INTRODUCTIONThe essential basis for determining the intensity of the magnetic field in which pottery was fired is the comparison of the thermoremanent magnetization then acquired, M A , with that acquired due to reheating in a known laboratory field, ML. As long as the intensity of the laboratory field, FL, is of comparable magnitude to the intensity of the ancient field, FA, then FA = (MA/ML)FL.(1)As was established by Thellier (1938) the magnetic grains that acquire a remanent magnetization on cooling through a certain temperature lose it at the same temperature on reheating; this is termed the blocking temperature (TB) for those particular grains. Within a ceramic sample there is usually such a variety of grains that the distribution of blocking temperatures is continuous from the Curie point (680°C for haematite, 575°C for magnetite) downwards. Thus equation (1) can be used to determine values for FA from a succession of laboratory reheatings to increasingly higher temperatures until the Curie point is reached. A given reheating temperature yields the value for FA as recorded by the grains having blocking temperatures in the range up to the reheating temperature. Such stepwise remagnetization is the essential feature of the method developed by Thellier more than two decades ago (Thellier and Thellier 1959) and it is still the method of accepted reliability whether for ancient ceramics or for volcanic lava. This reliability stems from the fact that the method allows detection of the reheating temperature at which mineralogical change interferes with the validity of equation (I), data from higher temperatures then being rejected. Recently an alternative technique for avoiding interference by mineralogical change has been established; this is the 'coercivity monitoring' method developed at the University of Liverpool by Shaw (1974). Its reliability is now well established and its importance for archaeomagnetism has been demonstrated in several applications (Shaw 1979, Gunn and Murray 1979, Games 1980. With the advent of SQUID* cryogenic magnetometers utilization of much smaller samples became feasible, and Walton developed a SQUID magnetometer specifically for paleointensity determinations on 3 mm cores extracted from pottery, using a modified version of the Thellier technique (Walton 1977(Walton , 1979(Walton , 1980. The smaller sample size allows an order of magnitude reduction in the time required for the stepwise reheatings. In the following pages we give an account of the procedures developed for use with this magnetometer that routinely take into account the effects of grain anisotropy (Rogers et al. 1979); such effects had not been * Acronym for Superconducting Quantum Interference Device. 53 54 noted in pottery before the use of this instrument but they are important in paleointensity determinations irrespective of method. ANISOTROPYMost types of pottery exhibit an intrinsic anisotropy in acquiring thermoremanent magnetization (TRM) that is much stronger than any shape anisotropy that arises from the dem...