also indicated by the presence of large anthropomorphic stelae in Macedonia (Andreou 2010). Access to distant contacts may have played an important role in interpersonal antagonisms as evidenced through finds in cemeteries (Andreou 2010; Pappa 2010; Akamatis 2009). Metal axes found in the region during the 3rd millennium bc point to a circum-Pontic tradition, the finds of which are either absent or rare in the rest of the Aegean (Andreou 2010). A rare find of a skepter, found at Sitagroi, indicates contact with the circum-pontic steppes (Renfrew et al. 1986) as the distribution of these skepters covers areas to the west, north and north East of the Black Sea (Kohl 2007). In the late Bronze Age, the region demonstrates connections to the north, and in particular the Danube, as well as to the Mycenaean south (Andreou 2010). Precious artifacts and imports or imitations of foreign products, such as gold, silver, amber, or glass jewelry, Mycenaean-type swords and seals, Mycenaean- and Danubian-style aromatic containers, and luxurious matt-painted ceramic drinking sets occurred often as prestige items in tombs. These contact networks in all likelihood contributed to the introduction of this new cereal crop, a late comer that was gradually adopted by 2nd millennium communities of Greece through complex processes related to differential access to long distance contacts, intra- and inter-community social relationships and the interplay of the cultural symbolism of 'traditional' foodstuffs and introduced culinary innovations, including new ingredients like millet (cf Valamoti 2007a, Boivin et al. 2012).

Millet and the horse

The timing of the introduction of millet to northern Greece coincides with that of the horse and may provide further evidence for contacts with the north. The introduction of the horse to Greece is thought to have taken place sometime during the end of the 3rd millennium bc while the earliest unambiguous evidence for its use is provided by the Linear B archives (Halstead 1984; Bökönyi 1986). In a recent review of horse finds in prehistoric Greece, Cantuel et al. (2010) suggest, on the basis of published archaeozoological data from Greece and the Balkans, that the horse was introduced in the north Aegean, in the region of Macedonia, during the Early Bronze Age, between approximately 2400–2100 bc as indicated by horse remains at Kastanas and Archondiko (Kostopoulos 2002) while in the south of Greece its presence is not certain prior to the Middle Bronze Age, found at Lerna (Argolid) in the south of Greece. The horse is encountered at Late Bronze Age Tiryns in the south, though it is still rare compared with the remaining components of the archaeozoological assemblage and it seems likely that horses in that period are found in funerary contexts, where horses are buried, a practice reminiscent of the strong symbolism of the horse, for certain prehistoric populations of the eurasian steppes (Cantuel et al. 2010; Cherlenok 2006; Cherlenok et al. 2006; Kuzmina 2006; Mallory 1981). Cantuel et al. (2010) also point out a possible link between horse possession and high status in Aegean Bronze Age society by analogy to the rest of the Balkans as discussed by Greenfield (2006), a point raised already by Cavanagh and Mee (1998). The horse is probably introduced to Greece from the north as horses are found north of the Balkans around 3300 bc in Baden communities and later they appear in the north Balkans (Greenfield 2006; Cantuel et al. 2010). However, these 4th millennium occurrences in the north Balkans are rather rare and sporadic and it is only during the early 3rd millennium bc in the Carpathian Basin and Central Balkans that horse remains seem to increase (Greenfield 2006; Bökönyi 1974, 1978). Horse domestication in the Old World is a controversial issue. The Eurasian steppes emerge as the area with the earliest evidence for horse domestication in the 4th millennium bc, a suggestion supported by archaeozoological evidence for milking and harnessing horses as well as genetic analyses on horse populations (Olsen 2006; Outram et al. 2009; Ludwig et al. 2009; Warmuth et al. 2012; Bendrey 2012). The steppe and steppe forest are more natural habitats for the horse (Olsen 1996) and the Balkan mountains may have acted as a natural obstacle for horses moving to southeastern Europe in postglacial times (Benecke 2006).

A possible association of millet and the horse, which for Greece appear to be introduced in the same period, is also observed further north as they appear together at sites of the Baden culture (3600–2900) where the horse is a regular find (Kohler Schneider and Caneppele 2009). The horse and millet also co-occur among nomadic or seminomadic communities of the Eurasian Steppes, known as the “pit grave kurganculture (Kohl 2007). The use of horses may have contributed to the transmission of crops and other cultural elements across the Eurasian Steppes in both directions as there are indications for the introduction of wheat and barley to East China via the northwest of China from across the Eurasian steppe during the 3rd and 2nd millennia bc in Altai, the Lower Yellow River valley and Lungshan (Crawford 1992; Crawford et al. 2005; Flad et al. 2010; Lee et al. 2007). It is, also through this area, that Sherratt sees China “reaching out to surrounding regions” on the basis of metalworking related artefacts (Sherratt 2006). It is possible that somewhere between China and the eastern steppes of Europe, millet, wheat and the people growing them encountered each other, each incorporating to their diets new cereals, cereals that contributed in varying ways to economies and societies over vast and different regions. The cereals of the “others,” millet for the west and wheat and barley for the east, crossed the Eurasian steppes into opposite directions perhaps with the aid of horses. An association of millet and the horse with a nomadic way of life would have made sense as horses facilitate transport and mobility besides being well suited to harsh steppe winters (Outram et al. 2012) and millet is a crop with a short growing cycle, easy to transport and with low growth requirements (Arber 1934; and see Weber and Fuller 2006; Barton et al. 2009 for an association of mobile lifestyles and millet in prehistoric China). Likewise, Pashkevich, associates millet to a nomadic lifestyle and dry climatic conditions in her review of archaeobotanical evidence from the forest-steppe and steppe zone of the Southern part of the East European plain, i.e. the areas of Moldova and Ukraine (Pashkevich 2000). Nevertheless, the significance of any association of millet and the horse among different communities of the Eurasian steppes remains to be demonstrated. The possibility that nomadic populations using millet and tending horses might have contributed to the rapid spread of these two westwards and to Southeastern Europe needs to be investigated further in the archaeological record taking into consideration artefactual, archaeobotanical, and archaeozoological and isotopic data with a finer resolution across a large geographical area, a task beyond the scope of this paper.

Bronze Age communities of northern Greece, through the long distance networks to the north and the northeast discussed above, would have come into contact with communities that had adopted millet and the horse, such as those of the Baden culture, or nomadic communities of the Eurasian steppes along the coasts of the Black Sea. These contact networks, manifested through various remnants of material culture in the archaeological record formulated the paths through which millet and the horse reached the north Aegean sometime around the end of the 3rd millennium bc. In the past innovations or changes in material culture and diet tended to be attributed to migrations of people or invasions. Recent discussions are dismissive of such interpretation models and consider more likely some form of adoption of cultural elements, including new crops, through contact networks between regions. Based on archaeobotanical remains alone, it seems from the crop repertoire that the Neolithic traditional foodstuffs continued into LBA and became enriched rather than replaced. Whether this means that indigenous people adopted new crops like broomcorn millet, via contacts with other cultures already consuming them, or that migrating people introduced their crops to the indigenous populations, is, on present evidence, difficult to tell. To address such issues, multidisciplinary projects are much needed, covering broad regions.


The review of the archaeobotanical record of broomcorn millet from prehistoric Greece presented in this paper has shown that broomcorn millet first appears in some considerable concentration in Greece during the end of the 3rd millennium bc Some early encounters with the plant during the end of the 5th and in the 4th millennia bc cannot be entirely ruled out but still remain ambiguous and unless directly dated, the vast majority of the early neolithic finds could have been intrusive. The archaeobotanical and isotopic data from Greece clearly indicate that this Far Easern crop after its introduction to southeastern Europe in the end of the 3rd millennium bc, became well established, in the north in particular, during the second half of the 2nd millennium bc which corresponds to the Late Bronze Age and the Mycenaen world of southern Greece. Earlier contacts involving millet exchanged as “exotica” cannot be ruled out; on the contrary, this may well have been the case. Whether millet represented a cereal staple used as human food or fodder still remains to be clarified. It is possible that millet consumption, either as food or as fodder may have underlined cultural identities or status or both as is indicated by variability at the individual and site level in terms of C4 plant intake. Broomcorn’s millet initial introduction to the north of Greece, then further to the south, and its adoption as a crop may have involved the introduction of the horse in that area and indeed the spread of millet from Central Asia to Europe may have been facilitated by horse transportation and the use of horses by nomads basing their subsistence on millet cultivation. There is ample archaeological evidence that the coastal areas and the islands of the north Aegean which have yielded evidence for broomcorn millet, were involved in trading networks connecting them to regions in the north, along the Danube and to the east and the Black Sea. Given the earlier occurrence of broomcorn millet to the north of Greece, a north south route seems very likely. Its trajectory/itinerary, however, still remains to be explored through a more refined examination of archaeobotanical, archaeozoological and artefactual evidence. Integrated research combining these and isotopic analysis of human and animal skeletal remains would contribute towards disentangling the complex processes of cultural transformation underway among Late Bronze Age societies of Greece that preceded the Iron Age, contributing, at the same time, new elements as regards diet and status, like millet and the horse.


The eight seeds from Mandalo correspond to a single find found in a fraction (1/8) sorted from the original sample.


The material from Skala Sotiros is currently under preparation for publication. The material from Archondiko is currently under study.


Unpublished C14 dates, Demokritos laboratory, Athens. With kind permission from Dr Ch. Koukouli-Chrysanthaki.


Until recently, it was prepared in Thessaloniki in northern Greece (I tasted it once) and sold in the premises of Hatzi pastry shops, of Albanian origin, yet, as the owner informed me early in 2012, they stopped preparing it as the traditional way of preparation is against E.U. food safety regulations.



The author is deeply grateful to Prof. Yo-Ichiro Sato for his kind invitation, the SATO project for financial support, and Leo Aoi Hosoya for her support and enthusiasm. Dorian Fuller and Leo are acknowledged for their patience and understanding for late manuscript submission. Prof. A. Papathimou and Giorgos Deliopoulos kindly allowed the publication of the Archondiko pyravnos reconstruction. Eleni Telioridou and Chryssa Petridou helped with sorting the samples from Archondiko. Angeliki Karathanou and Eugenia Gatzogia shared information on their unpublished Ph.D. material in relation to millet from Late Bronze Age and Iron Age contexts. Elena Ilade provided information on basha drink and millet mamalinga in Rumania and Elena Marinova on boza drink in Bulgaria. Dr Chaido Koukouli Chrysanthaki kindly provided the unpublished C14 dates on the millet samples from Skala Sotiros. Tassos Bekiaris and Nikos Katsikaridis prepared Fig. 1. Alan Alram, Stelios Andreou, Ioannis Akamatis, Aurelien Creuzieux, Giedre Motuzeite, Sevi Triantafyllou and Vera Warmuth kindly provided several references. The author would like to thank two anonymous reviewers whose perceptive comments helped in the improvement of the manuscript. Any remaining mistakes are that of the author’s.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013