Acknowledgements. We would like to thank the Greek authorities for their permission and for their willing assistance in our work in Greece. We would particularly thank Professor Christos Doumas, Professor Peter Warren and Dr Sandy McGillivray for their help and suggestions. Partial financial support was provided by the Science & Engineering Research Council and the study was carried out at the Nuffield Palaeomagnetic Laboratory of the University of Newcastle upon Tyne. We thank all colleagues for their assistance.
1 Warren, P.M., ‘Knossos Stratigraphical Museum excavations, 1978–80. Part 1’, Archaeological Reports for 1980–81, British School of Athens 27 (1981) 73–92.
2 Similar channelled kilns are known at other Minoan sites on Crete, e.g. Hagia Triada, Knossos (SE of the palace), Phaestos, Vathypetro and Kato Zakros.
3 Plastic disks are glued to the structure and their flat surfaces are orientated. Small pieces of the structure (c. 10 cm3) can then be removed while attached to each disk. Tarling, D.H., Palaeomagnetism (London 1983) 379.
4 Molyneux, L., ‘A complete result magnetometer for measuring the remanent magnetization of rocks’, Geophysical Journal of the Royal Astronomical Society 24 (1971) 429–34.
5 Fisher, R.A., ‘Dispersion on a Sphere’, Proceedings of the Royal Society A217 (1953) 295–305. k is the estimate of precision and α95 is the radius of a cone of 95% confidence.
6 Lowrie, W. and Fuller, M., ‘On the A.F. demagnetization characteristics of multidomain magnetization in magnetite’, Journal of Geophysical Research, 76 (1971) 6339–49.
7 Showing a repeatability of ±1°.
8 Mean direction 354.4°, 59.3° − N = 26, k = 84, α95 = 3.1°.
9 The field required to reduce the initial intensity of remanence by half.
10 Tarling, D.H. and Symons, D.T.A., ‘A Stability Index of remanence in palaeomagnetism’, Geophysical Journal of the Royal Astronomical Society, 12 (1967) 443–8.
11 It can be argued that sample 16 should be similarly discounted (see Fig. 3), giving rise to a mean direction of 355.4°, 61.3° (k = 343, α95 = 1.6°) but neither the mean direction nor statistics are changed much and the more conservative estimate, i.e. including sample 16, is accepted here.
12 In the northern hemisphere, such refraction effects7, 8 cause a flattening of the magnetic vector in the floors, and a westward deviation in the eastern walls, and an easterly deviation in the western walls. Aitken, M.J. and Hawley, H.N., ‘Archaeomagnetism; evidence for magnetic refraction in kiln structure’, Archaeometry 13 (1971) 83–5; Tarling, D.H., Hammo, N.B. and Downey, W.S., ‘The scatter of magnetic directions in archaeomagnetic studies’, Geophysics, 51 (1986) 634–9.
13 Kono, M. and Ueno, N., ‘Palaeointensity determination by a modified Thellier method’, Physics of the Earth and Planetary Interiors, 13 (1977) 305–14.
14 Downey, W.S., Magnetic Studies of Santorini Minoan Tephra and Cretan Archaeological Materials Ph.D. thesis (Newcastle upon Tyne 1983) 716.
15 Fox, J.M.W. and Aitken, M.J., ‘Cooling-rate dependence of thermoremanent magnetization’, Nature 283 (1980) 462–3.
16 Downey, W.S. and Tarling, D.H., ‘Archaeomagnetic dating of Santorini volcanic eruptions and fired destruction levels of late Minoan civilization’, Nature, 309 (1984) 519–23.