have, however, found a number of clay models of two-wheeled carts. Indirect evidence
comes from the conjunction of model wheels and Khirbet Kerak pottery: brightly
burnished red and black kraters and bowls that were made on a potter’s wheel. The
ware was first found by William Albright almost 100 years ago at the southern end of
the Sea of Galilee, where stands the tell called Khirbet al-Karak in Arabic and Tel Bet
Yerah in Hebrew. Khirbet Kerak ware was subsequently found at various sites, from
northwestern Iran to Jericho. Very different from (and superior to) Levantine ceramic
traditions, Khirbet Kerak ware is now known to have originated in southern Caucasia,
where it remained in common use through most of the 3rd millennium BC. Rafael
Greenberg, who has directed excavations at Tel Bet Yerah, has recently published ten
wheel models found at the site, one in stone and the others in clay, each with an axle
perforation and a pronounced nave. See Greenberg 2014. Greenberg points out that
we have no evidence for wheeled vehicles of any kind in the Levant until ca. 2800 BC,
when model wheels—all with axle perforations and pronounced naves—began to be
made and displayed, and that they show up only at sites and in strata where substantial
amounts of Khirbet Kerak ware are also present. An important site outside the Levant
is Norşuntepe, a copper-working center in eastern Anatolia, where “no less than 35
model wheels are catalogued” (Greenberg 2014, p. 95). The conjunction between model
wheels and Khirbet Kerak ware leads Greenberg to speculate that early in the 3rd
millennium BCpeople from the Kura-Araxes culture had not only the potter’s wheel
but also wheeled vehicles and therefore an advantage in mobility: with their ox-carts
they may have brought themselves and their pottery to places far away from south
Caucasia. It may be that wheeled vehicles were produced in southern Caucasia at a
fairly early date, but not early enough—so far as we now know—for southern Caucasia
to have been the homeland of PIE.
54 The map of the probable PIE homeland in Anthony 2007, Fig. 5.1, is approximately
what I have in mind. Anthony and Ringe 2015 present good arguments that PIE evolved
in this general area in the 4th millennium BC, but for both linguistic and archaeological
reasons I must disagree with their assumption that this was also the homeland of Proto-
Indo-Hittite (which they call “Early PIE”).
55 On the Dnieper-Donets culture see Mallory 1989, pp. 190–191; Anthony 2007,
pp. 174–182; and Dolukhanov et al. 2009, pp. 789–790.
56 Dolukhanov et al. 2009, p. 789.
57 See Ivanova 2012, pp. 2–3: “In Gräbern und Siedlungen des 5.Jahrtausend v.Chr.im
Nordkaukasus lässt sich eine materielle Kultur erkennen, die mit gleichzeitigen
archäologischen Komplexen aus dem nördlichen und westlichen Schwarzmeer
verwandt und in ein Netzwerk für „Prestige-Güter” eingebunden war. Es handelt sich
um einen entfernten Ableger der südosteuropäischen Kupferzeit.
58 On the dates of the lower Don culture see Telegin et al. 2003. The authors place the
beginning of the Mariupol Neolithic ca. 6500 cal BC, and conclude that it coexisted
with hunter-gatherers for a long time. For a lower date see Kotova 2010, p. 167:
The sites and cemeteries were excavated in the Northern Azov Sea region, which
covers the territory from the Lower Don to the Kalmius River, and are dated to
c. 6050–5200 cal BC. It disappeared during an arid period about 5200 cal BC, when
the Sredny Stog culture was formed.
59 On the Khvalynsk culture see Anthony 2007, Fig. 9.1 and pp. 182–186.
60 For close parallels between artifacts in the pre-Maikop sites and sites along the lower
Don see Ivanova 2007, p. 13.
61 Kohl 2007, pp. 72–86 has a full description of the Maikop kurgan and the Maikop
culture.
62 Ivanova 2007, pp. 10–13.
63 Courcier 2014, pp. 621–622.
Origins and spread of Proto-Indo-European 25