1. Introduction
2. Early discoveries
3. William Gilbert
4. Halley
5. Coulomb
6. Oersted and Ampere
7. The Lodestone
8. Gauss and Humboldt
9. Explorations and Surveys
10. Faraday's Lines of Force (field lines)
11. Faraday's Disk Dynamo
12. Sunspots
13. The Dynamo Process on the Sun
14. The Earth's Dynamo
15. Dipole Reversals and Plate Tectonics
16. Magnetic Storms and Ring Currents
17. The Magnetosphere
18. Magnetic Reconnection
19. Planetary Magnetospheres
20. Assessment

Alfred Wegener

    When lava flows out of a volcano and hardens, the result is often a black rock known as basalt. Basalt is slightly magnetic, and as it solidifies and cools, it becomes magnetized in the direction of the prevailing magnetic field. The process is somewhat similar to the capture of the prevailing magnetic field by a cooling bar of steel, noted by Gilbert (Figure 3).

    Volcanoes are often active for long periods, and by comparing the magnetization of their lava flows from different times (especially if such flows can be dated by some method) one can learn about changes in the direction of the local magnetic field. Bernard Brunhes (1867-1910) found ancient lava flows in France whose magnetization appeared to be reversed [Brunhes, 1906] . Other examples were then found, and the Japanese geophysicist Motonori Matuyama (1884-1958) examined the evidence and suggested that the magnetic signatures were evidence of actual reversals [Matuyama, 1929]. Matuyama proposed that long periods existed, in the past history of Earth in which the polarity of the magnetic poles was the opposite of what it is now.

    [See also "Brunhes' Research Revisited: Magnetization of Volcanic Flows and Baked Clays," by Carlo Laj, Catherine Kisel and Hervé Guillou, Eos (Transactions of Amer. Geophysical Union), 27 August 2002]

    Unfortunately, the situation was complicated by the work Louis Néel (1904-2000), later awarded the 1970 Nobel prize, and of John Graham, who showed that some materials exhibited spontaneous self reversal of their magnetization. Only in the early 1950s did Jan Hospers, starting with a study of Icelandic basalts, convince many in the geophysics community that most rocks with reversed magnetism were not self reversed but were relics of epochs when the Earth's had reversed magnetic polarity [Hospers, 1951; Cox et al., 1967; Cox, 1973; Frankel, 1987]

The Mid-Atlantic Ridge

    It was long ago noted that elevations on Earth were not smoothly distributed, but tended to cluster at two levels [Fig 1, McGill 1982] --the continents and the ocean floors, with a relative scarcity of in-between depths. Like others before him, Wegener noted that the coastlines of some continents, especially the ones bordering the Atlantic Ocean (see Figure) would fit together remarkably well (as was shown by Alex du Toit (1878-1948) and later by Carey and Bullard (1907-1980), the edges of the continental shelf fit even better [LeGrand, 1988, p. 204-5]). He then collected evidence showing that the fit between matching edges of continents extended to geological formations and even to fossils of animals and plants.

    His explanation was the theory of "continental drift." Geologists had previously proposed that continents were slabs of lighter rock floating on top of denser material below, the way ice slabs floated in the Arctic ocean. Wegener went one step further and proposed that such slabs could slowly move or "drift."

    His theory was vigorously attacked, especially by Sir Harold Jeffreys (1891-1989), Britain's leading theoretical geophysicist. Jeffreys argued that the layers on which continents floated were too viscous to permit such drifts, that their resistance to such motions would be too high. Wegener's had little support among geophysicists, and he even found it difficult to secure a university position in Germany (he ended up as professor in Graz, Austria). Wegener died on a polar expedition in 1930, at age 50, caught by bad weather while returning by dogsled from a supply mission to a weather station on top of the Greenland ice cap.

    A few die-hards kept Wegener's ideas alive [e.g. Du Toit, 1937], and by the 1950s, when the geophysics community finally accepted the reality of magnetic reversals, a variant of it emerged, the theory of polar wandering. By that theory--promoted by Keith Runcorn (1922-1995) and Thomas Gold--the crust of the Earth slowly slid around the interior, somewhat like one of Halley's layers, so that areas now in the polar zone once had temperate climate, and vice versa [Runcorn, 1955]. This explained phenomena such as coal seams in the polar islands of Spitzbergen (Svalbard) and evidence of glaciation in countries such as South Africa and India, as well as variations in the magnetization of crustal lava.

    The theory reached its peak in the early months of 1955, when it was debated in a seminar in Cambridge and then put to a vote by the participants: only one voice was raised against it [Ernest Deutsch, private communication, May 1985]. Soon however trouble arose in interpreting magnetic data: observations from the same period but from different continents disagreed in assigning the positions of ancient magnetic poles.

    The resolution of all these controversies came from the seafloor [Bullard, 1975; Kious and Tilling, 1996]. Oceanographic surveys had shown a "mid-Atlantic ridge," a raised ridge running roughly north-south, approximately halfway between the continents to its east and its west. The ridge was clearly volcanic, as shown by a concentration of earthquake epicenters along it and by volcanic islands located on it, such as the Azores. Other such linear features were found in the Pacific and Indian oceans, and all were linked in a worldwide pattern.

    Two prominent geophysicists, Robert Dietz (1914-1995), who was at the time with the Naval Electronics Laboratory Center in San Diego, and Harry Hess (1906-1969) of Princeton University--proposed that the seafloor might be spreading out from the ridges ( [Dietz 1961]; this started the term "seafloor spreading", though the idea began with Hess [Frankel, 1980] ). In their view new seafloor was formed by lava emitted from the ridges, while old seafloor descended again to lower levels at oceanic trenches, deep slots in the ocean floor, generally parallel to chains of islands.

    The convincing evidence, though, came from magnetic surveys. In the late 1950s electromagnetic magnetometers (developed in World War II) became available to geophysicists. Unlike the suspended-needle instruments, they could easily operate aboard ships, airplanes or even (a bit later) aboard satellites. Some were proton precession instruments, related to the nuclear magnetic resonance (NMR) method now extensively used in chemistry and for medical imaging. Others were fluxgate magnetometers, using substances which entered magnetic saturation at a stable and well-defined field value. These instruments were accurate enough to map the magnetism caused by local materials of the Earth's crust, by deriving the deviations from the smooth global field ("anomalies"). On land, where such instruments were used in the search for oil, the observed magnetic anomalies showed no consistent patterns. Above the ocean floor, on the other hand, the patterns were well-ordered, often in long stripes of opposite magnetic polarity.

    In 1962 Lawrence Morley [Morley, 1963 (1981), 1986 (1990); Lear, 1967] proposed an explanation, which independently also occurred to Fred Vine and Drummond Matthews [Vine and Matthews, 1963]. Their idea was that the seafloor indeed spread out from the mid-ocean ridges, to both sides, at about an inch a year. As the lava oozed out (a process later photographed from research submarines) it hardened, acquired the prevailing magnetization and was then was gradually carried away to either side of the ridge.

    Now and then, however, the main dipole reversed its direction, and the magnetic imprint reversed direction too. Each magnetic stripe, therefore, consisted of lava that cooled during an epoch of a certain magnetic polarity--wide stripes from long epochs, narrow stripes from short ones--and of course, all stripes tended to be parallel to the central ridge. The process was somewhat analogous to that of a tape recorder, with the seafloor near the mid-oceanic ridges behaving like a pair of magnetic tapes, slowly unrolling in opposite directions while faithfully recording the Earth's magnetic field at the time they emerged.

    For a long time Morley's contribution was not acknowledged, because the referees of "Nature" and "Journal of Geophysical Research" to which he submitted his article rejected it as too speculative. One referee wrote: "His idea is an interesting one--I suppose--but it seems more appropriate over martinis, say [rather] than in the Journal of Geophysical Research" [Morley, 2001; see also ,1973, p. 224]. It was finally cited in a non-scientific journal [Lear, 1967]. Even Vine and Matthews met resistance, until Jim Heirtzler [1968; Heirtzler, LePichon and Baron, 1966; Heirtzler et al., 1968] produced a 2-dimensional map of the striping of the Reykjanes Ridge near Iceland, where navigational radio aids allowed accurate determination of position. The map (Figure 10) graphically illustrated the remarkable symmetry of the striping.

Fig. 10     Map of magnetic striping of the seafloor near the Reykjanes ridge [Heirtzler, 1968]

    But where did the seafloor go at the edge of the continents? In some places--e.g. off Japan--it indeed descended into deep oceanic trenches. Elsewhere, however--for instance, off the eastern seaboard of the US--nothing remarkable seemed to take place. From this came the idea that just as the higher parts of the Earth's crust formed separate continents, so the underlying layer, the lithosphere, was divided into "plates" which constantly moved. Plate material was continually created at the oceanic ridges, then as the plates moved, the continents sitting on top of them were carried along, and at the trenches the plates descended again.

    This idea, later elaborated, became the theory of "plate tectonics" ("tectonics" means "building up"--in this case, of the Earth's crust). Wegener had the right idea, but nature carried it out differently from what he thought--the continents were not plowing through the lithosphere, the way ships plow through the ocean, instead they sat on top of conveyer belts that carried them along. Some plates also rotated, and some regions slipped alongside others ("transform faults"), but the overall size of the Earth, of course, did not change. (Note: This section is a very abbreviated and cursory summary of a broad subject on which an extensive literature exists, only in part related to geomagnetism. See for example Frankel [1982], LeGrand [1988] and Glen [1982].)