The Penokean orogen of Michigan's Upper Peninsula includes a belt of dome-and-keel structure presently defined by deep troughs, or “keels,” of Paleoproterozoic Marquette Range Supergroup strata between gneiss domes composed of Archean basement rock. Structural, metamorphic, and geochronological data from the Southern Complex indicates that dome-and-keel structure developed in two stages. The first stage involved rise (intrusion, possibly diapirically) of the 2.6 Ga Bell Creek Assemblage (a gneissic megacrystic granite) into the Twin Lake Assemblage (migmatitic mafic to felsic gneiss). Flow folding in gneisses and migmatites indicate that this Archean event involved plastic flow of basement. The second stage occurred after the ca. 1.8 Ga Penokean orogeny, subsequent to the formation of a fold-thrust belt involving Paleoproterozoic Marquette Range Supergroup strata. During this stage, deep, narrow troughs developed in the region that had been the fold-thrust belt. Analysis of structures bordering the Republic Trough indicates that Paleoproterozoic keel borders are shear zones; keel rocks moved down relative to dome rocks. In effect, the Paleoproterozoic keels are steep- to vertical-sided grabens, suggesting that the dome-and-keel architecture is a consequence of extensional faulting. Amphibolite facies metamorphism occurred in Paleoproterozoic keel strata along dome-keel borders. Peak-metamorphism developed adjacent to dome borders at the time keel-bounding shear zones were active. The relative timing of Paleoproterozoic keel formation supports the model that this stage reflects collapse of the Penokean orogen. Our results show that the present dome-and-keel structure of the Southern Complex region represents superposition of Paleoproterozoic collapse structures on preexisting Archean gneiss domes.

Many structural domes and anticlines rise from undeformed surroundings; the general crust of the earth is neither shortened nor extended by these local swells or blisters, but the area of the plate upfolded is increased. With plastic material the increase is effected largely by flow; with brittle material, largely by fracture. Many mining districts are associated with upfolds of this type. Districts here described fall into two categories: (1) those whose structural frame is a dome, and (2) those whose structural frame is an anticline. Domes may show fractures which radiate from the apex, or concentric fractures which are segments of circles, of varying diameters but with a common center, the apex of the dome. Both types may appear on the same dome; the fractures of each type aid enlargement of the plate during doming. Sunlight and Kirwin, Wyoming, are minor districts with dominant radial fracture patterns. Vein matter was deposited while the walls of the vein fractures were being pulled apart. With a radial vein system the only way in which all the vein walls could be simultaneously pulled apart is by stretching of the fabric of an expanding dome. The structural setting of the Ophir, Utah, lead-silver district, and of the Matehuala, Mexico, copper district is that of a half dome truncated by a normal fault with downthrow away from the domical apex. Displacement on the fault is greatest opposite the apex and decreases progressively in both directions, becoming zero at the spring line. No part of the dome ever existed on the down-thrown side of the fault, which was a fracture before doming. Maximum uplift was centered on one side of the fault and there produced the half dome; but the pre-existing fracture offered locally an easier mode of uplift by rise of the foot-wall block. At Ophir ore shoots followed intersections on the half dome of radial fractures with limestone beds. At Matehuala stretching during doming was effected largely by flow of limestone, but a monzonite stock intruded in the half dome was too brittle to flow. The limestone pulled away from the unyielding stock; fractures concentrated around the periphery of the stock localized the Dolores copper ore bodies. In the Silverton-Telluride district, Colorado, fractures radiate from a relatively large, roughly circular graben, along whose margin monzonitic stocks were intruded. Evidence suggests that this fracture pattern resulted from domical uplift, with the graben at the apex. Transfer of volcanic material from depth to the surface in the central area produced a sag which has eliminated the upward bulge of the dome. Mineralization advanced outward from the graben step by step with the outward growth of the radial fractures. Copper-silver pipes within the fault zone bounding the graben were formed first, followed successively outward by base-metal deposits as veins, which were reopened to admit gold and silver, and by precious-metal vein deposits in the outermost zone. The structural setting of La Plata, Colorado, is that of a dome, truncated, south of its apex, by a zone of high-angle faults striking eastward. Displacements on the faults are greatest opposite the domical apex. The doming is accentuated by a horseshoe-shaped hinge fold, open on the south. Along the fold dips of the strata steepen sharply; outside it, dips are gentle, whereas inside it, they are nearly flat. Several stocks were intruded along the fold, and others inside it. The steep flexural fold and abundant intrusions suggest upward shove of a flat-topped piston, perhaps a magma column congealed in its upper part, but fluid and under pressure below. Fractures are abundant along the horseshoe fold, and, with respect to the dome, fall into two classes, radial and concentric. Gold-silver deposits were concentrated chiefly within or near the horseshoe fold, and in the eastward-trending fault zone. Doming, which began during the intrusive epoch, persisted through the period of metallization, because at that time older fractures were reopened, new radial and concentric fractures were created, and fractures of both generations became loci for ore bodies. The structure at Rico, Colorado, is that of a dome with eastern elongation. Superimposed upon the major dome, toward its eastern end, is a doubly plunging anticline, also with easterly trend, cut by fractures which parallel its axis, and by fractures normal to the axis. Rich ore bodies were localized at a stratigraphic horizon originally occupied by a bed of gypsum which was dissolved, leaving silty material which the ores replaced. Ribbon-shaped mantos lay directly above fractures, both of the longitudinal and transverse sets. Ore solutions ascended these fractures to form the mantos during late stages of the doming. The Goldfield, Nevada, district lies on the southwest flank of a dome encircled by a belt of intense alteration, and of complex fracturing, which probably coalesces at depth into one or more persistent faults concentric with respect to the domical apex, and which formed the channel for altering and metallizing solutions. The volcanic rocks were brittle when first fractured by doming, but solutions rising along the fractures of the circular belt softened the rock by alunitization and kaolinization. Silica-bearing solutions then created the irregular silica “ledges” at horizons close to the then surface. The soft rock encasing the ledges flowed, as doming persisted, but the brittle ledges fractured. Ledges which had no “keel” below them were inaccessible to gold-bearing solutions; those with keels extending down to the main ore channel received the rich ore bodies. Many mining districts are associated with doubly plunging anticlines, which with brittle rock are broken by fractures which strike parallel or normal to the axis. At Creede, Colorado, older extrusive rocks were flexed into a north-trending anticline. Younger extrusive rocks do not share in the folding, but faults which strike parallel to the anticlinal axis and dip toward it cut and displace both older and younger rocks; they form a graben along the crest of the buried anticline. These faults originated as tension fissures, the result of arching of the older volcanic rocks. After extrusion of the younger volcanic rocks, renewed uplift was concentrated along the abutments of the arch, in the footwalls of the graben faults. These faults were propagated upward through the younger volcanic rocks. The eastern graben fault, the Amethyst, fingers out at its southern end. Most of the silver ore of the district came from the southern segment of the Amethyst vein. Here intense local uplift in the footwall, unable to utilize the split-up fault as a lubricated plane of movement, tore apart the walls to permit entry of the silver-bearing solutions. The structure at Bodie, California, is that of an irregular anticline upon which are superimposed several domes. The country rock is volcanic. Most of the faults and veins strike parallel to the anticlinal axis and dip toward it; but the Fortuna fracture, which carried the richest ore body, lies in anomolous relation to the anticline, for it is neither a longitudinal nor a cross fracture. It seems to have resulted from an earlier deformation, but to have been utilized by the uplift which formed the anticline in such a way that its flat segment gaped open to admit rich silver- and gold-bearing solutions. Guanajuato, Mexico, lies on the northeast flank of a major anticline which plunges southeast. The anticline carries a crestal graben. The graben fault on the northeast flank is the Veta Madre, with maximum displacement on the northwest; displacement decreases progressively southeastward, in the direction of plunge of the anticline. Like the Amethyst fault at Creede, the Veta Madre originated as a tension fissure, but became an antithetic fault when the arch broke into segments under continued uplift. Major silver-ore bodies on the Veta Madre were localized where differential movement of the walls brought shallow cups in the footwall surface opposite planar areas in the hanging-wall surface. At El Oro, Mexico, the attitudes of remnants of an andesite flow overlying shale, together with the fracture pattern, indicate deformation to form a broad anticline trending north-northwestward. The San Rafael vein lies along a normal fault striking parallel to the anticlinal axis, with downthrow on the west, toward the axis. Faulting had been completed by the time of mineralization. Early, low-grade vein matter welded the fault, but arching continued and with it an urge toward resumption of faulting, prevented by the welding. The resulting strong shearing strain produced a number of vertical feather-joint branches in the hanging wall of the fault. These were mineralized by solutions rich in gold and silver. The Mogollon, New Mexico, district lies on the west flank of a large anticline trending and plunging north-northeastward. The Pacific-Great Western and Queen faults strike parallel to the anticlinal axis and dip eastward toward the axial plane. The block between these faults contains a local bulge truncated on the east by the Queen fault. Displacement on the fault is greatest opposite the crest of the bulge and decreases progressively in either direction. The local uplift in the footwall increased the displacement on the Queen fault, but it took place in the hanging wall of the Pacific-Great Western fault. The original displacement was reversed in the segment affected by the bulge. The bulge has the form of a doubly plunging anticline trending northward, parallel to the Queen and Pacific-Great Western faults. The chief productive veins of the district occupy cross fractures normal to the anticlinal axis. Most of those north of the highest point on the up-bowed axis dip southward, whereas most of those south of that point dip northward. These fractures gaped open, under continued bulging, in time to receive the richest surge of silver-gold solutions. The following generalizations appear valid. Uplift in these districts was accompanied by development of tension fissures. Uplift and consequent stretching of the arching plate persisted through the period of mineralization, but by this time stretching in many areas had reached a stage at which Assuring could no longer facilitate it; Assuring was succeeded by graben and antithetic faulting. Ore deposition sometimes preceded this faulting but more often followed it. The fracture pattern on these domes and anticlines developed as uplift progressed. Solutions deposited vein matter in those fractures which were permeable at the time and accessible from the main solution channel. Mesothermal deposits associated with domes and anticlines fall into groups defined by age of mineralization and by metallographic provinces, but epithermal deposits are scattered from one end of the Cordilleran region to the other. They show, however, a preference for major uplifts. Silverton, Rico, La Plata, and Creede lie on a tectonic element marked by recurrent uplift from the close of the Paleozoic to the Pleistocene. Epithermal deposits in Mexico are concentrated on the site of the persistently positive Occidental geanticline. The crystalline basement lies deep throughout much of Nevada, but Goldfield, Tonopah, and other epithermal districts lie above or close to relative highs in the basement which are much larger than the local uplifts with which these districts are associated. These major uplifts were developing while epithermal metallization was taking place. The whole Cordilleran region was fast assuming its present shape. The local phenomena of uplift, Assuring, intrusion, and metallization were satellitic features superimposed on the uplift of the Cordilleran region as a whole. Because of this fact, a deep-seated origin for epithermal ores is suggested.

The magmatic and tectonic processes of the pre–2.5 Ga hot, young Earth differed profoundly from those of the modern planet. The ancient rocks differ strikingly in individual and collective composition, occurrence, association, and structure from modern rocks. Widespread forcing of Archean geology into plate-tectonic frameworks reflects unwarranted faith in uniformitarianism and in inappropriate chemical discriminants, and disregard for the lack of features that characterize plate interactions. Archean crust records extreme and prolonged internal mobility and was far too weak and mobile to behave as rigid plates, required, by definition, for plate tectonics. None of the geologic indicators of subduction, arc magmatism, and continental sundering, separation, and convergence have been documented. No Archean oceanic crust or mantle has been recognized, and the only known basement to supra-crustal rocks, including the thick basalts, high-Mg basalts, and ultramafic lavas that typify greenstone successions, consists of tonalite-trondhjemite-granodiorite (TTG) migmatites and gneisses. A thick global melabasaltic protocrust likely formed by ca. 4.45 Ga, and from it TTG suites were extracted by partial melting over the next 2 b.y. Delamination of the increasingly dense restitic protocrust enabled rise of lighter and hotter depleted mantle and hence more melting. The oldest known crustal materials are zircons, which scatter in age back to 4.4 Ga and are recycled in migmatites whose final crystallization was after 3.8 Ga, and in ancient sediments. Earth may have had a dense greenhouse atmosphere, not a hydrosphere, before 3.6 Ga, for the oldest proved supracrustal rocks are of that age, and older felsic crust may have been too hot to permit rise of dense melts. Rigid plates of lithosphere did not stabilize until a billion years after that and then were mostly small and local. Dense lavas erupted atop mobile felsic crust after 3.6 Ga produced a density inversion that was partly righted by sinking of the volcanic rocks and rising of the subjacent TTG. In some places, the early dense rocks retained cohesion and sank as synclinal keels between rising domiform diapiric batholiths. In others, the early dense rocks sank deep into mobile TTG crust, and only later in Archean time was the felsic substrate strong enough to enable dome-and-keel style. The TTG substrate rose slowly, with variable amounts of partial melting to generate more-fractionated melts and with additions of new TTG from the underlying protocrust, for hundreds of millions of years. The mantle beneath preserved cratons generated ultramafic melts that required a temperature ∼300°C hotter than modern asthenosphere ca. 3.5 Ga. Severe and prolonged lateral deformation was superimposed on large parts of some cratons during the era of volcanism and diapirism, obscuring dome-and-keel geology over broad tracts. Lower crust was at high temperature for prolonged periods and flowed pervasively, coupled discontinuously to the upper crust to produce lateral deformation therein. Rifting, separation, rotation, and collision of internally more rigid lithosphere fragments began ca. 2.1 Ga, but may have been dominantly intracontinental deformation, quite distinct from modern plate tectonics. The products of this regime differ greatly from those of Phanerozoic plate tectonics, and reflect a transitional era of erratically stiffening lithosphere. An early-depleted upper mantle has been progressively re-enriched, by delamination and subduction of crustal materials, while new “juvenile” crust derived from it has become progressively more depleted, during Pro-terozoic and Phanerozoic time.

Abstract How and when continents grew and plate tectonics started on Earth remain poorly constrained. Most researchers apply the modern plate tectonic paradigm to problems of ancient crustal formation, but these are unsatisfactory because diagnostic criteria and actualistic plate configurations are lacking. Here, we show that 3.5–3.2 Ga continental nuclei in the Pilbara Craton, Australia, and the eastern Kaapvaal Craton, southern Africa, formed as thick volcanic plateaux built on a substrate of older continental lithosphere and did not accrete through horizontal tectonic processes. These nuclei survived because of the contemporaneous development of buoyant, non-subductable mantle roots. This plateau-type of Archean continental crust is distinct from, but complementary to, Archean gneiss terranes formed over shallowly dipping zones of intraoceanic underplating (proto-subduction) on a vigorously convecting early Earth with smaller plates and primitive plate tectonics.