The last alloy to be considered here is steel, an alloy of iron and carbon (C). The incorporation of even small concentrations of carbon into iron can significantly increase its hardness. Iron with low levels of carbon (under c.0.1 percent C) is known as “ bloom iron” or “ wrought iron” and its strength and toughness are similar or inferior to that of standard tin- bronze (Rostoker and Bronson 1990:9; Rehder 1992). In contrast, alloys of iron with higher levels of carbon (c.0.1– 2.0 percent C) – i.e., “ steel ” – show dramatic improvements in strength and hardness over tin- bronze. Iron with c.2– 5 percent carbon is known as “ cast iron.” This is very hard, brittle, and unsuited to forging, and does not appear to have been produced in the ancient Near East (although see Moorey 1994 : 285)

It is clear that early iron smelters produced a heterogeneous bloom that had variable levels of carbon; some areas would have been wrought iron, whereas other areas could have absorbed enough carbon to be classed as mild steels (Avery 1982: 207; Rostoker and Bronson 1990 : 97– 8). Such heterogeneity could have survived bloom consolidation in relatively oxidizing atmospheres, leading to the unintentional production of malleable billets containing both wrought iron and steel. Alternatively, steel can be deliberately produced by carburizing or “ steeling ” wrought iron in a furnace under charcoal – a slow, solid- state diffusion process that can take many days. It can prove difficult for archaeometallurgists to distinguish between individual artifacts made from steel produced accidentally or in an uncontrolled manner in the iron bloomery and those that were intentionally carburized, although the metallurgical characteristics of larger object assemblages provide a better guide to intentionality (Maddin 1982 ). Early examples of carburized artifacts are known from Cyprus, Anatolia, and the Levant in the late 2nd millennium BC , where they occurred alongside uncarburized iron (Waldbaum 1999). Based upon analytical studies, Wheeler and Maddin (1980 :121) have suggested that “ smiths were carburizing intentionally on a fairly large scale by at least 1000 BC in the Eastern Mediterranean area.” Contemporary ferrous artifacts with variable levels of carbon are known from Luristan in Iran. However, these are interpreted as the unintentional result of heterogeneous bloom smelting (Moorey 1991a ; Rehder 1991 ).

The effects of carbon in steel can be dramatically increased with heat treatment involving forging (hammering), quenching (rapid cooling), and tempering (mild heating to remove brittleness) (Rostoker and Bronson 1990 : 4– 5). The archaeological evidence for the developed heat treatment of iron is relatively limited, partly due to the high corrodibility of iron and the consequent difficulty of analyzing corroded artifacts. Studies by Stech-Wheeler et al. (1981) , Maddin (1982) , Smith et al. (1984) , and Muhly et al. (1985) , amongst others, indicate that the heat treatment of iron was quite variable in the late 2nd and 1st millennia BC. Evidence for the intentional quenching and tempering of carburized iron comes from 11th century BC sites such as Lapithos, Idalion, Amathus, and Kouklia-Skales on Cyprus. This is matched by a 12th/11th century BC pick from Mount Adir in Israel that is made of quenched and tempered steel (Maddin 1982 ) and possibly by an even earlier, Middle Bronze Age blade from Pella (Jordan) of quenched steel, although its archaeological context is questionable (Smith et al. 1984). In other parts of the Near East, however, iron with inconsistent carburization and little or no evidence for developed heat treatment continued to be produced well into the 1st millennium BC (e.g., Curtis et al. 1979 ).