Encyclopedia of Geology. Editor: Richard C Selley, et al., Volume 3, Elsevier, 2005.
The theory of plate tectonics postulates that the Earth’s outermost layer, the lithosphere, which consists of the crust and the uppermost mantle, is about 100 km thick and is broken into rigid plates that slowly move and change their configuration in response to thermal instabilities in the mantle. The theory holds that the ocean floors are youthful due to their continual creation at spreading ridges and destruction as they plunge back deep into the mantle. The continents, in contrast, are buoyant sialic blocks with components of all ages, which ride passively on the surfaces of the plates. This theory, established in 1968, has unprecedented power for making quantitative calculations of past crustal motions and predictions of future ones. The immediate precursor of plate tectonics was the hypothesis of seafloor spreading proposed separately by two scientists in 1960 and 1961. Seafloor spreading was based primarily on information provided by geophysical explorations of the ocean basins during and after World War II. This article reviews the discoveries that led to plate tectonics, and the changing climates of opinion that led to a new era in the geo-sciences.
Post-War Explorations of the Ocean Basins
The geology of the ocean basins, which occupy more than 60% of the Earth’s surface, remained largely unknown until after World War II. Despite the early hypothesis of continental drift, a majority of geologists regarded the ocean basins, particularly the Pacific Basin, as primordial features, formed when the Earth’s crust first cooled. In 1946 the Dutch geologist, Philip Kuenen, calculated that if the Earth is 3 billion (3 × 109) years old, the ocean floors should be covered with a layer of sediments 5 km thick. In 1956, when radiometric dating showed the Earth to be ~4.55 billion years old, many expected the sediments to be much thicker. Indeed, as late as 1958, some of the scientists who were planning Project Mohole, to drill to the Mohorovičić discontinuity through the Pacific floor, anticipated that the cores would reveal the entire sequence of sediments deposited during the Lipalian interval from the end of the Precambrian to the beginning of the Cambrian Period—a time of severe erosion on the continents.
In the latter 1940s, the United States Navy provided generous funding for oceanographic research including ships, state-of-the-art equipment, laboratories, and scientists. Two of the four institutions credited with founding plate tectonics began sea-going explorations: the Scripps Oceanographic Institute at the University of California in San Diego, and the Lament Geological Observatory (now the Lamont-Doherty Earth Observatory) at Columbia University in New York. The other two institutions contributing to the establishment of plate tectonics were the Department of Geophysics at Cambridge University in England, and Princeton University.
Preeminent scientists directed both of the oceanographic institutes: Roger Revelle at Scripps, and Maurice Ewing at Lamont. Both began intensive programmes of mapping the submarine topography by echo depth soundings, displaying the layering of the ocean floors by seismic profiling, and collecting samples by piston coring. Scientists aboard Scripps vessels also measured heat-flow values, while those from Lamont focused more on measuring gravity at sea. Both organisations towed magnetometers behind their ships. At the beginning, these scientists were not testing hypotheses; they were performing the first comprehensive investigations of the last unknown domain of the Earth. There were many surprises in store for them.
The ship Horizon, which left Scripps in July 1950, with Revelle and several other now-famous scientists aboard, made its first discovery 300 miles south-west of San Diego. There, in deep water, its companion Navy ship fired the first explosives and Horizon’s seismic profiler showed that the Pacific sediments were not 5 km or more thick: they were only 260 metres thick. The profiler detected three layers in the ocean floor: Layer 1, consisting of sediments ~0.25 km thick; Layer 2, of consolidated sediments or volcanics or both, 1-2 km thick; and Layer 3, the authentic bedrock of the ocean floor, which ultimately was shown to maintain a uniform thickness of ~4.5km throughout the ocean basins.
From the beginning, the echo soundings showed the ocean floor to be unexpectedly crowded with hills, 10-30 or more km long, 2-5 km wide, and 50-500 m high. Further explorations showed that these so-called ‘abyssal hills’ occupy more than 30% of the floors of all oceans, placing them among the Earth’s most abundant topographic features.
Revelle and Arthur Maxwell measured heat-flow from the ocean floor and observed, to their astonishment, that the values averaged within 10% of those from the generally granitic and consequently more radioactive continents. For want of other possibilities, they suggested that the heat in the ocean floor must be carried there by rising limbs of convection cells in the mantle, an idea compatible with the earlier ‘drift’ model proposed by Arthur Holmes in the 1920s.
In the mid-Pacific, Horizon dredged the surfaces of guyots—volcanic cones with flat summits, submerged 1-2 km below sea-level. These curious features had been discovered and named by Harry Hess, of Princeton University, who located 160 of them while he was on naval duty during World War II. At first, Hess assumed they were Precambrian islands, erupted long before reef-building corals evolved, that had been eroded at sealevel and then drowned by the rising of water due to the deposition of 2-3 km of oceanic sediments. The rare opportunity to study and describe the dredged materials fell to Edwin Hamilton, to whom Robert Dietz, on the Horizon, had assigned this topic for his PhD thesis. The hauls brought up fragments of Cretaceous corals, only about 100 my old. Suddenly, the ‘ancient’ ocean basins were seen to be much younger than the continents.
In the final days of the cruise, the Horizon steered a course to allow the youthful Henry Menard, of the US Naval Electronics Laboratory, to determine the nature of the Mendocino escarpment, a long narrow feature, striking westward from Cape Mendocino in California. On a S-N traverse of ~60 km, the echo-sounder showed that, starting in deep water the ship passed over a low ridge tilted to the south, a narrow trough, a steep scarp 18 km high, a high ridge tilted to the north, a low swale, and then it arrived in relatively shallow water. This ‘fracture zone’ was a new class of submarine structure of which Menard and others would find several similar examples farther south in the eastern Pacific Ocean. They appeared to be strike-slip faults with a component of normal faulting, but they terminated before slicing into the American continents. Eventually, fracture zones, long and short, would be found in all the ocean basins.
Meanwhile, Lamont ships had begun producing the earliest detailed topographic charts of the mid-Atlantic Ridge and its linkage to what would prove to be a world-encircling succession of submarine mountain ranges, with various spurs, nearly 60000 km long, 1-3000 km wide at the base, and 2 km high, with peaks rising to 4 km above the ocean floor. In 1956, Maurice Ewing and his colleague, Bruce Heezen, reported finding a deep rift valley (see Tectonics: Rift Valleys) occupying the ridge crests. Subsequently, the rifts proved to be discontinuous but, where present, they were zones of shallow-focus seismicity, strong magnetic anomalies, and higher than average values of heat-flow. The ridges are offset horizontally into segments by transverse fracture zones, which were assumed to be strike-slip faults.
The Navy prohibited publication of contoured bathymetric charts, but two Lamont scientists, Heezen and Marie Tharp, began drawing ‘physiographic diagrams’ showing how the ocean floors would look with the waters drained away. These diagrams, published in the 1960s, astonished both non-scientists and scientists, who found them remarkably useful for interpreting the structures and history of the ocean basins. Heezen and Tharp both won gold medals for their work.
The East Pacific Rise (or Ridge) is unique in being a huge bulge (nearly 13 000 km long, 3000 km broad, and nearly 4 km high) with no sharp crest, no central rift, and no mid-ocean position—it lies close to the margins of the American continents. In 1960, Menard proposed that convection currents had moved the northern portion of the Rise beneath western North America where its presence would account for the high plateaus of Mexico and Colorado and the Tertiary Basin and Range topography of Utah and Nevada. It followed that the floor of the eastern Pacific Ocean, cut by the great E-W fracture zones, was the western flank of the buried Rise.
Many of the observations, made in the 1950s, suggested youthful, mobile ocean floors, but contrary explanations were offered in each case, puzzling scientists who were trying to piece together the dynamic history of the Earth. Then in 1960 and 1961, Hess and Dietz independently but simultaneously, proposed hypotheses that addressed global dynamics in terms of moving seafloors driven by convection in the mantle.
Sea-Floor Spreading: 1960 and 1961
In 1960, Hess circulated the preprint of a book chapter in which he described the ocean floors as the exposed surface of the mantle. He proposed that large-scale convection cells in the mantle create new ocean floor at the ridges, where the rising limbs diverge and move to either side until they cool and plunge down into the mantle at the trenches or beneath continental margins. He now explained guyots as volcanic peaks eroded on oceanic ridge crests and carried to the depths by the moving seafloor. Where limbs rise beneath continents they split them apart and raft the granitic fragments in opposite directions until grounding them over zones of down-welling. Hess suggested that inasmuch as granite is too buoyant to sink into the mantle, some fragments of continents have survived since the beginning of geologic time, while the ocean floors have been swept clean and replaced by new mantle material every 300-400 my.
Hess further theorized that the mantle consists of peridotite, an olivine-rich rock that becomes hydrated to serpentinite at the ridges by reaction with heated waters released from depth. He favoured serpentinite partly because of its ease of recycling and partly because he believed it would be impossible to achieve the uniform 4.5 km thickness of the ocean floors with dykes and lava flows. Hess acknowledged that mantle convection was regarded as too radical an idea to be widely accepted by geologists and geophysicists, but he pointed out that his model would account for many phenomena in a coherent fashion: the formation of the ridge-rift system and the trenches, the youth and uniform thickness of the ocean floors, the thinness of pelagic sediments, and moving continents. Hess called his chapter an essay in ‘geopoetry’.
In 1961, Dietz published a three-page article in Nature in which he, too, argued that convection in the mantle, moving at a few centimetres per year, could produce the overall structure of the ocean basins: the ridges form over sites of rising and diverging limbs; the trenches form at sites of converging and down-welling limbs, and the fracture zones are shears between regions of slow and fast creep. Dietz suggested that the mantle consists of eclogite, a dense pyroxene-garnet rock, and that the ocean floors are built of basaltic dykes and pillow lavas, formed by the partial melting of the eclogite at the ridges. But he added that the actual rock compositions were of less importance than the fact that the ocean floors must be recycled mantle material. Dietz pointed out, as had Hess, that rising convection currents rift continents apart and carry the sialic fragments en bloc to sites of down-welling and compression, where folded mountain ranges form on their margins. He added that the pelagic sediments ride down into the depths on the surfaces of the plunging oceanic slabs where they are granitized and welded onto the undersides of the continents—thus contributing to the persistent continental free-board, despite steady erosion of the continental surfaces toward base level.
Dietz called this process ‘sea-floor spreading’. This concept, he argued, requires geologists to think of Earth’s outer layers in terms of their relative strengths, so we should begin referring to the ‘litho-sphere’, a historic name for Earth’s outermost layer, which is relatively strong and rigid to a uniform depth of about 70 km (now generally taken as 100 km) under both continents and oceans. Beneath the litho-sphere lies the weaker, more yielding, ‘asthenosphere’ on which the lithosphere moves in response to convection currents. The asthenosphere had previously been hypothesized on the basis of seismic evidence.
Dietz made the first suggestion (later confirmed) that the oceanic abyssal hills are a chaos topography, developed when strips of juvenile sea floor have ruptured under stress as the floors move outward. Finally, he referred to two papers in press, one by Victor Vacquier et al., and one by Ronald Mason and Arthur Raff, reporting the discovery of linear magnetic anomalies on the Pacific floor. Some of the magnetic lineations appeared to be offset by up to 1185 km along the Mendocino fracture zone. Dietz suggested that the lineations are developed normal to the direction of convective creep of the ocean floor. Noting that neither the fracture zones nor the magnetic lineations impinge upon the continental margins, he suggested that both lost their identity as the Pacific floor slipped beneath the American continents.
Rarely are ideas subsequently seen as basic to a grand new system of thought in science, appreciated at full value when they first appear. That was the case with sea-floor spreading, which failed to catch the attention of more than a few readers at a time when most geologists and geophysicists were unprepared to take seriously the idea of mantle convection, much less that of the new notion of spreading seafloors. However, Dietz’s paper elicited a favourable letter to Nature from J Tuzo Wilson at the University of Toronto, and it inspired Ewing to redirect a large portion of Lament’s research into testing the sea-floor spreading hypothesis. Ewing outfitted two ships with upgraded seismic reflection profilers to measure the thickness of pelagic sediments across the oceans. He was to find virtually no sediments on the ridges and a modest thickening of the layers towards the edges of the continental shelves. Ewing took the earliest photographs of deep sea sediments and observed ripple marks, which, until then, had been used as diagnostic of shallow waters.
In the next few years, as data favourable to sea-floor spreading accumulated, a regrettable negative reaction developed toward Dietz. Even though his model differed from that of Hess, the belief spread that he had ‘stolen’ Hess’s basic idea and rushed it into print under his own name. This impression still persists to some degree, even though Dietz’s choice of basaltic rather than serpentinized ocean floors is the one universally accepted today.
In 1986, Menard reviewed this controversy in his The Ocean of Truth, a personal account of the sequence of events that led to plate tectonics. Menard, who had refereed the pre-publication manuscripts of both Hess and Dietz, wrote that he felt certain—and said so at the time—that Dietz was unaware of Hess’s preprint when he wrote his own paper. Nevertheless, Menard believed that only one person can have priority for an idea and it should be Hess, so he urged Dietz, if only for appearances sake, to add a footnote to his next publication on seafloor spreading conceding credit for the idea to Hess. Dietz did so, and the two men made their peace in print. Subsequently, Menard came to realise that, in fact, more than one person can be struck with the same idea, particularly at a time when new data are coming in and are being freely discussed. He cited several additional examples that took place during the race towards plate tectonics.
Finally, Menard remarked that the priority for this idea probably should go to Arthur Holmes in Britain, who had favoured convection in the mantle as a ruling factor in global tectonics from the late 1920s until he died in 1965. Holmes described a basaltic layer which becomes a kind of endless travelling belt as it moves from ridges to trenches carrying continental fragments along with it. Over the years, Holmes changed his diagrams somewhat, but they all show the limbs of a convection cell rising under a continental slab, which is stretched and pulled apart into two fragments that are rafted to either side leaving behind new ocean floors. Holmes did not depict the creation of new ocean floor at a spreading oceanic ridge; he showed the ridges as being wholly or partially sialic. In that important respect his idea differed from those of Hess and Dietz. Nevertheless, Holmes’s basic model was, without question, a predecessor of sea-floor spreading.
History books do not necessarily aid us in resolving disputes. In 1973, Allan Cox of the US Geological Survey omitted Dietz’s article from his collection of the landmark papers that led to plate tectonics. And in the entry on Hess in Volume 17 of The Dictionary of Scientific Biography, published in 1990, we read that his hypothesis of sea-floor spreading was the most important innovation leading to plate tectonics, but that it was given its name by Dietz who, “with Hess’s preprint in hand”, published the first article on it in 1961. Fortunately, others knew better. In 1966, the Geological Society of America would present its highest honour, the Penrose Medal, to Hess for his research on the petrology of ultramafic rocks and for his provocative tectonic hypotheses including that of the spreading ocean floor. In 1988, the Penrose Medal went to Dietz for his “world-class, innovative contributions in three divisions of the geosciences: sea-floor spreading, recognition of terrestrial impact structures, and the meteorite impact of the Moon’s surface”.
Magnetic Anomalies on the Ocean Floors: 1961, 1963
In August, 1961, the paper by Mason and Raff, of which Dietz had seen a preprint, was published by the Geological Society of America along with a companion paper by Raff and Mason that included a particularly striking diagram showing long, narrow, mostly vertical black and white ‘zebra stripes’ of alternating high and low magnetic intensities measured on the floor of the north-eastern Pacific Ocean. The authors suggested that the contrasting intensities might reflect structural ridges and troughs or a system of sub-parallel dykes, but subsequent topographical and gravity surveys failed to detect either one. This research project, which proved to be momentous, was undertaken when Mason, a visitor to Scripps, casually asked a seismologist over morning coffee if anyone had thought of towing a magnetometer behind a ship to gather data while the ship was engaged in other operations. Revelle overheard the question and offered Mason the assignment, then and there. Mason soon learned that Lamont already had towed a magnetometer behind a ship in the Atlantic, and he could borrow it pending acquisition of an instrument by Scripps.
One reader, Lawrence W Morley of the Geological Survey of Canada, who had conducted extensive aero-magnetic surveys over lands and seas, remained mystified by this pattern for nearly two years until he discovered Dietz’s paper on sea-floor spreading. Morley immediately wrote a short article and submitted it to Nature in February 1963, proposing a test of sea-floor spreading. He argued that the magnetic stripes form at the crests of spreading ridges where the erupting lavas acquire the magnetization of the Earth’s ambient field (a detail Dietz had not specified). Today, the Earth’s magnetic field is north-seeking, but in the 1950s it had become clear that some rocks have cooled during periods of south-seeking polarity, so Morley proposed that the parallel stripes on the moving sea-floor record the periods of normal and reversed polarity. And, inasmuch as Cretaceous rocks were the oldest yet recovered from the ocean floors, he speculated that such stripes have recorded the history of the ocean basins for the past 100 my or so. Morley tentatively calculated the rates of spreading and lengths of reversal periods. Nature rejected his paper saying they had no space available. Morley then sent it to the Journal of Geophysical Research, which kept it for some time and then rejected it, with a message from one referee saying that such speculation was more appropriate to cocktail party chatter.
On 7 September 1963, Nature published the now famous paper by Frederick Vine and Drummond Matthews of Cambridge University: ‘Magnetic Anomalies over Oceanic Ridges’. Like Morley, Vine and Matthews assumed that convection, sea-floor spreading, and reversals of magnetic polarity all occur, and that the ocean floors consist of basalt that becomes strongly magnetized at the ridge crests. Today, the Vine-Matthews paper is widely seen as the founding paper of plate tectonics, but at the time it was poorly received and largely ignored for the next three years. It was not even included by their Department of Geophysics at Cambridge University in its list of important contributions for 1963! Years later, when attitudes changed, and Morley’s story came out, the Earth science community began to speak of the Vine-Matthews-Morley (VMM) hypothesis. Morley then realized that he had gained more fame by having his paper rejected than he would have by its publication.
In 1963, all three of the basic assumptions listed by Vine and Matthews were suspect to geoscientists. Professor Harold Jeffreys, at Cambridge University, argued in every edition of his book, The Earth, beginning in 1924, that the mantle of the contracting Earth is too stiff to allow for convection. In his sixth and final edition of 1976, he added several pages arguing against plate tectonics. In 1963 and again in 1964, Gordon MacDonald, of the Institute of Geophysics and Planetary Physics of the University of California, published a detailed paper, The Deep Structure of the Continents’, in Reviews of Geophysics and a short version of it in Science. He intended this paper as a death-blow to continental drift and sea-floor spreading. MacDonald argued that isostasy prevails; thousands of measurements made on the surface and by satellites show that gravity over continents is equal to that over oceans—despite marked difference in their compositions and densities. This, together with the equal values of heat-flow determined worldwide indicated that the continents must have formed by vertical segregation of the mantle directly beneath them, which, therefore, must differ from the mantle under ocean floors to depths of 400 to 700 km. MacDonald concluded that no significant horizontal motion, due to convection or any other process, has occurred.
Magnetic Field Reversals, Isotopically Dated: 1964
Since early in the twentieth century, many advances had been made in techniques of measuring the remanent (permanent) magnetization of rocks and interpreting the results. In the 1950s, polar wandering curves had yielded evidence in support of continental drift. Nevertheless, inasmuch as Earth’s magnetic field varies in alignment, intensity, and polarity, many geoscientists still viewed palaeomagnetism as an incomprehensible property, studied by ‘black box’ techniques, and scarcely to be trusted. However, by the early 1960s, measurements had begun to show that whereas recent basaltic lavas from widely spaced sources behave as north-seeking compasses, those of Early Pleistocene age behave as south-seeking compasses. Clearly, reversals of the magnetic field have taken place and the search was on for reliable means of dating them.
In June 1964, Allan Cox, Richard Doell, and Brent Dalrymple of the US Geological Survey at Menlo Park, California, published a paper entitled, ‘Reversals of the Earth’s Magnetic Field’, in which they reported that they had dated rocks of normal and reversed magnetism by the K/Ar method. Their results documented the occurrence of two epochs of normal polarity and one of reversed polarity during the past 3.5 my. They named the epochs in honour of pioneers of palaeomagnetic studies: the Bruhnes normal epoch, from the present to 1.0 my ago; the Matuyama reversed epoch from 1.0 to 2.5 my ago; and the Gauss normale poch from 2.5 to 3.4 my ago. Both the Matuyama and the Gauss epochs were interrupted by short ‘events’ of opposite polarities—named, respectively, the Olduvai and the Mammoth events—each lasting 250-300 000 years. Later, in February 1966, Doell and Dalrymple revised this time-scale by shortening the present Bruhnes normal epoch to 700 000 years and adding the newly discovered ‘Jaramillo normal event’, which occurred about 0.85 my ago in the early part of the Matuyama reversed epoch. Since then, additional epochs and events have been dated for the past 180 my, back to the mid-Jurassic Period.
Cox and his group dated their basaltic samples using a mass spectrometer of the type designed in the late 1950s by John Reynolds of the University of California at Berkeley, which opened a new age in isotopic geochemistry. Reynolds’ all-glass instruments could be heated to evacuate all traces of atmospheric argon in order to yield accurate measurements of the small amounts of argon produced by the radioactive decay of potassium. The work by Cox and his colleagues persuaded many scientists that palaeomagnetism must be taken seriously. Once again, however, another group was doing the same research. Within weeks of its appearance, Nature published ‘Dating Polarity Geomagnetic Reversals’ by Ian McDougal and Donald Tarling, who had established a palaeomagnetism laboratory at the Australian National University, and were friendly rivals of the group at Menlo Park.
Transform Faults: 1965
In June 1965, Nature published a paper by Wilson entitled: ‘A New Class of Faults and their Bearing on Continental Drift’, in which he proposed what he called ‘transform faults’ as a test of sea-floor spreading. These are faults with large horizontal movements that appear, along with their seismic activity, to terminate abruptly. He wrote that instead of just terminating, these faults are ‘transformed’, at each end, into mid-ocean ridges, mountain ranges, trenches, or island arcs, which together make up the network of mobile belts that divide the Earth’s surface into large, rigid plates. Wilson cited the San Andreas as a transform fault on land with its southern end connected to the oceanic ridge in the Gulf of California and its northern end to one north of Cape Mendocino. But his clearest examples of transform faults were those that connect offset segments of spreading oceanic ridges. All the slabs of ocean floor on each flank of a ridge move side-by-side down-slope and across the seafloor. Nevertheless, along the transform fault plane connecting the ridge crests, and only there, the rock walls move in opposite directions. Wilson noted that in 1963 Lynn Sykes at Lamont had shown that along the fracture zones cross-cutting the ridges, seismic activity is strictly limited to these transform fault planes. Wilson attributed this to the relative motions of the rock walls and also to the marked differences in temperature, elevation, and age of the rocks across the fault—where hot rock beneath a ridge is juxtaposed with a cold wall of much older seafloor. How did the ridges become offset into segments? Wilson and Sykes both argued that the offsets formed in place when the lithosphere first split open at the beginning of the present cycle of ridge building. No strike-slip faulting separated them, nor has there been any along the fracture zones, which now were recognized as topographic continuations of transform faults that became aseismic when spreading carried them beyond the ridge crests.
Wilson’s paper led to a major change in the history of structural geology. Its reception was mostly positive as geoscientists looked at their problems from the new perspective. But in March 1965, Alan Coode, at Newcastle University, had sent a short note illustrated with a block diagram of ridge segments connected by a transform fault (although he gave no new name to it) to the Canadian Journal of Earth Sciences. That journal appears quarterly, so Coode’s article was published, to little notice, a week or two after Wilson’s paper appeared in Nature. Great honours would be showered on Wilson; none on Coode, although he had submitted his paper first.
The Eltanin Profile: 1966
Sea-floor spreading got off to a slow start partly because neither the Vine-Mathews nor the Mason-Raff papers showed obvious symmetry of linear magnetic anomalies across oceanic ridges. Early in 1966, James Heirtzler, Xavier Le Pichon, and J G Baron at Lamont published the results of an aeromagnetic survey over the Reykjanes Ridge, south of Iceland, which yielded a beautifully symmetrical pattern of magnetic stripes across a spreading ridge. They computed the spreading rate at slightly less than 2 cm per year of separation.
In December 1966, Walter Pitman III, and Heirtzler, at Lamont published the magnetic profiles recorded by the research vessel, Eltanin, during four passes over the Pacific-Antarctic Ridge, south of Easter Island. All four profiles gave similar results but the most southerly one, Eltanin-19, was so spectacularly symmetrical, both in the topography of the ridge and its record of magnetic reversals that it seems to have prompted mass conversions to sea-floor spreading, first at Lamont and then elsewhere. The profile showed each of the dated magnetic epochs of the past 3.4 my, for which it yielded a computed spreading rate of 9 cm per year of separation. Assuming a constant spreading rate within 500 km on either side of the ridge, the profile made it possible to extend the series of geomagnetic reversals back to 10 my. Pitman and Heirtzler documented a good match between the anomalies in the South Pacific and those on the Reykjanes Ridge, adjusted for the slower spreading rate in the Atlantic. Today the Eltanin-19 profile is ranked as one of the most important pieces of evidence in the history of geophysics. And further confirmation was at hand. Approaching the problem by an independent method, Neil Opdyke and his colleagues at Lamont, plotted the magnetic polarities of fossiliferous strata from deepsea cores of the South Pacific floor. Early in 1966, they found a definitive match with the dated record in the basaltic oceanic bedrock.
In February 1966, Vine visited Lamont where Heirtzler gave him a copy of the Eltanin-19 profile. Vine incorporated it into his paper ‘Spreading of the Ocean Floor: New Evidence’, which appeared in Science the following December, shortly after one by Pitmann and Heirtzler. In it, Vine presented six symmetrical profiles of magnetic anomalies across ridges in the Atlantic, Indian, and Pacific Oceans, and showed that the linear anomalies on the East Pacific Rise match those on the Juan de Fuca Ridge offshore from British Columbia, even though they lie 11000 km apart. His computed spreading rates for all the ridges ranged from about 2.0 to 3.0 cm of separation per year in the Atlantic and Indian Oceans to 8.8 cm per year across the East Pacific Rise. Vine speculated that the whole history of the ocean basins in terms of ocean-floor spreading must be ‘frozen-in’ as paired magnetic anomalies in the oceanic crust. Meanwhile, Vine had given a summary of his results in November at the annual meeting of the Geological Society of America, where it startled many geologists with their first serious introduction to sea-floor spreading.
A Matching of Continents: 1966
At the same GSA meeting, Patrick Hurley, a geochronologist at the Massachusetts Institute of Technology, presented a paper, with nine coauthors, offering radiometric evidence for continental drift. After attending the Symposium on Continental Drift, organised in 1964 for the Royal Society by Patrick Blackett, Edward Bullard, and Keith Runcorn, Hurley said: “I went to London a fixist and came home a drifter”. Bullard had displayed a map that had been programmed by a computer to apply Euler’s theorem to find the best fit of the continents across the Atlantic. The map showed an especially close fit between Brazil and Ghana, where Hurley knew that a sharp contact had been mapped between two rock provinces that were 600 million years old and 2000 million years old, respectively. If Bullard’s map were applicable, the same two provinces should occur near Sāo Luis at the easternmost tip of Brazil. Hurley gathered a group of Brazilian and other collaborators who obtained rock samples from the region of Sao Luis and dated them. The results showed an excellent match between the petrology and dates of the two rock provinces now on opposite sides of the Atlantic. Hurley’s research was crucial in persuading many American geologists to accept continental drift via seafloor spreading. In America, the doctrine of the permanence of continents and ocean basins, founded by James Dwight Dana at Yale in 1846, finally expired in 1967.
Plate Tectonics: 1967-1968
Of the many papers published during the emergence in 1967 and 1968 of plate tectonics from sea-floor spreading, four of them often are cited as milestones. The earliest of the four, ‘The North Pacific: An Example of Tectonics on a Sphere’, by Dan MacKenzie and David Parker, applied the concept of transform faults to motions on a sphere on which aseismic areas move as rigid plates. They, too, applied Euler’s theorem (with which Wilson acknowledged he had been unfamiliar when he wrote of the movement of rigid plates) to the effect that any displacement of a plate on a spherical surface may be considered as a rigid rotation about a fixed vertical axis. It follows that the slip vectors of moving plates must describe small circles around the Euler pole. McKenzie and Parker used fault-plane solutions for earthquakes in the North Pacific to work out the actual directions of the slip vectors and found that they wholly conformed to the sea-floor spreading model.
The remaining three papers, all of which review the evidence in broad fields, appeared early in 1968, in Volume 73 of the Journal of Geophysical Research. In his paper entitled ‘Rises, Trenches, Great Faults, and Crustal Blocks’, Jason Morgan at Princeton presented the first depiction of the entire surface of the Earth divided into rigid blocks, each having three types of boundaries: ridges where new crust is formed; trenches or mountain ranges where crust is destroyed by sinking or shortened by folding; and great faults. His diagram showed twenty rigid blocks, large and small, identified by the seismicity at their boundaries.
Morgan’s interpretation required that the seafloors must move as rigid blocks, for which the thin oceanic crust lacks the required strength. Therefore, he envisioned the outer layer of the Earth as a rigid ‘tectosphere’ about 100 km thick, sliding over the weak asthenosphere. He argued that the location of an oceanic ridge is determined, not by the activity of some deep-seated system of thermal convection, but by the motions of the blocks themselves: wherever tensional forces fracture the tectosphere to a depth of 100 km, hot mantle material wells up between the blocks and serves as a zone of weakness. Subsequently, each new intrusion of mantle material is injected into the centre of the most recent one, thus maintaining a median position for a spreading ridge in most cases. He argued that the emplacement of new ocean floor would be reflected in the symmetry of its magnetic anomalies and from these he charted the history of motions that brought the blocks to their present positions.
In an article titled: ‘Sea-floor Spreading and Continental Drift’, Xavier Le Pichon, then a visitor at Lamont, asked whether large crustal blocks do, in fact, move for significant distances without any deformation of the sea-floor sediments or the continental interiors. He adopted a simple model of six large blocks (rather than Morgan’s twenty), analysed their slip vectors, and found the results in reasonable agreement with the globe’s physiographic, seismic, and geological data. Le Pichon then extended his analysis back over the past 65 my, looking for patterns of continental drift and sea-floor spreading. He reported evidence for three major episodes of spreading—one in the Late Mesozoic, one in the Early Cenozoic, and one in the Late Cenozoic—with a major reorganization of the global pattern occurring at the beginning of each cycle. He correlated the slowing down of each cycle of spreading with the onset of mountain-building. Le Pichon’s analysis of surface motions provided a coherent and intelligible history of interrelated plate motions over the entire globe.
Lamont’s group of seismologists contributed ‘Seismology and the New Global Tectonics’, by Bryan Isacks, Jack Oliver, and Lynn Sykes. Sykes previously had shown that not only was seismicity limited to transform fault planes but that first motions on spreading ridges also confirmed Wilson’s predictions for transform faults. Now the group reviewed the full range of seismological evidence that supports plate tectonics. The spreading oceanic ridge-rifts are loci of shallow-focus normal faulting. The East Pacific Rise, which has no rift and few earthquakes, appears to be spreading rapidly by ductile flow. Deep-focus earthquakes occur only on planes dipping beneath continents and into trenches. At first the group had been surprised to observe that these quakes generate high-frequency seismic waves which travel faster and more efficiently along the seismic zones than had been expected. This led to their realisation that the quakes are generated within slabs of rigid lithosphere as they descend into the yielding asthenosphere-the process called subduction. These seismologists concluded that the sinking of cold, heavy plates plays a dominant role in generating plate motion. Their task of analysing earthquakes globally was aided immeasurably by the establishment by the US Coast and Geodetic Survey of the World Wide Standardized Network (WWSSN) of 125 stations from which data was continually transmitted to a central location for digitising and archiving. By the mid-1960s, the data were available on microfilm. The original purpose had been to distinguish between earthquakes and underground nuclear explosions, but the Network was an invaluable source of high-quality data for seismologists. The mid-1960s also saw the development of highspeed computers and of programs, capable of reducing reams of data from continuous recordings of many kinds. These, along with other technical advances, played an essential role in developing the theory of plate tectonics.
Plate Tectonics Today
Plate tectonics rapidly became the dominant model of geoscience. It simplifies our understanding of the three principal igneous rock types that make up the Earth’s crust: the basalts of the ocean floors are derived by partial melting of the peridotite mantle and erupted at the spreading ridges; the more siliceous andesites are derived by partial melting plus dehydration of descending slabs of lithosphere and are erupted over subduction zones, where they continuously contribute to island arcs and continental margins; the still more silica-rich continental rocks arose from the growth of andesitic landmasses and their subsequent reworking by metamorphism and metasomatism to form granitic plutons and rhyolitic volcanics, with no contribution from the mantle. Beginning with the first andesitic volcanoes, the buoyant continents have grown in area throughout geologic history while the oceans floors have lost area to subduction. By one estimate, the rapidly moving Pacific floor may disappear about 200 million years from now, when North America will collide with Asia.
Our understanding of the mechanism driving plate tectonics remains incomplete. Many scientists have abandoned the idea of rolling convection cells, which would not form a coherent system beneath the segmented oceanic ridges. However, we know that the asthenosphere is close to the melting point, so some scientists argue that global stresses rent the lithosphere into the pattern of mobile belts resembling the seam on a tennis ball. To them, the cracks came first and the basalts of the spreading ridges arose from decompressional melting of the mantle. They also see the gravitational pull of the cold, heavy slabs of sinking lithosphere as a prime factor in maintaining plate motion.
But the mantle is by no means a quiescent domain. It includes very deep-seated ‘hot spots’ from which vertical plumes arise and erupt as volcanoes. Each hot spot persists for millions of years. A dramatic example is the hot spot beneath the big island of Hawaii, which has remained fixed-in-place for 60 my while the Pacific plate has moved over it carrying a long line of islands and seamounts with ages steadily increasing toward the north-west. There is even a bend in the line showing that about 46 my ago the plate shifted its motion from northwest to almost due north. We now know of more than 100 mantle plumes under continental areas, oceanic ridges, and islands. Surely these long-lived sources of magma make a powerful contribution to the thermal instabilities involved in plate motions. Today, with or without our having to specify details of the causal mechanism, the theory of plate tectonics seems secure.
Is Plate Tectonics Unique to the Earth?
Only two of the more than 100 planets that we have discovered outside our solar system have actually been seen as black spots passing in front of their suns. All the rest are inferred from irregularities in the orbits of their suns. Therefore, a search for plate tectonics outside the Earth must be limited to our four rocky neighbours in the inner Solar System—our Moon, Mercury, Venus, and Mars—and some of the moons of the outer planets. We have analysed samples collected on the Moon, and analysed those rare meteorites that have been projected to Earth by impacts on the Moon and Mars. In addition, we have landed analytical instruments on the surfaces of the Moon, Venus, and Mars. All our results show that the crusts of these bodies consist of basalts and related rocks that are akin to those of our ocean floors. As yet we have found no evidence of granitic continents.
Scientists have searched in vain on images of these bodies for patterns indicative of plate boundaries: long ridges like spreading zones, chains of folded mountain ranges, or lines of volcanoes. Our Moon and Mercury are volcanically dead and pockmarked by impact craters. Venus, beneath its thick layer of clouds, appears to have been resurfaced with fresh basalts about 5 my ago, but it shows no linear pattern of features resembling ridges or subduction zones. Mars has the largest volcanoes in the Solar System, but they are scattered over a large area, and we see no evidence there of linear mountain ranges or trenches. Even less promising are the icy moons of the giant planets. We conclude that plate tectonics is unique to the Earth.
In this, we are fortunate, because plate tectonics is one of the essential factors that enable our planet to support complex forms of life. Volcanic eruptions continually renew Earth’s waters and its atmosphere, including ozone, that shields us from lethal solar radiation, and the greenhouse gases that moderate the temperature and maintain it within a livable range for mammals. The creation of new islands and splitting and rearranging of continental areas generate a great variety of habitats which, in turn, minimizes the effectiveness of mass extinctions from whatever cause. Were plate motion to cease, mountain building would cease, and uninterrupted erosion would begin to wear down the continents. Ultimately, the deposition of sediments in the oceans would presumably raise sea-level enough to flood much of the land surface. We literally owe our lives to plate tectonics.