Abstract This book considers the geology between North and South America. It contributes to debate about the area's evolution, particularly that of the Caribbean. Prevailing understanding is that the Caribbean formed in the Pacific and was engulfed between the Americas as the latter drifted west. Accordingly, the Caribbean Plate comprises internal, Jurassic–Cretaceous oceanic rocks, thickened into a Cretaceous hotspot/plume plateau, with obducted ophiolites and Cretaceous–Palaeogene, subduction-related, intra-oceanic volcanic arc and metamorphosed arc/continental rocks exposed on its margins. An alternative interpretation is that the Caribbean evolved in place. It consists largely of continental crust, extended in the Triassic–Jurassic, which subsided below thick Jurassic–Cretaceous carbonate rocks and flood basalts, and Cenozoic carbonate and clastic rocks. After uplift of ‘oceanic’ and volcanic arc rocks onto (continental) margins, the interior foundered in the Middle Eocene. Papers range from regional overviews and discussions of Caribbean origins to aspects of local geology arranged in a circum-Caribbean tour and ending in the interior. They address tectonics, structure, geochronology, seismicity, igneous and metamorphic petrology, metamorphism, geochemistry, stratigraphy and palaeontology.

Abstract We present an updated synthesis of the widely accepted ‘single-arc Pacific-origin’ and ‘Yucatán-rotation’ models for Caribbean and Gulf of Mexico evolution, respectively. Fourteen palaeogeographic maps through time integrate new concepts and alterations to earlier models. Pre-Aptian maps are presented in a North American reference frame. Aptian and younger maps are presented in an Indo-Atlantic hot spot reference frame which demonstrates the surprising simplicity of Caribbean–American interaction. We use the Müller et al. ( Geology 21 : 275–278, 1993) reference frame because the motions of the Americas are smoothest in this reference frame, and because it does not differ significantly, at least since c. 90 Ma, from more recent ‘moving hot spot’ reference frames. The Caribbean oceanic lithosphere has moved little relative to the hot spots in the Cenozoic, but moved north at c. 50 km/Ma during the Cretaceous, while the American plates have drifted west much further and faster and thus are responsible for most Caribbean–American relative motion history. New or revised features of this model, generally driven by new data sets, include: (1) refined reconstruction of western Pangaea; (2) refined rotational motions of the Yucatán Block during the evolution of the Gulf of Mexico; (3) an origin for the Caribbean Arc that invokes Aptian conversion to a SW-dipping subduction zone of a trans-American plate boundary from Chortís to Ecuador that was part sinistral transform (northern Caribbean) and part pre-existing arc (eastern, southern Caribbean); (4) acknowledgement that the Caribbean basalt plateau may pertain to the palaeo-Galapagos hot spot, the occurrence of which was partly controlled by a Proto-Caribbean slab gap beneath the Caribbean Plate; (5) Campanian initiation of subduction at the Panama–Costa Rica Arc, although a sinistral transform boundary probably pre-dated subduction initiation here; (6) inception of a north-vergent crustal inversion zone along northern South America to account for Cenozoic convergence between the Americas ahead of the Caribbean Plate; (7) a fan-like, asymmetric rift opening model for the Grenada Basin, where the Margarita and Tobago footwall crustal slivers were exhumed from beneath the southeast Aves Ridge hanging wall; (8) an origin for the Early Cretaceous HP/LT metamorphism in the El Tambor units along the Motagua Fault Zone that relates to subduction of Farallon crust along western Mexico (and then translated along the trans-American plate boundary prior to onset of SW-dipping subduction beneath the Caribbean Arc) rather than to collision of Chortis with Southern Mexico; (9) Middle Miocene tectonic escape of Panamanian crustal slivers, followed by Late Miocene and Recent eastward movement of the ‘Panama Block’ that is faster than that of the Caribbean Plate, allowed by the inception of east–west trans-Costa Rica shear zones. The updated model integrates new concepts and global plate motion models in an internally consistent way, and can be used to test and guide more local research across the Gulf of Mexico, the Caribbean and northern South America. Using examples from the regional evolution, the processes of slab break off and flat slab subduction are assessed in relation to plate interactions in the hot spot reference frame.

Abstract Caribbean Plate margins are assemblages of terranes located, since the Mid-Cretaceous, along transform boundaries between the Caribbean, North and South America and the Pacific and Atlantic oceans. Litho-stratigraphic, petrological and metamorphic features of the main units and their regional correlations allow definition of the main geotectonic elements (continental margins, oceanic basins, subduction zones, magmatic arcs) involved in the evolution of Caribbean Plate margins. They provide valuable constraints on plate evolution since the Jurassic. This involved proto-Caribbean ocean opening, thickening into an oceanic plateau, beginning of convergence in the Early Cretaceous, atypical evolution of a supra-subduction system during the Mid-Cretaceous, subduction of rifted continental margins, Late Cretaceous convergence related to eastward migration of two opposite triple-junctions and strike–slip tectonics. Using these data, we compare different models and suggest improvements.

Abstract Compiled and synthesized geological data suggest that the Caribbean Plate consists of dispersed continental basement blocks, wedges of ?Triassic–Jurassic clastic rocks, Jurassic–Late Cretaceous carbonate rocks, volcanic arc rocks, widespread, probably subaerial basalts and serpentinized upper mantle. This points to an in situ origin of the Caribbean Plate as part of Middle America, continuing the geology of the eastern North America margin in a more extensional tectonic setting. Extension increases from the Gulf of Mexico through the Yucatán Basin to the Caribbean.

Abstract Regional geological data and global analogues suggest Caribbean Plate geology continues that seen along the margin of eastern North America in a more extensional setting, between the diverging Americas. From west to east there are continental masses with Triassic rifts, proximal continental blocks with kilometres-thick Mesozoic carbonates, more distal areas of Palaeozoic horsts flanked by Triassic–Jurassic dipping wedges of sediments, including salt and overlain by Cretaceous basalts, and most distal areas of serpentinized upper mantle. Plate history began along with the Late Triassic formation of the Central Atlantic Magmatic Province and involved Triassic–Jurassic rifting, Jurassic–Early Cenozoic extension and Oligocene–Recent strike–slip. Great extension promoted volcanism, foundering, eastward growth of the plate by backarc spreading and distribution of continental fragments on the plate interior and along its margins. Hydrocarbons probably are present. Caribbean geology has important implications for understanding of oceanic plateaus, intra-oceanic volcanic arcs, the ‘andesite problem’ and genesis of ‘subduction’ HP/LT metamorphic rocks. The model can be tested by re-examination of existing samples and seismic data and by deep sea drilling.

Abstract Tightly curved mountain belts are prominent features of global topography. Typically, these ‘oroclines’ occur in areas of regional compression but enclose basins where extension has been contemporaneous with outward directed thrusting in the orogens. Examples of such basin–orogen pairs include the Alboran Sea–Gibraltar Arc, Tyrrhenian Sea–Aeolian Arc, Aegean Sea–Hellenic Arc and Pannonian Basin–Carpathian Arc, all in the western Tethys but matched in the eastern Tethys by the Banda Sea and Outer Banda Arc. The development of the basins has been variously explained by gravitational collapse of rapidly elevated mountain blocks and by extrusion prompted by asthenopheric flows, but it is not even universally agreed that similar processes have operated in all cases. Critics have cited gross differences in volcanic activity (absent from the Gibraltar Arc, modest in the Carpathians but intense in other examples) and in the presence or absence of recognizable Wadati–Benioff Zones. The superficial similarities between the Caribbean Sea–Antilles Arcs and typical oroclinal basin–orocline pairs have recently been invoked in support of an in situ Caribbean evolutionary model, even though the disputed origins of oroclines limit their reliability as analogues. The Caribbean's considerably greater area further emphasizes the need for caution, while the most obvious objection to identifying it as a member of the oroclinal group is its very long history. Oroclinal basins typically pass from initiation to effective stabilization in a few tens of millions of years, whereas the original Caribbean oceanic crust, which is now bounded to the east and west by active subduction zones, was probably formed in the Jurassic. Rather than invoking an overall common origin for the Caribbean and the Tethyan basins, it is more useful to look for shared causes of specific individual similarities. The impact of a rigid block might be as effective in imposing curvature on a mountain belt as rapid expansion in an adjacent area. However, it does seem that the case for the crust of the Caribbean being typical of oceanic large igneous provinces (LIPs) may have been overstated and, in the light of oroclinal analogues, that some features of the still poorly understood Beata Ridge and Lower Nicaragua Rise may be most easily explained by east–west extension promoted by the convergence between North and South America.

Abstract This work analyses the present-day principal strain orientation on the downgoing slab of the South America Plate (SAM) beneath the Sandwich Plate (SAND). The strain regime was deduced from the study of 331 earthquake focal mechanism solutions examined by fault population analysis methods. In the slab, the maximum horizontal shortening direction (ey) rotates in trend in a clockwise direction from NE in the north, to SE in the south. Based on this rotation, three different areas were defined according to the prevailing focal mechanism type: (1) the North Zone, with ey oriented N058°E and reverse and strike–slip focal mechanisms; (2) the Central Zone, with only reverse focal mechanisms and ey striking N080°E; and (3) the South Zone, with ey oriented N106°E and reverse and strike–slip focal mechanisms. The strain field in the North Zone of the SAND involves decoupling of the slab at approximately 70 km depth. In contrast, the South Zone edge slab exhibits no decoupling and it exhibits different geometry (hook-like shaped) from the North Zone. Finally, we define the dextral strike–slip component acting at the South Sandwich Fracture Zone (SSFZ), according to focal mechanism solutions and the regional tectonic configuration.

Abstract The geology of southwestern Mexico (102–96°W) records several synchronous events in the Late Oligocene–Early Miocene (29–19 Ma): (1) a hiatus in arc magmatism; (2) removal of a wide ( c . 210 km) Upper Eocene–Lower Oligocene forearc; (3) exhumation of 13–20 km of Upper Eocene–Lower Oligocene arc along the present day coast; and (4) breakup of the Farallon Plate. Events 2 and 3 have traditionally been related to eastward displacement of the Chortís Block from a position off southwestern Mexico between 105°W and 97°W; however at 30 Ma the Chortís Block would have lain east of 95°W. We suggest that the magmatic hiatus was caused by subduction of the forearc, which replaced the mantle wedge by relatively cool crust. Assuming that the subducted block separated along the forearc–arc boundary, a likely zone of weakness due to magmatism, the subducted forearc is estimated to be wedge-shaped varying from zero to c . 90 km in thickness; however such a wedge is not apparent in seismic data across central Mexico. Given the 121 km/Ma convergence rate between 20 and 10 Ma and 67 km/Ma since 10 Ma, it is probable that any forearc has been deeply subducted. Potential causes for subduction of the forearc include collision of an oceanic plateau with the trench, and a change in plate kinematics synchronous with breakup of the Farallon Plate and initiation of the Guadalupe–Nazca spreading ridge.

Abstract We present analogue models that illustrate the tectonic evolution of the continental margin of southwestern Mexico and the Early Cenozoic deformation of the Xolapa complex. Together with geological data they suggest that oblique convergence caused distributed deformation and mountain building near the present-day margin of southern Mexico in a general left-lateral transpressional regime. A similar deformation is also observed north of the Xolapa complex in Maastrichtian to Paleocene sedimentary and volcanic rock units. Since post-Oligocene exhumation of middle crust does not significantly affect Late Eocene to Oligocene volcanic rocks, we infer that the evolution of the transform margin led to the formation of discrete boundaries that eventually decoupled exhumed mid-lower crust from the onshore upper-crust sequences since the Late Eocene.

Abstract I propose a new seismotectonic model for the Chortís Block, at the northwestern corner of the Caribbean Plate. Shallow seismicity in the area clearly shows three zones of deformation: one along the North America–Caribbean Plate boundary and another along the Central America volcanic arc, and one in the area of the grabens of northern Central America. Analysis of Centroid moment–tensor solutions for shallow earthquakes in these three area show that T or tension axes are horizontal and trend away from the corner, and that P or compression axes for the plate boundary and the volcanic arc are also horizontal and trending towards the corner. Calculation of seismic moment release per unit volume reveals similar values for the volcanic arc and the plate boundary. The state of stress and similarity in seismic moment release suggest that the Chortís Block is being extruded towards the ESE. This is probably due to compression of the large North America and Cocos Plates that surround it.

Abstract The Caribbean Plate consists of a plateau basalt, formed probably in the Middle Cretaceous, complicated by a continental block, Chortís, several magmatic arcs, strike–slip motions along major fault systems such as the Motagua–Polochic fault zone in Guatemala, the pull-apart basin of the Cayman Trough and subduction zones below Central America and the Lesser Antilles. Five major collisional events have been identified: (i) Late Paleocene–Middle Eocene collision of the Greater Antilles with the Bahamas platform; (ii) Late Cretaceous collision of Chortís with the Maya Block; (iii) emplacement of nappes upon the Venezuelan foreland in the Cenozoic; (iv) collision of the Western Cordillera oceanic complex with the Central Cordillera of Colombia; and (v) Miocene collision of the eastern Costa Rica–Panama arc with the Western Cordillera. All these ‘orogenic events’ show an eastward movement of the Caribbean Plate relative to the Americas. Migration of the Jamaica Block from the Pacific caused obduction of the oldest ophiolites of Huehuetenango at the western end of the Polochic–Río Negro faults in Guatemala. South-southwest migration of the Chortís Block from west of Mexico and northward towards the Maya Block destroyed a trench associated with the Motagua–Jalomáx fault system and caused the Chuacús Orogeny, emplacing Guatemalan ophiolite complexes and metamorphosing the rocks from the Chuacús Series.

Abstract New structural, geochronological, and petrological data highlight which crustal sections of the North American–Caribbean Plate boundary in Guatemala and Honduras accommodated the large-scale sinistral offset. We develop the chronological and kinematic framework for these interactions and test for Palaeozoic to Recent geological correlations among the Maya Block, the Chortís Block, and the terranes of southern Mexico and the northern Caribbean. Our principal findings relate to how the North American–Caribbean Plate boundary partitioned deformation; whereas the southern Maya Block and the southern Chortís Block record the Late Cretaceous–Early Cenozoic collision and eastward sinistral translation of the Greater Antilles arc, the northern Chortís Block preserves evidence for northward stepping of the plate boundary with the translation of this block to its present position since the Late Eocene. Collision and translation are recorded in the ophiolite and subduction–accretion complex (North El Tambor complex), the continental margin (Rabinal and Chuacús complexes), and the Laramide foreland fold–thrust belt of the Maya Block as well as the overriding Greater Antilles arc complex. The Las Ovejas complex of the northern Chortís Block contains a significant part of the history of the eastward migration of the Chortís Block; it constitutes the southern part of the arc that facilitated the breakaway of the Chortís Block from the Xolapa complex of southern Mexico. While the Late Cretaceous collision is spectacularly sinistral transpressional, the Eocene–Recent translation of the Chortís Block is by sinistral wrenching with transtensional and transpressional episodes. Our reconstruction of the Late Mesozoic–Cenozoic evolution of the North American–Caribbean Plate boundary identified Proterozoic to Mesozoic connections among the southern Maya Block, the Chortís Block, and the terranes of southern Mexico: (i) in the Early–Middle Palaeozoic, the Acatlán complex of the southern Mexican Mixteca terrane, the Rabinal complex of the southern Maya Block, the Chuacús complex, and the Chortís Block were part of the Taconic–Acadian orogen along the northern margin of South America; (ii) after final amalgamation of Pangaea, an arc developed along its western margin, causing magmatism and regional amphibolite–facies metamorphism in southern Mexico, the Maya Block (including Rabinal complex), the Chuacús complex and the Chortís Block. The separation of North and South America also rifted the Chortís Block from southern Mexico. Rifting ultimately resulted in the formation of the Late Jurassic–Early Cretaceous oceanic crust of the South El Tambor complex; rifting and spreading terminated before the Hauterivian ( c . 135 Ma). Remnants of the southwestern Mexican Guerrero complex, which also rifted from southern Mexico, remain in the Chortís Block (Sanarate complex); these complexes share Jurassic metamorphism. The South El Tambor subduction–accretion complex was emplaced onto the Chortís Block probably in the late Early Cretaceous and the Chortís Block collided with southern Mexico. Related arc magmatism and high- T /low- P metamorphism (Taxco–Viejo–Xolapa arc) of the Mixteca terrane spans all of southern Mexico. The Chortís Block shows continuous Early Cretaceous–Recent arc magmatism. Supplementary material: Analytical methods and data, and sample description are available at http://www.geolsoc.org.uk/SUP18360.

Abstract Precambrian and Palaeozoic basements are present in southern Mexico and Central America, where several crustal blocks are recognized by their different geological record, and juxtaposed along lateral faults. Pre-Mesozoic reconstructions must take into account the nature of such crustal blocks, their geological history, age and petrology. Some of those crustal blocks are currently located between southernmost north America (the Maya Block) and Central America (Chortís Block).To better understand the geology of these crustal blocks, and to establish comparisons between their geological history, we performed U–Pb dating of both igneous and metasedimentary key units cropping out in central and western Guatemala. In the Altos Cuchumatanes (Maya Block) granites yield both Permian (269±29 Ma) and Early Devonian (391±7.4 Ma) U–Pb ages. LA-ICPMS detrital zircon ages from rocks of the San Gabriel sequence, interpreted as the oldest metasedimentary unit of the Maya Block, and overlain by the Late Palaeozoic Upper Santa Rosa Group, yield Precambrian detrital zircons bracketed between c . 920 and c . 1000 Ma. The presence of these metasedimentary units, as well as Early Devonian to Silurian granites in the Mayan continental margin, from west (Altos Cuchumatanes), to east (Maya Mountains of Belize) indicates a more or less continuous belt of Lower Palaeozoic igneous activity, also suggesting that the continental margin of the Maya Block can be extended south of the Polochic fault, up to the Baja Verapaz shear zone. A metasedimentary sample belonging to the Chuacús Complex yielded detrital zircons with ages between c . 440 and c . 1325 Ma. The younger ages are similar to the igneous ages reported from the entire southern Maya continental margin, and show proximity of the Complex in the Middle-Late Palaeozoic. The S. Diego Phyllite, which overlies high-grade basement units of the Chortís Block, contains zircons that are Lower Cambrian ( c . 538 Ma), Mesoproterozoic ( c . 980 to c . 1150 Ma) and even Palaeoproterozoic ( c . 1820 Ma). Absence of younger igneous zircons in the San Diego Phyllite indicates that either its sedimentation took place in a close range of time, during the Late Cambrian, or absence of connection between Chortís and Maya Blocks during the Early–Mid-Palaeozoic. The Precambrian zircons could have come from southern Mexico (Oaxaca and Guichicovi Complexes), or from Mesoproterozoic Massifs exposed in Laurentia and Gondwana. Palaeogeographic models for Middle America are limited to post-Jurassic time. The data presented here shed light on Palaeozoic and, possibly, Precambrian relationships. They indicate that Maya and the Chortís did not interact directly until the Mesozoic or Cenozoic, as they approached their current position.

Abstract The Mesozoic Proto-Caribbean Plate was consumed in the subduction zone of the Greater Antilles volcanic arc until the Campanian. At this time, volcanic arc magmatism ceased along Cuba. From Late Campanian to Danian, Cuba and its surroundings were a collision zone where the GAC accreted to the North American palaeomargin. In the Danian the almost east–west trending SE Cuba–Cayman Ridge–Hispaniola? volcanic arc was born. The related north dipping subduction zone acted as the SE North American plate boundary. From the Paleocene to Middle Eocene dense Caribbean lithosphere travelled northwards. The location, strike and subduction polarity of the assumed subduction zone are very different from those described by other models. Almost simultaneously the Cuban Orogeny developed in western and central Cuba. During the orogeny the northern ophiolite belt of Cuba and the Cretaceous volcanic rocks were thrust northwards tens of kilometres, onto the Mesozoic North American palaeomargin. In the Middle Eocene subduction stopped. Simultanously(?) a change in the regional stress field originated the near east–west trending sinistral Oriente fault zone, whose position and origin are probably tied to the weakened hot crust to the south of the Palaeogene volcanic arc axis.

Abstract The Benbow Inlier in Jamaica contains the Devils Racecourse Formation, which is composed of a Hauterivian to Aptian island arc succession. The lavas can be split into a lower succession of basaltic andesites and dacites/rhyolites, which have an island arc tholeiite (IAT) composition and an upper basaltic and basaltic andesite sequence with a calc-alkaline (CA) chemistry. Trace element and Nd–Hf isotopic evidence reveals that the IAT and CA lavas are derived from two chemically similar mantle wedge source regions predominantly composed of normal mid-ocean ridge-type spinel lherzolite. In addition, Th-light rare earth element/high field strength element–heavy rare earth element ratios, Nd–Hf isotope systematics, (Ce/Ce*) n-mn and Th/La ratios indicate that the IAT and CA mantle wedge source regions were enriched by chemically distinct slab fluxes, which were derived from both the altered basaltic portion of the slab and its accompanying pelagic and terrigenous sedimentary veneer respectively. The presence of IAT and CA island arc lavas before and after the Aptian–Albian demonstrates that the compositional change in the Great Arc of the Caribbean was the result of the subduction of chemically differing sedimentary material. There is therefore no evidence from the geochemistry of this lava succession to support arc-wide subduction polarity reversal in the Aptian–Albian. Supplementary material: References for data sources used in figures can be found at: http://www.geolsoc.org.uk/SUP18361.

Abstract Within the last decade, modern petrological and geochronological methods in combination with detailed studies of the field geology have allowed the reconstruction of tectonic processes in the northwestern part of the Caribbean Plate. The development of an oceanic Proto-Yucatán Basin can be traced from the Late Jurassic to the Mid-Cretaceous. From the Mid-Cretaceous onward, an interaction of this basin with the Caribbean Arc can be observed. Geochronological data prove continuous magmatic activity and generation of HP mineral suites in the Caribbean Arc from the Aptian to the Campanian/Maastrichtian. Magmatism ceased at least in onshore central Cuba at about 75 Ma, probably as the southern edge of the continental Yucatán Block began to interact with the advancing arc system. Similarly, the youngest recorded ages for peak metamorphism of high-pressure metamorphic rocks in Cuba cluster at 70 Ma; rapid uplift/exhumation of these rocks occurred thereafter. After this latest Cretaceous interaction with the southern Yucatán Block, the northern Caribbean Arc was dismembered as it entered the Proto-Yucatán Basin region. Because of the continued NE-directed movement, Proto-Yucatán Basin sediments were accreted to the arc and now form the North Cuban fold and thrust belt. Parts of the island arc have been thrust onto the southern Bahamas Platform along the Eocene suture zone in Cuba. Between the arc's interaction with Yucatán and the Bahamas ( c . 70 to c . 40 Ma), the Yucatán intra-arc basin opened by extreme extension and local seafloor accretion between the Cayman Ridge (still part of Caribbean Plate) and the Cuban frontal arc terranes, the latter of which were kinematically independent of the Caribbean. Although magmatism ceased in central Cuba by 75 Ma, traces of continuing Early Palaeogene arc magmatism have been identified in the Cayman Ridge, suggesting that magmatism may not have ceased in the arc as a whole, but merely shifted south relative to Cuba. If so, a shallowing of the subduction angle during the opening of the Yucatán Basin would be implied. Further, this short-lived (?) Cayman Ridge arc is on tectonic strike with the Palaeogene arc in the Sierra Maestra of Eastern Cuba, suggesting south-dipping subduction zone continuity between the two during the final stages of Cuba–Bahamas closure. After the Middle Eocene, the east–west opening of the Cayman Trough left the present Yucatán Basin and Cuba as part of the North American Plate. The subduction geometry, P–T–t paths of HP rocks in Cuban mélanges, the time of magmatic activity and preliminary palaeomagnetic data support the conclusion that the Great Antillean arc was initiated by intra-oceanic subduction at least 900 km SW of the Yucatán Peninsula in the ancient Pacific. As noted above, the Great Antillean Arc spanned some 70 Ma prior to its Eocene collision with the Bahamas. This is one of the primary arguments for a Pacific origin of the Caribbean lithosphere; there simply was not sufficient space between the Americas, as constrained by Atlantic opening kinematics, to initiate and build the Antillean (and other) arcs in the Caribbean with in situ models.

Abstract The Early Cretaceous island arc lavas in the Caribbean region are frequently assigned to the primitive island arc (PIA) series and not to the island arc tholeiite (IAT) series. However, this review demonstrates that the Caribbean PIA rocks have immobile trace element abundances, trace element ratios and Nd–Hf isotope systematics which are indistinguishable from modern IAT lavas. Thus, it is proposed that the term PIA series be discarded and that the Early Cretaceous island arc rocks in the Caribbean be classified as IAT rocks. Supplementary material: References for data sources used in figures can be found at: http://www.geolsoc.org.uk/SUP18362.

Abstract Multidisciplinary study of the Osa and Burica peninsulas, Costa Rica, recognizes the Osa Igneous Complex and the Osa Mélange – records of a complex Late Cretaceous–Miocene tectonic–sedimentary history. The Igneous Complex, an accretionary prism ( sensu stricto ) comprises mainly basaltic lava flows, with minor sills, gabbroic intrusives, pelagic limestones and radiolarites. Sediments or igneous rocks derived from the upper plate are absent. Four units delimited on the base of stratigraphy and geochemistry lie in contact along reactivated palaeo-décollement zones. They comprise fragments of a Coniacian–Santonian oceanic plateau (Inner Osa Igneous Complex) and Coniacian–Santonian to Middle Eocene seamounts (Outer Osa Igneous Complex). The units are unrelated to other igneous complexes of Costa Rica and Panama and are exotic with respect to the partly overthickened Caribbean Plate; they formed by multiple accretions between the Late Cretaceous and Middle Eocene, prior to the genesis of the mélange. Events of high-rate accretion alternated with periods of low-rate accretion and tectonic erosion. The NW Osa Mélange in contact with the Osa Igneous Complex has a block-in-matrix texture at various scales, produced by sedimentary processes and later tectonically enhanced. Lithologies are mainly debris flows and hemipelagic deposits. Clastic components (grains to large boulders) indicate Late Eocene mass wasting of the Igneous Complex, forearc deposits and a volcanic arc. Gravitational accumulation of a thick pile of trench sediments culminated with shallow-level accretion. Mass-wasting along the margin was probably triggered by seamount subduction and/or plate reorganization at larger scale. The study provides new geological constraints for seamount subduction and associated accretionary processes, as well as on the erosive/accretionary nature of convergent margins devoid of accreted sediments. Supplementary material: Supplementary material: Sample localities and analytical data can be found at http://www.geolsoc.org.uk/SUP18363.

Abstract La Désirade in the Lesser Antilles contains one of the rare fragments of Jurassic oceanic crust known on Caribbean islands. Others in the northeastern Caribbean occur on Puerto Rico and Hispaniola. These fragments each include radiolarian-bearing chert that has been linked to an origin in the Pacific Ocean. Of these, a fragment in Sierra Bermeja, Puerto Rico is clearly of Pacific origin as it contains Lower Jurassic radiolarians that predate the opening between North and South America. Red ribbon chert at El Aguacate, Dominican Republic is essentially identical to widespread radiolarites found in accreted material of the Pacific basin and from Pacific Ocean ODP Site 801. Extensive sampling in the Atlantic basin has produced no Jurassic radiolarites. Thus, based on age (the older of the Sierra Bermeja outcrops) and lithology (El Aguacate), two of these fragments are definitely of Pacific origin. Re-evaluation of the chert/pillow lava sequence on La Désirade in light of recent discoveries at spreading ridges has resulted in a revised interpretation of their probable origin. A wide range of features of these cherts indicate pelagic and hydrothermal sedimentation at an Upper Jurassic spreading ridge, one that almost assuredly existed in the eastern Pacific realm. These features include: the types of chert found on the island, lack of argillaçeous partings, small outcrop size, discontinuous chert bodies, presence of limestone squeeze-ups into pillow lavas and indications of hydrothermal activity, including epidotization of basalt migrating outward from pillow margins with chert rinds that record pelagic and hydrothermal sedimentation at an Upper Jurassic spreading ridge, one that almost assuredly existed in the extreme eastern Pacific realm.