The standard meme of structural geology is that orogenic belts are the sites of great crustal shortening and compression. English-speaking geologists adopted this meme from the 1815 observations of Sir James Hall who modelled folded strata with a stack of paper, and attributed the narrowing of the folded stack to the cooling contraction of the Earth. This was adopted by Hutton; Elie de Beaumont, Lyell, Dana, and Bailey Willis, and so it is the meme to this day.

Contemporaneously with the contracting-Earth model, a gravity-driven diapiric model was developed by German-reading European geologists: Gillet-Lamont (1799), Scrope (1825), Kuhn (1836), Naumann (1849), Herschell (1856), and Schardt (1898), but particularly by Reyer (1888,1892, and 1894), who made thorough mathematical analyses together with dimensionally valid experiments, adjusted for scale and time.

The German work was paralleled by Italian geologists, Bombicci (1882), Bonarelli (1901), Anelli (1923 and 1925), Signorini (1936), Merla and La Paz (1943), and more German work by Steinmann, and Haarmann, by the Netherlands leader van Bemmelen, and the Swiss under Argand and Wegmann and by Beloussov and his associates in Moscow, and more recently by Ramberg's experimental work in Sweden.

Too few American geologists read the German, Russian, and Italian literature, so the English-speaking concept has prevailed. Unfortunately, American and British certainty that they must be right has dominated the scene. The future will tell!

The Himalayan blunder

The greatest error of the last two centuries of geology has been the universal belief that the Himalaya was the site of colossal crustal compression. Everybody believed it. Most still do. I was taught it, believed it, and used to teach it. It ranks with the Ptolemy - universal belief that Earth was the stationary center of the solar system. Both were wrong. Universal belief does not itself imply validity.

The great altitude of the Himalaya and the Tibetan plateau is assumed to be due to double thickness of continent, either from the Indian craton under- thrusting Asia, or the mutual crumpling of both to double thickness. The great overthrusts of the Himalaya are used as a measure of the foreshortening.

A flawed meme

Himalayan geologists, thoroughly indoctrinated with the regional paradigm of intense compression, have been astonished that, apart from the Himalayan nappes, the common deformation through the Tibetan region is not thrusts but normal faulting, (Bohlin and Norin, 1960; Ni and York, 1978; Molnar and Tapponnier, 1978; Molnar and Chen, 1983; Armijo et al., 1986; Molnar and Chen, 1983; Molnar et al., 1987). By their creed, they would have interpreted faults as reverse, if they could.

The plate-tectonic model separates former India from Afghanistan and Iran by thousands of kilometers, before the Mesozoic. Yet Arthur A. Meyerhoff and Curt Teichert pointed out in 1971 that several ancient geologic features connect India and Iran:

Detailed mapping from the central Indian shield to south-eastern Afghanistan and Iran   show that several distinctive rock formations and faunal zones extend from central Iran and Afghanistan on to the Indian shield (Madhya Pradesh). These stratigraphic units and paleontological markers, which are continuous, include the Proterozoic-Cambrian Hormuz Salt series, two of the Productus-bearing beds, and the Aptian-Albian Orbitolina zone and associated warm-water carbonates.

The conclusive evidence which ties the Indian shield irrevocably to Asia is found in the westernmost Himalayas-around the Vale of Kashmir), the Hazara district, and the Salt Range-as well as in the surface and sub-surface sections of West Pakistan, Afghanistan, Iran, Saudi Arabia, Iraq, Syria, and Turkey.

India has been part of Asia since Proterozoic or earlier time.

The great height of the present-day Himalayas is quite unrelated to their original formation as a geological structure. Least of all is it directly related to Tertiary collision. Popular concepts of its being related to the speed of movement of Gondwanic India, or an alleged distance that unit moved, are geologically mythical. Gondwana glacigene and other facies, and Indian reptiles, extend far beyond the alleged "suture", into Sikkim and Tibet. . . . Tibet and Tarim were always part of India, and closely related to Australia, so any Asia-India plate boundary must have been as far north as the Ten Shan.

the Kun Lun-Altin Tagh northern front of the Tibetan plateau is a block-fault uplift, not a compressional thrust.

As no oceanic Tethys existed, and India and Angaraland did not collide, the Himalayas could not have been born of collision nor of subduction, but resulted from vertical uplift.

There is little evidence supporting the previous existence of such a large ocean between the northern boundary of the southern continents and the southern boundary of Eurasia. An orogenic belt existed in this region, but it is far from clear that it was sited on the borders of a once extensive ocean.

Plate-tectonic models have had India drifting thousands of kilometers, while all the field relations suggest India and Eurasia could never have been far apart.

The Tethys was primarily an epicontinental sea which in Palaeozoic time transgressed on an undivided Indo-Tibetan platform. Within this sea an axial oceanic trough separating India from Tibet came into being only in the Mesozoic. This oceanic trough reached its maximum development only in the Late Cretaceous. Consequently, the Indian Ocean did not open at the expense of pre-existing wide Tethys. The geology of the Himalaya fails to indicate a Tethys ocean of Palaeozoic - earliest Mesozoic age, and in this respect supports the theory of Earth expansion.

India has had a relatively close association with Eurasia throughout the Phanerozoic.

Normal faulting has been the dominant tectonic regime north of the Himalaya in the last 2 ± 0.5 my.

I fail to understand why some workers persist in using a model that is so much at odds with undisputed observations.

The Cornell Institute for the study of continents (Fielding et al., 1994) has produced high-resolution digital topography of Tibet on a three arc-second grid, compiled from data generated by the U.S. Defence Mapping Agency. They state that,

notwithstanding the high altitude, the flatness of Tibet implies that: Tibet has not undergone major compressional deformation during recent geologic time   (Pliocene-Quaternary) ... there has been little deformation (especially shortening) of the uppermost crust north of the graben systems during the late Cenozoic.

Paleomagnetic false lead

Most have assumed a gross northern migration of India, based on India's apparent polar wander path. Paleomagneticians measure only the angles sub-tended to the paleopole, but they plot them as distances to the paleopole, assuming present-radius Earth, thus compounding the compression myth. But the radius of the Earth has increased so their apparent polar distance increases logarithmically, which yields the false conclusion that India has migrated northward.

Bite the bullet

Contraction across Tibet is zero.  In fact it has widened!  The great Himalayan nappes can be added in too. The width between the Ganges valley to beyond the Tien Shan is greater now than in the early Cretaceous, before the great Himalayan orogenesis!

Geology must now go back a century to the findings of Meyer, and recognize that gravity governs tectonics, that diapiric motions dominate, that hotter regions expand and rise towards gravity equilibrium, and on nearing the surface equipotential they must spread laterally. Here, indeed, the structural patterns of Hutton, Lyell and Dana do dominate but they are superficial to the upper 10 km or so of the crust, as in Figures 12 and 16. Gravity nappe surfaces steepen downward to rising diapirs, rather than as superficial thin-skinned slides. Field geologists see the upper dozen kilometers of crust, where rising cores have been deeply eroded, so their observations remain mainly valid. But false models appear when the horizontal foreshortening they obscure is applied continentally.

Figure 12 is the paradigm for a simple, single-phase, diapiric orogen. For initial simplicity this is drawn symmetrically, although most orogens are multiphase and asymmetric. However, the Pyrenees and Urals seem to be symmetrically double-sided.

An orogen is initiated by crustal stretching through local higher temperature in the mantle, causing thermal expansion and expansion through phase change to less dense minerals, such as spinel to olivine or eclogite to gabbro. Hence convective gravity drive. The stretching means thinning crust at the floor of an active eugeosyncline, with faults, higher heat flux, hot water emissions (siliceous cherts and radiolarites), mantle-derived pillow lava, serpentinized peridotite, immature sediment, turbidity flows, slumping on various scales, and olistrostromes.

A transient negative gravity anomaly develops from serpentinization of peridotite and because the "viscosity" of the subcrust delays adjustment, and correction is not as local as the mass deficiency. The eugeosyncline tends to have a raised rim against a miogeosyncline; basement may even crop out there (compare the raised rim of rift valleys).

For the same reason, the crustal sag may extend hundreds of kilometers on each side of the rifting zone, hence a lateral miogeosyncline, with slower sinking and slower sedimentation, more mature sediments, without faulting volcanism or local unconformities, but discontinuities and diastems from non-deposition.

In contrast with the crustal shortening paradigm, which assumes crustal extension during the sedimentation cycle, followed by intense compression during the orogenic stage, the diapiric model implies continuing extension at aIl stages.

The heated zone below continues to rise from reduced density from thermal expansion and mineral phase changes, and also to correct the negative gravity anomaly which is increased because water and light sediment fill the trough in place of normal litho- sphere.

If the rising of the core exceeds the rate of sedimentation, this extruded zone may be eroded, then subside again from renewed stretching, with new locally unconformable sedimentation. Eventually, the rising of the core dominates, pushing above sealevel.

Nappes

Consider the conditions at a point like P (Figure 12, third diagram) which is squeezed between the rising ram of the diapir pushing up, and the weight of the sediment pile above bearing down. Material is driven laterally. A stable regime is reached where the rate of the rising ram maintains the weight of the sediment above, balanced by the rate of lateral extrusion.

The viscosity of the sediment is relatively low (~10¹⁵ poises or less). Flow is aided by water pressure and by lubricated serpentinite. If the rate of upward push of the diapiric ram increases, or the rate of lateral outflow of the sediment declines, the sedimentary pile rises higher to a new balance. The outward-thrusting sediment overrides the adjacent sediment. Lineations and thrust surfaces are in the direction of flow, inner zones always over riding those farther out.

The rising diapir pushes outward like a giant bulldozer, forming overthrusts and flattened folds, and chains of horizontally foreshortened folds like the Jura. In sections to depths of several kilometers, only horizontal compression is observed.

However, the foreshortening of concentric folds must vanish at depth (Figure 15). The lower limb of an overthrust, always suffers less translation than its over riding part.

But this does not happen, because the resistance to flow is proportional to the area of the base. Before the nappe has gone far this resistance halts it, the nappe breaks in the rear, and a new nappe is driven above (or even below). So nappe piles on nappe, perhaps to a dozen or more. New nappes above may cause lower nappes to move again, especially when fluid pressure builds up.

On our conservative figures the additive overthrust of stacked nappes could be 2,500 km. During this process the width of the orogen core has not narrowed - in fact it has probably widened! So much for the beloved foreshortening meme!

Figure 12 is, for physical simplicity, drawn symmetrically. But symmetrical orogenesis is rare, if it exists at all. Asymmetry is the norm. Earth rotates towards the east. Expansion has been much greater in the southern hemisphere than the northern. Orogens along the Asian Pacific margin overthrust toward the east, so do those along the west margin of North America, away from the ocean, whereas the Asian overthrusts are toward the ocean.

Intermontane basins

Continued crustal stretching and rise of mantle may result in mantle-derived material reaching the surface, medially as in Figure 12 or laterally, resulting in a medial or lateral basin with a floor of risen mantle three or four km below sealevel. As diapirs tend to become a chain of swells rather than a continuous ridge, the result is series of small seas (as in east Asia) or a greatly widened orogenic zone consisting of paired orogens separated by oval or kidney-shaped intermontane plains or small seas, as in Figure 11.

Although lagging behind their enclosing ridges, these basins are also diapirs because sub-Moho matter has risen 30 km - to 4 km below sea level.

"Zwischengebirge" is another category of diapir mantle-floored basins. As the enclosing fold mountains wrap around them, they have been misinterpreted as intermontane nuclear blocks against which the strata have been folded (e.g. the Lut of Iran, the Tarim, and the Ordos basin). This concept began with Stahl (1911) who interpreted Archean-looking metamorphics (which later turned out to be Jurassic) in the Persian High Plateau (the Lut) as an ancient intermontane nucleus. Kober (1921) called such 'nuclei' Zwischengebirge (literally 'inter-montane' sediments lying flatly on an ancient massif around which the bordering ranges had been folded, and extended the concept to the several European basins of Figure 11. Oil company geologists under Hugo de Bockh adopted this concept, which was generally accepted.

Folding

In the miogeosyncline, where temperatures are low, the viscosity differences between sandstone, shale, limestone, and conglomerate are large, so folding is dominated by bedding slip, and hence concentric folding (Figure 13).

Concentric folding is often called parallel folding because the bedding is normal to the radius of the fold. Similar folding is often called shear folding. But all folding involves shear, either parallel to the bedding or transverse to it. Such terms lack priority, are physically misleading and should be dropped.

Gravity rules the Earth. A meter cube of granite could stand indefinitely on its own base, but the pressure on the base of a kilometer cube exceeds its crushing strength. Unsupported laterally, it would crush or flow. As I emphasized in my 1962 paper on scale in tectonic phenomena, resistance to flow (elasticity, viscosity, rheidity) increases by the square of the scale but gravity driving force increases by the cube. Hence diapirism and flow phenomena generally must increase with scale.

All agree that crystalline glaciers flow like treacle, needing only time, and that crystalline salt forms cylinders that rise diapirically many kilometers, and may even extrude and flow like lava on the surface as do salt domes in Iran (Figure 14), first figured by Lees (1928), and that this happens more slowly with crystalline gypsum, and gneisses (Carey, 1953). Crystalline gneiss in the Goodenough gneiss domes described by Ollier and Pain (1981) (Figure 9, p. 30) is rising diapirically now more rapidly than intense tropical erosion can wear them down.

Appalachians

Let us apply this paradigm to some real orogens. First the Appalachians (Figure 20).

Complexly folded eugeosynclinal rocks are intruded by diapiric granite plutons. The overthrusting along the Appalachian front is greatest where the arcs bow outward, least where bowed inward, which is as expected from a chain of diapirs. But in compressional tectonics - how ??

A thick formation of salt several kilometers below a pile of denser geosynclinal sediments does not rise as a sheet, but as a pattern of cylindrical domes with a diameter of a couple of kilometers, more or less, surrounded by a rim syncline. A long cliff-line suffering landslide collapse does not slump as a single long body, but as a series of spoon-shaped masses.

Likewise, orogenic diapirs may initiate a long fracture belt, but soon consolidate to individual  diapirs a few hundred kilometers apart, or even less, which may surge independently. The total Paleozoic thickness is much the same all along the Appalachians, but the diapiric regurgitation peaked in the Middle Ordovician in North Carolina, in the Late Ordovician in Maine, in the Late Devonian in New York, and the Late Carboniferous in Georgia (Figure 21). Each of these diapiric surges spread thousands of meters more detritus over the neighboring terrain than was then accumulating elsewhere along the belt. Natural in a line of diapirs, but how, if caused by a cratonic vice? The general uplift of the Appalachian mountain belt was independent and later

Permian and Mesozoic miogeosynclinal beds lie flatly on the pre-Alpine Vosges, Black Forest, and Bohemian massifs. Nearer the Alps, they form the Jura trains of Mesozoic concentric folds on salt decollements. Next is the post-Alpine tensional molasse zone of the Swiss plain. Then the stack of miogeosynclinal limestone and shale of the Helvetic nappes, then reappearance of the up-turned pre-Alpine basement (Mont Blanc, Aiguilles Rouge, and Aar) before the change to the schistes lustres, then the orogenic core and the Ivrea zone. Far from the intense compression claimed for the Alps, they are probably a little wider now than before the orogenesis!

Let us assume that the surface structural section across the Himalayas as given by Gansser on the standard compressional assumption is correct, we can redraw it according to the diapiric model with all its nappes without any crustal shortening or collision of India against Asia (Figure 24).

Surrounded by such high country, it is surprising to find Turfan 154 m below sea-level, part of an 800 m trough of internal drainage. Issik-kul, 1,000 km west of Turfan and on the same general trough-line, is 1609 m above sea-level but 702 m deep, with overlooking peaks of 5,000 m.

Through the Paleozoic and early Mesozoic, the whole region was part of the South Asia-Australia-India-Arabia platform with many basin subsidences and uplifted plateaus, but no granites, no flysch, and no orogenesis. North of Mt Everest is a little- deformed sequence from Ordovician to Eocene. The Zanskar-Spiti region has 5 km of a relatively continuous sequence from the late Proterozoic to the Eocene. The Himalayan climate changed from the periglacials of the Permo-Carboniferous to the tropical conditions of the Triassic.

The Neogene granitoid domes, described by Le Fort (1988) as "a string of pearls" (Figure 26) are two-mica adamellite diapirs surrounded by a dense stockwork of aplitic and pegmatite domes, which extend for some 1300 km parallel to the Indus-Trangpo line and some 100 km south of it, are still rising. South of Lhasa is Yamdrock Tso (7,200 m), then Kingmar, Lhagoi Kangri (6,500 m), and finally Gurla Mandhata.

Figure 29 shows the arcs of granitic domes and the lens of Quaternary volcanism of the youngest diapirism. Lines of granitoid domes are characteristic of the central geo-anticlinal regurgitation of orogens generally (for example, the Puerto Rico plutons of Figure 52, the Paleogene plutons of Cuba, the line of plutons through Wales to the Scilly Isles, and the Devonian plutons of Tasmania).

New Guinea

When the Australian shield commenced to move away from India (pp. 59-95), oceanic crust began to isolate the Australian block on the west, north and east. Cretaceous-Paleocene Tethyan diapirs continued to develop but those to the east and west of the Australian block had oceanic crust on both sides, whereas the northern diapirs fronted the Australian block on the south, but new ocean floor to the north (Figure 30).

The western edge of the Australian shield crosses the New Guinea trunk, then runs along the southeast shore of Geelvinck Bay to Cape D'Urville. It marks a major change in trend, separates the Vogelkop diapir from the Star Mountains diapir, with a considerable change in facies. This Vogelkop oroclinal bend is equivalent to one of the dextral megashears of east Asia (Figure 30).

Another transverse lineament from the Strickland River to the mouth of the Sepik, separates the Star Mountain diapir from the Mt. Hagen diapir, marks a change in trend; and bounds the cluster of Pleistocene volcanoes in Papua.

The Aure transverse lineament continues the eastern edge of the Australian shield from Port Romilly to the Huon Gulf, separates the Mt. Hagen diapir from the Owen Stanley diapir, makes a change in trend, and a significant difference in the geologic history. The Solomons sphenochasm and the Coral Sea sphenochasm originate from it. The Aure trough rapidly filled with four km of Miocene sediments. Wedge-shaped rifts, such as the Omati trough cut north-westward into the edge of the Australian craton, to be very rapidly filled with 3,000 m of middle Miocene globigerine limestone.

The Star Mountains Diapir is 650 km long, 150 km wide at the center, bowed northward. The Nogolo-De Wal valley line is a major strike fault zone within the diapir, between Upper Mesozoic beds and a wide belt of schists and intrusive rocks. Another fault zone follows the Baliem valley, which seems to be the boundary between the folded miogeosyncline and the diapir. Neogene gold and copper bearing porphyry plutons intrude the axis of the diapir at Porgera, Ok Tedi., Nena, Tifalmin, Frieda River, Yandera and Ertsberg. The Antares range is a large quartz porphyry pluton.

The Mt Hagen Diapir is 400 km long and bowed to the southwest and 200 km wide at the center. Nappes and tight folds extrude to the southwest. The trough of the miogeosyncline trends north-westward along the Kikori River and Lake Kutubu, and subsided 4,000 m during the post-Triassic Mesozoic. The Cecilia anticline in Pliocene strata exposes a Neogene core, marking the beginning of the outward thrust from the diapir. The next in the section toward the diapiric core, the Mueller Range anticline, exposes the Paleogene, then the larger Levant anticline exposes Cretaceous, and on its flanks begin the schuppen slices of the diapir.

There are two stages in the Mt. Hagen diapir. The large copper bearing plutons from the Yuat and Jimmi Rivers and the Bismarck and Mueller Ranges, which is fronted by the Maramuni and Bismarck fault zones which mark out an earlier more deeply dissected diapir. Although separated by the younger Aure Trough and Aure and Sunshine faults, the large plutons of the Bulolo region extending southwest as far as the head of the Lakekamu River, probably belong to this group. while the Laigap and Telefomin fault zones front the younger diapir. Pre-Cretaceous basement re-emerges in the Kubor Paleozoic granite and metamorphics.

The cluster of Quaternary volcanoes indicate that this diapir is still active and migrating south-westward. The volcanoes of this group (Ae-Ba, Rentoul, Sisa, and Lalaigan; Aramia, Bosavi, and Kareva; Dimiea, Giluwe, and Mt Hagen; Kawbenaberi, Iavokia,  Murray, and Suaru; Aird Hills Duau, Favenc, and Karimui) seem to be aligned on transverse fractures parallel to the lines separating the diapirs.

The central diapir emerges in the Thurnwald Range, south of the belt of strongest shearing. The Laigap fault zone north of the Central Range brings up a belt of ultramafics, homologous with the Owen Stanley fault zone, north of the Owen Stanley Range. The northern boundary of the diapirs is the southern side of the Sepik-Ramu-Markham trench. Neogene gold-bearing plutons occur at Porgera, Ok Tedi, Frieda River, Yandera, and Ertsberg.

The Owen Stanley diapir (Figure 9) is 800 km long, consisting of a belt of lightly metamorphosed upper Mesozoic and Paleogene strata, overlain across a steeply dipping fault zone by the Papuan ophiolite belt. The central Owen Stanley mass extends a farther 300 km southeast to Misima Island and the Louisiade Archipelago. Before the great expansion of the Tertiary, it certainly was continuous with New Caledonia, and had ophiolite diapirs on both the northeast and southwest sides.

North Coast ranges

A second line of diapirs, the Bewani, Torricelli, Prince Alexander, Adelbert, Finisterre, and Saruwaged Ranges, (which earlier had included New Britain) emerge in the northern coastal ranges with numerous fault slices (strongly modified by the Tethyan torsion). The Tertiary sediments in these ranges rest on the old ocean floor, which emerges in frequent thrust slices. The petroleum geologists who explored this region had no interest in this "basement" and bulked it simply as "epidiorite". The Matapau oil seepage emerges along such an "epidiorite" fault. The Cyclops and Serra mountains may be a third diapir.

Goodenough diapirs

Most spectacular of all are the Goodenough diapirs (Figure 31 and 32), described by Ollier and Pain (1981) - a line of four mantled-gneiss domes, which are currently rising more rapidly than they can be attacked by rain-forest erosion.

These mantled-gneiss domes are currently emerging and shouldering aside the local strata. Foliation in the gneiss is concentric and parallel to the dome surface. The topographic surface of the domes is like a cone of triangular facets which are moderately cut by the radial drainage but become less and less dissected at lower levels which have only recently emerged. The lowest few hundred meters are almost pristine. Ollier and Pain point out that Goodenough Island is one of the steepest islands in the world-only 20 by 16 km, and 2500 m high, surrounded by a flat marine terrace 2 to 10 km wide.

Let the sea lap around Gurla Mandhata (Figure 31 ) so that the young horizontal sediments labelled T become a depositional shelf, and the analogy with Goodenough becomes complete.

The interior of the Goodenough dome consists of amphibolite-facies quartzo-feldspathic gneiss with minor amounts of calcic gneiss and amphibolite and a small intrusion of granodiorite in the center. The schistosity and boudinage in the surrounding sediments and the gneissisity within the dome are parallel to the dome surface, indicating the active rise of the diapir.

The minerals of the gneiss must continuously recrystallize during the diapiric rise, just as ice continuously recrystallizes during glacier flow, salt continuously recrystallizes during rise of salt domes, and serpentine minerals recrystallize during injection of serpentinite.