The Earliest Farmers in Macedonia
Wilkie, Nancy C., Savina, Mary E., Antiquity
Survey in southwest Macedonia provides a new field record of the pattern of Neolithic settlement. It prompts a further opinion in the enduring debate about the nature of the diffusion of agriculture, discussed most recently in ANTIQUITY by van Andel & Runnels (1995).
The demic diffusion model for the spread of agriculture as proposed by Ammerman & Cavalli-Sforza (1971; 1984: 63-84) suggested that early farming populations expanded into southeastern Europe gradually at the front of a wave-of-advance. Van Andel & Runnels (1995) see that migration as occurring in 'discrete steps', with settlers filling up only the attractive sites - the few large, fertile flood-plains of rivers and lakes - before leaping to the next such environment. They propose that the expansion of farming settlements into Yugoslavian Macedonia and southern Bulgaria was prompted by the 'saturation' of the Larisa basin in Thessaly at the end of the Middle Neolithic (c. 7500 BP). Our recent survey in the adjacent nomos (province) of Grevena (southwest Macedonia) shows that the initial migration north from Thessaly took place in the Early Neolithic (EN), earlier than Van Andel & Runnels propose. Moreover, EN sites in Grevena do not show that preference for flood-plain environments which Van Andel & Runnels suggest for the Larisa basin. Indeed, the data from the Larisa basin presented in Van Andel & Runnels (1995: [ILLUSTRATION FOR FIGURE 6 OMITTED]) show that the EN farmers in Thessaly, as in Grevena and elsewhere, settled in a wide range of environments near rivers, both on the flood-plain and on the adjoining hills and terraces. Even in the Larisa basin, a majority of the Neolithic sites were founded in environments outside the margins of active flood-plains (Van Andel & Runnels 1995: table 3).
The 'patchiness' of Early Neolithic settlement
Key to Van Andel & Runnels' argument is that the earliest agricultural settlements in Yugoslavian Macedonia directly resulted from the saturation of the Larisa basin during the Middle Neolithic (MN), not by gradual demic diffusion following the initial Neolithic occupation of Thessaly. Van Andel & Runnels (1995: 481) contrast the dense pattern of EN sites in Thessaly, with either a 'sparse scatter' (e.g. Boeotia, Attica, Epiros and the Peloponnese) or 'absence' (Macedonia) of EN sites elsewhere in Greece. A more complete picture of the Neolithic sites in Macedonia than Van Andel & Runnels' (1995: [ILLUSTRATION FOR FIGURE 1 OMITTED]) is seen in FIGURE 1 (adapted from Sakellariou 1983; [ILLUSTRATION FOR FIGURE 9 OMITTED], with the addition of sites from our own Grevena survey, all first occupied in the EN).
We believe that the number of known EN sites - fewer in Macedonia than in Thessaly - is a function partly of uneven exploration rates and partly of landscape differences. The large extent of research into the Neolithic of Thessaly during the past century is indicated in the recent atlas of prehistoric sites in East Thessaly (Gallis 1992). Research on prehistoric periods in Macedonia has been more recent, more sporadic and generally less thorough (Borza 1990; 1995a; Kokkinidou & Trantalidou 1991; French 1964). We do not accept that, because intensive surveys in central and southern Greece have not produced dense EN site distributions similar to those in Thessaly, the same would hold true for Macedonia. And as the areas intensively surveyed are small, distribution patterns cannot be extended from them to larger regions, let alone to Greece as a whole. Nor are all the surveys comparable, either in method, purpose or geological setting. The Thessalian plain has nearly ideal conditions for preservation of EN sites (Van Andel & Runnels 1995: 487), while other areas of Greece, deeply dissected like much of Macedonia, are less likely to preserve in situ deposits of EN material. When the nature of the data is taken into account, the 'patchiness' of the EN settlement pattern, as it is described by Van Andel & Runnels, loses its significance. …