Unraveling the Record
The entire history of the ocean basins in terms of oceanfloor spreading is contained frozen into the oceanic crust.
—FRED VINE, 1966
The Vine-Matthews-Morley concept was slow to catch on among marine geophysicists, and received little immediate attention from other sectors of the Earth Sciences community. In the next few years, the focus of United States marine geophysical work shifted from the west coast to Columbia University’s Lamont Geological Observatory, now the Lamont-Doherty Earth Observatory, in New Jersey, where Jim Heirtzler headed up operations. Heirtzler was a traditionalist, and would prove hard to convert to the theories of seafloor-spreading and polarity reversals.
Meantime, quite independently, the race to validate the field-reversal theory by establishing a global timescale for polarity reversals was in full swing among paleomagnetists. Earlier attempts had been hampered for want of a precise method of dating rocks. Traditional fossil-based dating methods were adequate for defining geological periods, which typically lasted hundreds of millions of years, but the speed of evolution and migration of species was too slow to resolve polarity intervals that lasted only a few million years at most.
During the 1950s, new dating techniques had been developed. These involved measuring the radioactive decay of certain isotopes of naturally occurring elements. At the beginning, the methods focused on isotopes that decayed very slowly; they had been applied to some of the oldest rocks and meteorite fragments to investigate the age of the Earth and of the solar system itself.
One such method involved the decay of an isotope of potassium, 40K. In 40K the nucleus of the atom has twenty-one neutrons in addition to the standard nineteen protons. Most naturally occurring potassium is 39K, in which atoms have twenty neutrons; only about one-hundredth of a percent of naturally occurring potassium is 40K. This tiny but incredibly evenly distributed fraction is radioactive: in 1250 million years half of the 40K nuclei in any sample will decay—89.5 percent to calcium (40Ca) and 10.5 percent to argon (40Ar). Potassium–argon dating relied on comparing the amount of 40Ar that had accumulated in a rock since its formation with the amount of 40K remaining. This ratio, together with an accurate knowledge of the half-life and details of the decay process, could be used to calculate the time since the minerals in the rock crystallized.
Although radiometric dating was a vast improvement on fossil-based methods, early potassium–argon age estimates still lacked enough precision. Uncertainties and errors centered on possible contamination by the small amount of argon—about one percent—present in Earth’s atmosphere, and the need to eliminate this from the mass spectrometers used to measure isotope ratios.
In the early 1960s, the pioneering laboratory in potassium–argon dating was at the University of California, Berkeley, where geologists Brent Dalrymple and Sherman Grommé and petrologist Ian McDougall, on an overseas fellowship from Australia’s Commonwealth Scientific and Industrial Research Organization, were engaged in graduate and postgraduate work. Meanwhile, at the United States Geological Survey in nearby Menlo Park, geophysicists Allan Cox and Richard Doell were also forging ahead in pursuit of a timescale for geomagnetic reversals.
In 1961 McDougall returned to the Australian National University in Canberra, stopping off in Hawaii en route to sample sequences of lava flows. In Canberra he set up his own potassium–argon dating facility, and teamed up with Ted Irving and Don Tarling, a new graduate student from Britain, to tackle the polarity timescale problem from Down Under.
The race was on. Over the next few years the lead bounced back and forth across the world as first the Californians, then the Australians, published new results, new age estimates and ever more evidence for sequences of field reversals. The story unfolded in a series of letters and papers, published mostly in Nature and Science, that described successive versions of what came to be called the Geomagnetic Polarity Time Scale, GPTS.
In 1963 Cox and Doell acquired samples from six Californian lava flows that Dalrymple had been dating using his high-precision, argon-free, mass spectrometer. They tested the suitability of the samples for paleomagnetic work and, when satisfied, measured the directions of their remanent magnetization. These six results, together with three from Italian lava flows published earlier by Martin Rutten, professor of geology at the University of Utrecht in The Netherlands, would form the basis of the first accurately dated polarity timescale.
Uncertainties in the estimates of age were about five percent, small enough that the samples could be unambiguously ordered. Three normally magnetized flows were dated at between 0 and 0.98 million years, three reversed flows at between 0.99 and 1.69 million years, and another three normal flows at between 2.4 and 3.2 million years. Although Cox and his colleagues were careful to point out that other scenarios were possible, their results supported the suggestion that first Matuyama and then Hospers had made— namely that the last reversed-to-normal polarity change had occurred sometime in the early Quaternary period, about a million years ago, and there had also been an earlier period of normal polarity in the late Tertiary period. They provisionally labeled the present period of normal polarity N1, the earlier one N2, and the intervening period of reversed polarity R1, and tentatively suggested that reversals might occur regularly, roughly every one million years.
Later in 1963, McDougall and Tarling announced that the dates and polarities of their Hawaiian samples were consistent with the results of Cox, Doell and Dalrymple, except for one reversely magnetized flow that they had dated at 2.95 million years, with an uncertainty of 0.06 million years, placing it in the middle of the late Tertiary normal interval N2.
The following year any semblance of tidy regularity disappeared for good when Cox and his colleagues published their next version of the timescale. Two short “events” had broken into the long polarity “epochs,” as well as an earlier reversed interval, R2. At Mammoth Lakes in California, they had found a reversely magnetized lava flow, which they dated at about 3.1 million years. Uncertainties in their estimate of age meant they could not be sure whether their Mammoth Lakes result matched McDougall and Tarling’s rogue one, or whether there had been two separate events: for some time a question mark would remain over the Mammoth event.
Early steps in unraveling the history of geomagnetic polarity reversals. In 1963 Allan Cox and his team suggested there had been alternating intervals of normal and reversed polarity, with each spanning about one million years. Later studies resolved shorter polarity “events” within Cox’s original “epochs” and improved the ages of the polarity boundaries. By 1966 the timescale bore little semblance of regularity. On the right is a modern version of the geomagnetic polarity timescale for the past six million years, with the main epochs named after famous geomagnetists, and the shorter events after the locations where they were first identified.
The 1964 version of the timescale also included a famous “event” that had first been noticed in Olduvai Gorge in Tanganyika (today’s Tanzania). The excavations at this East African location were well known for their early hominid fossils, for which accurate and precise dating was important. Between the fossil-bearing sediments, a volcanic “tuff” layer, originally dated at 1.9 million years, was found to be normally magnetized when “it should have been reversed,” according to Sherman Grommé and Richard Hay who reported it.
The possibility of self-reversal was never far from the minds of the early paleomagnetists, and at first it seemed that the Olduvai Gorge tuff might be another rare example to rank alongside Nagata’s Haruna lava. Cox and his coworkers now instigated a strict regime of tests to validate their results. To rule out self-reversal, they heated their samples in the absence of a magnetic field and cooled them again in the laboratory field. To identify and remove unstable secondary magnetizations, they progressively demagnetized the samples in alternating magnetic fields of increasing strength. To check for consistency, wherever possible they sampled different parts of a flow and baked contacts.
/>
By the end they had found no evidence of self-reversing rocks and no flows with ambiguous results. Each appeared to indicate a field of either normal or reversed polarity: the samples from Olduvai Gorge were unequivocally normal. Cox and his colleagues were happier still when they found a second record from a volcanic island in the Bering Sea that corroborated the Olduvai event.
By now the polarity timescale was made up of results from North America, Hawaii, Europe and Africa, ordered chronologically and interleaved. The regular polarity epochs were punctuated with just two short events: the Olduvai, 1.9 million years ago, and the Mammoth, about three million years ago.
The consistency of the results from three continents, together with the failure to find any evidence of self-reversal, provided the proof needed to convince almost all remaining doubters that the polarity of Earth’s main magnetic field had indeed reversed several times in the recent history of the planet. Cox and his team surreptitiously slipped in the suggestion that the epochs be named after pioneers in the field of geomagnetism, and so N1, R1, N2 and R2 became known respectively as the Brunhes, Matuyama, Gauss and Gilbert, names that have since survived all subsequent attempts to regularize them.
Cambridge University has always been something of an intellectual crossroads, and it was a fortuitous meeting of minds there in early 1965 that would give Fred Vine the incentive to develop the Vine-Matthews-Morley hypothesis to fruition. For eighteen months the idea had languished. Vine had been at a loss as to how to continue. He had been unable to find good magnetic anomaly records over ocean ridges that were believed to be spreading centers, and no reliable geomagnetic polarity timescale had been available for comparison. The patterns of magnetic anomalies recorded in the northeast Pacific were tantalizing, but it was not obvious where, or even whether, an active spreading ridge lay at their heart.
When Tuzo Wilson, a geophysicist from the University of Toronto, and Princeton’s Harold Hess arrived in Cambridge, both on sabbatical leave, almost everyone was on vacation. Fred Vine was holding the fort at the Department of Geodesy and Geophysics, and coffee-time conversation between the three naturally turned to seafloor spreading and associated questions.
Vine was to recall later that Harold Hess gave him a much-needed boost of confidence. “One of the first things he said was that he thought the Vine-Matthews hypothesis was a fantastic idea. No one had ever said that to me before,” Vine would tell William Glen, a geologist and science historian who researched the story for his 1982 book The Road to Jaramillo.
Tuzo Wilson would provide the last crucial piece of the puzzle. It was now known that new crust formed along ocean ridges, and that old crust returned to the mantle at trenches. However, something more was needed to make the whole process work on a spherical globe. Wilson postulated that Earth’s surface was a series of mobile plates, bounded by the ridges and trenches, and that between them were what he called “transform faults” along which the plates slid sideways, parallel to the faults.
These were a new-found type of fault, and the direction of motion along them would prove whether or not Wilson’s concept of “plate tectonics” was correct. But what interested Vine was Wilson’s suggestion that there was a whole series of these transform faults in the eastern Pacific Ocean, offsetting segments of the East Pacific Rise. In particular, the Rise underwent a major offset along the San Andreas transform fault, and in the northeast Pacific it consisted of three relatively short ridge segments, each offset from the other by more transform faults: the Juan de Fuca Ridge, named by Vine and Wilson, and the later-named Gorda and Explorer Ridges.
Wilson’s breakthrough sent Vine straight back to his computer. He now knew exactly where to place the spreading centers in his seafloor models. To calibrate for age he used the latest geomagnetic polarity timescale of Cox and his colleagues—major reversals at 1.0, 2.5 and 3.4 million years ago and short events at 1.9 (the Olduvai) and possibly 3.0 (the Mammoth) million years ago.
To start with, Vine computed the magnetic anomaly sequences you would expect if the seafloor were spreading at the rates of one centimeter a year and two centimeters a year. The first was too slow. The second, although too fast, did reproduce many of the details that were observed in the actual profiles. By judiciously adjusting the boundaries of the strips of normally and reversely magnetized seafloor in his model, he came up with a “best fit.” However, to reconcile his model with Cox’s timescale he had to accept significant variations in the rate of spreading, particularly around one million years ago. Although he tried to justify these variations, he was clearly not happy with them. However, he was not confident enough to suggest there might be a problem with the timescale.
Reading Vine’s paper “Spreading of the Ocean Floor: New Evidence,” published in Science in December 1966, you can imagine him cursing his prior lack of confidence. In the new version of the polarity timescale they released that year, Doell and Dalrymple had revised the age of the Matuyama–Brunhes boundary from 1.0 million to 0.7 million years, and added a new normal polarity event, the Jaramillo, near the top of the now expanded Matuyama Epoch, at 0.9 million years.
In effect, an interval of reversed polarity had been inserted into the old scheme between 0.9 and 0.7 million years ago. With this addition, Vine’s model for the Juan de Fuca Ridge system matched the geomagnetic polarity timescale with an almost constant spreading rate of 2.9 centimeters per year.
By now, Vine had left Cambridge and was a member of the teaching staff at Princeton University. His paper was a summary of his landmark PhD thesis—but it also drew on a remarkable new marine magnetic record from the South Pacific Ocean. From September to November 1965 the USS Eltanin, a United States naval ship operating under the auspices of the National Science Foundation, had collected magnetic and seismic data over the Pacific–Antarctic Ridge system. In particular, it had made measurements along two lines, Eltanin -20 and -21, to complement a previously measured line, Eltanin -19. In The Road to Jaramillo, William Glen writes of Vine’s excitement at spotting these Eltanin records when he visited Jim Heirtzler at the Lamont laboratory early in 1966. He instantly recognized the pattern of wiggles: it had been imprinted on his memory during long hours of work on the Juan de Fuca records. Here was inescapable, independent confirmation that the seafloor carried a record of the sequence of geomagnetic polarity reversals.
Vine requested and was given a copy of the Eltanin -19 profile by Heirtzler—unbeknown to the two young research students, Walter Pitman III and Ellen Herron, who had taken part in the Eltanin cruise. The students had just finished laboriously processing the Eltanin magnetic data, and had immediately realized its importance. Naturally, they were furious that at that very moment their supervisor had handed over their treasured discovery to their rival. Only after much heated correspondence was an agreement eventually hammered out with the publishers of Science: Pitman and Heirtzler’s presentation of the Eltanin data and their interpretation would appear two weeks before Vine’s “Spreading of the Ocean Floor: New Evidence.”
Vine’s paper included seafloor data from around the world. It all told the same story. At last scientists could unravel Earth’s magnetic history millions of years back into geological time, and move the debate on the source of this magnetism to an exciting new level.
Vine began by looking at a detailed low-altitude aeromagnetic survey of the Reykjanes Ridge, the part of the mid-Atlantic Ridge immediately south of Iceland. In the course of the survey, forty-nine parallel crossings had been made over the ridge crest, each separated by about eight kilometers. Vine showed profiles from four of these and the correlation was remarkable, as was the excellent fit with his computed models. This data had been collected at Heirtzler’s suggestion, but ironically when Heirtzler and his French coworker Xavier Le Pichon had first published it in June 1966 they had used it to argue against the Vine-Matthews-Morley theory. The obvious symmetry of the magnetic anomalies about the ridge axis convinced Vine that Heirtzler and Le Pichon were wrong. But in char
acteristically modest fashion, he simply described the remarkable correlation as “encouraging.”
A chart compiled from a marine magnetic survey over the Reykjanes Ridge, part of the mid-Atlantic Ridge to the south of Iceland. Black and grey shading show areas where the magnetic field is stronger than expected, above normally magnetized seafloor. White shading represents areas of lower than expected field intensity above reversely magnetized seafloor. The whole pattern is symmetric about the ridge axis.
He went on to show equally remarkable fits to magnetic anomaly profiles observed over almost all the known spreading ridges: the South Atlantic, the Carlsberg in the northwest Indian Ocean, the Red Sea, an update of the Juan de Fuca, and the Eltanin -19 profile from the East Pacific Rise. He estimated spreading rates that varied from one centimeter a year for the Reykjanes Ridge and the Red Sea to 4.4 centimeters a year for the East Pacific Rise. A faster spreading rate meant that a polarity interval was represented by a wider strip of seafloor, and so even the shorter polarity events were more likely to be detected by a ship-borne magnetometer some 2000 or 3000 meters above the seafloor. In particular, the Juan de Fuca and East Pacific Rise profiles revealed considerably more detail than the others.
No longer lacking confidence, Vine now argued that the seafloor anomalies held the key to extending the geomagnetic polarity timescale: the process of seafloor spreading was laying out a complete history of Earth’s magnetic field across the seafloor like a giant barcode in black and white, normal and reversed. Once you knew the spreading rate, you could read the code.
He tested this idea by assuming that his spreading rate of 4.4 centimeters a year was good for the whole 500 kilometers of the Eltanin -19 profile, and decoded the entire record. It turned out to be almost identical to the one Pitman and Heirtzler had published two weeks earlier. Apart from some slight revisions in dates, this timescale, covering the past eleven million years, is still accepted today.
North Pole, South Pole: The Epic Quest to Solve the Great Mystery of Earth's Magnetism Page 18