A series of pottery fragments from an early Neolithic site in Iran is analyzed to determine the original baking temperatures. Two magnetic techniques are used: remanent magnetization decay, and hysteresis loop analysis. The results are complementary, although some are, at first sight, contradictory. The causes of these apparent contradictions are explained. THe analyses confirm that the pottery was baked under highly variable conditions, probably over open fires. Control of baking was poor, and there was considerable variability not only from one vessel to another but even from one zone to another on the same vessel. Our results suggest the need for caution in interpreting often ambiguous findings in this field of ancient technology.