Abstract
The lodestone is an extremely rare form of the mineral magnetite (Fe3O4) that occurs naturally as a permanent magnet. It therefore attracts metallic iron as well as fragments of ordinary ‘inert’ magnetite. This ‘magic’ property was known to many ancient cultures, and a powerful lodestone has always commanded a high price. By the eleventh century AD the Chinese had discovered that a freely suspended elongated lodestone would tend to set with its long axis approximately north–south, and utilized this property in the magnetic compass. They also appear to have discovered that this invaluable characteristic could be handed‐on to a steel needle if the latter were contacted with, or stroked by, a lodestone. The magnetism of the lodestone was scientifically investigated by William Gilbert in the sixteenth century, when he defined its ‘poles' and the well‐known rule that ‘like poles repel, unlike attract’. He also studied ‘inclination’ and ‘variation’, and means to aid the preservation of magnetic power. How to concentrate it by ‘arming’ the lodestone with caps or pole‐pieces of soft iron was discovered in the same century. These methods have been repeated, confirmed, and improved. The lodestone occupies a vital place in the history of magnetism, but little beyond Gilbert's work can be reached by historical studies because vastly improved steel or alloy permanent magnets, and electromagnets, replaced it before quantitative measurements were developed. These techniques have therefore been applied retrospectively to both museum specimens and contemporary natural lodestones. A good source of the latter was found to be the igneous complex known as Magnet Cove, Arkansas, and this material has been used as the ‘type example’. All specimens were discrete, well‐rounded, rusty brown pebbles found near the surface. Their unweathered interiors were black titanomagnetite. No significant trace element or crystallographic differences could be found between the lodestones and the magnetically inert material that always accompanied them. The magnetic moment per unit volume (J v) of the ‘as‐found’ Magnet Cove lodestones varied between 6.5 and 11.6 emu cm−3, which compares poorly with the hundreds of units characteritzing modern permanent magnets. Hysteresis loops gave a saturation intensity (J s) of 27–51 emu cm−3, suggesting that intensity has diminished since formation. This agrees with general experience of magnets, especially in the absence of a ‘keeper’. The initial volume susceptibility of Magnet Cove magnetite was about 0.18 for low fields, and always remained J v at weekly intervals for 100 days. The decay curves visually resembled exponential functions, but mathematical tests proved that they were not strictly so. Interpretation was therefore difficult, but a pragmatic procedure involving excessive extrapolation suggested an ‘age’ of about 3500 years.