Manley had undoubtedly believed in geomagnetic polarity reversals, Hospers and Runcorn were equally convinced, but on the other side of the Atlantic, where the Carnegie Institution’s Department of Terrestrial Magnetism had become something of a stronghold for secular variation studies, objections were mounting and a rival theory was brewing.
Before the war, scientists at the department had started a program to investigate whether soft, unconsolidated seabed sediments might carry continuous records of changes in Earth’s magnetic field. After the war, a keen new research student called John Graham took up this work and extended it to older, consolidated sedimentary rocks. Graham kitted out a truck as a mobile field station and laboratory, and with a group of coworkers mounted several military-style expeditions across the United States to sample sedimentary rocks, some flat-lying and some deformed by folding.
Graham’s primary goal was to assess the stability of the magnetization retained in these ancient sediments. Decades on, his legacy to paleomagnetism would be the field tests he devised to ascertain whether magnetization carried by a rock dated back to the time of its original formation, or had been reset during subsequent phases of deformation.
Graham’s “fold test” was conceptually very simple. He imagined a flat-lying sedimentary rock that had been stably and uniformly magnetized at the time it was deposited. When such sediments were folded, he figured, the rocks on one side of the fold would be tilted one way, and those on the other would be tilted the other way. If the magnetization were stably locked into the rocks, it too would tilt so that it ended up in different orientations on different sides of the fold. Working backwards, Graham decided that if he could measure the angles through which the rocks had been tilted, he should be able to figure out by geometrical calculations the original direction of magnetization. If the clusters of magnetization directions from the two sides converged to one, he could conclude that the magnetization pre-dated the folding process, and could take this single direction as the original paleomagnetic field direction.
By the end of 1949 Graham had collected a set of data that, in other hands, could have led to all the groundbreaking discoveries later made by Hospers and his successors at Cambridge. But perhaps because he was conditioned by the conservative adherence to secular variation at the Carnegie Institution, he missed the clues and failed to follow the right leads. He came very close to the geocentric axial dipole hypothesis when he commented that most of his results from rocks spanning the past sixty million years clustered around geographic north, rather than the present field direction, but without the rigor of Fisher’s statistics the claim lacked force. In any case, he used it more to argue that the field had been stable and had retained a normal polarity over that period of time than to question the difference.
In his fold-test studies, however, Graham had come up with one disconcerting piece of data. This involved the magnetization of the Silurian Rosehill Formation in Virginia, which dated back about 430 million years. After he had applied tilt corrections to samples from different parts of this fold, he found that the magnetization directions separated into two distinct clusters—one in a northwesterly and upward direction, and the other southeasterly and downward.
This was probably Graham’s best example of a successful fold test, but how could he explain the resulting directions? There was not one cluster but two in roughly opposite directions, and neither was remotely close to the present magnetic field direction in Virginia, which was northerly and downward. Finally, after consulting his Carnegie Institute colleagues, Graham concluded that his rocks must have been magnetized during a period of anomalously high-amplitude secular variation. He wrote:
My physicist colleagues … expect to find evidences … for stronger secular variation foci in remote epochs when the interior of the Earth may have been hotter and the disturbing current systems nearer to the surface … The corresponding current systems now appear to be roughly at 1200-kilometer depth.
However, Graham next considered another explanation— whether there was some fundamental way rocks could become magnetized in a direction opposite to the prevailing magnetic field. He consulted Louis Néel on the matter. Ernest Rutherford is reputed to have commented that if you ask a geologist to describe a stone he will conjure up the history of the entire Earth, but if you give the same stone to a physicist he will describe the minutest details of its atomic structure. Néel was the archetypal theoretical physicist: rather than credit the polarity-reversal hypothesis, he came up with several sub-microscopic processes by which such “self-reversal” might occur and described them in meticulous detail. Wholesale disruption of the Earth’s inner constitution was apparently no longer required to explain reversely magnetized rocks.
All of Néel’s self-reversal mechanisms required the combination of rather special physical conditions and unusual chemical compositions. How common were such conditions and compositions in nature? A logical test would be to take a sample of rock that has been found to carry a reversed natural magnetization, heat it above the Curie temperatures of its minerals, cool it in a known laboratory field and see which way it ends up magnetized. One of the first such experiments was carried out by a Japanese paleomagnetist, Takesi Nagata, in Tokyo and reported in Nature magazine in 1952. Nagata had found that samples of lava ejected from Mount Haruna, an active volcano in eastern Honshū, became magnetized in the opposite direction to the local field when cooled in temperatures between 440° and 250°C.
This was enough to reinforce the idea of self-reversal in many minds, including Graham’s. In 1953, while Hospers was arguing that at least two polarity reversals had occurred in the past few million years, Graham wrote:
… there is reason to doubt that inverse magnetization of rocks prove a reversal of the Earth’s magnetic field; indeed, fuller understanding of the mechanism of their magnetization may be crucial in showing the constancy of the Earth’s field.
And in 1954 he argued that:
… during Paleozoic time, the Earth’s magnetic field retained approximately its present orientation and, except for possible brief excursions, its present sense.
Graham had thrown the proverbial cat among the pigeons.
Back in Cambridge a frenzy of activity was underway. Edward (Ted) Irving, who was in the thick of it, would later describe these years as “fluid, even chaotic … No one was really in charge.”
Inspired by Hospers’ results, and imagining new and novel possibilities for the paleomagnetic method, Runcorn had hired Irving, a research student, and in 1951 the pair set out to sample the fine-grained, red Torridonian Sandstone of Scotland in search of records of ancient secular variation. On their way back to Cambridge they stopped off in Manchester to measure their samples using Blackett’s now famous magnetometer. Through many unsuccessful attempts to verify Blackett’s theory of the origin of the geomagnetic field, it had evolved into by far the most sensitive instrument available.
Like Graham’s results, Irving and Runcorn’s contained two opposite clusters of directions: northwest and upwards and southeast and downwards. Viewed either way, they were, according to Irving, “miles away from that expected if Scotland had not moved.”
Irving and Runcorn were clearly in no doubt that the axis of the geomagnetic dipole had been always, on the average, aligned with the rotation axis. In other words, the average positions of the geomagnetic poles and geographic poles had always coincided.
Implicitly, they were assuming that Hospers’ geocentric axial dipole hypothesis was good not just for the past few million years, but right back to the Cambrian—the past 500 million years. The fact that neither set of directions pointed in the geocentric axial dipole direction suggested that over the past 500 million years Scotland had rotated with respect to the poles, and carried the magnetization of the Torridonian Sandstone with it. This much had to be true, whether the opposite nature of the two clusters of directions was due to field-reversal or self-reversal.
All ideas of looking for secular var
iation went out of the window. Graham may have thought such results disconcerting, but to Irving and Runcorn they were the next exciting clue in the interlocked mysteries of geomagnetic reversals and continental drift. Was the long-awaited breakthrough just around the corner?
Before long the Cambridge group was augmented with the arrival of two new postgraduate students, Kenneth Creer and David Collinson. A Blackett-type magnetometer was built and, asIrving reported, “Ken raced through the Paleozoic”—sampling and studying the Devonian Old Red Sandstone of South Wales, the Permian Exeter volcanic rocks, and many other rock formations, all of which fell chronologically between Irving’s Torridonian sandstone and Hospers’ Tertiary lavas.
Edward (Ted) Irving, whose studies of the Scottish Torridonian Sandstone led to confirmation that the polarity of the Earth’s magnetic field had “flipped” many times, and that the continents were drifting. The Cambridge group of which he was a member also included Keith Runcorn and Kenneth Creer.
It came as no surprise that not one of the results pointed towards the geographic north pole. Creer later recalled that when their results were presented at a meeting of the Royal Society in May 1954:
… one visitor asked for the latitude and longitude of the geomagnetic poles of Irving’s Torridonian, which was our most thoroughly measured rock formation. In order to provide a quick answer to this question we had the Royal Society premises searched to find a world globe. Then to locate the ancient pole position quickly we improvised by using a piece of string … We found one of these paleopoles to be in the central Pacific and the opposite one in Ethiopia. But we did not have logarithmic tables nor a slide rule to be able to carry out numerical calculations there and then.
Later that summer, with Runcorn overseas, Creer was asked to take his place at the annual meeting of the British Association for the Advancement of Science in Oxford and speak about the work of the Cambridge group. By now Creer and Irving had calculated paleomagnetic poles (“paleopoles”) for every one of their rock formations.
What did these paleopoles represent? Two hundred years earlier Gauss’s mathematics had shown that a geocentric dipole accounted for by far the biggest part of Earth’s magnetic field, and this had been amply confirmed by later calculations. Hospers had suggested that, averaged over hundreds of thousands of years, the net effect of the secular variation, particularly of the remaining non-dipole part of the field, came to nothing; this had led to his hypothesis that the time-averaged magnetic field since the Tertiary was equivalent to a geocentric axial dipole. Irving and Creer had calculated average magnetic field directions from each of their more ancient geological formations. These also represented long time-averages, but unlike Hospers’ average directions they did not point towards the geographic pole.
Hence the question, where did the magnetic pole corresponding to the 500-million-year-old Torridonian paleomagnetic direction lie? Creer and Irving knew it would lie along the direction indicated by the declination, at an angular distance determined by the inclination: this was the basis of their globe and string construction. Only afterwards did they calculate the positions using geometry and spherical trigonometry.
When ordered chronologically, the paleopoles from the various geological formations showed an unmistakable trend: the youngest pole, calculated from Hospers’ directions, coincided with the present geographic pole, but the older poles lay further and further away. Inspired by a diagram entitled “Paths of the north pole (relative to the continents)” that he found in Beno Gutenberg’s 1951 book The Internal Constitution of the Earth, Creer plotted the path of the paleomagnetic poles from the pre-Cambrian to the present.
Creer had his first “polar wander path” ready in time for the meeting of the British Association for the Advancement of Science in September 1954. He would later recall that those present were “receptive although not at all enthusiastic about continental drift or polar wander.” However, media coverage of the meeting told a different story: The Times, The Manchester Guardian and even Time magazine all ran excited articles.
Despite adopting the name “polar wander path,” Creer and his coworkers did not really believe that either the paleomagnetic or geographic poles had wandered. The geocentric axial dipole hypothesis implied that they coincided, and physical realities made wholesale movement of the rotation axis all but impossible.
Time magazine’s 1954 depiction of the apparent movement of the north pole, as seen from Britain, from 700 million years BC to the present—as deduced by the Cambridge paleomagnetism group of Creer, Irving and Runcorn. This work led to the confirmation of continental drift and the theory of plate tectonics.
So what exactly was wandering? The only other option was that the rocks, and hence the land mass to which they were fixed, were moving with respect to the poles. Here at last, then, was evidence that the British Isles, and likely as not the whole continent of Europe, were drifting: at the time the Torridonian Sandstone was laid down it was Britain, not the pole, which had sat close to the equator and sweltered. The polar wander path was really a record of continental drift.
Creer’s polar wander path confirmed that the British Isles had moved significantly over the globe in the 600 million years since the Precambrian period. However, it did not answer the question of whether America and Europe had ever been closer together, as the fit of the coastlines suggested. Had the whole crust of the Earth rotated over the inner parts of the globe? Or had different parts of the crust moved independently?
This could be tested. If the whole crust had moved intact, the polar wander paths of different continents would overlie one another. If parts of the crust had moved independently, the paths would have evolved differently as the continents drifted with respect to each other. To solve the mystery, what were needed were polar wander paths for all the continents. Leaving Creer and Irving to tidy up at home, Runcorn set off to the United States to fire up paleomagnetists there and forge collaborations with them.
By 1957, the Cambridge trio had amassed enough data to construct a polar wander path for North America, as well as to improve the British path. The comparison was breathtaking: for the past few tens of millions of years the two polar wander paths had coincided, but before this recent and relatively short period of Earth’s history the two had been offset by an almost constant longitude difference of about twenty-five degrees. To see this effect, have a look at a world map. If you move South America east by about twenty-five degrees, it slips right into the big bend in the West African coastline, give or take a margin of continental shelf.
The coincidence was too good not to be true. In fact, it was difficult to explain a systematic offset like this in any other way. Creer and company certainly tried: they checked for problems with the magnetic recording process, for the presence of secondary components of magnetization, and for errors in the assignation of ages. However, in each case it was hard to see why all the rocks on one continent should be affected, while none on the other were.
As they would state in Philosophical Transactions of the Royal Society of London, they kept coming back to the conclusion that:
… in pre-Jurassic times, Europe and North America were several thousand miles closer than they are today. Such a displacement … is described in the literature as continental displacement or drift. In view of the unsatisfactory nature of the other explanations, it is postulated that in Paleozoic and early Mesozoic times Europe and North America were very much closer together, and at some time prior to the mid-Tertiary they moved apart to their present positions.
Here was evidence that for most of Earth’s geological history Europe and North America had been joined. They had moved apart only “at some time prior to the mid-Tertiary.” And since the polar wander paths were recorded in the magnetization of the rocks, the paths too had separated by an amount equal to the eventual separation of the continents.
By 1957 when this was published several members of the Cambridge group had moved on and taken paleomagnetic research to new shore
s, but one thing was clear: the group was convinced that polarity reversals were real, as was the geocentric dipole nature of Earth’s magnetic field and the overall correspondence of the geomagnetic and geographical axes—the so-called geocentric axial dipole hypothesis. Their statement was unequivocal:
Because the magnetic poles reverse there are two possible positions for the north geographical pole for any geological period.
The question now, of course, was what caused such reversals. Runcorn, who, according to his students, “knew no geology” and “was best kept out of the laboratory,” had emerged as the group’s theoretician. Since his early experiences with Blackett, he had become a dedicated convert to the theory that Earth’s magnetic field was generated by dynamo action. Unlike Graham, who seemed happy to hypothesize that electric currents were flowing in the mantle, he firmly believed that the field-generating currents would prove to be in Earth’s liquid outer core.
Accordingly, he reasoned that two things were the case. First, the coincidence of the geomagnetic and geographical axes was a natural consequence of Earth’s rotation acting on the fluid core. Second, polarity reversals could be brought about by “minor changes in the pattern of convection in the core.” And so:
The case for special causes of reversals of magnetization is not strong when reversals of the main field give a simple and general explanation for the varied phenomena observed.
The Cambridge group had also noted that polarity reversals and polar wander operated on different timescales. Reversals appeared to be relatively frequent—Hospers had suggested they occurred at intervals of about half a million years—while polar wander was very slow and gradual. On the basis of this, the group argued that the two phenomena must come about through different processes. But if reversals originated in fluid motions in Earth’s core, what caused polar wander?
North Pole, South Pole: The Epic Quest to Solve the Great Mystery of Earth's Magnetism Page 16
