1. The excavators identified two of the sampled objects as “slag” (HDM 1021 and 1114). Both objects are rather small (a few hundred milligrams) and broken into a number of pieces. In the case of HDM 1021 the shape of one larger fragment suggests that rather than a slag, the object might be a completely corroded piece of a metal pin. Such a suggestion is corroborated by the low iron content (see below). For the present purpose we have retained the identification of HDM 1114 as “slag”, but have classified HDM 1021 as metal. Both classifications are ambiguous, however. Five of the specimens have been assigned probabilities of between 93 and 98% of belonging to Troia I. These are the unalloyed copper pins, or fragments thereof, HDM 1009, 1013, 1108, and 1111 as well as the only arsenical copper in our suite of samples, a needle with double-conical head (HDM 1116; Pfeffer 1990).
2. An alignment along lines of constant element abundance ratios results if different ore charges contained different amounts of an accessory phase which carried the two elements in a constant ratio. The silver-gold diagram suggests, e.g., that the bronze objects dating to Troia VI received their silver and gold from a carrier phase that varied in amount about tenfold, but contained the two elements in an abundance ratio Ag/Au - 5. The unalloyed copper from Troia I, on the other hand, shows about the same tenfold spread in silver and gold contents as the Troia VI bronze objects but in this instance the Ag/Au abundance ratio in the putative carrier phase of the two elements appears to have been several hundred, rather than five as for the Troia VI bronze.
3. At Thermi, from the oldest periods I-III, dating to Troia I, there are just 4 objects out of 49 with high 206Pb-normalized abundance ratios, from Thermi IV 1 out of 10, but from settlement period V there are 3 out of 8 with high ratios. Likewise, among the 11 artifacts from the 2 oldest copper-bearing periods at Poliochni,”azzurro” and “verde” (Bernabo-Brea 1964, 1976), there is not a single one with a high ratio; from the next one, period “rosso”, it is 1 out of 22, but from period “giallo” it is 11 out of 54. In the present context we define “high” 206Pb-normalized abundance ratio as such with 208Pb/206Pb 2.10. Although somewhat arbitrary, any other reasonable choice for the separation line will change the numbers, but not the conclusion.
4. The sites in question are Dsogancilar (TG 133) and Yuvalar (TG 138) in Çanakkale Province, Kozçagiz (TG 142) and Serçeörenköy (TG 192) in Balikesir Province, Keles (TG 154) and Tahtaköprü (TG 156) in Bursa Province. Three more sites with copper slags or copper ores, but without evidence for ancient mining activities, are Avcilar (TG 128) and Hahlar(TG 144) in Balikesir Province and Camyurt (TG 136, Çanakkale Province). Detailed descriptions of the sites as well as lead isotope abundance ratios have been reported in part in Wagner et al. (1986).
5. For HDM 165, 1108 isotopically matching copper ores exist at Serçeörenköy and matching lead ores at Balya, for HDM 1020,1132 at Tahtaköprü, and for HDM 1118, 1125 there are matching copper slags at Dogancilar and at Kozçaglz (Fig. 8).