Research on the origins of the obsidian was carried out on 52 geological and archaeological samples, by means of multi-element chemical analyses; 27 major elements and trace elements were examined using spark, flame and plasma emission spectrometry (ICP). These analyses were exploited by using models of the geochemistry of magmas : partial fusion, fractionated crystallisation, partition coefficients. Their specific characteristics lead to particular focus on the elements, called "hygromagmaphiles", which are concentrated in a magma during its evolution. Here these are four lanthanides, La, Ce, Yb, Lu, as well as Y, Zr, Nb and Zn.
Five samples, chosen from the geochemical families of Bingöl, were moreover tested by thermoluminescence, tests which led to a relative chronology in three parts : two successive outflows with, for the oldest, an age probably much more than I million years, and a sample artificially heated at about 8 000 BP. The method leads us to consider two types of criteria which permit the attribution of objects to the same geological site. I) Identity of chemical compositions, particularly in hygromagmaphile elements : provenance of the same outflow. 2) Different chemical compositions, but conservation of relations between hygromagmaphile elements : provenance of successive outflows coming from the same magma. It is thus that the two families of Bingöl are confirmed and explained : they are comagmatic and correspond to outflows of different ages, which are confirmed by thermoluminescence tests. One sample indicates that it could come from a third outflow at Bingöl, which would be interesting to investigate in the field.
These analyses show the particular importance of the Anatolian stratum at Bingöl for the supply of obsidian to neolithic (P.P.N.B.) and chalcolithic villages, not only in the Taurus and in upper Mesopotamia, but also well to the south, in an oasis of the Syrian desert (El Kowm).
