Search

International audience; Prominent landforms, either buried or preserved at the seafloor, provide important constraints on the processes that led to the opening and present-day configuration of the Dover Strait. Here, we extend previous investigations on two distinct landform features, the Fosse Dangeard and Lobourg Channel, to better understand the poly-phase history of their formation and inferences for the opening and Pleistocene evolution of the Dover Strait. The Fosse Dangeard consist of several interconnected palaeo-depressions. Their morphology and spatial distribution are interpreted to be the result of plunge-pool erosion generated at the base of north-eastward retreating waterfalls. Their infills comprise internal erosional surfaces that provide evidence for the occurrence of several erosional episodes following their initial incision. The Lobourg Channel comprises various sets of erosional features, attesting to the occurrence of several phases of intense fluvial and/or flood erosion. The last one of these carved a prominent inner channel, which truncates the uppermost infill of the Fosse Dangeard. The morphology of the Lobourg inner channel and the erosional features associated with its incision strongly resemble landforms found in megaflood-eroded terrains, indicating that this valley was likely eroded by one or several megafloods. Our study therefore corroborates the existence of waterfalls in the Dover Strait at least once during the Pleistocene Epoch. It also provides evidence of the occurrence of multiple episodes of fluvial and flood erosion, including megafloods. Finally, this study allows us to establish a relative chronology of the erosional/depositional episodes that resulted in the present-day morphology of this region. © 2018 Elsevier Ltd

Thick packages of lavas forming seaward‐dipping reflectors (SDRs) are diagnostic features of volcanic passive margins. Despite their significance to continental breakup studies, their formation mechanism is still debated. We use ~22,000 km of high‐quality, depth‐migrated, seismic data to document the three‐dimensional geometry of SDRs offshore South America. We find two types: Type I are planar and occur as fault‐bounded wedges. Type II are characterized by reflections that become more convex‐upward in the downdip direction and terminate against a subhorizontal base. We interpret the transition from Type I to Type II SDRs to represent a continuum from continental rifting to full plate separation with formation of new, subaerially generated, magmatic crust. Type I SDRs formed in half grabens during the stretching of continental crust, while Type II lavas infill the space produced by flexing of the crust due to the solidification of the underlying feeder dikes as the magmatic crust moved away from the spreading center. Type II SDRs vary in length and thickness along the margin. In the north, close to the Paraná flood basalts, they are long (tens of kilometers), reach thicknesses of up to 15 km, and have an across margin width of up to 600 km. To the south the Type II SDRs are thinner with lava lengths of

We present new constraints on the opening of the South Atlantic Ocean from a joint interpretation of marine magnetic anomaly grids and forward modelling of conjugate profiles. We use 45,000 km of recently collected commercial ship track data combined with 561,000 km of publically available data. The new data cover the critical ocean–continental transition zones and allow us to identify and downgrade some poorly navigated older ship tracks relied upon in earlier compilations. Within the final grids the mean cross-over error is 14 nT computed from 8,227 ship track intersections. The forward modelling used uniformly magnetised bodies whose shapes were constrained from coincident deep-seismic reflection data. We find the oldest magnetic anomalies to date from M10r (134.2 Ma, late Valanginian) north of the Falkland-Agulhas Fracture Zone and M3 (129.3 Ma, Barremian) south of the Rio Grande Fracture Zone. Hence, assuming the GPTS used is correct, continental breakup was contemporaneous with the Parana and Etendeka continental flood basalts. Many of the landward linear anomalies overlap seismically mapped Seaward Dipping Reflectors (SDRs). We interpret this to mean that a significant portion of the SDRs overlay crust formed by subaerial seafloor spreading. Here crustal accretion is envisaged to be similar to that at mid-ocean ridges, but sheet lava flows (that later form the SDRs) rather than pillow basalts form the extrusive component. Segmentation of the linear anomalies generated implies that this stage of continental breakup is organised and parallels the seafloor spreading centre that follows. Our results call into question the common assumption that at volcanic continental margins the first linear magnetic anomalies represent the start of conventional (submarine) oceanic crustal generation.

A network of large, bedrock-incised valleys is preserved on the seabed of the English Channel. Based on analysis of a 30 × 30 m bathymetric grid, the morphology of the valleys was interpreted to be a consequence of erosion by catastrophic flood processes from overspill of a large proglacial lake in the Southern North Sea. The significance of the “megaflood features” has since been recognized by the UK Government with the designation of their protected status in one of three Marine Protected Areas (MPAs) within the palaeovalley in the central English Channel. Here, we analyse recent multibeam bathymetry data (2 × 2 m DEM) from these MPAs, together with backscatter and high-definition seabed camera imagery. The new data allow us to ground truth and refine the earlier interpretation and recognize previously undiscovered finer features. Streamlined valley margins, streamlined islands and metres-deep scours eroded into the valley floor are described at higher detail, while new subtle features on the valley floor such as kilometre-long, sub-parallel inner channels and streamlined bedrock ridges are identified for the first time. These features are consistent with a high energy erosion origin. We also identify isolated large boulders (>1 m length) on flat seabed on the flanks of the palaeovalley, which are consistent with deposition from megaflood processes, although wave action during transgression/regression cannot be ruled out. Our new results enable more robust morphological evidence to support the influence of catastrophic flooding on bedrock valley incision in the English Channel.

Many sedimentary rock successions contain plan-view circular structures, such as impacts, diapirs and carbonate build-ups. When remotely sensed, it can be difficult to discriminate between their formation mechanisms. Here we examine this problem by assessing the origins of circular structures imaged in high-resolution multibeam bathymetric data from Weymouth Bay, UK. The imagery shows 30–150 m across, concave-down structures within the upper Purbeck Limestone Group on the southern limb of the Purbeck Anticline. Similar structures have not been identified in the extensive outcrops around the bay. The morphology and geological setting of the structures are consistent with three different interpretations: carbonate mounds, periclinal folds and evaporite diapirs. However, none of these structures has been previously recorded in the upper Purbeck Limestone Group outcrops of this internationally renowned geological region. We apply a scoring system to 25 features of the circular structures to discriminate between these three alternative interpretations. This analysis indicates that evaporite diapirs are the least likely and carbonate mounds the most likely origin of the structures. The presence of carbonate mounds revises the upper Purbeck palaeofacies distribution in its type area and provides an analogue for the exploration for hydrocarbon reservoirs in lacustrine mounds. Supplementary material: The methods used in this paper and metrics of the circular structures are available at https://doi.org/10.6084/m9.figshare.c.4103840

New seamless onshore to offshore bedrock (1:10 k scale) mapping for the Lyme Bay area is used to resolve the westward termination of the Purbeck–Wight Fault Zone (PWFZ) structure, comprising one of the most prominent, long-lived (Variscan–Cimmerian–Alpine) structural lineaments in the southern UK. The study area lies south of the Variscan Frontal Thrust and overlays the basement Variscide Rhenohercynian Zone, in a region of dominant E-W tectonic fabric and a secondary conjugate NW-SE/NE-SW fabric. The PWFZ comprises one of the E-W major structures, with a typical history including Permian to early Cretaceous growth movement (relating to basement Variscan Thrust reactivation) followed by significant Alpine (Helvetic) inversion. Previous interpretations of the PWFZ have been limited by the low resolution (1:250 k scale) of the available offshore BGS mapping, and our study fills this gap. We describe a significant change in structural style of the fault zone from east to west. In the Weymouth Bay area, previous studies demonstrate the development of focussed strain associated with the PWFZ, accompanied by distributed strain, N-S fault development, and potential basement uplift in its hangingwall. In the Lyme Bay area to the west, faulting is dominantly E-W, with N-S faulting absent. Comparison of the newly mapped faulting networks to gravity data suggests a spatial relationship between this faulting variation and basement variability and uplift.

Four hundred square kilometres of 1 m binned, full coverage swath bathymetry data, integrated with similar resolution onshore topography, have been used to generate a seamless onshore to offshore bedrock map covering an extensive area adjacent to the ‘Jurassic Coast’ World Heritage site. Analysis of these data provides new insights into the structural development of the Purbeck Monocline Cenozoic inversion structure; in particular, variations in the expression of strain between the hanging-wall block and the fault inversion zone. The footwall to the basin-bounding faults compartmentalized deformation and uplift, and acted as a buttress to compression. The data also show a limited thickness changes within the major lithostratigraphical divisions, and a notable absence of basin-related extensional faulting in the offshore area that is in marked contrast to the more extensively studied onshore region. This indicates that prior to inversion, the basin evolved by intermittent activity on a few major extensional faults. This improved understanding of the development of the basin and inversion structures results from our ability to integrate and quantitatively manipulate these high-resolution and spatially extensive offshore and onshore datasets.

International audience; Recognising ice-age catastrophic megafloods is important because they had significant impact on large-scale drainage evolution and patterns of water and sediment movement to the oceans, and likely induced very rapid, short-term effects on climate. It has been previously proposed that a drainage system on the floor of the English Channel was initiated by catastrophic flooding in the Pleistocene but this suggestion has remained controversial. Here we examine this hypothesis through an analysis of key landform features. We use a new compilation of multi- and single-beam bathymetry together with sub-bottom profiler data to establish the internal structure, planform geometry and hence origin of a set of 36 mid-channel islands. Whilst there is evidence of modern-day surficial sediment processes, the majority of the islands can be clearly demonstrated to be formed of bedrock, and are hence erosional remnants rather than depositional features. The islands display classic lemniscate or tear-drop outlines, with elongated tips pointing downstream, typical of streamlined islands formed during high-magnitude water flow. The length-to-width ratio for the entire island population is 3.4 ± 1.3 and the degree-of-elongation or k-value is 3.7 ± 1.4. These values are comparable to streamlined islands in other proven Pleistocene catastrophic flood terrains and are distinctly different to values found in modern-day rivers. The island geometries show a correlation with bedrock type: with those carved from Upper Cretaceous chalk having larger length-to-width ratios (3.2 ± 1.3) than those carved into more mixed Paleogene terrigenous sandstones, siltstones and mudstones (3.0 ± 1.5). We attribute these differences to the former rock unit having a lower skin friction which allowed longer island growth to achieve minimum drag. The Paleogene islands, although less numerous than the Chalk islands, also assume more perfect lemniscate shapes. These lithologies therefore reached island equilibrium shape more quickly but were also susceptible to total erosion. Our observations support the hypothesis that the islands were initially carved by high-water volume flows via a unique catastrophic drainage of a pro-glacial lake in the southern North Sea at the Dover Strait rather than by fluvial erosion throughout the Pleistocene

Many continental flood basalt provinces are spatially and temporally linked with continental break-up. Establishing the relative timing of the two events is a key step in determining their causal relationship. Here we investigate the example of the Deccan Traps and the separation of India and the Seychelles. Whilst there has been a growing consensus as to the age of the main phase of the Deccan emplacement (65.5 ± 1 Ma, chron 29r), the age of the rifting has remained unclear. We resolve this issue through detailed seafloor magnetic anomaly modeling (supported by wide-angle and reflection seismic results) of the north Seychelles and conjugate Laxmi Ridge/Gop Rift margins, and geochemistry and 40Ar/39Ar geochronology of rocks from the north Seychelles margin. We show that syn-rift volcanics offshore the Seychelles Islands in the form of seaward-dipping reflectors were most likely erupted during chron 28n, and the first organized seafloor spreading at the Carlsberg Ridge also initiated during this chron at 63.4 Ma. The severing of the Seychelles occurred by a south-eastward ridge propagation that was completed by the start of chron 27n (~ 62 Ma). A brief, pre-28r phase of seafloor spreading occurred in the Gop Rift, possibly as early as 31r–32n (~ 71 Ma). Initial extension at the margin therefore preceded or was contemporaneous with the Deccan emplacement, and separation of the Seychelles was achieved less than 3.5 Ma afterwards. This is the shortest time interval between flood basalt emplacement and break-up yet reported for any continental flood basalt-rifted margin pair. A contributing factor to the apparently short interval in the Deccan case may be that rifting occurred by a ridge jump into already thinned continental lithosphere. However, we conclude that external plate-boundary forces, rather than the impact of a mantle plume, were largely responsible for the rifting of the Seychelles from India.

The rifting of the Seychelles microcontinent from India involved two phases of extensional activity. The initial separation of the Laxmi Ridge from India was accompanied by extensive magmatism but the later separation of the Seychelles from the Laxmi Ridge was only weakly magmatic. The causal mechanisms linking continental flood basalts, lithospheric extension and mantle plumes, as well as the relative timing of extension and volcanism, are controversial1,2,3,4. The eruption of the Deccan flood basalts was approximately contemporaneous with the separation of the Seychelles microcontinent from India. However, between these continental blocks lies the enigmatic Laxmi Ridge, and the sequence of extensional events that formed these various tectonic elements is poorly understood. Here we present wide-angle seismic data along a profile across Laxmi Ridge that permit delineation of offshore igneous bodies associated with the Deccan magmatism; these bodies are similar to those associated with flood-basalt volcanism and rifting in the Atlantic region and elsewhere1. From the geometry of these bodies, we infer that there were two periods of extension. The first phase, which involved extension between Laxmi Ridge and the Indian subcontinent, was accompanied by significant Deccan-related magmatism. Full development of a continental margin was achieved during the second phase of weakly magmatic extension between Laxmi Ridge and the Seychelles. We suggest that between these rifting events the region passed beyond the reach of lateral flow from the source region of the Deccan flood basalts.

Four phases of cross-cutting tunnel valleys imaged on 3-D seismic datasets are mapped within the Middle–Late Pleistocene succession of the central North Sea basin (Witch Ground area). In plan the tunnel valleys form complex anastomosing networks, with tributary valleys joining main valleys at high angles. The valleys have widths ranging from 250 to 2300 m, and base to shoulder relief varying between 30 and 155 m, with irregular long-axis profiles characteristic of erosion by water driven by glaciostatic pressures. The youngest phase of tunnel valleys are smaller and have a thinner infill than the older generations. The fill of the larger valleys comprises three seismic facies, the lowermost of which has high amplitudes and is discontinuous. The middle facies consists of wedge-shaped packages of low-angle dipping reflectors and is overlain by a facies characterised by sub-horizontal reflectors, which onlap the valley margins. The seismic character, and comparison with lithologies identified in other northwest European Pleistocene tunnel valleys both onshore and offshore, suggests that the lower two seismic facies are most likely sand and gravel-dominated, while the uppermost facies consists of glaciolacustrine and marine muds. The 3-D morphology of the valley margins combined with the geometry of the infill packages suggest that episodic discharge of subglacial meltwater was responsible for incising the valleys and depositing at least some of the infill. Proglacial glaciofluvial deposits are inferred to account for some of the fill overlying the subglacial deposits. Glaciolacustrine and marine muds filled remaining valley topography as the ice sheet retreated. The preserved valley margins are shown to be time-transgressive erosion surfaces that record changes in geometry of the tunnel valley system as it evolved through time, implying that valleys associated with each ice-sheet advance/retreat cycle were dynamic and probably long-lived. Within the constraints of the existing stratigraphy the oldest tunnel valleys in the Witch Ground area of the central North Sea are most likely to be Marine Isotope Stage (MIS) 12 (Elsterian, ca. 470 ka) in age and the youngest pre-MIS 5e (last interglacial, ca. 120 ka). If each tunnel valley phase was formed during the retreat of a major ice sheet then four glaciations with ice coverage of the central North Sea are recorded in the pre-Weichselian, Middle–Late Pleistocene stratigraphy. Copyright © 2006 John Wiley & Sons, Ltd.

We have applied full waveform inversion to wide-aperture seismic reflection data from the southern East Pacific Rise near 14°S. The data contain clear compressional wave and doubly converted shear wave arrivals that provide good constraints on the P and S-wave velocities (Vp, Vs, and hence Poisson's ratio σ) and seismic attenuation (Qp, Qs) structure of seismic layer 2. Layer 2A is highly attenuating (Qp = 18–30 and Qs = 8–15) and layer 2B is moderately attenuating (Qp = 30–50 and Qs = 20–25). Our results show high σ at the seafloor and in layer 2A (σ = 0.48). Across the top of the 2A/B transition the rapid increase in Vp is accompanied by a sharp drop in σ to 0.25 within just 200 m of the seafloor. We perform simple calculations to gain an insight into the porosity and crack distribution with depth. These calculations suggest that porosity is in excess of 30% in layer 2A but reduces to 6–7% at the top of the 2A/B transition and to about 5% at a depth of 600 m below seafloor within layer 2B. Our results suggest that there is an increase in the average aspect ratio with depth across the 2A/B transition. The most likely explanation is that numerous thin cracks either mechanically close or are infilled at depth. Our results show there to be an abrupt change in the pore structure across the 2A/B transition which is consistent with a lithologic transition from extrusives to dykes but is equally consistent with a transition (mechanical or hydrothermal) within the extrusive pile.

We have applied waveform inversion to multichannel seismic reflection data collected at the East Pacific Rise at 9°40'N in order to determine the precise velocity structure of the magma body causing the axial magma chamber reflection. Our analysis supports the idea of a molten sill as previously suggested from forward modeling of seismic data from this location. Our inverted solution has a 30-m-thick sill with a P wave seismic velocity of 2.6 km s -1 . Although not well constrained by the data we believe that the S wave velocity in the sill is not significantly different from 0.0 km s -1 . The low P- and S- wave velocities in the sill imply that it contains less than 30% crystals. The molten sill is underlain by a velocity gradient in which the P wave velocity increases from 2.6 to 3.5 km s -1 over a vertical distance of 50-m. The shape of our velocity-depth profile implies that accretion of material to the roof of the sill is minor compared to accretion to the floor. The underlying velocity gradient zone may represent crystal settling under gravity. We suggest that only material from the 30-m-thick layer can erupt.

Multibeam echosounder surveys and seismic reflection profiling were conducted in the southern Lau back arc basin (SW Pacific). The bathymetric survey covered a 130-km-long section of the active spreading ridge (3 cm/yr half spreading rate), the N-S trending Valu Fa Ridge (VFR), while seismic profiles concentrated on a 35-km-long ridge section. Analysis of these data shows a morphologically segmented ridge which is underlain by a continuous magma chamber. The ridge shows similarities, but also differences, to the segmentation model developed for mid-ocean ridges. Analogous to this model the section of the VFR studied displays third- and fourth-order segmentation. The amplitudes of the along-axis undulations, however, are a factor of 5 to 10 greater than at mid-ocean ridges. Additionally, the VFR has numerous small volcanic cones along the axis, which have previously only been observed at much slower spreading ridges. In cross section the VFR has a very steep, narrow triangular shape which is thought to result from the high viscosity of its andesitic lavas. The axial magma chamber appears to vary in width (1–4 km) and location relative to the axis beneath different morphological segments. Perhaps the most significant deduction from our combined data is that the magma chamber is widest beneath the most prominent axial discontinuity, a small overlapping spreading center (OSC). This is counter to observations from mid-ocean ridges. Off-axis traces of ridge segments allow us to reconstruct the evolution of this OSC for the past 220 kyr. We conclude that this back arc spreading axis shows a segmentation pattern which is similar to mid-ocean ridges, but it is influenced and modified by the geochemistry of its lavas and its proximity to the island arc.

Rifting and magmatism are fundamental geological processes that shape the surface of our planet. A relationship between the two is widely acknowledged but its precise nature has eluded geoscientists and remained controversial. Largely on the basis of detailed observations from the North Atlantic Ocean, mantle temperature was identified as the primary factor controlling magmatic production, with most authors seeking to explain observed variations in volcanic activity at rifted margins in terms of the mantle temperature at the time of break-up. However, as more detailed observations have been made at other rifted margins worldwide, the validity of this interpretation and the importance of other factors in controlling break-up style have been much debated. One such observation is from the northwest Indian Ocean, where, despite an unequivocal link between an onshore flood basalt province, continental break-up and a hot-spot track leading to an active ocean island volcano, the associated continental margins show little magmatism. Here we reconcile these observations by applying a numerical model that accounts explicitly for the effects of earlier episodes of extension. Our approach allows us to directly compare break-up magmatism generated at different locations and so isolate the key controlling factors. We show that the volume of rift-related magmatism generated, both in the northwest Indian Ocean and at the better-known North Atlantic margins, depends not only on the mantle temperature but, to a similar degree, on the rift history. The inherited extensional history can either suppress or enhance melt generation, which can explain previously enigmatic observations.

High-resolution bathymetry is used to derive a new structural interpretation of a submerged inverted graben on the English Channel shelf (northwest Europe). The bathymetry provides a continuous plan-view image of the bedrock geology and resolves fine structural detail, particularly in areas of steeply dipping strata where traditional seismic imaging fails. The imagery, combined with shallow core, deep borehole, and two-dimensional seismic reflection data, shows a marked asymmetry in the bedrock outcrop and structural style, both across and along the axis of the basin. For example, the oldest synrift rocks crop out in a complex south-verging anticline along one margin, whereas more massively bedded younger rocks are folded into a simpler anticline at the other. Comparison of our observations with both recent analog modeling and a neighboring inverted basin highlights the importance of the mechanical properties of the basin stratigraphy in controlling the structural development. We conclude that the inversional asymmetry developed because of the distribution of strong competence contrasts within the synrift fill.

We analyze a dense grid of seismic normal incidence and wide-angle data collected over an active back arc spreading center in the Lau Basin of the southwest Pacific. Our survey area covers the whole of a 35-km-long morphological segment and a small overlapping spreading center. A bright reflector, which is coincident with a velocity inversion at a depth of 3.2 ± 0.2 km below the seafloor, is observed on every one of the 40 across-axis profiles. We interpret this reflector as being the roof of a crustal magma chamber. The widest magma chamber reflector occurs beneath the overlapping spreading center, where it extends up to 4 km, being imaged beneath both ridges and, in places, beneath the overlap basin. Elsewhere the width of the reflector varies between 0.6 and 2.3 ±0.4 km. The narrowest reflectors are observed beneath deviations (devais) of the ridge axis. In addition to being observed on each of the kilometer-spaced across-axis profiles the reflector is also seen as a continuous event for at least 10 km on a profile along the ridge axis. The reflection coefficient of the event is between −0.34 and −0.65. This requires the melt at the top of the chamber to have a P wave velocity less than 2.7 km/s if the interface is planar or less than 3.8 km/s if the interface is layered.

The granitic islands of the Seychelles Plateau have long been recognised to overlie continental crust, isolated from Madagascar and India during the formation of the Indian Ocean. However, to date the extent of continental crust beneath the Seychelles region remains unknown. This is particularly true beneath the Mascarene Basin between the Seychelles Plateau and Madagascar and beneath the Amirante Arc. Constraining the size and shape of the Seychelles continental fragment is needed for accurate plate reconstructions of the breakup of Gondwana and has implications for the processes of continental breakup in general. Here we present new estimates of crustal thickness and V P / V S from H–κ stacking of receiver functions from a year long deployment of seismic stations across the Seychelles covering the topographic plateau, the Amirante Ridge and the northern Mascarene Basin. These results, combined with gravity modelling of historical ship track data, confirm that continental crust is present beneath the Seychelles Plateau. This is ∼ 30 – 33 km thick, but with a relatively high velocity lower crustal layer. This layer thins southwards from ∼ 10 km to ∼ 1 km over a distance of ∼ 50 km , which is consistent with the Seychelles being at the edge of the Deccan plume prior to its separation from India. In contrast, the majority of the Seychelles Islands away from the topographic plateau show no direct evidence for continental crust. The exception to this is the island of Desroche on the northern Amirante Ridge, where thicker low density crust, consistent with a block of continental material is present. We suggest that the northern Amirantes are likely continental in nature and that small fragments of continental material are a common feature of plume affected continental breakup.

We present a model of the northern Seychelles continental margin derived from controlled source, wide-angle seismic traveltime inversion and teleseismic receiver functions. This margin has been widely cited as a classic example of rifting in association with a continental flood basalt province, the Deccan Traps. However, we do not find the typical set of geophysical characteristics reported at other margins linked to continental flood basalts, such as those of the north Atlantic. The oceanic crust formed immediately after breakup and throughout the first 3 Ma of seafloor spreading is just 5.2 km thick, less than half that typically seen at other volcanic margins. The continent-ocean transition zone is narrow and while two packages of seaward-dipping reflectors are imaged within this transition they are weakly developed. Beneath the thinned continental crust there is an approximately 4 km thick layer of high-velocity material (7.5–7.8 km/s) that we interpret as mafic material intruded and underplating the lower crust. However, we believe that this underplating most likely happened prior to the breakup. Overall the observations show that the rifting of India from the Seychelles was characterized by modest magmatism. The spatial extent of the Deccan flood basalt province is therefore smaller than previously thought. We speculate that either the lateral flow of Deccan-related hot material beneath the breakup region was hampered, perhaps as the rifted margins did not intersect the center of the Deccan source, or there was incomplete melt extraction from the wide melting region that formed between the rapidly diverging plates. If the latter explanation is correct, then the rate of plate separation, as indicated by the initial seafloor-spreading rate, is more important in controlling the volume of magmatism generated during continental rifting than has been previously recognized.

Two of the main factors that determine the nature of a rifted continental margin are rheology and magmatism during extension. Numerical models of lithospheric extension suggest that both factors vary with extension rate; yet until now extension rates of studied margins, as indicated by the rate of initial seafloor spreading, are mostly less than -30 mm/yr on each margin. This article presents the first geophysical results from the Seychelles-Laxmi Ridge conjugate pair of rifted margins which separated at -65 mm/yr. The Seychelles, with its spectacular exposures of Precambrian granite, was the earliest scientifically recognized microcontinent and arguably remains the classic example of one [Wegener, 1924; Matthews and Davies, 1966]. However, it is still unknown whether microcontinents result from plumes, changes in plate-boundary forces, lithospheric heterogeneity, or a combination of these factors.