Clastic Domains of Sandstones in Central/Eastern Venezuela, Trinidad, and Barbados: Heavy Mineral and Tectonic Constraints on Provenance and Palaeogeography

Please do not cite until published. Expected early 2009. In: James, K., Lorente, M. A. & Pindell, J. (eds) The geology and evolution of the region between North and South America, Geological Society of London, Special Publication. Clastic domains of sandstones in central/eastern Venezuela, Trinidad, and Barbados: heavy mineral and tectonic constraints on provenance and palaeogeography

James Pindell1,2, Lorcan Kennan1, David Wright3, and Johan Erikson4

1. Tectonic Analysis Ltd., Chestnut House, Duncton, West Sussex, GU28 0LH, UK 2. Also at: Dept. Earth Science, Rice University, Houston, TX 77002, USA 3. Department of Geology, University of Leicester, Leicester, LE1 7RH, UK 4. Department of Natural Sciences, St. Joseph’s College, Standish, ME 07084, USA

Corresponding first author: [email protected]

Supplementary material: Location maps and detailed heavy mineral data tables are available at http://www.geolsoc.org.uk/SUP00000

Abstract: Current models for the tectonic evolution of northeastern South America invoke a Palaeogene phase of inter-American convergence, followed by diachronous dextral oblique collision with the Caribbean Plate, becoming strongly transcurrent in the Late Miocene. Heavy mineral analysis of Cretaceous to Pleistocene rocks from Eastern Venezuela, Barbados and Trinidad allow us to define six primary clastic domains, refine our palaeogeographic maps, and relate them to distinct stages of tectonic development: (1) Cretaceous passive margin of northern South America; (2) Palaeogene clastics related to the dynamics of the Proto-Caribbean Inversion Zone before collision with the Caribbean plate; (3) Late Eocene–Oligocene southward-transgressive clastic sediments fringing the Caribbean foredeep during initial collision; (4) Oligocene–Middle Miocene axial fill of the Caribbean foredeep; (5) Late Eocene–Middle Miocene northern proximal sedimentary fringe of the Caribbean thrustfront; and (6) Late Miocene–Recent deltaic sediments flowing parallel to the orogen during its post-collisional, mainly transcurrent stage. Domain 1–3 sediments are highly mature, comprising primary Guayana Shield- derived sediment or recycled sediment of shield origin eroded from regional Palaeogene unconformities. In Trinidad, palinspastic restoration of Neogene deformation indicates that facies changes once interpreted as north to south are in fact west to east, reflecting progradation from the Maturín Basin into central Trinidad across the northwest-southeast trending Bohordal marginal offset, distorted by about 70 km of dextral shear through Trinidad. There is no mineralogical indication of a northern or northwestern erosional sediment source until Oligocene onset of Domain 4 sedimentation. Palaeocene–Middle Eocene rocks of the Scotland Formation sandstones in Barbados do show an immature orogenic signature, in contrast to Venezuela-Trinidad Domain 2 sediments, this requires: (1) at least a bathymetric difference, if not a tectonic barrier, between them; and (2) that the Barbados deep-water depocentre was within turbidite transport distance of the early Palaeogene orogenic source areas of western Venezuela and/or Colombia. Domains 4–6 (from Late Oligocene) show a strong direct or recycled influence of Caribbean Orogen igneous and metamorphic terranes in addition to substantial input from the shield areas to the south. The delay in the appearance of common Caribbean detritus in the east, relative to the Palaeocene and Eocene appearance of Caribbean-influenced sands in the west, reflects the diachronous, eastward migration of Caribbean foredeep subsidence and sedimentation as a response to eastward-younging collision of the Caribbean Plate and the South American margin.

Current kinematically rigorous Cenozoic models for the evolution of northeastern South America invoke a Palaeogene tectonic phase due to inter-American convergence followed by diachronous dextral oblique collision with the Caribbean Plate (Pindell et al. 1991, 2006; Perez de Armas 2005), which became strongly transcurrent in the Late Miocene (Pindell et al. 1998, 2005). In eastern Venezuela and Trinidad, the collision and subsequent shear between the Caribbean Plate and South America (Fig. 1) juxtaposed, imbricated and offset former palaeogeographic elements, hindering

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 1 -68° -67° -66° -65° -64° -63° -62° -61° -60° -59° -58° -57° -56° -55° -54° 16° 16° Lesser Edge of Carib. crust Caribbean Antilles Tiburón Rise Arc 15° Plate 15° Proto-Caribbean Inversion Zone 14° Grenada 14° Venezuela S. Carib. Foldbelt intra-arc Post-Aptian Basin Aves Basin BARBADOS oceanic crust 13° Ridge 13°

“Atlantic” fracture 12° Bohordal zones (post-95 Ma) 12° Fault Greater Tobago Margarita Demerara Fracture Zone Cariaco Paria 11° Basin pull-apart Caribbean Prism 11° TRINIDAD

Columbus 10° V. de Cura Serrania 10° Urica F. Oriental Basin Pre-Aptian Guarico Belt Guyana Transformoceanic crust Cret. shelf edge blind Carib Maturín Basin Los Bajos F. 9° underthrust Orinoco Delta 9° Guarico Basin VENEZUELA Guyana Shield 8° 8° -68° -67° -66° -65° -64° -63° -62° -61° -60° -59° -58° -57° -56° -55° -54° Fig. 1. Key tectonic features of the Eastern Caribbean region, including Central and Eastern Venezuela and Trinidad, and key features referred to in the text. The background for the map is the satellite gravity of Smith and Sandwell (1997). Today, Caribbean crust lies east of Tobago and under Barbados. This crust and the Caribbean accretionary prism have over-ridden much of the proposed Proto-Caribbean Inversion Zone, which is only exposed today in the area south of Tiburón Rise. The positions shown on this map are based on three-dimensional restoration of features mapped on seismic tomography (after Pindell et al. 2007a).

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 2 reconstruction of former depositional systems. Correct reconstruction of these is important because certain palaeogeographic aspects pertain directly to petroleum exploration. For example, fluvial- depositional systems and turbidite fairways may contain good quality reservoir sandstone packages that can drive hydrocarbon exploration efforts. Previous work has demonstrated three general principles regarding the provenance of sandstones in northeastern South America (see Figs 2–4 for stratigraphic summaries): (1) South America was the predominant source for most clastic sediments, which are typically quartz prone and mineralogically mature; (2) a less mature Caribbean mineral association began in the Caribbean foredeep basins at the following times: Palaeocene in western Venezuela, Eocene in Central Venezuela, and Oligocene in Eastern Venezuela-Trinidad, indicating the diachronous advance of the Caribbean Plate from the west; and (3) the Palaeogene clastic units of Barbados were derived from relatively high-grade metamorphic (e.g. sillimanite-bearing) rocks, probably from the South American craton or Andes, but also include minerals of probable Caribbean origin such as glaucophane (Senn 1940; Gonzales de Juana et al. 1980; Kasper & Larue 1986; Socas 1991; Kugler 2001). These observations allow the regional stratigraphic units to be understood in the context of the basic Caribbean-South America collisional model (e.g. Dewey & Pindell 1986; Kasper & Larue 1986; Pindell et al. 1988, 1998). However, there has been no systematic attempt to use heavy minerals to test such concepts as the diachroneity of collision, or the possible existence of a Proto-Caribbean prism/thrustbelt (proposed by Pindell et al. 1991, 2006). This study attempts to fill this void, and to define the cause-and-effect relationship between the regional stratigraphic units and tectonic evolution. We have collected and analysed the heavy mineral content of 118 sandstone and siltstone samples from the Cretaceous and Cenozoic of Eastern Venezuela, Trinidad and Barbados, supported by petrographic examinations of sand grain composition of these and many other samples, to determine sandstone composition variations through the stratigraphic column. In addition, we have studied thin sections and obtained X-ray diffraction data (T. Rieneck & W. Maresch, Universität Bochum, Germany pers. comm. 2007) on field and core samples from the three countries on distinctive red, rounded pebbles and less rounded rip up clasts, informally referred to as “cherries”. In Trinidad, these are particularly characteristic of the Cretaceous (Barremian–Albian) Cuche Formation in the Central Range and probably the similarly aged Toco Formation on the north coast, and are also found in (albeit less-oxidized) coarse intervals of the younger Gautier Formation. They are also found in the (possibly) Late Eocene to earliest Oligocene Plaisance Conglomerate and the basal part of the Mount Harris section of the eastern Central Range in Trinidad, and we have found them in one outcrop of the Oligocene–Early Miocene Nariva Formation sandstone. In Venezuela, they occur in the (possibly) Early Oligocene Lechería beds north of Barcelona. Additionally, we have examined (XRD, thin sections) the dark “clasts” of similar size and shape in the Galera Formation shales that have been previously considered possible precursors for the “cherries” of younger formations (Higgs 2006, 2009). The new heavy mineral and petrographic analyses augment those previously published, and along with the XRD results and previously published modal point count data are integrated with plate kinematic data and regional structural relationships to constrain Mesozoic–Tertiary clastic distribution patterns and, in turn, palaeogeographic evolutionary models for northern South America.

HEAVY MINERAL STUDY RATIONALE AND LABORATORY METHODS

Background to Heavy Mineral Studies

Heavy mineral analysis allows the efficient reconstruction of source area lithology and provides information on sand provenance and the direction of sand supply, all key to palaeogeographic reconstruction. Diagnostic minerals provide clues to the correlation of sequences linked by a common provenance, and the differentiation of those that were derived from different source lithologies. During the sedimentary cycle, original heavy mineral assemblages may undergo changes controlled by various modifying factors in the sedimentary environment, such as: (1) hydraulic processes during transport, producing preferential sorting according to size, shape and density related to the differing

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 3 TIME Traditional Southern Shelf Platform Trinidad (this paper) Barcelona Trough (deep)

Quaternary S Cedros N NW Erin SE W E Erin Talparo Talparo M. L’Enfer Morne L’enfer Pliocene Springvale Springvale G. Morne Forest G. Morne For. Manz. Guiaco Manzanilla L Miocene Cruse Cruse Karamat/Lengua Karamat/Lengua Trough imbricated and filled to M Miocene Herrera Tamana Tamana Herrera Cunapo become integral with the rest of Brasso Cunapo Brasso E Miocene Upper Cipero Nariva Trinidad at this time Nariva Change in orientation SW NE L Oligocene Nariva Nariva, shallows up into Cunapo/Brasso Silty Cip.Cipero Lower Cipero Nariva E Oligocene S Fernando Cipero Olig. P-a-P Mt. Harris Angostura absent Plaisance L Eocene San Fernando ?? Cret. olistoliths? Hospital Hill marl M Eocene Navet ?? Navet Charuma Navet E Eocene Navet ?? Pt-a-Pierre Pt-a-Pierre channels L Palaeoc. ?? Lizard Springs Chaudière Tarouba Chaudière E Palaeoc. Lizard Springs Soldado S Joseph’s Lizard Springs possible Creta- Maastricht. Guayaguayare Guayaguayare ceous olistoliths Upr Nap Hill U Nap Hill material of this age is Campanian ?? highly condensed or L Nap Hill Lr Nap Hill bypassed on a slope Ceno.–Sant. Gautier ss Gautier shale ?? Gautier Cuche Albian Cuche (shelf) Maridale Maridale Maridale rubble? (Cuche River) Cuche Cuche (shelf) Aptian Cuche sands (Mt HarrisCuche well) ?? Barranquín ? u/c ? Neocomian Couva evaporite Couva evaporite ? Laventille reef talus? (from Laventille LS; not seen) marine ? M-L Juras. ???? red beds ? oceanic crust? or very thinned cont crust

Fig. 2. Stratigraphic chart for Trinidad showing relationships between key formations both across and along strike, and contrasting our view with more traditional stratigraphic schemes (e.g. Carr-Brown & Frampton, 1979; Saunders et al. 1998). Ages for some of the formations are shown as different to the published literature. The revisions are based on our own unpublished faunal ages, our observations of field relationships and our interpretations of unpublished, proprietary, seismic lines. Ages assigned to most of our samples are based on this framework. Some samples were assigned to units different to those in published maps based on distinctive heavy mineral or petrographic characteristics. The chart shows the contrasts between the southern Trinidad Platform and the eastern end of the Barce- lona Trough (see Figs 11, 12 for location) to the northwest, which formed during the Palaeogene above the former passive margin. The platform overlies less stretched continental basement and lies to the north of the early Creta- ceous shelf edge, probably marked by a reef trend at the palinspastically restored position of the footwall of the Central Range. The trough lies to the north of the platform, over highly stretched continent, transitional crust, or possibly oceanic crust of Late Jurassic age and from its inception was significantly deeper. The stratigraphy of the trough is now found in the highest thrust sheets of the Central Range. Locations for key stratigraphic sections, sample sites and major morphotectonic divisions in Trinidad are shown on Supplementary Figure 1.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 4 TIME (along Urica Fault) Serrania del Interior Oriental, Venezuela (Urica to El Pilar) NW SE SW NE Quaternary La Mesa Note: Early workers (e.g. Hedberg 1950) Pliocene subaerial since end of Las Piedras suspected Eocene uplift. Most work in Middle Miocene, due to 80s–90s presumed all erosion was Mid- L Miocene Serranían orogeny La Pica dle Miocene and younger. However, Late Eocene-earliest Oligocene erosion may M Miocene Carapita have reached the Albian toward the northeast, possibly to Barranquín, E Miocene Carapita Areo Upr Naricual Merecure Naricual such that Miocene erosion is Areo superposed upon the Paleo- L Oligocene Lr Naricual Areo basal Caribbean gene unconformity. E Oligocene Los Jabillos foredceep onlap Los Jabillos Lecheria? L Eocene Paleogene unconformity, Pre-Late Eocene Tinajitas probably present M Eocene due to long-term sea level fall and Proto- until Late E Eocene Caratas Caribbean hanging wall Eocene-earliest uplift. Oligocene L Palaeoc. Vidoño Vidoño erosion Vidoño lst Vidoño lst E Palaeoc. San Juan San Juan San Juan Maastricht. Campanian San Antonio sandy San Antonio Guayuta

Ceno.–Sant. Querecual Querecual Chimana Albian El Cantil Garcia Aptian marine Barranquín fluvial Barranquín Barranquín Neocomian evaporites? evaporites? marine M-L Juras. red beds red beds

Fig. 3. Stratigraphic chart for Venezuela showing relationships between key formations both across and along strike for a NW-SE profile from the Barcelona area to the foreland near Urica, and for a SW-NE from Urica towards the northeastern Serranía. The approximate location of the profile is shown on Supplementary Figure 2.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 5 NNW SSE

? Basal Complex south Basal Complex north

Inter-prism basin floor intermediates? and melange? “Intrusive” contact Mud diapir (Joe’s River Fm) Unconformity Major fault Quaternary limestones/alluvium Conset Marl (Middle Miocene overlap assemblage) Bathsheba Intermediates; other intermediates include Prism Cover, Bissex Hill, Cambridge beds, Kingsley beds, Woodbourne Fm Oceanic Nappes (note: structurally emplaced in Miocene Basal Complex (Scotland Fm)

Fig. 4. Schematic structural cross-section of Barbados, summarizing possible stratigraphic relationships. Few reliable ages are available for pre-Miocene strata. According to Senn (1940), the Scotland Group section from top to bottom included the members: Belle Hill/Mount All; Windy Hill/Chalky Mount; Walkers; Murphy‘s and Morgan Lewis. Speed (2002) aggregated all these into an undifferentiated “Basal Complex”. Further, we suggest that there may be a juxtaposition of the once-separate Caribbean and Proto-Caribbean accretionary prisms (Basal Complex north and south, verging south and north respectively). Locations for key stratigraphic sections and sample sites are shown on Supplementary Figure 3.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 6 densities of the individual species; (2) post-depositional dissolution, due to the low resistance of the majority of heavy minerals to either prolonged acidic or alkaline geochemical conditions either preceding, or following burial. Mineral persistence during diagenesis is directly related to their chemical stability and there is typically a progressive decline in the abundance and diversity of heavy mineral species with increasing depth of burial and increasing age. Ultimately, sediments reach a stage of high mineralogical maturity, where the heavy mineral fraction comprises only ultrastable detrital minerals such as zircon, tourmaline, rutile and apatite. These modifying factors need to be evaluated in any heavy mineral study.

Sample preparation

Analytical work was performed in the geochemical laboratory of the Department of Geology in Leicester University, using the methods described in Mange & Maurer (1992). Sample preparation involved: (1) disaggregation of the consolidated sandstones by rock crushing, using a mortar and pestle; (2) removing drilling mud from the cuttings by wet sieving, using detergent, followed by drying; (3) acid digestion to dissolve carbonates both in cores and cuttings by means of 10% acetic acid which leaves acid-sensitive apatite intact; (4) wet sieving using a sieve of 0.063 mm mesh to remove remaining clay and silt particles; (5) drying, followed by standard sieving, retaining the 0.063– 0.210 mm size fractions; (6) oil stained samples and cuttings rich in organic particles were cleaned using chloroform; and (7) heavy mineral separation was performed in bromoform (specific gravity of 2.89) using the centrifuge and partial freezing method.

Microscopy

A split of each heavy mineral sample was mounted in liquid Canada balsam on a glass slide for the microscopic investigation. The viscous consistency of liquid Canada balsam facilitates rolling the grains to obtain the required orientation that helps identification and accurate observation of grain morphology. Grain counting was made along parallel bands on the slide, described as the “ribbon counting” method by Galehouse (1971). Components of the non-opaque heavy mineral suite were counted, excluding micas. However, the presence and abundance of various micas, opaque grains and authigenic phases, together with associated lithic fragments, organic particles etc. was recorded. The proportion of garnet largely depends on the grain size of the particular deposit, the effects of dissolution processes during diagenesis, mechanical fracture along cleavage planes producing several ‘grains’ from one original, and, to a certain extent, accidental fracturing of the often large grains during rock crushing. It is therefore advantageous to count garnet separately, thereby avoiding the masking effect of varying garnet quantities on the associated minerals. Colourless, and orange to pink varieties were distinguished and their frequencies recorded as a percentage of the total number of grains counted per sample. Anatase, a predominantly authigenic phase, was treated similarly. During the first stage of grain point-counting the number of individual heavy mineral species was recorded (conventional, species-level analysis) in parallel with the point counting of selected varietal types of zircon, tourmaline and apatite, allocated to relevant categories (high-resolution heavy mineral analysis). When the total of the individual species, excluding garnet, reached 200, counting of the varieties continued until 75–100 grains each of zircon, tourmaline and apatite varietal types were recorded respectively. This permits a reliable estimate of heavy mineral abundance. For recording the grain counting, a specially designed HYPERCARD program was used. Two datasheets were completed for each sample, one for the species-level (overall) mineralogy and one for the varietal study. The raw data were transferred to spreadsheets and recalculated to number percentages.

HEAVY MINERAL DATA

Trinidadian heavy mineral assemblages are summarized in Tables 1 and 2, Venezuelan heavy mineral assemblages are summarized in Tables 3 and 4 and Barbadian heavy mineral assemblages in Table 5. Detailed heavy mineral data and location maps are included in the online supplementary data. We have augmented our data with published sources and the limited older industry data that have been

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 7 Table 1. Summary of heavy minerals from Trinidadian formations – this study. Formation Age Heavy minerals

Cuche sandstone Cretaceous ZTR only

Gautier sandstone Cretaceous ZTR, garnet, apatite, possible spinel Toco Cretaceous ZTR, garnet, apatite

Naparima Hill sandstone Cretaceous ZTR only Chaudière Palaeocene–Eocene ZTR, possible rare apatite

Pointe-a-Pierre Palaeocene–Eocene ZTR only

Charuma phacoids Palaeocene–Eocene (?) ZTR, apatite, garnet, chloritoid Plaisance Conglomerate Early Oligocene ZTR, rare epidote

Mt. Harris sandstone Early Oligocene ZTR, rare epidote

Nariva Late Oligocene ZTR, garnet, chloritoid, staurolite, kyanite, glaucophane, apatite, epidote, corundum, monazite Cunapo Miocene ZT only (extremely low recovery of heavy minerals)

Brasso Miocene ZTR and chloritoid. Characteristic blue tourmaline

Herrera Miocene ZTR, staurolite, epidote, garnet, apatite, sphene, monazite, anatase. Rare hornblende, kyanite Basal Manzanilla* Miocene ZTR only

Lower Manzanilla Miocene ZTR only

Upper Manzanilla Miocene ZTR, apatite, epidote, clinozoisite, kyanite, chloritoid, chlorite

Cruse Miocene–Pliocene ZTR, apatite, staurolite, garnet, kyanite, chloritoid

Talparo Pleistocene ZTR, staurolite, sphene, kyanite, glaucophane, epidote, sillimanite, xenotime

* Appears to have been mismapped by Kugler (1996) as Cunapo Conglomerate.

Table 2. Summary of heavy minerals from Trinidadian formations – other sources. Formation* Age Heavy minerals Reference

Naparima Hill Argilite Cretaceous ZTR, trace garnet, staurolite, kyanite, epidote Edelman & Doeglas 1934

Lr. Lizards Springs Palaeocene ZTR, trace garnet, staurolite, kyanite, epidote Edelman & Doeglas, 1934

San Fernando Early Oligocene ZTR, garnet in some samples, trace staurolite, kyanite, epidote Edelman & Doeglas 1934

Bamboo or Flat Rock Silt Early Oligocene ZTR, trace garnet, staurolite, epidote, glaucophane in one sample Edelman & Doeglas 1934

Moruga Miocene ZTR, apatite, staurolite, garnet, blue topaz Kugler 2000

Forest Pliocene ZTR, apatite, garnet, chloritoid, epidote, staurolite, kyanite, Kugler 2000 andalusite, glaucophane, spinel.

Morne L’Enfer Pliocene ZTR, epidote, garnet, chloritoid, staurolite, kyanite, andalusite, topaz, Kugler 2000 anatase, glaucophane

Erin Pliestocene ZTR, epidote, staurolite, kyanite, andalusite, amphibole, topaz, Kugler 2000 anatase. Rare glaucophane

* There appears to be no public-domain heavy mineral data from the Mayaro or Springvale Formations or from the Angostura sandstones.

Table 3. Summary of heavy minerals from Venezuelan formations – this study. Formation Age Heavy minerals

Barranquín Cretaceous Zircon, tourmaline, rutile (hereafter ZTR). Minor staurolite, epidote

San Juan Cretaceous ZT only Caratas Eocene ZT only

Lechería Early Oligocene ZT only

Los Jabillos Early Oligocene ZT only Lr. Naricual Early Oligocene ZT, rare kyanite

Areo Late Oligocene ZT, minor garnet

Quebradón Miocene ZTR, apatite, staurolite, garnet

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 8

Table 4. Summary of heavy minerals from Venezuelan formations – other sources. Formation* Age Heavy minerals Reference

Pre-Cretaceous Palaeozoic, ZTR, hornblende, tremolite, glaucophane (or non-HP/LT Escalona 1985 Jurassic blue-grn amph?). One report of chloritoid

Canoa Cretaceous ZT, epidote, zoisite, magnetite, ilmenite. Minor kyanite, Escalona 1985 staurolite, other amphiboles

Tigre Cretaceous ZT, epidote, kyanite, staurolite, glaucophane (or non- Escalona 1985 HP/LT blue-grn amph?). Minor magnetite, ilmenite, other amphiboles Mito Juan Cretaceous ZTR, garnet, chloritoid, ilmenite, leucoxene PDVSA 2005

Garrapata Eocene (?) Amphiboles, pyroxenes in association with volcanic rock PDVSA 2005 fragments

Los Cajones Palaeocene–Eocene (?) ZTR, epidote, apatite, magnetite, leucoxene, in PDVSA 2005 association with volcanic and schist fragments Guárico Palaeocene–Eocene ZTR, garnet, trace chloritoid, anatase. Kamen-Kaye 1942 (Guarumen-Ortiz sandstone)

Misoa (El Mene) Eocene ZTR, staurolite (upper member only), rare garnet Feo-Codecido 1955

Cobre Eocene ZTR, garnet, staurolite PDVSA 2005

La Pascua Early Oligocene ZTR. Minor or trace staurolite, kyanite, hornblende, other Escalona 1985 amphiboles

Lr. Roblecito Late Oligocene ZTR. Minor staurolite, kyanite Escalona 1985

Upr. Roblecito Late Oligocene ZTR, staurolite, kyanite, glaucophane Escalona 1985

Merecure Late Oligocene–Miocene ZTR, minor garnet, minor chloritoid PDVSA 2005; Feo-Codecido 1955

Carapita Late Oligocene–Miocene ZTR, epidote, glaucophane Feo-Codecido 1955

Capaya Miocene ZTR, staurolite, glaucophane PDVSA 2005

Chaguaramas Miocene ZTR, staurolite, kyanite, andalusite, sillimanite, PDVSA 2005 glaucophane, chloritoid

Oficina Miocene ZTR, staurolite, kyanite, garnet, chloritoid PDVSA 2005 Hedberg et al. 1947

Freites Miocene ZTR, some garnet, chloritoid, staurolite, kyanite, Feo-Codecido 1955 glaucophane

La Pica Miocene ZTR, epidote, some garnet, chloritoid, staurolite, kyanite, Feo-Codecido 1955 glaucophane

Las Piedras Pliocene ZTR, kyanite, chloritoid, corundum, hornblende. Minor PDVSA 2005 sillimanite, staurolite, epidote, garnet, kyanite, Hedberg et al. 1947 glaucophane La Mesa Pleistocene ZTR, staurolite, kyanite, andalusite, sillimanite, magnetite Hedberg et al. 1947

Rio Caroni Holocene ZTR, ilmenite, staurolite Wynn 1993

* We recovered no heavy minerals, nor are there reports in the literature of heavy minerals from the Palaeocene-Eocene Vidoño Formation.

Table 5. Summary of heavy minerals from Scotland Group members – this study. Member Heavy minerals

Morgan Lewis ZTR, epidote, clinozoisite, apatite, garnet (abundant), chloritoid, kyanite, staurolite

Walkers ZTR, clinozoisite (very abundant), epidote, apatite, garnet, chloritoid, kyanite (abundant), staurolite, sphene; rare tremolite, chrome spinel, lawsonite Chalky Mount ZTR, epidote (minor), clinozoisite, apatite, garnet (abundant), chloritoid (abundant), kyanite (abundant), staurolite Windy Hill ZTR, epidote, clinozosite (abundant), garnet, chloritoid, staurolite

Mount All ZTR, epidote (rare), clinozosite (abundant), garnet, chloritoid, kyanite, staurolite Belle Hill ZTR, epidote (rare), clinozosite, garnet, kyanite, staurolite

Bathsheba ZTR, epidote, clinozoisite

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 9 released. Both sources are generally non-quantitative. A limited number of modern quantitative studies remain proprietary. Where sediment ages are well-constrained, the results are discussed by age, from oldest to youngest. Figures 2, 3 and 4 summarize the stratigraphy of Trinidad, Venezuela and Barbados, respectively.

Trinidad

Trinidadian samples were collected from numerous outcrops and from several well cores. Although exposure is generally poor, there is good palaeontological age control at many outcrops, and field data from past industry mapping and augering (down to 20 m) was made available to us. Where possible (usually), samples were collected from active stream cuts, or from recent quarries, in order to get below the worst of the tropical weathering. The Trinidad data (Tables 1 and 2) appear to be immediately separable into two age dependent assemblages, with only three possible exceptions as noted below. The large number of samples and the availability of absolute and relative age control allow the data to be displayed as a synthetic stratigraphic column (Fig. 5) that accentuates the contrasts between the two assemblages. Early Rupelian (Early Oligocene) and older rocks are characterized by mature zircon, tourmaline and rutile (ZTR) assemblages. Some garnet and apatite was also found in core samples from the Late Albian–Cenomanian Gautier Formation and in one sample of the similar-aged Toco Formation on the north coast. In contrast, probably Late Rupelian and definitely Late Oligocene, and younger rocks are consistently dominated by an immature assemblage of labile minerals, including staurolite, aluminium silicates and glaucophane, in addition to some apatite. Abundant garnet and chloritoid are particularly characteristic. Tropical weathering of outcrops cannot explain this apparent abrupt maturity contrast. Within a given formation, similarities between borehole and field samples, notwithstanding weathering and/or diagenesis, indicate there is a primary compositional difference, and therefore a difference in provenance, between Early and Late Oligocene sediment source rocks. In addition, we sampled both younger and older formations with identical weathering effects, some from immediately adjacent outcrops, and always found dramatic contrasts in heavy mineral assemblages. In some formations for which we have many samples from both outcrop and wells (e.g. Late Oligocene–Early Miocene Nariva Formation), it is clear that there is a geographical grouping of samples in which there seems to be a correlation between increasing tropical weathering and decreased content of certain unstable heavy minerals (Fig. 6). There may also be an element of subtle variation in original heavy mineral content within the Nariva. However, the primary contrast between mature and immature sands is still clear. Higgs (2006, 2009), reported that previous authors had identified staurolite and other non- ZTR minerals throughout the Palaeogene strata in Trinidad (relying heavily on the synthesis of Suter 1960) and proposed a northern, rather than Guayana Shield, source for both Palaeogene and Neogene strata. However, our examination of unpublished industry heavy mineral studies (Petrotrin archive files mostly from the 1930s–1950s) show the same mature-immature contrast that we have found, as also reported by Kugler (1996, but based on a 1950s synthesis) and Illing (1928). Trace to low abundance staurolite, kyanite, chloritoid and very rare glaucophane have been found in a few samples (and many were exotic blocks associated with mud volcanoes; their origins and age were thus poorly known), and there is little justification for considering these minerals as characteristic of the Palaeogene as a whole. There are three potential exceptions in our data to the apparent Oligocene character change for Trinidad. First, the samples from the Middle Eocene “Charuma Silt Member” of the Pointe-a-Pierre Formation at its type locality yield an immature heavy mineral signature (Table 1). Although this is the type locality of this stratigraphic unit, exceptional recently cleared exposure on the day of our collection showed that the Charuma section comprised sheared, sandstone phacoids within a scaly clay gouge zone, with sections and rafts of silty clay. There is no undisrupted bedding in the outcrop and the silty clay carries a Middle Eocene “Gaudryina” fauna. There are no fresh outcrops of this formation and attempts to separate heavy minerals from other sites where stratigraphy and field relationships are clear (e.g. from Piparo Gorge, where adjacent beds are mapped as typical Pointe-a- Pierre sandstone and show zircon and tourmaline only) failed to yield any, much less immature, heavy minerals. Although this unit has been drilled in several wells, there are no longer any cuttings available and the drilled section rarely if ever included sandstones (John Keens-Dumas pers. comm.).

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 10 PERCENTAGE 0 20 40 60 80 100 CAROLINA LK06-07 MAMORAL LK5-09 SVT LK5-41 2x GALFA PT. LK5-70 CR MAMORAL LK06-08 CITRUS GROVE LK5-16

LK5-13 MANZ 2x CUNAPO S. ROAD LK Sun 2C MD34 LK5-53 BP 347/3 3X PENAL BARRACKPORE BP 347/2 BP 347/1 BALATA CENTRAL BC 1/4 BP 362/3 3x PENAL BARRACKPORE BP 362/2 BP 344/2 BALATA CENTRAL BC 1/3

PENAL BARRACKPORE BP 362/1 HERRERA BALATA CENTRAL BC 1/2 PENAL BARRACKPORE BP 344/1 CA 40/1 2x CATSHILL CA 40/2 BALATA CENTRAL BC 1/1 NESTOR MAMON ROAD BR1 CUNAPO S. ROAD AUG03/01 PIPARO MUD VOLCANO LK5-18 LK5-07 2x SANDSTONE TRACE LK5-06 COLDAN QUARRY LK5-11c WILLIAMSVILLE LK5-12 CORBEAU HILL LK5-01 ABM54 WELL LK5-55 LK06-02 3x CHAUD. R. “NARIVA” LK06-03 LK06-04 PLUM MITAN AUG03/06 LK7-02 3x CHARUMA LK Sun2 LK ANT 1 PROVENANCE BREAK LK5-44 2x PLAISANCE QUARRY LK5-43 LK5-37 2X WEST MT. HARRIS LK5-32 MT. HARRIS PICNIC SITE LK5-14 3x CHAUDIÈRE RIVER LK06-06 LK06-16 MT. HARRIS LK06-19 NORTH MT. HARRIS LK7-20 KEY: WEST MT. HARRIS LK5-33 EAST MT. HARRIS LK5-20 Garnet PIPARO GORGE T45-1 NAVET DAM TTM2 Epidote Gp BRAKE FACTORY LK5-45 SAN FABIEN RD LK5-29 TOP CRETACEOUS Staurolite GALERA LK5-24C

MD34 LR. NAP. HILL LK5-52 UK Others ME15 GAUTIER LK5-54 SNAKE R LK5-68 Rutile LK5-65 TOCO ANDRE PT LK5-63 Apatite LK06-26 3x CUCHE RIVER LK06-24 Tourmaline LK06-22 LK5-51 2x MT. HARRIS 1 CUCHE TOCO/G.Zircon PAP MT. H. CHAUD/PAP PL CHAR NARIVA BR LK5-50 Fig. 5. Heavy mineral varieties from Trinidad samples, sorted by relative age. Note the very mature assemblages in Cretaceous through Early Oligocene rocks; trace staurolite and kyanite has also previously been reported from a few samples of this age. Regional facies patterns and provenance data both indicate a broadly “southern provenance”. There is a very pronounced provenance break in late Early Oligocene and younger samples, marked by the appearance of immature mineral assemblages derived from in part from the Caribbean Orogen. The orogen was to the west at the time but sediment of this age was transported down the axis of an orogen-parallel drainage system, much like the present-day Orinoco. In both wells and outcrops it is possible to find adjacent examples of these two assemblages in rocks which appear to have similar weathering or diagenetic characteristics, indicating that they are primary differences. Late Middle Miocene and younger strata show a more mixed provenance, with Caribbean detritus diluted either by new south-derived input into the basin or recycled Palaeogene rock which had been incorporated into the edge of the Caribbean orogen. Sediment recycling may also contribute to the cleaner signature. Our few Late Plio- cene and Pleistocene samples (Mamoral sample from Springvale Formation and Carolina sample from Talparo Formation), are strikingly more immature than slightly older Herrera and Cruse samples. Abbreviations: UK Upper Cretaceous, PAP type Pointe-a-Pierre, MT. H. Mount Harris sandstone, CHAUD/PAP mapped as Chaudière/Pt.-a- Pierre but probably equivalent to or younger than MT. H., PL Plaisance Conglomerate (possibly equivalent to basal MT. H.),Pindell CHAR et al., Charuma in press 2009, “Silt”, PREPRINT BR Brasso, MANZ Manzanilla Formation in Caroni Basin, CR Cruse Formation in SouthernTrinidad, Basin, Venezuela. SVT undifferentiated Barbados heavy minerals Springvale-Talparo. and palaeogeography Page 11 Nariva Formation “Striped” sands

Outside in: LK06/4 LK06/3 LK06/2 All Chaud. River

Nariva Formation “Sealed” sands

Outside in: LK05/55, ABM54 LK05/12, Williamsville LK05/18, Piparo MV

Nariva Formation Sandstone Trace

Outside in: LK05/1, Corbeau Hill LK05/11, Guaracara KEY: Garnet Nariva Formation Epidote Gp Other outcrops Staurolite Others Rutile Outside in: Apatite LK05/6 LK05/7 Tourmaline Zircon

Fig. 6. Variations in heavy mineral abundance in the Nariva Formation. Significant changes in relative abundance of key mineral in Nariva samples probably reflect post-depositional leaching by acidic fluids (basinal or tropical soil origin) superimposed on some original, unquantifiable, but probably limited compositional variation. Fairly fresh “striped” sandstones from the Chaudière River show intermediate amounts of garnet and some chloritoid and kyanite, but “sealed” samples from outcrops, core or mud volcanoes where intense surface weathering has not hap- pened show very high levels of garnet, with some apatite and staurolite also preserved. The Williamsville Quarry sample is the only one with preserved glaucophane (very unstable). Weathered outcrop samples in the Piparo area show moderate preservation of chloritoid but no garnet and only trace apatite. The Sandstone Trace samples show intense feldspar breakdown and pervasive clay cement but otherwise do not show deep weathering. Chloritoid is particularly abundant but garnet is rare and apatite and staurolite are absent. The relative ages of these samples are not clear. All are associated with earliest Miocene shales in adjacent outcrops.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 12 Faults juxtaposed the sampled outcrops with the Navet marls and Nariva sandstones, the latter with the same complex mineral signature and similar textural characteristics in thin-section. We suspect contamination is possible in our Charuma sample due to the shearing; the sample could be Oligocene Nariva contaminated by Eocene Charuma or Navet fauna, or it could be true Charuma (Eocene) contaminated by Nariva minerals. If this were not the type section of the Charuma beds, we would consider the outcrop as structurally disrupted enough to pay it little attention. Alternatively, the chloritoid, kyanite and garnet present in the samples could represent a rare flush of first cycle labile minerals from the shield into the Trinidad area. Trace kyanite and garnet are both reported from Palaeogene sections close to top Cretaceous and Late Eocene unconformities in Trinidad (Edelman & Doeglas 1934) and all these minerals are present on the shield to the south (see below). The second exception is a lignite-bearing sandstone collected from an isolated fault-bounded outcrop in the Chaudière River (on the north flank of Mount Harris), which also has a characteristic Nariva heavy mineral signature. Although mapped as Palaeocene Chaudière Formation by Kugler (1996), bedding in the outcrop dips in the opposite direction to proven Palaeocene beds nearby, and is closely associated with poorly exposed breccias, suggesting it may be fault-bounded. Abundant lignite is not found in formations older than Nariva. “Chaff-like plant remains and wisps of carbonaceous matter” (Kugler 1996) are described from the well-cemented sandstones at the Pointe-a-Pierre type locality, but heavy minerals from this site were extremely mature (other than one unpublished report of a single kyanite grain in one of several samples). In hand specimen, these samples are friable compared to adjacent Chaudière sandstones, and in thin section, these samples are texturally identical to Nariva sandstones collected elsewhere, being fine-grained, well-sorted, with poorly rounded grains. Nariva is mapped along strike several hundred metres west of this outcrop and we suspect it continues unmapped along the north flank of Mount Harris. The third exception is a sample from some 200 m south of the Mount Harris picnic site of the eastern Central Range, also mapped as Chaudière or Pointe-a-Pierre Formation (Algar 1993; Kugler 1996). We have identified chloritoid and blue tourmaline here, both characteristic of the younger Brasso Formation. Thin sections are texturally indistinguishable from Nariva or Brasso Formation sandstones collected elsewhere, and quite unlike nearby sandstones mapped as Pointe-a-Pierre Formation. Furthermore, a sample of interbedded claystone yielded a single foraminifer no older than earliest Miocene (J. Frampton pers. comm. 2006). We provisionally interpret this sample as Brasso sandstone in the footwall of a thrust carrying Early Oligocene Mount Harris sandstone in its hanging wall. There are mapped Brasso outcrops within a few hundred metres of this site. From the above, the three apparent exceptions to the Oligocene heavy mineral character change appear to be explicable by faulting and mis-mapping of fault-bounded Neogene rocks as Palaeogene. This is not surprising, as up to 150 km of dextral shear must have passed through Trinidad since 10 Ma (Pindell & Kennan 2007a).

Venezuela

Signature minerals such as kyanite, staurolite and glaucophane appear, at first glance, more common in Venezuelan samples (Tables 3 and 4). However, our samples show a fundamental difference between Eastern Venezuela versus Central and Western Venezuela: in the east, as in Trinidad, Late Maastrichtian through Early Oligocene sandstones (i.e. San Juan, Caratas, and Los Jabillos Formations) are highly mature (ZTR-dominated, Fig. 7). There is an apparent break in sandstone provenance along the Palaeogene South American margin at the Gulf of Barcelona/Urica Fault. East of this break, the South American shelf and slope section did not know about the impending arrival of the Caribbean Plate until at least the latest Early Oligocene. This accords with the concept of eastward younging oblique collision between the two plates that was first identified by the eastward migration of Caribbean foredeep subsidence along the margin (Pindell 1985; Pindell & Barrett 1990). Only trace glaucophane and staurolite have been reported from pre-Cretaceous and Late Cretaceous samples from wells in the foreland south of Caracas. Even in Central Venezuela, where orogenesis has commonly been thought to have begun in the Cretaceous, signature minerals are in much higher abundance in Late Oligocene and younger strata, such as the Upper Roblecito and Quebradón Formations, than they are in earlier autochthonous and para-autochthonous Caribbean

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 13 PERCENTAGE 0% 20% 40% 60% 80% 100% Quebradón V05-19 Domain 5

10/05-10F 2x Naricual Domain 4 10/05-10D PROVENANCE BREAK V05-08

Areo- 10/05-5B Lr. Naricual 10/05-5A Domain 3 Lecherias 10/5-12A

10/05-4A 2x Los Jabillos V05-15

Peñas Blancas V05-03

10/05-2B

10/05-2A Domain 2 4 x Caratas V05-06 KEY:

V05-05 Garnet BASE EOCENE TOP CRETACEOUS Epidote Gp 10/05-11A Staurolite V05-24A Others 4 x San Juan V05-18 Rutile Apatite V05-13 Tourmaline El Cantil V05-32 Zircon

V05-30 Domain 1

V05-28 4 x Barranquin V05-14

V05-04

Fig. 7. Heavy mineral varieties from East Venezuela samples, sorted by apparent relative age. Trace kyanite and staurolite are seen in the Barranquín Formation, but all other Cretaceous to Early Oligocene samples are highly mature. Kyanite reappears in the Areo, and garnet and staurolite in Naricual and younger samples.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 14 foredeep strata such as the Guárico Formation. The Early Oligocene La Pascua and Lower Roblecito Formations, which onlap an earlier Palaeogene hiatus, are characterized by a mature ZTR assemblage with only trace kyanite and staurolite suggesting, as in Trinidad, an important provenance break of intra-Oligocene age. In Central Venezuela, the volcanic, serpentinitic, and metamorphic content of the Garrapata and Los Cajones Formations (Early Eocene; Macsotay et al. 1995) indicate that Caribbean uplift, erosion and redeposition of clastic materials in a trough between the two plates was underway at that time, but these units are entirely allochthonous by an uncertain distance as there is uplift, rather than subsidence, of that age in the foreland. Table 4 appears to suggest that the Guárico Formation was receiving northern/Caribbean-derived detritus as well. However, Peirson’s (1965) original mapping of the Los Cajones and Garrapata as members of the Guárico is invalid, based on more recent field studies. Vivas & Macsotay (1997) propose the Los Cajones and Garrapata units as distinct formations, with no syn-sedimentary intercalation. Further, Perez de Armas (2005) states that the contact between the internal (Los Cajones/Garrapata) and external (Mucaria/Guárico) zones of the Guárico Fold-Thrust Belt is always a fault. Our own field studies support this view entirely, and suggest that the three isolated mapped occurrences of Los Cajones strata within the Guárico Belt south of the Don Alonso and Guárico faults (as shown by Bellizzia & González 1971) are not Los Cajones Formation, but rather sections of Guárico that have been intensely deformed with two intersecting cleavages, giving the false appearance of sedimentary rubble in shaly matrix. Thus, the Guárico and the Los Cajones/Garrapata Formations appear to have been initially deposited in different depocentres that merged over time as the Caribbean thrustbelt advanced upon South America. To our knowledge, it is not permissible with current data to consider the Garrapata and Los Cajones strata as members of the Guárico Formation, nor to claim that the volcanic, serpentinite, and metamorphic clasts and minerals of the Los Cajones/Garrapata units have anything to do with the Guárico Formation south of the Don Alonso-Guárico Fault. Our thin sections from Guárico samples show no such contamination, and we find nothing in the literature to counter this mature characterization of the Guárico. However, with the approach of the Leeward Antilles Arc in the Palaeogene, it would not be surprising if the Guárico were eventually shown to contain crystals of airfall tuff, as is the case with the Oligocene in Trinidad (Algar et al. 1998). We also highlight the occurrence of chloritoid in the Maastrichtian–(possibly) Palaeocene Mito Juan Formation of the southwest Maracaibo area, which we believe is the only known occurrence of this mineral in Cretaceous through Early Oligocene strata in Venezuela (PDVSA 2005). However, it is characteristic of Late Oligocene and younger sandstones in Eastern Venezuela- Trinidad, with rare exceptions.

Barbados

In Barbados, Palaeogene terrigenous turbidites of South American affinity (bearing mainly quartz with high-grade metamorphic clasts and minerals) occur in the Scotland District (stratigraphic and structural relations are summarized in Fig. 4). Senn (1940) and Poole & Barker (1982) subdivided the “Scotland Beds” into Lower and Upper Scotland units. Senn further considered the Lower Scotland to comprise the Morgan Lewis and the Walkers sections, and the Upper Scotland to comprise the Murphy’s, Chalky Mount, and Mount All sections. Speed (2002) considered all these Scotland beds as packets of accretionary prism material in his “Basal Complex”. The age range of the Scotland beds is Late Palaeocene through Middle Eocene, based on foraminifera and radiolaria (Speed 2002) and pollen (Pindell & Frampton 2007, citing D. Shaw 2007). Senn (1940) listed the following heavy minerals (citing an unpublished report by Hollis Hedberg) from 113 samples of the Scotland Formation: black opaque minerals, leucoxene, zircon, tourmaline, garnet, staurolite, sillimanite, kyanite, andalusite, topaz, glaucophane (in 5 of the samples only), epidote, zoisite-clinozoisite, rutile, anatase, brookite, chloritoid, hypersthene, augite, titanite, and corundum. Unfortunately, the locations of the samples were not specified and thus the data, as reported by Senn at least, cannot be used to attribute certain heavy mineral signatures to specific Scotland sections. In our data (Table 5 and Fig. 8), characteristic metamorphic minerals are found in all but the Bathsheba sample, which was considered as the “Intermediate Unit” (possible Early Miocene) by

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 15 PERCENTAGE 0% 20% 40% 60% 80% 100%

JP 12/06-07 BATHSHEBA

JP 12/06-02 BELLEHILLEAST

H1 RAGGEDPOINT

G2 BELLEHILL

JP 12/06-08b

JP 12/06-08a 3x MOUNT ALL

JP 12/06-01

JP 12/06-06 WINDYHILL

C4 2x CHALKY MT. POTTERIES C2

B2 2x BARCLAYS PARK B1

A2 3x CAMBRIDGE KEY: A1 Garnet

D6 Epidote Gp Staurolite D5 4x OIL QUARRY Others D2 Rutile D1 Apatite

JP 12/06-04 WINDMILL Tourmaline Zircon E4 THECHASE

JP 12/06-05 2x MORGAN LEWIS BEDS JP 12/06-03

E2 2x MORGAN LEWIS BEDS E1

Fig. 8. Heavy mineral varieties from Barbados samples, sorted by apparent relative age. All but the youngest sample, from Bathsheba, show mixed South America and Caribbean provenance. The Bathsheba sample is swamped by epidote group minerals but is otherwise mineralogically mature, with only zircon, tourmaline and rutile.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 16 Barker et al. (1986) rather than true Scotland. Apatite is absent in Mount All, Belle Hill and Bathsheba samples, and chloritoid is only absent in the Belle Hill and Bathsheba samples. The various Barbados lithostratigraphic units do not appear to be characterized by very distinct heavy mineral signatures. However, there is tremendous diversity in relative proportions within and between units that suggests the mixing, during transportation, of two end member sediment sources, one mineralogically and texturally mature and one immature. This would be expected if the Scotland Formation (Basal Complex) were fed by a foreland trunk river system flowing between the developing Caribbean Orogen and the Guayana Shield (Kasper & Larue 1986). As will be seen, given the Late Palaeocene– Middle Eocene age of the Scotland beds, the source area was probably situated in Colombia and/or Western and possibly Central Venezuela. The almost ubiquitous chloritoid, common staurolite, kyanite and glaucophane are in striking contrast to strata of the same age in east Venezuela and Trinidad, which show none of the mixed provenance seen in Barbados. If the Basal Complex of Barbados is not very far travelled (i.e. < 300 km), then its depocentre must have been deeper than, if not isolated from, the deep-water deposits of Eastern Venezuela and Trinidad, and also been depositionally downstream of Western or Central Venezuela, where Caribbean allochthons were being unroofed. However, the Basal Complex may be farther travelled than 300 km and the mature south- derived component of the sands may derive from Colombia or Western Venezuela, where, for example, sillimanite-bearing basement was exposed between the times of Guaduas (Early Palaeocene) and La Paz (Late Eocene) Formation deposition (Pindell et al. 1998), coeval with Scotland deposition.

POTENTIAL SOURCE AREAS FOR CHARACTERISTIC HEAVY MINERALS

Our approach to the heavy mineral data has been to identify the signature minerals that characterize particular formations or groups of formations. Our sampling is too sparse to look in detail at the ratios of particular minerals, or to more than qualitatively interpret the relative abundances of particular mineral varieties or grain shapes (typical of detailed studies on a particular well or formation). Instead, we attempt to tie these minerals to well-described possible sediment source regions in Colombia, Venezuela and Trinidad (Fig. 9 and Table 6). Identification of sediment source regions would be greatly aided by modern fission track studies of both potential source areas and of detrital grain populations in foreland sediments (an approach used in, for example, Ecuador by Ruiz 2004, 2007). Modelled unroofing history of detrital grains could then be compared to that of sediment source areas and combined with heavy mineral data could lead to less ambiguous identification of sediment sources, where there is more than one possible source for particular minerals, but where those source regions have different unroofing histories. The heavy minerals present in the Late Cretaceous to Palaeogene Venezuelan, Trinidadian and Barbadian sediments can derive from old metamorphic or igneous source areas, or may be eroded and recycled from Cretaceous or Tertiary strata. Sufficient information is available to constrain potential sediment sources, especially when combined with additional information such as sedimentary, structural or radiometric constraints on age of uplift, cooling and development of associated unconformities. Potential source areas for selected key heavy minerals are readily identified (Table 6). Some minerals, such as sphene, epidote and apatite are so ubiquitous in greenschist or amphibolite facies terranes or widespread plutons and associated aureoles as to be non-diagnostic of sediment source. Others, such as garnet, chloritoid, staurolite and kyanite, are geographically widespread but absent in a few critical localities of significance for palaeogeographic reconstruction. They may be present in small quantities in sediments of Palaeogene and older age, but dramatic changes in their abundance also correlate with the appearance of distinctive larger clasts, such as circum-Caribbean volcanic rocks, which suggest they are palaeogeographically diagnostic. Some minerals, such as glaucophane, appear to be very strongly associated with Caribbean provenance. Other useful indicators include the presence of rare but distinctive species (e.g. chrome spinel and lawsonite) and mineral associations indicative of a particular paragenesis (e.g. clinozoisite with glaucophane, epidote and chrome spinel is indicative of a metabasic blueschist source lithology). The major possible sediment source areas can be classified based on distinctive mineralogies and discrete tectonic origins. Four of these classes are unambiguously associated with terranes at the leading edge of the allochthonous Caribbean Plate and four appear to belong to autochthonous or

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 17 -80° -75° -70° -65° -60° -55° -50° Interior 1,4 Caribbean allochthons 5 Caribbean Plate 1/4 2/3 5/6 1 5 7 10° MAR 7 10° LMV 5 7 Maturín 5/6 5 1/4 Guárico 6 Barinas 5 6 8 Panama and W. MMV 1-4 7 5° Colombia were Llanos 5° not sediment GUAYANA SHIELD sources for E. 6 5 5 8 Venezuela and UMV Trinidad 6 Putumayo 0° 0°

6 South American Amazonas 5 Marañon- para-autochthonous 6 Oriente Solimões 6 thrust sheets -5° 5 -5° 5 BRAZILIAN SHIELD 7

-80° -75° -70° -65° -60° -55° -50°

Fig. 9. Maps showing location of sediment source areas classified in the text. Caribbean allochthons (cross-hatch pattern) comprise: (1) Palaeozoic-Mesozoic igneous and sedimentary protoliths with greenschist to amphibolite facies metamorphism; (2) Precambrian-Palaeozoic protoliths, HP/LT metamorphism; (3) Mesozoic mixed protoliths, HP/LT metamorphism; and (4) Mesozoic mixed protoliths, greenschist-facies or lower grade metamorphism. Ande- an Para-autochthons, shield areas (grey) comprise: (5) Precambrian-Palaeozoic protoliths, greenschist-amphibolite facies, locally granulite; (6) Mesozoic mixed protoliths, greenschist-facies or lower grade metamorphism; (7) Mesozoic sediments, unmetamorphosed, Cenozoic uplift and unroofing; and (8) Precambrian meta-igneous and meta-sedimentary rocks.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 18 Table 6. Possible source areas for selected characteristic heavy minerals. Mineral Sediment source area Reference Glaucophane Yaritagua complex (Lara) PDVSA 2005 Cordillera de la Costa HP/LT belt Sisson et al. 1997 Antimano Formation PDVSA 2005 Villa de Cura Group Smith et al. 1999 El Copey Formation, Araya Peninsula PDVSA 2005 Guayana Shield (very doubtful, location not specified) Kamen-Kaye 1937 Chloritoid Pastora, Botanamo Proterozoic greenschists and metavolcanics, Sidder & Mendoza 1995 Guayana Shield Margarita Island HP/LT belt Stöckhert et al. 1995 Manicuare Formation, Araya Peninsula Schubert 1971 Cordillera de la Costa HP/LT belt Sisson et al. 1997 Santa Marta Metamorphic Belt (Colombia) Maya-Sánchez 2001* Cordillera Central, Colombia (west of Palestina Fault) Maya-Sánchez 2001 Amaime-Cauca blueschist belt (Colombia) Maya-Sánchez 2001 Lawsonite Villa de Cura Group Smith et al. 1999 Bocas Complex (offshore Venezuela) PDVSA 2005 Staurolite Margarita Island HP/LT belt (Juan Griego) Stöckhert et al. 1995 Cordillera de la Costa HP/LT belt Sisson et al. 1997 Los Torres Association, Merida Andes PDVSA 2005 El Aguila Formation, Merida Andes PDVSA 2005 Iglesias Complex, Merida Andes PDVSA 2005 Macuira Formation gneisses, southeast Guajira (Colombia) Maya-Sánchez 2001 Santa Marta Metamorphic Belt (Colombia) Maya-Sánchez 2001 Silgara Formation, Santander Massif (Colombia) Maya-Sánchez 2001 Cordillera Central Colombia (west of Palestina Fault) Maya-Sánchez 2001 Bakhuis granulite belt, Suriname Delor et al. 2003a Greenschist belts, French Guiana Delor et al. 2003b Rio Caroní sediment (Guayana Shield) Wynn 1993 Kyanite Eastern Guayana Shield PDVSA 2005 Manicuare Formation, Araya Peninsula Schubert 1971 Margarita Island Stöckhert et al. 1995 Cordillera de la Costa (Central Venezuela) Sisson et al. 1997 Iglesias Complex, Merida Andes PDVSA 2005 Silgara Formation, Santander Massif (Colombia) Maya-Sánchez 2001 Caldas Gneiss (Central Cordillera, Colombia) Maya-Sánchez 2001 Sillimanite Imataca Archaean granulite gneisses, Pastora, Botanamo Proterozoic Sidder & Mendoza 1995 greenschists, overlying Roraima sediments, Guayana Shield Cabriales Gneiss (Central Venezuela) PDVSA 2005 Santa Marta Metamorphic Belt (Colombia) Maya-Sánchez 2001 Silgara Formation, Santander Massif (Colombia) Maya-Sánchez 2001 Guaicaramo, Garzón Massifs (Eastern Cordillera, Colombia) Maya-Sánchez 2001 Cordillera Central Colombia (west of Palestina Fault) Maya-Sánchez 2001 Guiana Shield (southeast Colombia) Maya-Sánchez 2001

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 19 Andalusite Cerrajón Schist, El Baúl Uplift PDVSA 2005 Cerro Azul, El Aguila Formations, Mérida Andes PDVSA 2005 Iglesias Complex, Mérida Andes PDVSA 2005 Pastora, Botanamo Proterozoic greenschists and metavolcanics, Sidder & Mendoza 1995 Guayana Shield Roraima and Mapare Guayana Shield quartzites PDVSA 2005; Sidder & Mendoza 1995 Northwest Guajira schists (Colombia) Maya-Sánchez 2001 Santa Marta Metamorphic Belt (Colombia) Maya-Sánchez 2001 Cordillera Central, Colombia (west of Palestina Fault) Maya-Sánchez 2001 Guayana (or Guiana) Shield (southeast Colombia) Maya-Sánchez 2001 Garnet Shield amphibolite facies terranes PDVSA 2005 Hato Viejo Formation (Cambrian, Guárico wells) PDVSA 2005 Mérida Andes amphibolite facies terranes PDVSA 2005 Cordillera de la Costa HP/LT belt Sisson et al. 1997 El Tinaco Allochthon Oxburgh 1966† Las Brisas, Formation, Caracas Group PDVSA 2005 Sphene Ubiquitous in metagranites and schists from Araya-Paria and west, not PDVSA 2005; source specific Bellizzia 1985 Apatite Ubiquitous, not source specific PDVSA 2005 Blue tourmaline El Baúl Arch “cornbunianites” (or tourmalinites) PDVSA 2005 Guayana Shield “cornubianites” (or tourmalinites) PDVSA 2005 * Text to accompany Maya-Sánchez and Vásquez-Arroyave, 2001. † This area originally interpreted as Las Mercedes Formation, Caracas Group but now mapped as part of the El Tinaco “basement” complex (e.g. Bellizzia, 1985).

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 20 parautochthonous South America (Fig. 9). There are important mineralogical contrasts (the reader is referred to the references in Table 6 for mineral-specific data sources) between these two groups of potential sediment source areas that support the palaeogeographic models we discuss later. The regional plate tectonic context and detailed local geological evolution of many of the Andean source areas discussed below is discussed in more detail in companion papers (Kennan & Pindell 2009; Pindell et al. 2009).

Class 1: Far-travelled Caribbean greenschist and amphibolite-grade metamorphic terranes with mostly Palaeozoic or older sedimentary (primary), and plutonic or volcanic (secondary), protoliths

These terranes all lie outboard, north or west, of accreted Caribbean oceanic or arc terranes. Where dated, they typically have a Palaeozoic initial age of metamorphism related to the final assembly of Pangaea. Initial cooling from a subduction zone-related orogenic event is Albian to early Late Cretaceous (Stöckhert et al. 1994; Sisson et al. 2005), indicating an origin to the west of Colombia because only passive margin conditions existed at that time along northern South America (Villamil & Pindell 1998). Class 1 rocks were rifted from South America and then incorporated into the leading edge of the Caribbean Plate west of Colombia in the Early Cretaceous and comprise the basement to the “Great Arc of the Caribbean”. Examples include the schists and gneisses of the northwest Guajíra Peninsula, parts of the Santa Marta Massif, and the Arquia Complex of western Colombia. We also interpret the El Tinaco Complex as a far-travelled Palaeozoic basement fragment (an idea proposed by Bellizzia 1985), because its geological history is typical of circum-Caribbean terranes and unlike that of the South American autochthon in Central Venezuela. The associated Tucutunemo Formation includes Permian limestones that are not local to Central Venezuela (Benjamini et al. 1986, 1987). These rocks are the westernmost potential sediment source area for sillimanite and andalusite, prior to eastward migration with other Caribbean terranes.

Class 2: Far-travelled Caribbean high-pressure, low-temperature (HP/LT) metamorphic terranes with mostly Palaeozoic or older major sedimentary (primary), and plutonic or igneous (secondary), protoliths

Class 2 rocks are situated in the same areas as Class 1. Well-known examples include the Juan Griego high-pressure rocks of Margarita, some Cordillera de la Costa metasedimentary protoliths and possibly much of the Manicuare Formation (Schubert 1972) of the Araya Peninsula. The garnet and glaucophane schists of the Yaritagua Complex of Lara, Venezuela (protolith age unknown) may also belong in this grouping (since pre-Caribbean HP/LT rocks are unknown in northern South America). All are intimately associated with HP/LT meta-igneous rocks and are often near ophiolite remnants inferred to be part of the Caribbean suture zone. By 100–120 Ma they were juxtaposed with younger igneous rocks (Class 3, below) and share the same metamorphic, structural and exhumation history (Stöckhert et al. 1995; Pindell et al. 2005). On the north side of the Venezuela and Trinidad foredeep basin, Margarita and the Cordillera de la Costa are the easternmost identified potential sources for staurolite and the Manicuare Formation is the easternmost identified source for garnet and kyanite. The limited drainage from these areas into the Orinoco suggests that they are secondary sources for foredeep staurolite, kyanite and garnet, at least for the last few million years, compared to the Mérida Andes (Class 5 below), which drains directly into the headwaters of the Orinoco.

Class 3: Far-travelled Caribbean HP/LT metamorphic terranes with Cretaceous volcanic and volcano-sedimentary protoliths and initial age of metamorphism

Class 3 rocks include the Rinconada metabasic rocks (Margarita), the Villa de Cura blueschist belt, and the Jambaló, Barragán and Pijao blueschists of Colombia. The Cretaceous protolith age is constrained by the oldest U-Pb zircon ages from Margarita (for a very detailed review of the geological evolution of Margarita, see Maresch et al. 2009). Detailed metamorphic and geochronological studies indicate protolith eruption followed rapidly by juxtaposition with continental-affinity rocks, high-pressure blueschist-eclogite metamorphism at 100–120 Ma, onset of

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 21 exhumation and cooling in the mid-Cretaceous, and intrusion at mid-crustal levels by granitoids by 85–90 Ma. These two HP/LT rock types are the only identified sources of the glaucophane that characterizes Late Oligocene and younger sediments in the study area and are also likely sources of kyanite and chloritoid. The easternmost glaucophane source is found at Tres Puntas on the Araya Peninsula, mapped as part of the El Copey Formation. The Villa de Cura, Cordillera de la Costa and Manicuare Formation are the easternmost identified chloritoid sources. There is a report of possible lawsonite in the Bocas-1 well (Escalona 1985), just north of the Paria Peninsula.

Class 4: Far-travelled Caribbean lower grade metamorphic terranes with Cretaceous volcanic and volcano-sedimentary protoliths and initial age of metamorphism

Class 4 rocks include the schists of the Quebradagrande Complex and northwest Guajíra Peninsula, both in Colombia, and the schists and meta-igneous rocks of Tobago (Snoke et al. 2001). The Cretaceous rocks which unconformably overlie the allochthonous El Tinaco “basement” of Central Venezuela also belong in this grouping. The Cojedes (with a basal conglomerate on El Tinaco Complex basement), Pilancones and Araguita Formations comprise Aptian–Albian clastic sediments and carbonates, with intercalated Albian and younger volcaniclastic and extrusive rocks (Bellizzia 1985). In the offshore Bocas-1 well along the northern Paria Peninsula, limestones thought to be Albian overlie metavolcanic schists (Ysaccis 1997), and the well may also contain a thin section of Mejillones Formation arc volcanic rocks. The association of low-grade metasedimentary rocks unconformably overlain by limestones and pillow basalts is typical of many Aptian–Albian circum- Caribbean arc terranes (e.g. Lebron & Perfit 1993; Pindell et al. 2005, 2006) and indicates a far- travelled Caribbean origin for the El Tinaco and the Bocas rocks. Low-grade metavolcanic rocks mapped as El Copey Formation on the Araya Peninsula may also belong in this class. They are spatially associated with fragments of mid-Cretaceous basalt that structurally overlie low-grade continental-affinity Mesozoic metasedimentary rocks (Class 6 below) and may be the remnants of the sole thrust of the Caribbean accretionary prism. The Sans Souci basalts of northern Trinidad may also be of this origin (Algar & Pindell 1993). Metamorphism is typically prehnite-pumpellyite, greenschist or lower amphibolite facies and fossils are sometimes well-preserved. The terranes are potential sources of pyroxene, hornblende, apatite, chlorite and apatite. There are few metapelites in these terranes, and no reports of chloritoid, garnet, staurolite or aluminum-silicates.

Class 5: Parautochthonous and authochthonous greenschist to amphibolite facies metamorphic terranes with Precambrian to Palaeozoic protoliths and typically Late Palaeozoic (Carboniferous–Permian) initial metamorphism

Rocks sharing the same Pangaea-assembly origin as Class 1 are common inboard of the Caribbean allochthons. They are found within thrust sheets driven ahead of accreted Caribbean rocks or uplifted during Late Oligocene and younger “Andean orogeny”, as “basement” slices beneath Mesozoic metasedimentary rocks, and as basement arches in the foreland. As such, they typically have significantly younger cooling ages than similar rocks associated more closely with the Caribbean Arc, and those ages become younger from west to east. Examples include: • Metamorphic belts intimately associated with the accretion of Caribbean terranes, such as parts of the Central Cordillera, Santa Marta Massif and Guajíra Peninsula of Colombia in the west. Garnet, sillimanite and kyanite-bearing schists are common in these belts. In the west, these terranes started to exhume following Late Cretaceous onset of subduction of Caribbean lithosphere beneath Colombia and accretion of Caribbean allochthons, while in the east Caribbean-associated unroofing is of Miocene or younger age. • Greenschist and amphibolite facies belts which started to unroof at the Late Oligocene onset of regional Andean orogeny in western South America, which include parts of the Central Cordillera, the Eastern Cordillera of Colombia, the Santander Massif and the Mérida Andes. With the exception of the Colombian Central Cordillera, which was unroofed in the Middle Eocene and then mostly re-buried in Late Eocene–Oligocene time (Pindell et al. 1998), there is no evidence for earlier Palaeogene unroofing of these belts. Andalusite is commonly associated with low-

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 22 pressure contact metamorphism adjacent to Late Palaeozoic plutons. A large area of the Mérida Andes drains directly into the upper Orinoco and this may have been the primary source for Late Oligocene and younger staurolite and Al-silicates. In contrast, these minerals are restricted to the Cordillera de la Costa of Central Venezuela and, at least at the present day, this area does not drain into the Orinoco. In the past, the Caribbean Mountains source areas may have been larger and may have drained south into a palaeo-Orinoco. • Sheared granitoids and metasedimentary rocks such as the Sebastopol Gneiss (Central Venezuela) and the Dragon Gneiss of the Paria Peninsula (Eastern Venezuela) are inferred to be the basement of Caracas Group and to structurally underlie far-travelled Caribbean rocks. Mega-feldspar augengneisses are associated with greenschist facies (chlorite, biotite, muscovite bearing) schists and phyllites, some granites and associated hornfels contact metamorphic rocks. Reported protolith ages (mostly Rb/Sr) range from c. 450–166 Ma and K-Ar and fission track cooling ages are as young as Miocene. There are no reported occurrences of garnet or staurolite within these “basement” gneisses and sillimanite is only reported in one place, adjacent to the Cordillera de la Costa. There is no evidence that Palaeogene unroofing was sufficient to fully exhume these rocks and their mineralogy indicates they are not a source for higher grade metamorphic minerals (in contrast to the model of Higgs 2006, 2009). • Poorly dated Palaeozoic rocks are also known from the foreland El Baúl Arch, where low- pressure metavolcanic and metasedimentary rocks, intruded by Carboniferous or Permian granitoids, crop out within the foreland basin. Andalusite is typical of contact metamorphic aureoles. The age of uplift of this arch is uncertain and it seems likely to be only a secondary source of andalusite, possibly only in the Late Neogene. • On the north side of the Guayana Shield there is an abrupt transition from Precambrian and Palaeozoic rock. Within the Palaeozoic terranes there may be one or more sutures related to Pangaea assembly which could be the source of some blue amphibole (identified in probable error as glaucophane) and chloritoid, found in small quantities in both pre-Cretaceous (possibly Jurassic) and Late Cretaceous passive margin strata. Such rocks have not been identified in the relatatively few deep wells and, thus, this idea remains unproven. Low-grade or non-metamorphic Palaeozoic clastic rocks are also reported from the Guárico Basin subsurface (Cambrian Hato Viejo and Carrizal Formations, PDVSA 2005). Heavy mineral assemblages in these rocks are mature, with some garnet. The strata are notably micaceous. These Palaeozoic rocks appear to be absent from the basement of the Maturín Basin, east of 65°W.

Class 6: Parautochthonous greenschist-facies or lower grade metamorphic terranes with mostly Mesozoic protoliths

In the Caribbean Mountains of Central Venezuela (66°–68°W, Fig. 1), the Caracas Group comprises greenschist facies schists in an anticlinal structural window beneath the Villa de Cura HP/LT allochthons. Limited K-Ar and fission track data indicate that peak metamorphism pre-dates the arrival of Caribbean allochthons dated by nearby foredeep subsidence (Pindell et al. 1991). Similar rocks are found in the Araya and Paria Peninsulas north of the Serranía Oriental and in the Northern Range of Trinidad, where cooling appears to be Late Oligocene and younger. Quartzites and marbles are common. Greenschist facies phyllites and schists are characterized by muscovite, epidote, chlorite and graphite. There are no reports of staurolite and chloritoid or Al-silicates. Garnet is not known east of the westernmost Araya Peninsula. Thus, as with the “basement” gneisses above, they are not a viable northern source region for sediments with higher-grade metamorphic minerals.

Class 7: Mesozoic sediments, unmetamorphosed, Cenozoic uplift and unroofing ages

Unmetamorphosed Cretaceous strata of the Colombian Eastern Cordillera and the Subandean fold-thrust belts of Ecuador and Colombia unconformably overlie metamorphic and plutonic rocks of Jurassic and older age. They uplifted significantly for the first time during the Neogene (e.g. Villamil 1999) and may have contributed significantly to Early Miocene and younger rocks in the study area. There are no published heavy mineral data on these rocks, but we expect they are broadly similar to rocks of the same age (i.e. mineralogically mature) examined in the course of this study.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 23 Class 8: Heterogeneous Precambrian metamorphosed sedimentary and igneous rocks in the Guayana Shield

A wide variety of minerals are reported from the amphibolite and granulite facies rocks of the Guayana Shield in Venezuela (alternatively spelled Guyana or Guiana in neighbouring countries), and from associated granitoids and their metamorphic aureoles, including all three Al-silicates, garnet, cordierite and chloritoid (Sidder & Mendoza 1995; Schruben et al. 1997). Glaucophane is reported, but no locality given, by Kamen-Kaye (1937) but is otherwise unknown anywhere in the shield (Salomon Kroonenberg pers. comm. 2008). Riebeckite and other blue-green amphiboles, not characteristic of HP/LT metamorphism, are present (Schruben et al. 1997) and we suspect that these may have been mistaken for glaucophane. Tropical weathering and laterite formation results in a low diversity heavy mineral assemblage in river sediment regardless of rock type in the drainage basin. We expect garnet to be more stable than staurolite, chloritoid (in acid water conditions), Al-silicates and, especially, glaucophane. None of these minerals are likely to survive more than two cycles of sedimentation. There appears to be only one published report on recent river sediment on the shield. The Río Caroní, which flows north from the Guayana Shield at c. 63°W, crosses a wide variety of metasedimentary and metavolcanic terranes. Heavy mineral residues comprise ZTR, ilmenite and, notably, staurolite (Wynn 1993). Staurolite is not reported in situ from the Venezuelan portion of the shield, but is present to the southeast in contact aureoles around granitoids within the greenstone belts of French Guiana (Delor et al. 2003a), continuous with those of Guyana (Cole & Heesterman 2002) and southeastern Venezuela, and in the high-temperature rocks of the Bakhuis Metamorphic Complex of Suriname (Delor et al. 2003b). Chloritoid is widespread, and appears to be a product of staurolite breakdown during near isobaric cooling. Although the shield must have been an important sediment source on the Cretaceous passive margin (below), we suspect that it only made a relatively small contribution to the characteristic unstable heavy minerals found in the Neogene basins because of the effects of tropical weathering. Although areas of low elevation have much lower denudation rates than areas of higher elevation (e.g. Wilkinson & McElroy 2007), the enormous area of the shield resulted in a large volume of sand, deposited in southwest to northeast-flowing fluvial-deltaic systems fringing the foreland basin in Venezuela and Trinidad up to the present day, and these sediments constantly dilute the heavy minerals derived from more exotic peri-Caribbean terranes.

PALINSPASTIC RECONSTRUCTION, TECTONIC ELEMENTS, AND PLATE BOUNDARY/THRUSTBELT EVOLUTION

In order to understand original facies distributions and the relationships of sediments to possible sediment source areas, it is essential to palinspastically restore plate movements and structural deformations back in time, and to portray former sedimentation patterns on appropriate palinspastic basemaps. Post-depositional deformation can juxtapose sediments and facies of very different origins, rotate palaeoflow indicators, and change the orientation of sandstone fairways. In northeast South America, there have been three superimposed deformation phases since the Jurassic creation of the passive margin, from youngest to oldest: (1) east-west-oriented Caribbean-South American dextral strike-slip since about 10 Ma; (2) Early and Middle Miocene southeastward dextral oblique collision between the Caribbean and South American crusts; and (3) the Palaeogene development of the Proto-Caribbean Inversion Zone (Pindell et al. 1991, 1998). A map restoring phase 1 deformation should be used to show late Middle Miocene facies belts, and a map restoring phases 1 and 2 deformations should be used to show Oligocene facies belts, and so on. From 2001 to 2007, Petrotrin provided the opportunity for us to work with much of the relevant seismic and well data in and around Trinidad and Tobago in conjunction with our field studies in Eastern Venezuela, Trinidad, and Barbados. We have critically examined most structures in the region to assign them to the correct phase of deformation, and have tested and refined estimates of shortening and strike-slip offset. This is important because most workers have previously combined the structures of phases 1 and 2 into a single model of ongoing transpressive collision between the Caribbean and South America, thus blurring the superposition of distinct structural styles.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 24 Our palinspastic maps address the stratal level at which deposition was occurring for the indicated age of the map. We account for the depth to fault detachment and only apply the restoration to strata within fault hanging walls or to terranes that are allochthonous relative to South America. This principle has some important consequences for understanding geological development, particularly in southern Trinidad where Late Miocene through Pliocene eastward extension soled into a detachment above previously deformed Middle Miocene and older strata, while a similar magnitude of dextral shear soling into an intra-Cretaceous or base-Cretaceous detachment was occurring on the Point Radix-Darien Ridge fault zone through central Trinidad. Thus, palinspastic grids appropriate for Cretaceous levels show large magnitude offset across the Point Radix Fault, while grids for late Neogene strata show much smaller offset, although the deformation is of late Neogene age. The methods of palinspastic map construction in northern South America have been described elsewhere (e.g. Pindell et al. 1998, 2000), and the detailed strain estimates used to construct the maps shown here are discussed in Pindell & Kennan (2007a). We have constructed palinspastic latitude- longitude grids for c. 12 Ma (Fig. 10), at the transition to phase 1 strike-slip-dominated plate boundary movements and for c. 25 Ma (Fig. 11), restoring the effects of transcurrent motions and oblique collision of the Caribbean Plate in eastern Venezuela and Trinidad. The estimated position of the crystalline leading edge of the Caribbean Plate is shown in both maps. These reconstructions are based on careful assessment of shortening, extension and strike-slip offsets, but reconstructions for times before 25 Ma are subject to greater uncertainty, as are strain estimates in central and western Venezuela. However, south of the deformation front shown for 25 Ma (Fig. 11) the basemap for older reconstructions remains the same. A general tectonic elements map of the Caribbean-South America collision zone (Fig. 12), using the 25 Ma palinspastic restoration, includes the migrating Caribbean trench and arc, the migrating Caribbean foredeep on the South American margin, and the Proto-Caribbean Inversion Zone of northern South America. The concepts of arc-passive margin collision (Speed 1985) and of a migrating Cenozoic arc-passive margin collision (Dewey & Pindell 1986; Pindell et al. 1988) have been well accepted, but the existence of a Proto-Caribbean Inversion Zone ahead of the Caribbean Plate remains more speculative, and is one of the features on which this paper may shed some light. The Proto-Caribbean Inversion Zone was probably initiated in the latest Maastrichtian, and certainly by the Palaeocene, by which time motion between North and South America had become convergent (Pindell et al. 1988; Müller et al. 1999). Based on seismic tomographic images (Van der Hilst 1990) of subducted Atlantic/Proto-Caribbean lithosphere below the eastern and southern Caribbean Plate, we have previously suggested (Pindell et al. 1991, 2006; Pindell & Kennan 2001, 2007a) that the convergence was accommodated, prior to the arrival of the Caribbean Plate from the west, at a newly formed south-dipping “Proto-Caribbean Inversion Zone” beneath northern South America. Today, only the eastern end of this inversion zone has not yet been subducted beneath the Caribbean Plate and remains visible at the earth’s surface, projecting east from the Lesser Antilles trench. There, an ENE- trending ridge (south) and trough (north) pair with some 3 km of buried basement relief is situated between the Caribbean crystalline limit and the Late Maastrichtian western Atlantic magnetic anomaly 30, and cuts across the previously formed regional pattern of Atlantic fracture zones (Speed et al. 1984). The linearity, basement relief, and cross cutting relationship of this structure are suggestive of a north-vergent thrust/inversion zone plate boundary. Beneath the Caribbean Plate, the seismic tomography suggests that this plate boundary once continued west-southwest to at least the Golfo de Triste along Venezuela, at the northeast end of the Mérida Andes. From there, the plate boundary may have continued along the limit of the continental crust to the north of Maracaibo and northern Colombia. It is not yet clear if deformation in the Mérida Andes of possibly Palaeogene age tied into this structure as well. Accepting our projection of the Proto-Caribbean Inversion Zone below the present day Caribbean Plate (Fig. 12), we suggest that the Proto-Caribbean hanging wall would have had positive but submarine bathymetric expression created by basement uplift about 2–3 km, as basement does today in the not-yet-subducted area south of Tiburón Rise (see basement structure map in Speed et al. 1984; Pindell & Kennan 2007a). We propose the existence of the “Barcelona Trough”, situated north of the South American margin and south of the Proto-Caribbean Inversion Zone and accretionary prism, that formed during the Palaeocene (Fig. 12). The trough was the site of significant Palaeogene clastic deposition, initially deep water but shallowing upward by Late Oligocene time. It was also the site of initial Oligocene

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 25 -64 -63 -62 -61 -60 12 12 Ma Backthrusted Proto- CARIBBEAN Caribbean Ridge Blanquilla Basin backthrust PLATE

Testigos ? ANGOSTURA TOPAZ 11 11 EMERALD MTF 200 km SPITFIRE DIAMOND COUVA MT. HARRIS Urica Transfer PLAISANCE SAN FERNANDO BRIGHTON

10 10 Limit Villa de Cura Pirital Fault Contrast present (light Deformation Front grey) & past (dark) Leading edge of blind shapes of Trinidad deep Caribbean basement slice 9 9 MTF = Margarita Transfer Fault (S. Carib Foldbelt lateral ramp) 100 km

-67 -66 -65 -64 -63 -62 -61 -60

Fig. 10. Palinspastic reconstruction for 12 Ma, close to the end of Middle Miocene orogeny showing the geometry of the orogen before Late Miocene and younger deformation? Distorted latitude-longitude grids account for cumu- lative deformation and the resulting map is analogous to a restoration of a balanced structural cross-section. Restor- ing even Late Miocene and younger deformation dramatically distorts the shape of Trinidad (palaeoshape shown with bold lines). To make this map we used relatively conservative shortening and shear estimates south of El Pilar, Caroni Fault. The position of the Northern Range is harder to constrain because of uncertainties in the amount and age of internal strain. The map provides a geographic framework for plotting and reconstructing Middle Miocene deformation.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 26 -62 -61 -60 25 Ma CARIBBEAN Continent-Ocean Boundary? 12 PLATE 11 Margarita PROTO-CARIB RIDGE OVERTHRUST TRACE OF PROTO-CARIBBEAN TRENCH NORTHERN DEPOCENTRE (BARCELONA TROUGH) 11 Wedged Carib. crustOVERTHRUST CARIB PRISM

SOUTHERN Buried 10 PLATFORM early Cret. shelf edge. Future Pirital Fault 10 Overthrust Villa de Palaeoshape Cura nappe of Trinidad

9 9 Espino hanging wall Espino half-graben 100 km

-67 -66 -65 -64 -63 -62 -61 -60

Fig. 11. 25 Ma palinspastic reconstruction, showing the geometry of the orogen before Early and Middle Miocene deformation. Note that the distortion in the shape of Trinidad indicates that apparent north-south facies changes between the Central Range and Southern Basin in present-day geographic coordinates are in fact NW-SE facies changes parallel to the Bohordal Escarpment. One important effect of the retro-deformation is to place the northern depocentre or Barcelona Trough adjacent to the Serranía Oriental, such that Palaeogene sands can be derived from the west or southwest without crossing the southern Trinidad Platform pelagic shelf, where carbonates prevailed. Large Early Cretaceous clasts (such as in the Plaisance conglomerate) can derive from either/both the northeast- facing Bohordal slope or the north-facing Central Range slope through incision of bypass surfaces with little Late Cretaceous or Palaeogene cover. The map supports the hypothesis of a point source for sediment from what we refer to as the “Espino-Maturín River”. The proposed “Espino-Maturín River” was almost certainly the source of the San Juan Sand Lobe of Maastrichtian age (see Fig. 15 below).

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 27 -74° -72° -70° -68° -66° -64° -62° -60° -58°

Venezuelan Former arc Tiburón Rise Basin (ODP 171) South Caribbean Foldbelt: Axial 14° E-ward tearing backthrust, taking turbidites 14° up Carib-SoAm shortening after Aves Ridge Proto-Caribbean Prism (submarine) choking of the collision in the Tobago

south, which is older to west) Grenada Basin Tobago Trough Lesser Antillles Arc Proto-Caribbean Ridge 12° 12° Margarita “Barcelona Trough” Bohordal Present Leeward Antilles Ridge Bohordal F. COB shelf edge Re-entrant Caribbean Bonaire Basin Caribbean Prism Urica Fault crust, refer- (stretched former forearc) ence only Tinaco, V de Cura, Serrania Los 10° Guarico Belt 10° not yet exhumed Quebradón del Interior Bajos shelf edge La Pascua/Roblecito Oriental marine foredeep Fault Former positions El Baúl of present coasts Chaguaramus 8° fluvial plain 8° 250 km Incipient Mérida Andes -74° -72° -70° -68° -66° -64° -62° -60° -58° Fig. 12. Mid-Tertiary tectonic elements of the Tertiary Caribbean-Proto-Caribbean trench-trench collision, com- piled from features identified and defined in Pindell et al. (1998, 2006). This reconstruction shows a v-shaped remnant of the Proto-Caribbean Seaway in the northeast, bounded by the Caribbean and Proto-Caribbean accretionary prisms. The Proto-Caribbean Ridge is thought to have developed as the South American Plate ramped up over underthrust Proto-Caribbean crust and initiated prism formation. It acted as a bathymetric barrier to sediment flow between the Barcelona and Proto-Caribbean basins, still visible on basement structure maps to the east of the present day Caribbean Prism and south of the Tiburón Rise. From Guajíra to the Serranía Oriental (overridden by the Caribbean by the time of this reconstruction), the hanging wall of the trench was continental, and from Serranía Oriental eastwards the trench was intra-oceanic. It is not clear whether a narrow oceanic or a thinned continental forearc existed north of the Serranía or in the Bohordal re-entrant. Caribbean collision with the Proto-Caribbean Inversion Zone was diachronous from west to east, reaching the Guajíra area in the Maastrichtian-Palaeocene, Maracaibo in the Eocene, Central Venezuela in the Oligocene, and Eastern Venezuela in the Miocene.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 28 south-vergent structural shortening in Trinidad, as the originally north-vergent Proto-Caribbean Ridge (hanging wall of Proto-Caribbean Inversion Zone) was incorporated into the greater migrating Caribbean orogen, and then backthrust, with respect to the Proto-Caribbean Inversion Zone, southeastward toward the southern Trinidadian platform. The Central Range of Trinidad restores to a former position in the vicinity of the present day Paria Peninsula. The northern part of the Early Cretaceous carbonate platform (subsequently drowned) in the Serranía Oriental restores some 100 km farther northwest of the Early Cretaceous southern Trinidad platform (Cuche Formation); there must have been a significant marginal offset and escarpment to the east of the northern part of the Serranía Oriental. This northwest-trending feature has been referred to as the Bohordal Escarpment (Pindell et al. 1991; Pindell & Kennan 2001). However, because the strata that defined it have been incorporated into the Serranía-Nariva fold-thrust belt, its original position can only be estimated from structural restoration of facies changes (e.g. El Cantil carbonate in Venezuela to Cuche shale in Trinidad). Any model for the tectonic history of northern South America (Fig. 13) must accommodate Cenozoic convergence between the Americas, which increased westward from zero in the area of magnetic anomaly 30 (c. 66 Ma), east-northeast of Barbados, to some 450 km at Guajíra, Colombia. At crustal levels, as the Caribbean lithosphere arrived in any north-south cross section during its eastward migration relative to the Americas, inter-American convergence was partly taken up at the southern Caribbean plate boundary, giving the southern Caribbean its predominantly compressive character. However, at depth beneath the Caribbean Plate, the original Proto-Caribbean lithosphere must also be shown to balance in north-south cross sections from North to South America (Pindell & Kennan 2007a). About 100–150 km of shortening had already accumulated at the Proto-Caribbean Inversion Zone at any point by the time it was overridden by the Caribbean Plate and about 70 km of underthrusting had been achieved near Eastern Venezuela between latest Maastrichtian and latest Eocene time. Such significant magnitudes of convergence must be manifested in the geology of northern South America, including the heavy mineral data. If the Mesozoic passive margin slope sedimentary section at the position of the Proto- Caribbean Inversion Zone at the time of its end-Cretaceous inception was a typical 10 km thick, then 100–150 km of north-south convergence by Late Eocene time would have produced a significant accretionary prism at least 15 km thick (possibly partially subaerial) with incipient metamorphism at deep levels (Fig. 13a). The slope and rise strata include the Late Jurassic–Cretaceous Caracas Group of the Caribbean Mountains and the Paria-Northern Range terrane (Speed 1985, 2002; Algar & Pindell 1993). Note that the rocks of the Araya Peninsula differ from those of Paria in that they were metamorphosed in the Cretaceous (Sisson et al. 2005) rather than in the Oligocene (see below), and thus are allochthonous of Caribbean origin, thrust onto the western Paria terrane. The position of this accretionary belt, with an isostatically estimated relief of perhaps 3 km above the Proto-Caribbean seafloor to the north (i.e. 1–2 km subsea), can be crudely reconstructed from seismic tomography data (see Pindell & Kennan 2007a). As the Caribbean and South America hanging walls converged, the peripheral bulges ahead of them would become yoked (Fig. 13b) and no longer roll or migrate ahead of either. In that case, further convergence would lead to hanging wall uplift in both margins, generating low stand fans along each (Fig. 13c). Further convergence would progressively load the Proto-Caribbean until it sank into the mantle, allowing collision between the Caribbean and South American crust (Fig. 13d). In northern South America, the Caribbean forearc initially overrode South America, but later, after accretion, thrust polarity has reversed, especially in the west (Fig. 13e). 40Ar-39Ar cooling ages on F1 metamorphic micas of 35–40 Ma from the Caracas Group (Sisson et al. 2005), 14–30 Ma from the Paria terrane (Speed et al. 1997), and 23–26 Ma from the Northern Range (Foland et al. 1992) are interpreted by those authors to record the peak of metamorphism coincident with the development of F1 foliation, from which the rocks have progressively cooled by unroofing due to erosion judging from zircon and apatite fission track studies (Algar et al. 1998; Sisson et al. 2005; Cruz et al. 2006). In at least the Paria Peninsula and the Northern Range, this F1 foliation dips south and carries an east-west stretching lineation (Algar & Pindell 1993; Cruz et al. 2006), and in Paria, at least, the early deformation involved top-to-the-west shear (Cruz et al. 2006). The F1 foliation in the Northern Range has been related to north-vergent deformation (Potter 1973; Algar & Pindell 1993), consistent with a north-facing accretionary prism.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 29 Caribbean and Proto- A Proto-Caribbean Caribbean forebulges yoke as B Guarico “hanging plates converge Proto-Caribbean accretionary prism wall” basin forebulge A Proto-Caribbean SAM Carib. Prism Proto-Caribbean Prism Guayana Transform B Caribbean Barcelona TroughBohordal reentrant Serranía A’ South Shelf edge Guarico Oriental Trinidad Active Basin B’ Buoyant South American collision crust

Yoked Caribbean, Accretionary C Proto-Caribbean prisms forebulges (lowstands) Hanging wall A unconformities A’

Caribbean SAM Caribbean Proto- D Caribbean Continued hanging Foredeep Basin forebulge B wall uplift (La Pascua/Roblecito) B’

Caribbean SAM

Caribbean Proto-

South Caribbean Future Leeward Antilles, Bonaire Basin, Villa E and Blanquilla MorónFt, de Cura wedge in Early Miocene time thrustbelts & uplift

Caribbean SAM

Caribbean Proto-

Fig. 13. Schematic history of Caribbean-South American interactions. (a) Onset of Proto-Caribbean underthrusting prior to Caribbean arrival, producing Proto-Caribbean accretionary prism. Guárico Basin lies in hanging wall of South America and thus may be tectonically active but not receiving Caribbean erosional detritus, except possibly airfall tuffs). (b) Map of Caribbean-South American oblique convergence above Proto-Caribbean lithosphere for late Middle Eocene time. The obliquity causes both the hanging wall uplift (position shown by fine circles) and the Caribbean forebulge (heavy grey line) to migrate eastwards. Because the trends of the hanging wall uplift and Caribbean forebulge are different, complex interference patterns will result. Here, the Guárico Basin hanging wall uplift is already transgressed by the Caribbean foredeep, but the Serranía Oriental is undergoing hanging wall uplift. Due to the Bohordal re-entrant, the southern Trinidad platform is too far from the Proto-Caribbean Inversion Zone to experience hanging wall uplift. With further convergence, hanging wall uplift will affect the Bohordal re-entrant (shown in dashed circles), and the Caribbean forebulge will pass through Trinidad in the Earliest Oligocene. (c) Mechanism of diachronous hanging wall uplift (Caratas, Cautauro, and Peñas Blancas/Tinajitas Formations record shallowing up, but Middle Eocene in Guárico Basin is removed by erosion) and tectonic production of lowstand wedges above an intermediary plate (Proto-Caribbean lithosphere). The two forearcs load the Proto-Caribbean and cause it to sink into the mantle, but not without concurrent uplift of the forearcs. Line of section is A-A‘ in part b. (d) Result of continuing convergence, with consumption of Proto-Caribbean lithosphere, and the Caribbean overthrusting South America. Migrating Caribbean foredeep develops on South America with Caribbean forebulge beyond the foredeep. Line of section is B-B’ in part b. (e) Future (Late Oligocene) tectonics of section B-B‘; South America chokes the north-dipping trench, and backthrusting develops outboard of the Leeward Antilles Arc, which takes up most subsequent convergence between the two plates.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 30 This may also be the case for the Paria Peninsula, but Cruz et al. (2007) have derived a different (and much younger) explanation for the south-dipping F1 foliation there, the timing of which we question. The peak metamorphic ages noted above closely match the time of Caribbean oblique arrival at our estimated position of the Proto-Caribbean prism, slightly older to the west (i.e. 100–300 km to the northwest of the present positions of these terranes based on palinspastic reconstructions and our estimated position of the Proto-Caribbean Inversion Zone). The Caribbean forearc basement (Villa de Cura complex and its eastward continuation in the western Gulf of Barcelona subsurface; Ysaccis 1997) and the Caribbean accretionary pile (e.g. the allocthonous Caramacate, Los Cajones, Garrapata Formations of the Caribbean Mountains; the allocthonous Manicuare and Copey Formations of the Araya Peninsula, and the Sans Souci allochthon of the Northern Range) were thrust onto the pre- existing Proto-Caribbean prism. In Central Venezuela and the western Gulf of Barcelona, the Villa de Cura forearc basement wedge itself was thrust over the Proto-Caribbean prism by some 50–100 km, whereas to the east in the Paria-Northern Range, where prism-prism collision began later and in a more oceanic setting, only the Caribbean prism, and not the forearc basement, appears to have been thrust onto the Proto-Caribbean prism. In Central Venezuela, the forearc Villa de Cura/Tinaco Nappe overthrust the Caracas Group and accreted it to its base. With continued shortening, the Villa de Cura/Tinaco/Caracas composite nappe terrane was thrust southward onto, or wedged into, the Guárico Belt of the outer South American shelf margin in the Middle Eocene, which in turn shortened during Middle Eocene time, when peak metamorphism was reached in the Caracas Group (35–40 Ma). In the Late Eocene–Early Oligocene (30–35 Ma), the imbricated Guárico Belt was entirely detached from its basement and thrust to the south, driving subsidence in the foredeep Guárico Basin (La Pascua-Roblecito Formations). In the Early Oligocene, the overthrust South American crust began to choke the collision zone and further advance of the Caribbean nappe pile slowed considerably. The bulk of continued shortening between the Caribbean and South America transferred to new structures in the north, including the north-vergent South Caribbean Foldbelt, and the Margarita Fault or Transfer Zone (Ysaccis 1997) served as the lateral ramp along which that shortening was transferred northwest (on Fig. 12 the Margarita Fault is shown linking to the eastern end of the Villa de Cura Nappe at this time). Both the South Caribbean Foldbelt detachment and the Margarita Fault sole into an intra-crustal detachment that deepens south beneath the Leeward Antilles arc. In Central Venezuela, this south- dipping detachment cut the entire Caribbean lithosphere, allowing the lower footwall of the Caribbean to begin underthrusting the crust of the South American hanging wall (Fig. 13e). About 100 km of further shortening has occurred on this structure since the Oligocene, and a slab of Caribbean lithospheric mantle (not crust) is visible on Central Venezuelan seismic tomographic images beneath the Caribbean Mountains (Van der Hilst 1990; Pindell & Kennan 2007a). To the east, rather than the Caribbean forearc basement overthrusting the Proto-Caribbean prism, it appears that the Tobago forearc terrane was wedged beneath the Paria-Northern Range terrane. This Early Oligocene event imparted or enhanced the south-dipping F1 foliation in the Paria- Northern Range prism, along with its east-west stretching lineation and top-to-west sense of shear (dextral, north-vergent backthrusting). As a result of this wedge insertion and backthrusting, cooling in Paria-Northern Range was probably initiated by a combination of cooling from below, related to the cold Caribbean forearc slab beneath Paria and the Northern Range, and the surface uplift and erosion that ensued above the wedge. Concurrently, south-directed thrusting of the older Proto-Caribbean prism was initiated, resulting in Late Oligocene syn-tectonic deposition of Upper Naricual Formation of the northwest Serranía, and initial imbrication of Early Oligocene strata in the Bohordal re-entrant (possibly including the Angostura Formation, N. Evans, oral presentation to 4th Geological Conference of the Geological Society of Trinidad and Tobago, 2007). On the east side of the Margarita Fault, backthrusting within the Caribbean crust also developed in the Early and Middle Miocene at the same time as Serranía thrusting; the Blanquilla foldbelt between Margarita and La Blanquilla shortened some tens of kilometres prior to 10 Ma (Ysaccis 1997), constricting the original southern end of the Grenada Basin. At 10 Ma, the Blanquilla backthrust was positioned entirely west of the Bohordal transfer zone, and backthrusting farther east (northwest flank of the Tobago Ridge) is only hinted at locally; it is apparent that to the east of the Bohordal transfer zone (Fig. 12), the Bohordal re-entrant with its thinner continental or oceanic crust and thick overlying shaly Palaeogene clastic section offered less resistance to south-directed thrusting than the Serranía to the west.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 31 In the light of this dynamic history and our heavy mineral data, we propose the existence of six clastic domains associated with distinct tectonic settings in northern South America. Four domains are present on a mid-Cenozoic reconstruction (Fig. 12), but two are younger. Three of the domains are not geographically fixed but sweep diachronously from west to east with the migrating Caribbean- South American collision, causing strata of different domains to be superimposed. The six clastic domains and their respective proposed source areas are: • Cretaceous passive margin clastics, derived from the Guayana Shield to the south; • Palaeogene clastics associated with development and dynamics of the Proto-Caribbean Inversion Zone and the Barcelona Trough, derived from the pre-existing passive margin section and/or shield; • cratonic side of the migrating Caribbean foredeep fill, derived from reworked existing marginal strata above the craton, with some first cycle input from the craton; • orogenic side of Caribbean foredeep fill, derived from the subaerial parts of the migrating Caribbean-South America collision belt and deposited down depositional dip in the foredeep basin axis; • proximal, syn-orogenic detritus, derived from the uplifting Caribbean-South America collision zone, and deposited adjacent to the mountain front, in wedge top and piggyback basins; • post-orogenic foreland fill, derived from the east-flowing Orinoco River.

CLASTIC DOMAINS

Domain 1: Cretaceous passive margin

Northern South America persisted as a passive margin until the Maastrichtian, following Jurassic rifting from eastern Yucatán and Neocomian dextral motion along the Guyana Transform (Pindell & Kennan 2007a). Sequence- and litho-stratigraphic patterns indicate a northward-deepening palaeobathymetry with marginal offsets: the passive margin (Fig. 14) faced the Proto-Caribbean Seaway, an arm of the Central Atlantic (Pindell & Barrett 1990; Erikson & Pindell 1998a, b). Overall, as expected in a thermally subsiding passive margin, successive transgressions step farther into the continent through time, and along the Guyana-Suriname margin former prograding shelf edges have been drowned to form two or three packages of strata back-stepping landward from the continent- ocean boundary. Seismic, well and sequence stratigraphic data (e.g. Erlich et al. 2003; Erlich & Keens-Dumas 2007) indicate that the Early Cretaceous shelf edges were close to the present-day Central Range in Trinidad, and the Guyana Transform offshore from Trinidad to Suriname. Subsequently, the shelf edges in younger Cretaceous strata stepped south (or landward), being located in Maastrichtian time near the south coast of Trinidad and just offshore from Venezuela to Suriname. The heavy mineral signature of Trinidadian Cretaceous rocks is dominated by zircon, tourmaline and rutile. The paucity of labile minerals (common in the rocks of the eastern Guayana Shield) indicates intense weathering in the sediment source region. Of particular interest are the red “cherries” which characterize the Albian Cuche and Toco Formations in Central and Northern Ranges in Trinidad, respectively, and similar red mudstone clasts in the Cenomanian Gautier and Santonian Lower Naparima Hill Formations in the subsurface of the southern basin. In thin section these contain angular silt to fine quartz sand with some muscovite embedded in an iron oxide matrix, and the most likely source is from laterites or intensely oxidized coastal plain mudstones and siltstones, with access to the Shield for mica and fine-grained angular quartz clasts, eroded by streams or coastal incision. Limited XRD data indicates that calcareous nodules within the shales of the Galera Formation shales are neither the same nor sufficiently iron-rich to be precursors to “cherries” in various younger Tertiary formations and sections. The presence of coarse sandstones and cherries in the Cuche Formation of the Central Range has been used by Higgs (2006, 2009) to propose a northern sediment source as old as Aptian. However, the Cuche strata of the Central Range are somewhat older (Barremian–Early Aptian) than shaly “Cuche Formation” reached in the few deep wells in southern Trinidad, which may be correlative with the late Aptian to early Albian Garcia shale in Venezuela. The sediments of the

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 32 -67 -66 -65 -64 -63 -62 -61 -60 12 12 Proto-Caribbean oceanic crust Water depth increases north and east Cuche shale drowns both shelf edge carbonates and south-derived conglomeratic ssts (e.g. Cuche River) Cuche, Garcia trans- which were deposited during earlier lowstands Approx. COB gresses underlying Olistostromes? N. Range platform carbs, ssts (Chancellor) Guyana Transform 11 11 Urica Escarpment Bohordal Escarpment Early Cret. top of slope

El Cantil (reef & sandy limestones) El Cantil on Barranquin? 10 Canoa 10

El Cantil sandy in SW Serrania due to input from Espino-Maturín River Top of slope backsteps into Canoa? southern Trinidad by Santonian

9 9

Aptian-Lower Albian, 110 Ma 100 km -67 -66 -65 -64 -63 -62 -61 -60

Fig. 14. Palaeogeographic reconstruction for Aptian to Early Albian time. Northern Venezuela and Trinidad were part of the Domain 1 passive margin at this time, facing the Proto-Caribbean Seaway, with continent-ocean boundary (COB) probably located near the palinspastically restored position of the Araya-Paria Peninsulas and the North- ern Range. A former shelf (represented by the Barranquín and Morro Blanco Formations in Venezuela) was drowned by the Garcia Shale transgression and only in Venezuela did a widespread carbonate shelf re-establish itself (El Cantil Formation). In Trinidad, early Cuche sandstones, limestones and conglomerates were buried by the Cuche Shale and succeeding Gautier Shale in the Central Range, also near a drowned shelf edge. We interpret the outcrops in the Cuche River as south-derived, part of a pre-transgression reef or reef-fringing trend, analogous to similar south-derived facies of Cenomanian-Santonian age in southern Trinidad. They are overlain conformably by the Late Aptian Maridale Marl (Garcia equivalent). Strata older than the Maridale Marl have not been drilled in southern Trinidad, but to the south in Venezuela there is a thick sand and conglomerate clastic fringe of early Aptian and older age overlying the Guayana Shield.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 33 Central Range Cuche could be derived from a southern sediment source if the shelf edge was at or near their site of deposition and subsequently back-stepped to the south. The Cenomanian Gautier and Santonian Lower Naparima Hill Formations have been extensively cored in southern Trinidad and include wedges of sandstone and conglomerate very similar to the coarser Cuche facies (Erlich & Keens-Dumas 2007). These were deposited during lowstands or periods of wetter climate in their sediment source areas. Facies variations and the geometry of slump folds and slump faults, fold and fault vergence, ripples, and pebble orientation and imbrication in oriented cores all indicate southward derivation and that palaeobathymetry deepened to the north/northwest (Larue 1995; Sprague et al. 1996; unpublished industry data). To the south of Trinidad, these sands and intervening shale- dominated formations give way to a sand and conglomerate updip fringe (di Croce et al. 1999). We interpret the coarse Cuche facies as south-derived flushes of material encased in shale. Barranquín and younger Cretaceous clastic sections were sourced from the Precambrian rocks of the Guayana Shield, and from the now-buried Palaeozoic rocks (Hercynian, related to the assembly of Pangaea) and early Mesozoic rift fill (Espino Graben and equivalents). The trace staurolite in the Barranquín can be tied to the Shield where it is found today in Río Caroní sands (Wynn 1993). Only to the west, in the Guárico foreland basin, do Cretaceous formations show small amounts of minerals such as kyanite, chloritoid and glaucophane. The kyanite and chloritoid are both proven in Guayana Shield rocks to the south. The origin of the glaucophane is unclear, but may be from suture zones within the Precambrian, or between the shield and heterogeneous Palaeozoic terranes that underlie the foreland basins of western and central Venezuela. Cretaceous sandy sediments of Late Albian– Cenomanian age in Domain 1 occasionally include garnet and apatite (e.g. Gautier and Toco Formations). This coincides with a well-developed regional lowstand (Late Cenomanian, Villamil & Pindell 1998) and may indicate faster transport of minerals from source to basin, preserving the intermediate-stability garnet and apatite, but not less stable species.

Domain 2: Palaeogene clastics associated with development and dynamics of the Proto-Caribbean Inversion Zone and the Barcelona Trough

Inter-American convergence began at the proposed Proto-Caribbean Inversion Zone in the latest Cretaceous (Fig. 15) and this inversion zone and its associated accretionary prism probably led to the development of a submarine ridge that separated the original Proto-Caribbean oceanic basin and the Barcelona Trough. By Palaeocene–Early Eocene time (Fig. 16), the Barcelona Trough narrowed westward and tapered out in the Guárico Trough of onshore Venezuela, and was open to the deep Atlantic Ocean toward the east. Unlike the younger Caribbean foredeep there is no diachronous shift in age of its sedimentary fill with respect to South America. The Proto-Caribbean Inversion Zone continued westward, along the northern flank of the structurally growing Northern Range, Paria, and Caracas Group accretionary belt, until the point, which migrated east through the Cenozoic, where it was progressively overthrust by the Caribbean trench. The Proto-Caribbean Seaway to the north of the prism, but not the Barcelona Trough, received early Palaeogene to orogen-derived heavy minerals and lithic clasts from the Caribbean-South America collision zone updip to the southwest. The two trenches bounding the Proto-Caribbean basin (Figs 16, 17) were logical sites for long distance down-dip axial transport of Caribbean orogenic clastic sediments, but much of that detritus may have spread out across the Proto-Caribbean interior as well. The Late Palaeocene–Middle Eocene Scotland Formation or Basal Complex of the Barbados accretionary prism, with its glaucophane, sillimanite, and other minerals typical of arc-continent collision, undoubtedly relates to the Proto- Caribbean clastic province somewhere far to the west of Barbados today, as coeval sandstones in eastern Venezuela and Trinidad bear no such minerals. A question that is not yet resolved is whether, or how much of, the Basal Complex was accreted at the Proto-Caribbean prism versus the Caribbean accretionary prism (Fig. 17) prior to their collision and merger in the Middle Miocene (Pindell & Frampton 2007). The occurrence of glaucophane in the Eocene–Oligocene section in DSDP wells on the Tiburón Rise (Mascle et al. 1988a, b) suggests that some orogenic heavy minerals were able to reach even farther northeast within the Proto-Caribbean/Atlantic basin than the original depositional site of the Barbados Scotland beds. The Palaeocene–Early Eocene Matatere Formation of the accreted Lara Nappes of Western Venezuela also carries an orogenic signature (PDVSA 2005), and probably represents a more proximal portion of the Caribbean orogenic detrital fairways.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 34 -72° -71° -70° -69° -68° -67° -66° -65° -64° -63° -62° -61° -60° -59° 14° 14° Caribbean Plate ` Coastlines are in palinstastically was 250 km west Proto-Caribbean seafloor restored position, accounting for subse- of here at this time quent strike-slip and shortening offsets.

13° 13° Note: at this time, Maracaibo, Inversion zone initiated as sinistral Guajira, Paraguaná were to west. transform in Late Cretaceous time Incipient Proto-Caribbean Accretionary Prism Incipient Proto-Caribbean Inversion Zone 12° Palaeoshape 12° of Trinidad

Outer Slope

11° 11° Slope Shelf-Slope Break Trinidad San Juan Slope Paleoposition of Caracas Gp., Araya-Paria-Northern Range Platform 10° Sand Lobe 10° Outer Shelf Inner Shelf Outer Shelf Fluvial-deltaic? Inner Shelf

9° 9° Fluvial-deltaic? 100 km Late Maastrichtian, 68 Ma ESPINO-MATURIN RIVER 8° 8° -72° -71° -70° -69° -68° -67° -66° -65° -64° -63° -62° -61° -60° -59°

Fig. 15. Palaeogeographic reconstruction for the Maastrichtian (68 Ma) of northeastern South American continental passive margin (Domain 1). Present day coastlines are shown with fine grey lines and restored positions of coastlines are shown with heavy black lines. The palaeogeographic map is drawn for the period immediately prior to the onset of underthrusting on our proposed “Proto-Caribbean Inversion Zone”. The map shows the buried positions of the Early Cretaceous shelf-slope break beneath the Central Range of Trinidad, which continued to support a relatively shallow “Trinidad Platform”. Sedimentary facies indicate the existence an inner slope close to the south coast, but the ultimate drop to abyssal depths lay north of the buried older shelf edge. The buried “Bohordal Fault” is inferred to bound the Serranía Platform on the east, and the ancestral Urica Fault is thought to separate the Serranía from Central Venezuela. Both were active as sinistral transfer faults during Early Cretaceous and older rifting. The continent-ocean boundary to the east of Trinidad is well-constrained from present-day Columbus Channel to Suriname. Prior to Jurassic opening of the Proto-Caribbean Seaway, the deep basement of Cuba and the Bahamas Platform lay northeast of this line. Sedimentary facies are shown as fluvial fringe, shallow marine inner shelf, outer shelf or upper slope (light grey) and mid-slope and deeper (dark grey). The interpretation is based on well data and our own field work. The southern fluvial fringe is mostly eroded and onlapped by La Pascua and younger formations (Domain 3, southern onlap edge of Caribbean foredeep, see Fig. 17). Of particular note is the San Juan Sand Lobe (isopachs based on Di Croce 1989), which appears to originate from the axis of the Jurassic Espino Graben. The southern boundary faults of this graben (inactive by this time) may have controlled topography and the course of an “Espino-Maturín River”. The faults were subsequently inverted during the late Cenozoic (as the Anaco Thrust).

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 35 In contrast to the trench-bounded Proto-Caribbean province, the Barcelona Trough was, beginning in the Maastrichtian, shallower than the Proto-Caribbean basin floor and a logical site for deposition of South America-derived progradation of mature clastics as a result of long-term and/or episodic relative sea level fall. Our data suggest that the Palaeogene sands of Eastern Venezuela and Trinidad derive from the Shield or are reworked from Cretaceous passive margin sediments such as the Barranquín, Canoa and Temblador Formations. Recycled Cretaceous sediment is likely to be cleaner than first-cycle sediment sourced directly from the shield, explaining the typical lack of staurolite, garnet or apatite in the Palaeocene through Early Oligocene strata. Suter (1960) reported staurolite in some Palaeogene samples, drawing on unpublished oil industry work by J. C. Griffiths (the original reports can no longer be found). Our own examination of the few available unpublished industry reports from that era show that trace staurolite and kyanite has occasionally been found in southern Trinidadian Late Cretaceous (Naparima Hill) to Eocene or Early Oligocene (San Fernando) strata. We found only one report on a single kyanite grain from Pointe-a-Pierre grits at the type section. Palaeocene–Early Eocene strata of the Barcelona Trough (Fig. 16) include the shale dominated fine distal turbidites of the Guárico Formation of Central Venezuela, the northern facies (Pindell & Kennan 2001) of the Vidoño Formation, and the Chaudière Formation of central Trinidad, long noted to have lithological and faunal similarities (e.g. Salvador & Stainforth 1968). The Guárico, northern Vidoño and Chaudière Formations carry detrital muscovite, but apparently no heavy minerals other than ZTR. Muscovite is the only mica capable of surviving lateritic weathering intense enough to breakdown staurolite, garnet and other metamorphic minerals (Anand & Gilkes 1987), so a South American source for the Barcelona Trough sediment is entirely reasonable, the most likely being the Guayana Shield or perhaps reworked Cretaceous Canoa Formation incised inner margin strata, both to the south of the Barcelona Trough. In the northern Serranía Oriental, the northern Vidoño facies beds carry 1–5 m blocks of chert/hemipelagite of probable olistostromal origin. The blocks resemble the underlying Campanian–Maastrichtian San Antonio/Río Chavez Formation of the Serranía platform, and thus are likely south derived. On the eastern flank of Mount Harris, eastern Central Range of Trinidad, either the Chaudière or the Plaisance (field relations do not allow determination of which) also carries 1–5 m probably allochthonous blocks, identified as Maastrichtian Guayaguayare and Upper Cretaceous Naparima Hill and Gautier Formations (Kugler 2001), again well known to the south; transport direction was presumably northwards into the deeper parts of the basin, possibly via shelf-edge canyons or even shelf edge slump complexes. Nothing in these formations suggests or requires a northward derivation of material for these northward deepening units. Over much of the Serranía Oriental, the glauconitic Vidoño Formation (upper slope/outer shelf) coarsens upward into the glauconite-bearing shelf sands and shales of the Early to Middle Eocene Caratas Formation, which, in turn, is capped over much of the western and southern Serranía Oriental by the Tinajitas Limestone. Because this shallowing upward trend appears to occur over only about 100 m of section, we suspect the shallowing was tectonically driven (i.e. due to hanging wall uplift and/or the Caribbean forebulge). Unfortunately the age of this event is poorly constrained. Mixed assemblages Middle- and Late Eocene larger foraminifera are common (Sageman & Speed 2003; O. Macsotay pers. comm. 1995; T. King pers. comm. 2008) and the oldest age of overlying strata permits a latest Eocene or Early Oligocene age. To the west of the Urica Fault trend in Central Venezuela, the Guárico Formation (south of Don Alonso–Guárico Fault) appears not to extend above the Early Eocene level, yet the foredeep subsidence south of the Guárico Formation does not begin until the latest Eocene or Early Oligocene (La Pascua Formation onlap, younging southeastward). Middle Eocene conditions here are thus not clear. This may be due to extreme telescoping at the frontal Guárico thrust such that a more distal Middle Eocene section is entirely overridden. However, by analogy with the Serranía Oriental, it may be that the Middle Eocene was a time of uplift and erosion in Central Venezuela, such that the imbrication of the Guárico Formation by Caribbean collision post-dated the Proto-Caribbean hanging wall/Caribbean forebulge unconformity. Further, deformations during this pre-Caribbean uplift may have produced a reported angular unconformity beneath the late Middle Eocene Peñas Blancas limestones at the easternmost end of the Guárico Belt, especially near the Urica fault trend, a Jurassic transfer fault repeatedly rejuvenated in the Cenozoic. Our own observations of this reported unconformity along the Piritú ridge, southwest of the Gulf of Barcelona suggest Late Miocene shale

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 36 -72° -71° -70° -69° -68° -67° -66° -65° -64° -63° -62° -61° -60° -59° 14° 14° Caribbean Went compressional Plate & Proto-Caribbean Seafloor during Palaeocene Prism

13° Sands spreading across seafloor 13° from axis of Caribbean foredeep Proto-Caribbean Inversion Zone Domain 2 is SW-NE trending trough, bounded 12° on north by Proto- 12° Proto-Caribbean Accretionary Prism Caribbean Domain 2 Caribbean Ridge, to south Foredeep Axis Base of Slope by shelf and slope bypass Chaudiere/Pt.-a-Pierre zones, open to NE BARCELONA TROUGH 11° 11° Outer Shelf Lizard Springs Caratas Tarouba Slope GUARICO TROUGH 10° Soldado 10° Misoa shelf probably flanked a Inner Shelf Slope bypass Shelf- trough axis to NW, with much and canyons Slope sand bypassing shelf, feeding Fluvial-deltaic? Break 9° early Cajones, Barbados prism 9°

ESPINO-MATURIN RIVER 100 km Early Eocene, 55 Ma 8° 8° -72° -71° -70° -69° -68° -67° -66° -65° -64° -63° -62° -61° -60° -59°

Fig. 16. A reconstruction for Earliest Eocene time (c. 55 Ma) shows the effect of the initiation of Proto-Caribbean underthrusting on the north side of the former passive margin. Clastic Domain 2 developed to the south of the uplifted Proto-Caribbean Ridge and associated accretionary prism, to the north of the former Cretaceous slope in the Central Range, and east of the Bohordal Fault, with sedimentary facies and thickness indicating substantially greater waters within the Barcelona Trough than in the Serranía to the southwest. At this time, the Caribbean Plate lay about 800 km to the west and was not influencing subsidence and sedimentation in eastern Venezuela and Trinidad. There is no indication of a northern source for any of the sediment in the Barcelona Trough. Instead, reconstructions indicate a point source in the embayment between the Bohordal Fault and Central Range slopes, which may have been fed by the “Espino-Maturín River” (which previously sourced the San Juan Sand Lobe). The area of the future northeast Serranía unconformity is shown dashed. At this time basement was probably uplifted immediately south of the Proto-Caribbean Inversion Zone, resulting in deposition of a thin and shallow Caratas section, with exposure and erosion only occurring during the latest Eocene to earliest Oligocene.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 37 diapirism and differential deformation of the mechanically rigid limestone and the underlying and more deformable terrigenous beds. Thus it remains unclear to us how much significance should be placed on this reported “Eocene angular unconformity”. The late Middle or early Late Eocene may then have been the time of initial tectonic emplacement of the overlying Los Cajones/Garrapata allochthonous nappe onto the Guárico strata, the subsequent shortening of which: (1) imbricated the Querecual, Mucaria, and Guárico Formations by perhaps 100 km (Perez de Armas 2005) with no Middle Eocene sections known at any of the Guárico imbricates; and 2) produced the autochthonous latest Eocene–Oligocene La Pascua/Roblecito flexural foredeep basin. The Late Eocene Orupé Formation conglomerates of numerous lithologies in red sandy matrix from the Caribbean Mountains is a heavily eroded molassic orogenic fringe, now known in local positions, small outliers, and remnant erosional blocks (Macsotay et al. 1995). In central Trinidad, Early–early Middle Eocene deep-water turbidite fans and channels of the glauconite-free Pointe-a-Pierre Formation are spatially associated with the shales and silts of the Chaudière Formation, which probably served as the basin floor on which the Pointe-a-Pierre channels and fans developed. This setting was situated north of the coeval pelagic/shaly southern shelf/upper slope margin of Trinidad (Lizard Springs and Navet Formations), and was deeper than that of the northern Vidoño Formation. Unlike the exposed Caratas Formation of the Serranía, the Caratas interval of the Maturín subsurface is gritty and lacks glauconite (unpublished oil industry data). Since sand could not have derived from southern Trinidad where coeval pelagic sediments and shale prevailed, the Maturín subsurface is the logical equivalent and updip source for the Pointe-a-Pierre sands of central Trinidad. These relationships support the existence of a Bohordal re-entrant as a palaeogeographic feature. One final note concerning the Caratas Formation (and equivalents in Trinidad) is that the complete lack of a Caribbean influence in the sandstone mineralogy argues against the proximity of the Caribbean Plate during the Eocene. The Caratas is not “flysch” in the orogenic sense as portrayed by James (2006), but rather records the reworking of sands along a north facing continental margin.

Diachronous Caribbean basal foredeep unconformity in the northeastern South American margin

In the Guárico (foreland) Basin south of Caracas and west of Urica Fault, a hiatus separates eroded Upper Cretaceous passive margin section from southwardly transgressive ?latest Eocene– Oligocene La Pascua sandstones, which shale upward into the Oligocene Roblecito Formation. In the Serranía Oriental and Maturín Basin to the east, all but two known sections have a regional hiatus separating the Maastrichtian San Juan, the Palaeocene Vidoño, and the Early to Middle Eocene Caratas Formations from the overlying southwardly transgressive Oligocene Los Jabillos/Merecure sandstones, which deepen up into the Areo and/or Carapita Formations (Caribbean foredeep). In the south flank of the Maturín subsurface, this basal foredeep onlap may be as young as Early Miocene (Di Croce et al. 1999). In southern Trinidad, a hiatus separates variably eroded/incised Campanian- early late Eocene Naparima Hill, Guayaguayare, Lizard Springs, and Navet Formations from the Early Oligocene San Fernando and the late Early Oligocene Basal Cipero Formations. Because of their overlap in age (Late Eocene is common to each), these hiatuses clearly correlate with each other, the main difference being the eastward-younging time of southward marine transgression. Rarely do these hiatuses display any discernable angularity. The foredeep onlap is clearly diachronous eastward and records the eastward relative migration of the Caribbean foredeep basin and plate along the South American margin (Dewey & Pindell 1986; Pindell & Kennan 2009; Pindell et al. 1988). These authors interpreted the underlying hiatus as being due to peripheral bulge uplift within a passive margin, which is typically about 5% of foredeep subsidence in continental crust (i.e. about 250 m for a 5 km foredeep deflection). However, the recognition that South America was probably an active rather than a passive margin before Caribbean arrival (Pindell et al. 1991, 1998, 2006) suggests a modification to this explanation for the unconformity. If South America were an active margin before the collision (i.e. the foot of the margin was either a site of inversion or incipient subduction), homoclinal hanging wall uplift of both the South American and Caribbean margins, as the Caribbean and South American plates converged above and loaded the intervening Proto-Caribbean lithosphere, could have generated greater uplift and

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 38 -72° -71° -70° -69° -68° -67° -66° -65° -64° -63° -62° -61° -60° -59° 14° 14° Caribbean Grenada, Tobago Tobago Scotlands deposited across Plate Basins Proto-Caribbean seafloor

13° Arc 13° Axis

Proto-Caribbean Ridge Proto-Caribbean Prism 12° Margarita 12° Caribbean Accretionary Prism Domain 2 Pt.-a-Pierre SW-derived sand BARCELONA TROUGH

11° 11° Villa de Cura Emerging? Garrapata, Cajones Navet ? Slope Uplifted former 10° Caribbean 10° Guarico Trough Lst fringe? ? Foredeep Axis Domain 3 Domain 4 Marly Slope bypass Inner Caratas sst, 9° and canyons Shelf 9° Mature SoAm detritus Tinajitas lst ESPINO- MATURIN RIVER 100 km Late Middle Eocene, 42 Ma 8° 8° -72° -71° -70° -69° -68° -67° -66° -65° -64° -63° -62° -61° -60° -59°

Fig. 17. A late Middle Eocene (c. 42 Ma) reconstruction shows a possible interpretation of the depositional context for the Scotlands of Barbados, if we accept Middle Eocene foraminfera and pollen in the Scotlands as non- reworked and honour the apparent absence of any Late Eocene fauna. The majority if not all of the Scotland sands were probably deposited about 600-700 km downstream of the onshore end of the Caribbean foredeep axis ahead of the leading edge of the Caribbean Plate. It is also possible that some of the Scotland strata were deposited to the north of the Proto-Caribbean Inversion Zone, particularly if the Proto-Caribbean Inversion Zone axis was overfilled, such that accommodation space was only available to the east along the Proto-Caribbean Trench. The presence of trace glaucophane as far as Tiburón Rise to the east of the present-day Barbados Prism shows that Caribbean derived debris may have been spreading over the entire Proto-Caribbean seafloor. A trench associated with a Proto- Caribbean Inversion Zone seems a logical place to expect accumulation of Caribbean-derived turbidites. In the southern Scotlands there is a coarser fraction (conglomerates containing what appear to be Rio Chavez shales and carbonates) that is clearly of South American provenance, and there is also abundant mature quartz sand together with Caribbean immature orogenic heavy minerals. Some of this South American component could come from the proposed northeastern Serranía unconformity. Alternatively, south-derived sediment from the western Guayana Shield may have been transported onto Proto-Caribbean Seafloor along the foredeep axis of the Caribbean. Carbon- ate and shale clasts may be derived in this case from South American margin rocks already incorporated into the leading edge of the Caribbean orogen in western Venezuela. The map also shows the Caribbean forebulge rolling from west to east. Associated uplift drives a diachronous tectonic lowstand, in this case uplifting the former Guárico Trough. Detritus uplifted off the forebulge may have been deposited to the northwest, in the northeastern Barcelona Trough, or to the south and transported to the Trinidad region by the proposed Espino-Maturín River.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 39 fluvial incision than a peripheral bulge within a passive margin (Pindell et al. 1991, 2006; Pindell & Kennan 2007a). Further, although hanging wall uplift would also be diachronous, it would be of greatest magnitude toward the trench, although an initially deep water setting may still have remained submarine, and it would predate peripheral bulge uplift because South America could not be flexed by the Caribbean Plate to produce a forebulge until the Caribbean lithosphere had actually begun to cross the position of the Proto-Caribbean Inversion Zone, by which time the hanging wall uplift would necessarily have been erased by sinking of the Proto-Caribbean slab into the mantle beneath it (Fig. 13). Thus, in cross section, the forebulge can be expected to emerge some distance landward from the zone of earlier hanging wall uplift. When the hanging wall uplift produces emergent conditions that are continuous with the interior of the continent, the forebulge will simply be superposed on and enhance a pre-existing unconformity, making incision deeper and non-marine conditions last longer. Where the hanging wall uplift has not exposed the depositional surface to erosion, such as when a marine continental shelf lies too far from the trench to be uplifted sufficiently, then the Caribbean forebulge may be the first and only mechanism of uplift at that shelf. This may have been the case in southern Trinidad, where the Caribbean basal foredeep unconformity spans the least time and incises the least section (San Fernando Formation on the Navet Hospital Hill member), and which lay some 300 km southeast of the probable trace of the Proto-Caribbean Inversion Zone, which is too distant for hanging wall uplift to have operated there. Due to the tectonic uplift involved with these mechanisms, a Type-1 unconformity is expected in most parts of the shelf margin (eastern Guárico Basin, Serranía Oriental and Trinidad), where the shoreline retreated to near or past the original shelf edge.

Lowstand wedges correlative with the Caribbean basal foredeep unconformity

With the development of the Caribbean basal foredeep unconformity, coeval lowstand channels and fans carrying South American detritus from the shield and marginal section can be expected outboard of the unconformity, and this detritus should carry material ranging in age from Precambrian to the depositional age. Obtaining the actual depositional age is therefore challenging, especially as so much of it is coarse, but can be achieved by (1) identifying the youngest zircons in sandstones (Algar et al. 1998); (2) bracketing the age of the detritus between adjacent datable shales or carbonates; and (3) dating intra-fan or intra-channel shales that record a true depositional age. Dating efforts thusfar point strongly to an earliest Oligocene depositional age for the bulk of the low-stand detritus. Along the Central Range trend of Trinidad, lowstand fan candidates include the Plaisance Conglomerate (mostly Cretaceous–Palaeogene South American margin clasts with coarse, quartz sand matrix; the 300 m thick Mount Harris section of the eastern Central Range; and the 250–300 m thick Angostura Sandstone (Early Oligocene according to BHP-Billiton 2003) in the eastern offshore projection of the Central Range. The Plaisance Conglomerate is traditionally considered a member of the San Fernando Formation (Kugler 1996, but see below). All three units lie entirely north of the Central Range Fault, and carry the distinctive red mudstone “cherries” discussed earlier for Cretaceous formations dating back to the Albian, and thus are South American derived, probably denoting direct fluvial communication with lateritic weathering sites on a coastal plain. In addition, many, but not all, of the ZTR quartz arenitic sandstone outcrops once mapped in the Central Range as Early Eocene Pointe-a-Pierre Formation (Saunders 1998) are probably earliest Oligocene equivalents of the Mount Harris, Plaisance and Angostura beds, if we accept the 31–33 Ma detrital zircon fission track ages from fresh euhedral crystals of probable volcanic ash origin as the depositional age (Algar et al. 1998). Of particular interest is the report of entirely Proterozoic U-Pb ages from detrital zircons (Kevin Meyer, oral presentation to 4th Geological Conference of the Geological Society of Trinidad and Tobago, 2007), which points strongly to a southerly South American derivation for this and all the noted similar units, whereas a provenance from hypothetical uplifted northern terranes (such as the model of Higgs 2006, 2009) should be dominated by Palaeozoic–Mesozoic ages. South of the Central Range Fault in Trinidad’s Central Range, the Plaisance Conglomerate does not occur, but the traditional definition of the San Fernando Formation would demand that the other three members (Glauconitic Sandstone, Glauconitic Siltstone, and the Vistabella Limestone members) be considered as possible low-stand candidates like the Plaisance. Additional low-stand candidates south of the Central Range Fault are the Marabella Conglomerate and the ?Late Eocene

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 40 “limestone debris” beds of Soldado Rock. In places these units can be shown to occupy incised valley positions; at Marabella, the Glauconitic Sandstone Member fills a trough that cuts through the Navet and into the Palaeocene Lizard Springs. Elsewhere in southern Trinidad the San Fernando, if present, may lie on rocks extending down to the Santonian. The Glauconitic Sandstone member of the San Fernando Formation contains Cassigerinella chipolensis (Jenkins 1964), indicating an Early Oligocene age (P18 biozone). This is the youngest determined age known to us for the San Fernando, but it could be a maximum age as some of the beds appear to be reworked. In addition, at least one locality shows that the San Fernando can be truncated by incision; at Marabella Village near San Fernando City, the Glauconitic Sandstone member is incised and overlain by the Marabella Conglomerate consisting of 95% rounded clasts of Naparima Hill argillites and 5% Maastrichtian sandstone, most less than 8 cm across, the latter of which is only known to the south and west (San Juan Formation, Eastern Venezuela). Where San Fernando members are absent, mainly in southern onshore Trinidad, the medial Oligocene Lower Cipero Formation is the first unit to be deposited above the Eocene–Late Cretaceous formations, such as on the Naparima Hill Formation in southeasternmost Trinidad. The Lower Cipero clearly belongs to the Caribbean foredeep basin, which continues up section through the Middle and Upper Cipero, until about 10 Ma (end Middle Miocene). Given the above, erosional incision is deepest and the duration of missing time is greatest (potentially the whole of the Late Eocene and some of the Oligocene, some 7 or 8 Ma) at the basal San Fernando unconformity. We consider this unconformity to be correlative with that of the Serranía Oriental, indicating regional incision of the Late Eocene shelf and upper slope. We view the three southern members of the San Fernando as post-incision transgressive valley fill and onlap (thus defining the basal Caribbean foredeep basin) following low-stand basinal deposition associated with the unconformity, as is suggested by the high glauconite content of the southern San Fernando (i.e. condensed section). This in turn suggests that the low-stand wedge of the Plaisance Conglomerate and its correlatives to the north of the Central Range Fault (Mount Harris Formation) are older than the southern San Fernando members and that the Plaisance (and Mount Harris) should not be considered members of the San Fernando Formation. This should not be surprising given that there is virtually no lithological similarity between these formations. In this interpretation, Caribbean foredeep onlap is marked by the southern members of the San Fernando Formation, which is transitional to the basal Cipero. The local occurrence of the Marabella Conglomerate, which cuts into the San Fernando, is probably a local submarine channel where scour and fill occurred in an otherwise onlapping and deepening setting, perhaps the last vestige of high-energy fluvial discharge as the foredeep migrated south. The proximity of the Marabella Conglomerate to the Saint Joseph Conglomerate (Early Palaeocene basal Lizard Springs deposited on incised Naparima Hill argillites, presumably from the nearby shoreline to the south) suggests the area was the site of repeated channel incision and filling. The following points highlight the reasons for removing the Plaisance Conglomerate as a member of the San Fernando Formation: 1. The Plaisance Conglomerate occurs only in the Central Range trend, north of the Central Range Fault, and there is no stratigraphic contact with the other members to the south. Our palinspastic reconstructions restore the Plaisance some 70 to 100 km to the WNW of its occurrences today, putting its westerly extent adjacent to the eastern flank of the reconstructed Serranía Oriental, and its entire occurrence within the deep water Bohordal re-entrant. Highly quartzose formations with ZTR heavy mineral assemblages (southern non-glauconitic Caratas, the San Juan, and possibly the Barranquín Formations) were exposed in the Serranía as sources for the quartz matrix in the Plaisance, and these are not known in the area of the Trinidadian portion of the basal foredeep unconformity. Further, a river may have bypassed the Maturín Basin when it was subaerial and delivered its load directly to the Bohordal re-entrant. 2. The Plaisance carries well-rounded limestone and sandstone boulders up to 50 cm across of Serranía-type passive margin strata dating down to the Albian; such incision in the Late Eocene is possible in the Serranía, but there is no indication of such incision in southern Trinidad. Employing the concept of the hanging wall uplift during Caribbean convergence, uplift and subaerial exposure in the Serranía may have been greatest in the northern and eastern Serranía (possibly eroding as deep as the Albian level), homoclinally decreasing southward (Pindell & Kennan 2001). It is possible, given today’s geological map pattern and the cooling history of the

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 41 Barranquín Formation in the Serranía, which suggests that uplift was greatest in the north and began in the Eocene (Locke & Garver 2005), that southward flowing streams carried such boulders to a Maturín trunk river which then carried them to the Bohordal re-entrant. 3. The southern San Fernando members do not contain red “cherries” or significant quartz as the Plaisance does, suggesting different source areas. Besides reworked limestone debris, the main clastic component of the southern members is granular glauconite, suggesting a “Vidoño-like” source area where the Caratas was never deposited, possibly along the Columbus Channel. In the absence of palinspastic reconstructions of shortening and strike-slip, this lithological difference has caused some workers (e.g. Senn 1940; Kugler 1953; Higgs 2009) to infer the existence of a Palaeogene northern subaerial clastic source area. In terms of heavy minerals, the San Fernando contains rare zircon and tourmaline and trace kyanite and staurolite, but no other signature minerals (Edelman & Doeglas 1934, unpublished oil industry report). Higgs (2006) proposed a northern source area (e.g. his Margarita Coastal Nappe) for Palaeogene Central Range clastics based on the presence of staurolite, but the chlorite, biotite and glaucophane also found in this “Margarita Coastal Nappe” are entirely absent in Trinidadian Palaeogene strata. Data from the Caroní River on the Guayana Shield (Wynn 1993) clearly shows that trace staurolite in some Palaeogene samples can be derived from the Guayana Shield. Furthermore, structural and geochronological data unequivocally demonstrate that the igneous and Palaeozoic basement elements of the HP/LT metamorphic terranes of Margarita had been juxtaposed by about Albian time, and were both intruded by c. 85 Ma plutons of the Caribbean Arc (Maresch et al. 2009); their HP/LT metamorphism and P–T–t paths are similar to other Caribbean HP/LT complexes associated with the Caribbean Arc. Thus we conclude that they lay far to the west in Palaeogene time, and that the Juan Griego rocks are not a viable source for staurolite in the Barcelona Trough. In addition, there are no tuffs in the South American Cretaceous that would be expected if Margarita and nearby island arc rocks (the Patao-Mejillones-Tobago High, to the north of Araya-Paria) were not completely allochthonous (Pindell 1993; Stöckhert et al. 1995). The red “cherries” in the Plaisance unit are indistinguishable (XRD results) from those in Aptian to Cenomanian Formations from southern Trinidad and thus it is most reasonable to accept a southern derivation for all post- Maastrichtian “cherries”. The reappearance of cherries in Early Oligocene sections is not evidence of a new orogen to the north of Trinidad. Our tests on dark clasts within the Maastrichtian Galera Formation, which we entertained as possible precursors to “cherries” developed on a northern Trinidad high, indicate that they do not contain sufficient iron to produce the cherries (W. Maresch pers. comm. 2007). Higgs also suggested that the Espino-Maturín River hypothesis was unviable because the coeval Caratas sands in the Serranía Oriental is highly glauconitic; however, this is not true of the Caratas in the Maturín subsurface (unpublished industry data), which is coarse-grained, gritty and mineralogically mature, similar in character to the Pointe-a-Pierre sands of Trinidad. To summarize the low-stand wedge for Trinidad, we propose a west to east progradation of fans in latest Eocene to Early Oligocene (Mount Harris Formation, with Plaisance, “Oligocene Pointe- a-Pierre”, Mount Harris sandstone, and Angostura sandstone members). Olistostromes of former outer passive margin (slope) material from the northern flank of the southern Trinidad shelf (drowned by Aptian time, Fig. 14) and slope may be expected to be interbedded with the primary west-to-east clastic trough fill, although from a different (southern) origin. The intensification of quartz-sand clastic input in the earliest Oligocene may relate to a coeval subaerial exposure of the Serranía Oriental surface, such that shield-derived sands that had been coursed onto a marine shelf now bypassed the Serranía platform and Maturín area to a delta setting (point source) at the southwest corner of the Bohordal re-entrant. In Eastern Venezuela, low-stand wedge candidates are limited to three depositional units; the Lechería beds (informal), the Lower Naricual Formation (Socas 1991), and the “reworked Tinajitas” seen in the Tinajitas Syncline at Via Alterna. All of these are in the vicinity of Barcelona, suggesting that this area may be the only part of the Serranía that remained below sea level during the low-stand. The Lechería unit (informal usage) is a structurally isolated occurrence of micaceous, thin to thick-bedded, fine to coarse-grained quartzose turbidites with large intra-formational rip up clasts/bedsets, interbedded with shales, and characterized by distinctive tabular to rounded red mudstone clasts (or “cherries”). It is remarkably similar to the Plaisance unit in Trinidad, but was mapped by Creole petroleum industry geologists in the 1950s as Oligocene–Early Miocene Naricual

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 42 Formation, despite great differences with that formation, because there was no other formation in the Serranía Oriental to which correlation could be made. The Lechería beds have a mature heavy mineral assemblage (ZTR with minor garnet, unlike the Naricual which is orogenic with basalt fragments), and an Oligocene fauna (T. King pers. comm. 2001). We infer an Early Oligocene age because the Lechería beds lack the lignite/coal and mineralogical immaturity typical of the Late Oligocene Naricual and younger formations nearby. Socas (1991) studied the Naricual Formation near Barcelona and proposed a new sub-unit, the Lower Naricual member of the Naricual Formation, comprising undated sandstones with a mature heavy mineral assemblage in the immediate area of the well-established “Upper Naricual”. The two sub-units are not in apparent stratigraphic contact. Lignite and coal are absent, diagenesis is better developed, and bedding plane dips are often greater in the Lower Naricual. In these respects, it is more tempting to correlate the Lower Naricual and Lechería, but we did not see the haematitic “cherries” in the outcrops of Lower Naricual we visited. At Via Alterna in the Tinajitas Syncline, the position of the Tinajitas limestone is occupied by limestone of a redeposited nature, comprising mass flow bed sets mainly of mobile rhodoliths (R. Higgs, for Tectonic Analysis Ltd, 2001). This occurrence of the Tinajitas differs from the more typically platformal and more probably in situ occurrences of the Peñas Blancas/Tinajitas limestone. Thus, although the “Tinajitas” here is in normal stratigraphic position, its age is not necessarily late Middle Eocene (top of Caratas Formation) as is the in situ Tinajitas. The Via Alterna Tinajitas beds may thus be (1) an unique version (carbonate only) of the lowstand wedge (latest Eocene or earliest Oligocene), (2) a basal bed of the overlying Los Jabillos Formation (base of onlapping Caribbean foredeep section), or (3) an end Middle Eocene debris flow derived directly from a Tinajitas or Peñas Blancas platform carbonate. We favour the last option; the lack of “cherries” and of coarse quartz hints against the lowstand under discussion herein, especially when such lithologies (Lechería) occur very nearby, and the lack of quartz argues against the Los Jabillos option. To the west, along the foothills of the eastern parts of the Caribbean Mountains, descriptions of the Early Oligocene Tememure Formation (Vivas & Macsotay 1995) are very similar to our observations of the Lechería beds. Although we have not studied the Tememure ourselves, we consider the two units to be equivalent and deposited along strike from each other, probably as lowstand fans, a proposal that O. Macsotay considered entirely feasible (O. Macsotay pers. comm. 2008). In summary for Eastern Venezuela, the Lechería beds north of Barcelona probably represent a low-stand wedge of Early Oligocene age. These beds are not seen farther inland, because the rest of the Serranía and Guárico Basin were subaerial at the time of deposition. Palinspastically, the Lechería beds lie along with the northern fringe of the Serranía WNW to the northwest part of the Gulf of Barcelona. Given that position, the source area for the Lechería beds could be the pre-La Pascua unconformity of the northeastern Guárico Basin just as well as the Serranía Oriental, although the apparent absence of staurolite or kyanite distinguishes it from the La Pascua Formation (see discussion of Domain 3, the Caribbean foredeep onlap fringe, below). The Lechería beds may give us a control point for the occurrence of latest Eocene–Early Oligocene lowstand wedges well west of Trinidad, hinting that similar wedges once flanked the northern fringe of the Serranía Oriental. Further, because such beds are absent at the Late Eocene hiatus in all the Serranía Oriental river sections, they also provide further evidence that the Serranía Oriental was subaerially exposed for at least some if not all of the Late Eocene. Sandstones containing ferruginous mudstone clasts (which we interpret as “cherries”) are also reported from a well east of the Serranía in the Venezuelan part of the Gulf of Paria (unpublished oil company data), probably providing another control point for this low-stand trend.

Domain 3: southern onlap assemblage of the Caribbean foredeep

The diachronous collision of the Caribbean Plate with northern South American involved first the accretion of the Caracas Group, Paria and Northern Range slope and rise strata followed by further convergence that led to the emplacement of allochthonous terranes and the accreted strata onto the South American margin (Fig. 18). In this paper, we will not explore the thrust history of this orogen in detail (see Pindell et al. 2009), treating it simply as “the allochthon”.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 43 -68 -67 -66 -65 -64 -63 -62 -61 -60 -59 14 14 Caribbean Barbados Possible Scotland-like sediment east of Barbados? Plate Active Grenada Arc Tobago Basin Trough Proto-Caribbean Prism 13 Barbados 13 Proto-Caribbean Ridge (today)

Caribbean Prism Extinct Outer Shelf Arc (testigos) Tobago REMNANT 12 ?? 12 BARCELONA Domain 2 TROUGH Northern Range Angostura (SW-derived) Mt. Harris Paria Olistoliths? (S-derived) Olistoliths? 11 11 Plaisance Slope V. de Cura Outer Shelf Lechería Plata Tememure Palaeoshape of Trinidad 10 Roblecito Shale Axial mixing Shelf- 10 Transgressive Max. NE-ward extent of sub-San Fer- Slope alluvial fringe Domain 2/Domain 3 Domain 4 nando u/c here could be a slightly Marly Break Boundary La Pascua Sand younger Caribbean forebulge effect Inner Shelf

9 Domain 3 Former Espino River uplifted on crest Fluvial- 9 Abandoned of Caribbean forebulge and abandoned, deltaic? Espino-Maturín with upstream drainage capture by River Espino Graben Caribbean foredeep to north 100 km Early Oligocene, 32 Ma 8 8 -68 -67 -66 -65 -64 -63 -62 -61 -60 -59

Fig. 18. A reconstruction for Early Oligocene time (c. 32 Ma) captures the end of sedimentation in Domain 2 (Barcelona Trough) and the onset of sedimentation at the leading edge of the southeast-migrating Caribbean foredeep in the Serranía. By earliest Oligocene time a significant area of subaerial exposure had developed in the northeast Serranía and sediment eroded from this surface was deposited directly over the Bohordal shelf edge into the Barcelona Trough, or transported via tributaries of the proposed “Espino-Maturín River” and discharged into the Trough at a point source upstream of the restored position of the Plaisance Conglomerate. The onset of Caribbean foredeep subsidence drowned this non-angular unconformity. In central Venezuela the onlap strata are assigned to the La Pascua Formation, younging south and east into the Los Jabillos Formation of the Serranía (Latest Eocene in the northwest to Late Oligocene in the El Furrial area). South of its onlap edge there must have been a subtle flexural forebulge developing which probably resulted in capture of much of the drainage of the Espino River into the foredeep. The Plaisance conglomerate and Angostura sandstones, which were abruptly buried by shales, represent the culmination of the Espino drainage and the uplift of its drainage basin on the crest of the forebulge and the ensuing shaly sedimentation is the eastern distal and entirely submarine equivalent to the onlap seen in the Serranía. As the forebulge crossed the trace of the “Espino-Maturín River” its drainage would have been captured by the Caribbean foredeep axis and mature south-derived sediment deflected to the northeast. We tentatively show the Lechería as comprising detritus from this captured drainage system. This map shows a tentative interpretation of the youngest of the Scotlands of Barbados, honouring a tentative Early Oligocene faunal age for the coarsest facies. In this case the Caribbean foredeep may have been channelling sediment comprising both Caribbean-derived and shield derived sediments, including mixed heavy minerals, distinctive red “cherries” and limestone and shale clasts towards Barbados at this time shortly before being caught in the vice between the waning Proto-Caribbean prism and the advancing Caribbean accretionary prism. Alternatively, if we accept only older Scotland ages, Caribbean foredeep sediment may have been accumulating only south of the Proto-Caribbean Accretionary Prism by this time.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 44 In central Venezuela, west of the Urica Fault, the Caribbean forearc basement (Villa de Cura Nappe) and its accretionary prism (e.g. Palaeocene–Middle Eocene Caramacate, Escorzonera, Los Cajones, Morro del Faro, and Garrapata units south of Villa de Cura) approached the Proto-Caribbean Prism (Caracas Group) in an offshore position that was outboard from the Guárico Trough behind the Proto-Caribbean Prism. Hanging wall uplift elevated the Proto-Caribbean Prism and Guárico Trough in the late Middle Eocene, eroding any Middle Eocene section. Caribbean advance then overthrust the Proto-Caribbean prism with peak metamorphism occurring at 40–35 Ma. The composite Caribbean forearc, accretionary prism and the subcreted Proto-Caribbean Prism, together advanced southeastward onto the Guárico Belt over South American basement in the Late Eocene, initiating foredeep subsidence in the Guárico Basin (La Pascua Formation). Further motion continued to shorten the Guárico Belt, below which the detachment was probably at the Albian level thus imbricating the Querecual, Mucaria, and Guárico Formations, as well as the Oligocene through Middle Miocene foredeep basin further south (Perez de Armas 2005). Domain 3 is represented in the Guárico Basin by the latest Eocene/Early Oligocene onlapping sands of the La Pascua Formation and by the overlying earliest Oligocene Lower Roblecito shales. The heavy mineral assemblage in the La Pascua is dominated by ZTR, with minor staurolite and kyanite, which are readily sourced from Palaeozoic or shield rocks of South America, or more likely recycled from the Cretaceous Canoa or Tigre Formations, over which it transgressed. There is no sign of the more labile heavy minerals seen in the underlying Cretaceous, such as glaucophane (reported from the Tigre Formation) or garnet (reported from the Gautier Formation). In the Serranía Oriental, the Domain 3 basal foredeep onlap is marked by the Early Oligocene Los Jabillos Formation inner shelf sandstones, which unconformably overlie the Caratas Formation and deepen upward into the Upper Oligocene Areo Formation shale. Our heavy mineral data for the Areo Formation (Fig. 7) shows ZTR and trace kyanite, but the Los Jabillos Formation shows only ZT, all south-derived, as this landward flank of the Caribbean foredeep was situated south of the foredeep axis. The transgression was diachronous to the south, reaching the Maturín subsurface in the earliest Miocene (Merecure Formation: di Croce et al. 1999; PDVSA 2005), advancing east at about the same 15–20 km/Ma rate as the Caribbean Plate. Foredeep subsidence following the passage of the forebulge may have been accompanied by instability on the southern flank of the basin. Thus, although we are proposing the Lechería beds comprise a low stand wedge deposited before the Los Jabillos foredeep transgression, they could instead post-date the Los Jabillos Formation, and be derived from the craton as the foredeep developed but before the Naricual orogenic fringe sediment had arrived from the north. The depositional instability may reflect thrust belt advance, rather than gravity-driven slumping during a lowstand. “Cherries”, which are found in strata as old as Albian (Gautiér Formation) and as young as Early Miocene (a single outcrop of Nariva Formation) certainly seem to be an indicator of southern provenance. However, the overall sedimentary character, clast content and heavy minerals of the Lechería are so like the Plaisance Conglomerate and other units of the Mount Harris Formation, as well as the Tememure Formation, that we prefer to equate them and assign a pre-Los Jabillos age. In southern Trinidad, the Late Oligocene “Lower Cipero”, “Arenaceous Cipero” (referring to Foraminifera) or “Silty Cipero” overlies the San Fernando Formation and rocks as old as the Naparima Hill argillites without angular unconformity. These strata appear to lie at the base of the Caribbean foredeep section, and locally overlie the San Fernando Formation incised channel fill, which we thus include the San Fernando at the base of the Domain 3 Caribbean foredeep section. The limited sand content of both the San Fernando and the Lower Cipero in most places reflects the lithology of the underlying Navet, Lizard Springs, and Upper Cretaceous shales and marls across which erosion, incision and subsequent foredeep transgression took place. Cipero facies include several shelly limestone “reefs”. As elsewhere, we would expect Domain 3 sediments in Trinidad to be mineralogically mature. We were unable to sample any sands from proven basal Cipero Formation, and one sand sample from the San Fernando which was either San Fernando Formation (most probable), or possibly overlying Lower Cipero, yielded only trace zircon and tourmaline. However, Edelman & Doeglas (1934) collected two samples from the Bamboo Silt member of the Lower Cipero from a location assigned to the P19 biozone (ex. G. ampliapertura biozone, latest Early Oligocene) by Kugler (2001). Their heavy mineral residues were dominated by ZTR, as expected, but the heavy mineral residues also contained about 5% chloritoid, 3% garnet, trace kyanite and a single

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 45 glaucophane. This signature is more like the Nariva of Domain 4 (see below). The sampled section is highly faulted, and observed bedding often contradicts the mapped faunal biozones and its stratigraphic age was the subject of debate by early workers (e.g. Stainforth 1948). Given the proven presence of chloritoid, kyanite and garnet on the Guayana Shield to the south, it is possible, though we consider it unlikely, that this is a unique occurrence of unweathered labile minerals in the Domain 3 transgressive fringe. Alternatively it is possible that the P19 fauna has been reworked into sediment of Nariva age, because the Bamboo Silt can be interpreted as stratigraphically overlying marls of the P21 (ex-G. opima opima biozone, early Nariva time). Given the local complexity of bedding and poor exposure we also cannot rule out the possibility that Nariva sand has been mixed with Cipero marl during faulting. Sand is not characteristic of the Bamboo Silt, which mostly comprises foetid-smelling clay and marl with abundant large foraminifera (Stainforth 1948). In summary, the loading of the South American crust by the advancing Caribbean forearc and growing orogen caused diachronous west-to-east migrating drowning and southward marine onlap of the pre-existing hanging wall and/or forebulge unconformity (Fig. 18). Clastic Domain 3 comprises the strata associated with this onlap, and its detritus is south-derived.

Domain 4: distal syn-orogenic detritus, axial feed along the Caribbean Foredeep

Domain 4 comprises the axial foredeep tract between the Domain 3 transgressive belt (distal foreland) to the south and the advancing Caribbean thrust belt and its associated proximal clastic fringe to the north (Fig. 18). Sandstone petrography and heavy mineral data support the view that the Guayana Shield and its sedimentary cover continued to contribute detritus to the migrating foredeep, probably off of a migrating forebulge ahead of the basin. However, Domain 4 also shows a significant contribution from arc volcanic, metamorphic, and ophiolitic terranes of the Caribbean Plate and/or of the collision zone with South America. The first consistent appearance of characteristic Caribbean orogenic detritus in autochthonous positions along the margin youngs from west to east, beginning in the Maastrichtian in the western parts of the Colón Formation in northern Colombia and in the Palaeocene–Early Eocene in the western parts of the Misoa Formation in the Maracaibo Basin (Van Andel 1958), continuing in the latest Early Oligocene Upper Roblecito of central Venezuela (Escalona 1985), in the Late Oligocene Naricual Formation of the Serranía Oriental and Late Oligocene–Early Miocene Nariva Formation of Trinidad, a diachroneity spanning some 40 million years and reflecting the diachronous emplacement of Caribbean allochthons onto the South American autochthon. In contrast to the typically mineralogically mature, shield-derived or reworked sediments of Domains 1– 3, those in Domain 4 include lithic fragments (chert, basalt, metavolcanic and metamorphic rocks) and diverse heavy minerals. Glaucophane, in particular, appears to be diagnostic of a Caribbean provenance from unroofed HP/LT terranes in the collision zone. It first appears in our study area in the Upper Roblecito and Chaguaramas Formations (Venezuela) and in the Nariva Formation (Trinidad). It is unlikely to survive recycling from the trace glaucophane reported from some, but not all, Late Cretaceous sediments by Escalona (1985). The reappearance of minerals such as garnet, staurolite, kyanite, chloritoid and apatite in Late Oligocene and younger sediments strongly suggests it derived from Class 1 and Class 5 terranes of the Caribbean Orogen (see also Bellizzia & Dengo 1990) and the associated parautochthonous rocks in the flanking thrust belts during collision. The lack of these minerals in Domain 3 strata also argues against a southern provenance for Domain 4 minerals.

Barbados as the downstream eastern continuation of Domain 4

The texturally mature, fine to coarse quartz sandstones and minor conglomerates of the Scotland beds of Barbados carry the signature glaucophane and high-grade continental metamorphic minerals of Domain 4. The age of the Scotland turbidites has been debated. Favouring a Palaeocene– Middle Eocene age are numerous arenaceous foraminifera (some reworked, some probably in situ, T. King pers. comm. 2008) in shales and pelagic radiolaria from hemipelagic beds (Speed 2002) and pollen analyses on 8 dispersed samples (Pindell & Frampton 2007). Favouring an Oligocene depositional age are Oligocene zircon fission track cooling ages (Baldwin et al. 1986) and two tentative Oligocene faunal ages (Pindell & Frampton 2007, citing T. King pers. comm. 2008) from conglomerate clasts collected by us below Chalky Mount on the east coast. Further, these same

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 46 conglomerates carry red “cherries” that are similar to, but less oxidized, than the “cherries” in Late Eocene–earliest Oligocene Plaisance and Lechería beds of Venezuela and Trinidad. Probably the strongest data are the radiolaria and pollen ages (Late Palaeocene–Middle Eocene), but the question remains open and it is possible that the Scotland unit has scattered Oligocene packets of sand/conglomerate within dominantly older material, either tectonically interleaved or deposited within incised channels. Indeed, our palaeogeographic maps show no specific reason to expect clastic supply to be shut off from the Scotlands, other than possible bathymetric building up of prism morphology, in Late Eocene time. Concerning depositional position, the Scotland material clearly did not bypass Eastern Venezuela or Trinidad, nor does it bear any similarity with the shale dominated, south-derived Guárico Formation of Central Venezuela other than age. Neither could the huge volumes of highly mature quartz derive from the Caribbean arc, where quartz is practically lacking. Thus, the Scotland turbidites were derived from western Venezuela and/or Colombia, which is the nearest position from which quartz sands with an orogenic influence could derive. The structural style of the Scotlands is that of an accretionary prism ahead of a trench (Speed 2002). In single-trench models for the Caribbean (e.g. Dewey & Pindell 1986; Speed 2002), and retracting the present day plate configuration back in time, the Scotlands represents the Caribbean Prism only. However, in double-trench models with Caribbean and Proto-Caribbean prisms (e.g. this paper), it becomes a question to which prism, or both, the Scotlands were accreted. Pindell & Frampton (2007) took this issue up, and we will not deal with it here because heavy minerals do not provide an answer; both prisms should have the orogenic influence, as they both flank the Proto- Caribbean basin. Suffice it to say that Caribbean orogenic material was carried into the Proto- Caribbean basin from northwest South America, where Domain 4 was first initiated in the Maastrichtian, by a significant river draining probably the Palaeogene Andean foredeep (Pindell et al. 1998), and was accreted at one or both of the accretionary prisms flanking the Proto-Caribbean basin. In the two prism model (this paper), a suture between the Caribbean and Proto-Caribbean prisms should exist somewhere in the greater Barbados Ridge today (Pindell & Frampton 2007).

Oligocene jump in Domain 4 deposition from Barbados towards Trinidad

Upon Early Oligocene prism-prism collision north of the Serranía Oriental (contrast Figs 18, 19), the Proto-Caribbean inter-prism trough was closed. Proto-Caribbean prism rocks and the Serranía Oriental Cretaceous passive margin were incorporated into an enlarged composite Caribbean Prism and thrust towards the southeast. As a result, the river that had formerly fed detritus down the Caribbean foredeep axis and into the Proto-Caribbean basin was deflected abruptly eastward into the Serranía Oriental (Naricual Formation of the northwestern and western Serranía) and Trinidad downstream (Nariva Formation), defining a new position for the palaeo-Orinoco River. Thus, further deposition near Barbados comprised reworking of the older prism materials (“Intermediate” or prism cover strata; Barker et al. 1986; Speed 2002) as the two prisms collided. Syndepositional structural growth in the northwestern Serranía resulted in slight angular unconformity between the Upper Naricual Formation deltaic sediments and the older strata of Domains 1–3. More regionally, the deflection of Caribbean foredeep axial drainage into the Serranía Oriental and Trinidad (Fig. 19) superimposed clastic Domain 4 sands of the Upper Naricual, Upper Areo, Carapita, and Nariva Formations) onto the earlier south-derived Domain 2 and 3 sediments, probably by downlap, or sidelap (as the Guayana Shield was an important sediment source), during west to east initial progradation (Pindell et al. 1998). Heavy mineral assemblages for all Domain 4 sediments in the Serranía and Trinidad show an unambiguous Caribbean orogen signature, including characteristic glaucophane, chloritoid, staurolite, garnet and Al–silicates.

A reinterpretation of the Nariva Formation of Trinidad

Most of the Nariva Formation in Trinidad comprises unctuous mud. Where exposed at the surface by tectonism, it is usually highly sheared and tightly folded, but in subsurface sections such deformation is far less pervasive. The general lack of planktonic foraminifera indicates turbid water conditions consistent with the muddy character, but the depth of deposition remains subject to

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 47 -68 -67 -66 -65 -64 -63 -62 -61 -60 -59 14 14 Edge of Scotland gives way to finer Aves Grenada Caribbean grained sedimentation Ridge Basin Crust Active Arc Proto-Caribbean Prism 13 Tobago 13 Trough E-verging thrust belt driven by Caribbean Plate over- N. limit of Domain 2 riding Proto-Carib. Ridge. at this time defined by yoked Proto-Caribbean Ridge and Caribbean 12 Blanquilla Basin Future Pirital thrust forebulges 12 Extinct Tobago Arc Caracolito Basin Domain 2 Margarita Slope Nariva 11 Backthrust belt 11 Urica F.UnderthrustAllochthon” by Carib. crust Domain 5 Nariva V. de Cura unroofing Outer Shelf Naricual Areo Lower Cipero Palaeoshape Guarico FTB of Trinidad 10 10 Domain 4 Merecure Shelf- Quebradón Slope Break Chaguaramus Domain 3 Marly Forebulge Espino Inner unconformity 9 Graben Shelf 9 in Oficina area Fluvial- deltaic 100 km Late Oligocene, 25 Ma 8 8 -68 -67 -66 -65 -64 -63 -62 -61 -60 -59

Fig. 19. Reconstruction for Late Oligocene time (c. 25 Ma). The leading edge of the Caribbean crust had migrated sufficiently far to the southeast to interact strongly with eastern Venezuela. East of the Urica Fault, rather than overthrusting the margin, the leading edge of Caribbean crust seems to have wedged between crystalline basement (below) and passive margin sediment (above) and started to drive shortening to the southeast in the passive margin section. The onset of shortening in the Serranía pushed the foredeep axis to the south far enough that the downstream end of the foredeep now drained into the former Domain 2 Barcelona Trough. This in turn led to the aban- donment of coarse-grained sedimentation in the Scotland of Barbados shortly before the accretion of the former Proto-Caribbean and advancing Caribbean prisms. We see important along-strike and across-strike facies variations in the Caribbean foredeep. Domain 3 south-derived mature sandstones are initially buried by axial foredeep (Domain 4) shale dominated sections, with minor sandstone content. The section becomes sandier approaching the active thrust front (Domain 5). To the west, in shallower marine to alluvial settings foredeep axis sediment tends to be sandier, and the transition from coarse to fine facies in the foredeep axis migrates from west to east as the trough fills in. In Trinidad, sandstone facies in the Nariva formation become more abundant in northern or western outcrops, and younger in the formation, relative to shale and marly facies developed towards the east end of the foredeep axis. These sandstones carry a distinctive heavy mineral signature indicating input from high-pressure metamorphic terranes that is absent from Domain 2 and Domain 3 sediments.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 48 considerable debate. Conventional literature (e.g. Kugler 2001) implies a thickness up to 1500 m and a 1000–2000 m deep basinal setting. However, the Nariva is intensely imbricated and our studies of seismic and well data suggest that depositional thickness is closer to 500 m. The turbid water conditions implied by the foraminifera could pertain to fairly shallow shelf or upper slope environments if sedimentation occurred in front of a strong fluvial influence, such as the Upper Naricual delta. Turbid waters appear to have been confined to the axis of the Nariva depocentre. To the southeast in Trinidad, turbidity waxed and waned, resulting in turbid Nariva facies interfingering with the open marine marly Middle Cipero facies of coeval age. If coastal currents along the Guayana margin were northwest-directed, as they are today, this may also have helped to maintain clearer water conditions in southern Trinidad, deflecting Nariva sands and muds towards the area southeast of present-day Barbados. Nariva-aged mudstones are of structural importance, serving as a major décollement surface in onshore central Trinidad, and in parts of the Barbados Prism, including the Trinidadian ultra-deep exploration blocks. The onset of Nariva Formation mudstone and sandstone deposition appears to be gradual rather than related to sudden, regionally synchronous, relative sea level change. Thus, we consider that the best definition for the base of the Nariva Formation is the first consistent appearance of signature heavy minerals such as chloritoid, glaucophane, garnet, kyanite, and staurolite in medial Oligocene time. “Trinidadian” Nariva deposition probably began in the Bohordal re-entrant of the Barcelona Trough in the middle Oligocene downstream from the Upper Naricual delta and Areo Formation shelf. By the Late Oligocene, sediment deposition and imbrication of Nariva and older strata at the leading edge of the advancing Caribbean orogen resulted in the Barcelona Trough beginning to shallow upward. The approach of the tectonic load and flexure of the crust to the southeast allowed Nariva facies to encroach diachronously southward into the clear-water Cipero marls of the Southern Basin. The growing orogen built a northern subaerial margin to the Naricual-Nariva depocentre, but no subaerially eroded conglomerates are known from the Nariva in the basin axis, suggesting that Nariva sands may largely be axially derived from the Naricual Delta. The first indications of north-derived detritus are in the Cunapo and Brasso Formation conglomerates of the Central Range which downlap onto and bury the Nariva (Kugler 2001; our own data). Claystones within the Brasso have foraminiferal assemblages characteristic of clear water (Wilson 2003, 2004), suggesting that the Brasso depocentre was a perched piggyback basin and that the turbid water mudstones of the axial Nariva derive from sediment plumes downstream of the Naricual delta. Nariva Formation sandstones form only a small proportion of the whole formation, and sandy sections are only locally more than a few tens of metres thick, such as in the Brighton Marine and Nariva Hill areas. The distinctive immature Caribbean orogen heavy mineral assemblage of the Nariva Formation is seen in: (1) white sands which in outcrop appear to be very clean, with little mica and no plant material; (2) in outcrop and well samples of “dirty” sands with lignite and some disseminated mica; (3) sheared and broken blocks of sandstone within sheared Nariva muds which have often been interpreted as olistostromes; and (4) highly sheared phacoids in the type section of the Charuma Silt member of the Pointe-a-Pierre Formation. Intense deformation and sediment remobilization mean it is often difficult to determine if multiple Nariva sands are individual beds or structural repetitions of the same bed. Outcrops of this formation are generally poor quality. Our field observations show some features not easily reconciled with the conventional deepwater view, such as trough-cross-bedding (Corbeau Hill), thin red claystone layers, and perfectly preserved leaves in extremely leaf-rich beds (Guaracara Road Quarry), perhaps suggesting a shelf-depth setting above storm wave base. Lignite beds up to a metre thick also occur, but no associated rootlets or palaeosols have been found, ruling out an in situ shoreface or lagoonal origin. Disseminated lignite and plant fragments are a feature of almost all Nariva sandstones and may have been transported as suspended load from the coaly Naricual Formation to the west. We have also found “cherries” in the Nariva at the Guaracara Road quarry and in the Naricual at Río Aragua Este near San Francisco, probably the most easterly sandy occurrence in the Naricual Formation sensu stricto in the Serranía Oriental. The Nariva Formation is conventionally considered to be olistostromal (e.g. Kugler 2001), with exotic blocks as old as Aptian occurring within Nariva muds. No pre-Nariva sandstones carry the “orogenic” signature, and there are no associated Nariva conglomerates (other than the aforementioned “cherries”), and none of the exotic blocks in the Nariva show evidence of rounding during transport. Thus, the “orogenic” sandstone blocks at the Williamsville “building site” and

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 49 elsewhere are almost certainly of intraformational origin, and we favour a tectonically sheared origin for the sand bed disaggregation, which is not seen in the subsurface away from fault zones. The shaly matrix at Williamsville is highly sheared and transposed to the point where original sedimentary laminations are only rarely seen. We have found outcropping “floating blocks” of pre-Nariva lithologies embedded in Nariva mudstone only in areas that are clearly intensely tectonically sheared. Exotic blocks are seen neither in unsheared Nariva outcrops, nor in well core. “Slip masses” of pre- Nariva lithologies (such as the limestones along Plum Road or at Stack Rock, Cuche sandstones, Naparima Hill argillites and Chaudière Formation shales) are widely reported in the literature and in older oil company reports and were used to infer an adjacent, uplifting and eroding northern highland. Our studies of well data, published maps (Kugler 1996) and the original field sheets on which these maps were based, and our own fieldwork show that many or all of the “olistostromes” were interpreted from the existence of exotic blocks at the surface only. Blocks are commonly embedded in Nariva- derived regolith and are much smaller than drawn on the maps. The blocks are often found downstream of in situ outcrops of older formations that could be the source of some of the reported “olistoliths”. Large, deep-rooted, mud volcanoes in central Trinidad associated with major thrusts and shear zones also carry blocks which were subsequently winnowed at the surface of the mud volcano and transported downstream of their site of “eruption”. True olistostromes may exist in the Nariva Formation but we have yet to find a convincing example either in the field or in a well where a sedimentary origin is visible through tectonic shearing overprint, or where shearing is absent. We are confident that we have identified several isolated Nariva “orogenic” sandstone blocks encased in a sheared, silty matrix from which Eocene benthonic foraminifera are reported (Charuma Silt type section along the Central Range Fault Zone, Kugler 1996, 2000). We have also recovered chloritoid grains from the matrix at this site, but have not been able to recover fauna. In this area, clastic facies of Eocene through Early Miocene age are fault- juxtaposed with multiple slivers of Eocene Navet Marl and Middle Miocene Biche limestone. Significant dextral strike-slip fault displacement has occurred on this fault zone in the Pliocene- Quaternary.

Domain 5: Proximal, syn-orogenic wedgetop and piggyback deposits on the evolving orogen

Domain 5 comprises syn-orogenic detritus subaerially eroded from the rising orogen and deposited on the northern, proximal fringe of the Caribbean foredeep basin and in piggyback basins within the growing orogen (Figs 20, 21). As with Domains 3 and 4, Domain 5 is diachronous, younging from west to east and north to south as turbidite and alluvial channels and fans propagated. Domain 5 strata were deposited in bathymetrically shallower conditions due to thrust imbrication and foredeep infilling beneath them. Clastic delivery was probably orthogonally downslope into the foredeep basin at first, but those that reached the basin axis became merged with Domain 4 strata, where basin floor currents may then have carried them along axis to the east for some unclear distance. In the western Guárico Basin north of El Baúl, Domain 5 is represented by the now deeply eroded Orupé Formation (Macsotay & Vivas 1995), a possibly Late Eocene (but also possibly younger) multi-component conglomerate in sandy matrix with abundant rounded chert, volcanic rocks, and metamorphic clasts. In the eastern Guárico Basin, Domain 5 is marked by the Upper Oligocene– Middle Miocene marine Quebradón Formation, but as the foredeep became filled it was taken over by the non-marine, eastwardly prograding and shallowing Chaguaramus Formation. Nearer to the Urica Fault, Domain 5 is marked by the Middle Miocene Quiamare Formation, especially its El Pilar Member in the north (Vivas & Macsotay 1989). In the southern flank of the Serranía Oriental, Domain 5 strata include the middle and upper parts (Middle Miocene) of the Carapita Formation, and the Chapopotal turbiditic member of the Carapita in the Maturín subsurface which forms a coarser reservoir than the surrounding Carapita shale (Lamb & Sulek 1968). The Morichito Basin above the Pirital thrust was once thought to be Middle Miocene, but now is considered Late Miocene (Roure et al. 2003) and is now assigned to Domain 6 (below).

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 50 -68 -67 -66 -65 -64 -63 -62 -61 -60 -59 14 14 Edge of CARIBBEAN Aves PLATE Caribbean Ridge Grenada crust Basin Tobago 13 13 25 Ma Active Trough 18 Ma Arc

12 Blanquilla 12 High Testigos High Foredeep Axis Extinct Tobago Arc

11 11 Cunapo Lr. Brasso End Nariva Slope Carapita Chapapotal Retrench Begin Upper Cipero 10 Domain 5 Domain 4 10 Guárico FTB Shelf- Margarita-Urica Outer Shelf Slope Axial drainage Transfer Zone Paleo-shape of Trinidad Domain 3 Break

9 Marly 9 Inner El Baúl Fluvial-deltaic Shelf Arch Caribbean 100 km Latest Early Miocene, 18 Ma Forebulge SOUTH AMERICAN PLATE 8 8 -68 -67 -66 -65 -64 -63 -62 -61 -60 -59

Fig 20. Reconstruction for Earliest Middle Miocene time (c. 18 Ma) shows the Caribbean Prism as having overridden the proto-Caribbean prism north of Tobago. This was followed by wedging of the crystalline Caribbean basement under the merged prisms, forming the characteristic “oceanics” blind roofthrust over backthrusted structural style seen from Tobago to Barbados. To the south, the Caribbean crust had accreted parts of the Proto- Caribbean Ridge in front of Tobago and was probably driving thin-skinned basement thrust sheets within the Serranía and Gulf of Paria resulting in a dramatic shoaling of the Serranía and Trinidad thrust belts, where unconformities developed in the north and former deep marine thrust belts (Nariva) were uplifted above imbricated thrust sheets of Cretaceous strata. The shallow marine Brasso Formation was deposited above the thrust belt, and was dominantly carbonate on structural highs and conglomeratic in structural lows, defined by transtensional lateral ramps. By this time, the Caribbean forebulge had swept through Trinidad sediments of Late Nariva age were deposited in central and southern Trinidad, while Brasso facies were deposited north of the thrust front in what is now the Central Range.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 51 In central Trinidad and in the southern Caroni Basin subsurface, Domain 5 strata include the Early and Middle Miocene Cunapo Conglomerate (alluvial and submarine fans), the late Early through Middle Miocene Brasso (shelf carbonates with clastic channels), and middle and late Middle Miocene Tamana (reefal and redeposited limestones) Formations, and the Retrench and Herrera sandstone fairways of the Cipero Formation. Black chert conglomerates and sandstones are characteristic of the Cunapo, Brasso and Tamana Formations. Downstream to the southeast, the Chapopotal, Retrench and Herrera sandstones and minor conglomerates reached the Middle Miocene foredeep axis and are interbedded with the younger elements of Domain 4. None of these units contain metamorphic clasts to our knowledge, as the Paria and Northern Range were not yet unroofed. The Middle Miocene Serranía-Naparima thrust belt of Eastern Venezuela and Trinidad was a continuous thrustbelt prior to the low-angle detachment faulting that has formed the Gulf of Paria basin since the Late Miocene. The black chert pebbles/grains of the Cunapo Conglomerate, Chapopotal turbidites, Brasso and Tamana channels, and Herrera turbidites derive from the San Antonio Formation in the Serranía Oriental. It may also have derived from a possible former Upper Cretaceous section above the Northern Range/northern Caroni Basin, but as the Northern Range/northern Caroni Basin restores palinspastically back to a position north of the Serranía Oriental in the Middle Miocene (Fig. 10), this point is moot. We concur with Erlich et al. (1993) that subaerial highlands to the northwest fed conglomerates and finer detritus to the Brasso/Tamana shelf, with bypass channels carrying some of this detritus farther south into the Upper Cipero foredeep basin, where they fine and become the characteristic “salt and pepper” Herrera turbidites, the “pepper” being largely chert, the “salt” largely quartz, plus other lithic grains. At least the northern Serranía Oriental (and hypothetical cover of Northern Range) had become subaerial in the Early Miocene, supplying the rounded Cretaceous chert pebbles in Early Miocene portions of the Cunapo Conglomerate. Further, thrust imbrication had caused littoral to neritic depositional depths by the late Early Miocene in Trinidad, as shown by the Brasso and Tamana Formations (Erlich et al. 1993). Accommodation space for these strata was locally provided by axis- parallel extensional structures (half grabens) cutting obliquely (ESE-trending) across the orogen (Pindell & Kennan 2007b). Tamana limestone thicknesses reach over 100 m yet much of its depositional environment is probably in less than 20 m water depth. Thus, Tamana depocentres must have subsided during the southeastward thrusting which otherwise produced uplift. The Tamana may thus be used to estimate positions where thinning of the thrust belt was achieved by low-angle detachments; this process is a function of the oblique collision and must happen to some degree in all highly oblique thrust belts. These low angle detachment basins crossing the thrust belt were also probably the means by which conglomeratic channels reached the southern foredeep basin. At the downstream ends of some of these extensional structures (e.g. Middle Miocene stage of extension on the Los Bajos Fault: Pindell & Kennan 2007b) there are unusually coarse Herrera facies (e.g. Herrera conglomerates elevated by diapiric mud in the core of the Southern Range anticline at Galfa Point). The heavy mineral signature of the Herrera sandstones is broadly similar to the slightly older Nariva Formation, but with a lower proportion of exotic minerals to ZTR. Although there was probably continued first cycle sediment input from far travelled Caribbean terranes at the back of the thrustbelt, it is possible that weathering and recycling of Nariva sandstones which were incorporated into the advancing thrust wedge resulted in some loss of labile minerals. A further cause for dilution may be capture of drainage on the Guayana Shield that had previously been directed south or east. The absence of slivers of Herrera-aged sandstones in the Nariva outcrop belt of central Trinidad strongly suggests that the Nariva had already been incorporated into the orogenic wedge by Herrera time, such that the Herrera turbidites mostly bypassed the Nariva Belt.

Domain 6: Orinoco Delta overlap assemblage

Oblique collision between the Caribbean Plate and South America culminated by about 10 Ma (Fig. 21) when the azimuth of Caribbean Plate motion relative to South America changed from ESE- directed to east-directed, after which east-west transcurrent (simple shear) tectonics took over (Pindell et al. 1998; Pindell & Kennan 2007a, b). However, foredeep loading continued in the Maturín and Columbus basins after 10 Ma, even in the absence of significant south-directed thrusting, because progressively thicker parts of the Caribbean lithosphere continued to move east, further loading the

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 52 -68 -67 -66 -65 -64 -63 -62 -61 -60 -59 14 14 CARIBBEAN PLATE Aves Ridge Grenada 18-12 Ma Basin

13 Tobago 12 Ma 13 West of MUTZ, South Caribbean Trough deformation Foldbelt takes up SoAm-Carib N-S Active front? convergence from 25-12 Ma once V. Arc de Cura reaches final position. Domain 5 Brasso, Herrera. Note that Nariva foldthrust 12 belt has now climbed up 12 above former Cretaceous Blanquilla shelf edge below Cent. Range. High Testigos High Margarita-Urica Tobago Transfer Zone (MUTZ) Extinct 11 Arc 11 Margarita Domain 4 Nariva buried Slope by Herrera, some sourced proximally from NW

Outer Shelf 10 Shelf- 10 Herrera Upper Cipero Slope Quiamare Break Carapita Domain 3 trans- Marly gression slows Inner 9 Shelf9 With 12 Ma transition to dominant strike- Fluvial-deltaic slip, the Orinoco Delta (Domain 6) will Caribbean prograde towards present day shelf edge. Forebulge 100100 km Late Middle Miocene, 12 Ma 8 8 -68 -67 -66 -65 -64 -63 -62 -61 -60 -59

Fig. 21. Reconstruction at the culmination of Middle Miocene orogeny in Venezuela and Trinidad at about 10-12 Ma shows substantial areas of the former foreland basin incorporated into the orogen. Proximal fine through coarse- grained clastic and carbonate facies accumulated on the north side of the foreland basin, deposited in shallower water conditions than in the axial trough to the south. Sands in Domain 5 were sourced from nearby pre-Miocene outcrops, including reworked foreland basin sediment. Significant east-west extension in the orogen is indicated by thick carbonate and clastic sediments deposited in waterdepths far less than the sediment thickness. True basement subsidence is indicated by continued foredeep subsidence beyond the deformation front, and additional accommodation space resulted from thinning of the allochthons. With the end of southeast-directed relative plate motion we see continued subsidence driven by distant loads of the Caribbean Plate but not in most areas by active continued foreland shortening. As a result, Domain 6 Orinoco sediments, although of the same origin as older foredeep axis sediments are able to overstep and bury the thrust front active during deposition of Domain 5 sediments.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 53 northwest-dipping, down-flexed South American lithosphere beneath the Maturín Basin and Trinidad (Pindell & Kennan 2007b). This subsiding basin has been fed primarily from west to east by the Orinoco River or precursors (which may have been located closer to the Caribbean Mountains and had a higher portion of sediment input from there) since about 10 Ma, with lesser contributions from north or south. Primary Late Miocene–Pliocene west to east shifts in palaeoenvironment were controlled by changes in tectonic loading in the north, eastward gravitational collapse of the orogen and eustasy (Wood 2000). The sands of the Orinoco trunk drainage system are highly variable, carrying detritus from multiple first-cycle and second-cycle sources including the Serranía Oriental and Caribbean Mountains to the north, El Baúl Arch and Mérida Andes in the west, the Guayana Shield to the south, and the cratonic cover of the interior plains. Domain 6 sediments post-date the primary accretion of the Caribbean Orogen to South America but remain strongly controlled by subsequent (post-10 Ma) dextral strike-slip, transtensional basin growth, and dextral transpression. Domain 6 sediments are particularly thick in the Gulf of Paria transtensional pull-apart basin, and in the Eastern Columbus Channel, where eastward-migrating gravity driven normal faulting provides sediment accommodation space. The influence of our proposed Proto-Caribbean Ridge is also clear up to the present. Thick distal Domain 6 sediments of Plio–Pleistocene age from the Orinoco are largely confined to the south of the Proto-Caribbean basement ridge ENE of Barbados (Speed et al. 1984; Dolan et al. 2004). The incorporation of these into the leading edge of the Barbados accretionary ridge has resulted in the prism being dramatically wider and thicker to the south of the Proto-Caribbean basement ridge than to the north. Our samples of Domain 6 strata are few and widespread but suggest some issues that should be the subject of future work. First, the Lower Manzanilla in Eastern Trinidad appears to lack an orogenic signature of chloritoid and other minerals seen in the Brasso. A nearby conglomerate, currently mapped as Brasso or Cunapo has a similar signature and is probably part of the Lower Manzanilla section. Second, sand beds at Plum Mitan Road citrus grove, currently mapped as Eocene Pointe-a-Pierre, have a clean ZTR assemblage but sedimentary features in common with younger strata and a possible Late Miocene foraminifer; we have assigned these sands to the Lower Manzanilla Formation. Third, the Upper Manzanilla shows minor staurolite input, and we have provisionally assigned undated sandstone at Mapapire Road to this or to the slightly younger Springvale Formation. The Middle Cruse Formation sandstones at Galfa Point have a hybrid signature consistent with their age. The Pleistocene Talparo Formation is distinct, with a stronger orogenic signature and reappearance of glaucophane, possibly signifying erosional downcutting to older levels in the Guárico/Maturín foreland to the west. We suspect that the cleaner samples represent the deposits of delta lobes which were sourced more directly from the southwest in the Guayana Shield, while more orogenic signatures represent the deposits of delta lobes which had a significant input from streams directly draining the Caribbean Allochthons.

SUMMARY

Heavy mineral and sandstone petrographic data have been applied to existing tectonic concepts and palaeogeographic models to refine our understanding of basinal and palaeogeographic evolution in Eastern Venezuela, Trinidad, Barbados, and the southeast Caribbean. The result is an internally consistent and stratigraphically dynamic regional evolutionary framework involving plate motions that can be used for understanding other aspects of evolution, as well as for extending or predicting elements of hydrocarbon plays in petroleum exploration. Six clastic domains of different genetic origin are identified, each of which pertains to a primary stage of palaeogeographic evolution. Domain 1, the Cretaceous passive margin, comprises clastic sediments, including distinctive red mudstone clasts, derived entirely from the Guayana Shield or its sedimentary cover (Precambrian and Palaeozoic “basement”, Mesozoic cover) on its south side. Intense tropical weathering results in mature ZTR assemblages, with local traces of staurolite, garnet and apatite, all of which are found in the shield to the south, or in present-day rivers crossing the shield. Red ironstone clasts (“cherries”) have been found in a number of Cretaceous formations and are thought to derive from laterites or oxidized palaeosols developed over older rocks or Cretaceous alluvial plain sediments to the south.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 54 Domain 2 comprises Palaeogene, largely redeposited sands, conglomerates, and shales of continental origin associated with basin formation behind the Proto-Caribbean Inversion Zone along northern South America. We propose the “Barcelona Trough” paralleling the margin from Central Venezuela to northeast of Trinidad as a clastic basin along northern South America that was distinct from the rest of the Proto-Caribbean basin, the two separated by the accretionary and basement hanging wall ridge of the north facing Proto-Caribbean Inversion Zone. Continental erosional detritus (reworking, and lowstand fans) produced by initial Caribbean interactions also fit within this domain, the interactions being (1) hanging wall uplift as the Caribbean and Proto-Caribbean prisms converged; and (2) forebulge uplift ahead of the migrating Caribbean foredeep basin. Domain 3 comprises the cratonic side or distal foreland of the migrating Caribbean foredeep basin. Sediments are reworked from existing cratonic cover and share the mature heavy mineral signature of Domains 1 and 2. There is little sign of exotic shield derived minerals, despite Caribbean forebulge rejuvenation of relief, pointing to the continuing efficiency of tropical weathering. Domain 4 occupies the axis of the migrating Caribbean foredeep basin, the sands being of dual provenance from both the Shield/cratonic cover to the south of the basin and also the developing orogenic topography to the north or northwest. These sands include the first sediments for which an explicit northern source, both mineralogically and geographically, can be justified, albeit reworked and mixed with south-derived sands along the basin axis. They carry a distinctive, immature heavy mineral assemblage including minerals derived from high-pressure blueschists and accreted Palaeozoic greenschist and amphibolite facies rocks. Sediments of Palaeocene to Middle Eocene (and possibly Oligocene) in Barbados fit into this domain, lying down dip (although on the ocean floor) of the Palaeogene Caribbean foredeep basin axis. During Late Oligocene time, accretion of the Serranía Oriental into the greater Caribbean accretionary belt closed off the Caribbean foredeep-Proto- Caribbean basin connection, deflecting the axis of the Caribbean foredeep southwards into the Serranía Oriental and Trinidad. These Caribbean foredeep sediments are superimposed (by downlap or sidelap) on the mature sandstones of Domain 3 in the Serranía Oriental and Trinidad. Domain 5 consists of proximal, syn-orogenic detritus on the advancing orogenic wedge and in piggyback basins, derived from the uplifting and diachronous Caribbean-South America collision zone. These strata are generally coarse grained immediately adjacent to the subaerial part of the orogen, but tend to fine to the southeast. They carry first-cycle detritus from the Caribbean Orogen and also probably reworked Domain 4 sediments caught up in the thrust belt, but these are diluted with reworked debris from mature sediments within the orogen. Clasts of organic rich chert from uplifted Cretaceous strata are common. As with Domains 3 and 4, these sediments are diachronous from west to east, but are not younger than c. 10 Ma. The olistostromal Early Eocene turbidites of the Los Cajones and Garrapata Formations of Central Venezuela can be considered in Domain 5 but at a time when the Caribbean forearc had not yet collided with the central Venezuelan autochthon. Domain 6, post-collisional, but syn-transcurrent, foreland fill of the Orinoco fluvial, deltaic, and pro-deltaic settings, prograding from west to east. To the south, these sediments merge with sediments from minor rivers draining the Guayana Shield, and towards the north they are confined by and contributed to by the Caribbean Mountains, the Serranía Oriental, the Paria Peninsula and the Northern Range of Trinidad. Domain 6 strata show a very mixed provenance, with variable but continued input from the Caribbean Orogen or recycled Domain 4 and 5 sediments upstream, together with possible first cycle material coming from the Shield. Domains 1–3 are mature and derived from South America, while Domains 4–6 show affinity to lithologies of the allochthonous Caribbean Plate and its accreted terranes. The transition in Eastern Venezuela and Trinidad occurred in the Oligocene, marking the onset of Caribbean collision with South America in that position. This transition in northern Colombia and western Venezuela occurs in Palaeocene, attesting to the diachronous nature of the Caribbean-South America collision, which is itself a consequence of the Pacific origin of the Caribbean Plate. However, in the mantle reference frame, the relative motion is due to the westward drift of the Americas relative to a Caribbean Plate that has undergone little west to east motion. The correct interpretation of the palaeogeography of all these domains depends critically on robust palinspastic restorations, with well constrained estimates of shortening and strike-slip offsets derived from interpretations of regional seismic, well, and field datasets. Palaeogeographies built from facies distribution in present-day coordinates are misleading at best, and may lead to highly erroneous

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 55 predictions of sand fairway orientation and fining direction. In some cases, the misinterpretation of structural mixing as olistostromal deposition may lead to the expectation of non-existent proximal facies (and land masses). The palaeogeographic development outlined in this paper could be greatly improved with additional dating of sediments using fauna or direct dating of detrital grains. Future combined studies of the thermal history of grain populations in the sediments and the multiple source areas for characteristic heavy minerals may reduce the ambiguity in current interpretations.

We cheerfully thank John Frampton and Barry Carr-Brown (Biostratigraphic Associates, Trinidad), John Keens-Dumas and Chris Lakhan (Petrotrin), T. King and Maria Bolivar (Biostratigraphic Associates, Venezuela) for providing palaeontological age data on field and core samples over many years; Walter Maresch and Thomas Reinecke (Universität Bochum, Germany) for XRD lab work on siltstone “cherries”; and Maria Mange for second opinions on certain heavy mineral identifications. We also are indebted to Marco Antonio Navas, Raj Maraj, Kelly Latter, Leslie Barker, and Roger Higgs for field assistance in Venezuela, Trinidad, or Barbados during this work, and to Tony Ramlackhansingh for discussion of the structural context of many of the samples. Salomon Kroonenberg provided valuable insights into the geology of the Suriname portion of the Guayana Shield. This paper derives from an industry funded research program by Tectonic Analysis Ltd. that was supported in stages from 2000–2007 by BP, BHP-Billiton, Chevron, Total, Anadarko/Kerr, Primera, Marathon, Repsol, Petro-Canada, TED, Phillips, Hess, Petrotrin, ENI, Shell, Venture, Talisman and Tullow, for which we are grateful, as well as from funds provided by the BOLIVAR study program (NSF grant EAR-0003572) at Rice University. We are particularly indebted to Petrotrin for providing a comprehensive seismic and well database and other files from which we built the palinspastic reconstructions so important for the palaeogeographic reconstructions presented in this paper. Keith James and Maria Antonieta Lorente organized the Sigüenza conference in 2006 which inspired us to write this paper.

References

Algar, S. 1993. Structure, Stratigraphy, and Thermo-chronologic Evolution of Trinidad. PhD thesis, Dartmouth College. Algar, S. T. & Pindell, J. L. 1993. Structure and deformation history of the northern range of Trinidad and adjacent areas. Tectonics, 12, 814-829. Algar, S., Heady, E. C. & Pindell, J. L. 1998. Fission-track dating in Trinidad: Implications for provenance, depositional timing and tectonic uplift. In: Pindell, J. L. & Drake, C. (eds) Paleogeographic Evolution and Non-glacial Eustacy: North America. SEPM (Society for Sedimentary Geology), Special Publication, 58, 111-128. Anand, R. R. & Gilkes, R. J. 1987. Muscovite in Darling Range bauxitic laterite. Australian Journal of Soil Research, 25, 445-450. Bellizzia, A. 1985. Sistema montanosa del Caribe – una Cordillera aloctona en la parte norte de America del Sur. Memorias del VI Congreso Geologico Venezolano. Sociedad Venezolana de Geologia, Caracas, 10, 6657-6835. Bellizzia, A. & Dengo, G. 1990. Caribbean mountain system, northern South America. In: Dengo, G. & Case, J. E. (eds) The Caribbean Region. Geological Society of America, The Geology of North America, H, 167-176. Bellizzia, A. & González, L. A. 1971. Geological Map of San Juan de los Morros (Sheet 6745, Scale 1:100,000, First Edition). Ministerio de Energia y Minas, Caracas, Venezuela. Benjamini, C., Shagam, R. & Menéndez, A. 1986. Formación Tucutunemo. Memorias del VI Congreso Geologico Venezolano. Sociedad Venezolana de Geologia, Caracas, 10, 6551-6574. Benjamini, C., Shagam, R. & Menéndez V, A. 1987. (Late?) Paleozoic age for the "Cretaceous" Tucutunemo Formation, northern Venezuela; stratigraphic and tectonic implications. Geology, 15, 922-926. BHP-Billiton 2003. Briefing Paper – Angostura Development, Trinidad and Tobago. World Wide Web Address: http://www.bhpbilliton.com/bbContentRepository/News/RelatedContent/Angostura120303.pdf Carr-Brown, B. & Frampton, J. 1979. An outline of the stratigraphy of Trinidad, Transactions of the 4th Latin American Geological Congress, 7th–15th July 1979, Port of Spain, Trinidad, 7-19. Cole E. & Heesterman, L. J. 2002. Geological map of Guyana (Scale: 1:3 500 000). Guyana Geology and Mines Commission, Georgetown, Guyana. Cox, D. P., Wynn, J. C., Sidder, G. B. & Page, N. J. 1993. Geology of the Venezuelan Guayana Shield. In: Geology and Mineral Resource Assessment of the Venezuelan Guayana Shield, USGS Bulletin, 2062, 9-17. Delor, C., Lahondère, D., Egal, E., Lafon, J. M., Cocherie, A., Guerrot, C., Rossi, P., Truffert, C., Theveniaut, H., Phillips, P. & Avelar, G. V. D. 2003a. Transamazonian crustal growth and reworking as revealed by the 1:500 000 scale geological map of French Guiana (2nd edition). Géologie de la France/Geology of France and surrounding areas. 2-3-4, 5-57. Delor, C., de Roever, E. W. F., Lafon, J. M., Lahondère, D., Rossi, P., Cocherie, A., Guerrot, C. & Potrel, A. 2003b. The Bakhuis ultrahigh-temperature granulite belt: II. implications for late Transamazonian crustal stretching in a revised Guiana Shield framework. Géologie de la France/Geology of France and surrounding areas. 2-3-4, 207-230. Dewey, J. F. & Pindell, J. L. 1986. Reply to comment on "Neogene block tectonics of Eastern Turkey and northern South America: continental applications of the finite difference method". Tectonics, 5, 703-705. Dewey, J. F. & Pindell, J. L. 2006. Tectonic significance of the Caribbean Plate as a reference frame since 100 Ma: the “Backbone” explained, “Backbone of the Americas”, 3rd–7th April 2006, Mendoza, Argentina. Geological Society of

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 56 America Abstracts with Programs, Speciality Meeting No. 2, 118. World Wide Web Address: http://gsa.confex.com/gsa/06boa/finalprogram/abstract_100635.htm Dolan, P., Burggraf, D., Soofi, K., Fitzsimmons, R., Aydemir, E., Senneseth, O. & Strickland, L. 2004. Challenges to Exploration in Frontier Basins – The Barbados Accretionary Prism. American Association of Petroleum Geologists International Conference, 24th–27th October 2004, Cancun, Mexico. World Wide Web Address: http://aapg.confex.com/aapg/can2004/techprogram/A90080.htm D.O.S. 1983. Geology of Barbados (Geological Map, Scale: 1:50,000). Directorate of Overseas Surveys, UK. , D.O.S. Series, 1229. Di Croce, J. 1989. Analisis Sedimentologico de la Formacion San Juan en la Cuenca Oriental de Venezuela (Estados Anzoategui y Monagas). M.Sc. thesis, Universidad Central de Venezuela. Di Croce, J., Bally, A. W. & Vail, P. 1999. Sequence stratigraphy of the Eastern Venezuelan Basin. In: Mann, P. (ed.) Caribbean Basins. Elsevier Science, Amsterdam, Sedimentary Basins of the World, 4, 419-476. Edelman, C. H. & Doeglas, D. J. 1934. Report on the heavy minerals of some Cretaceous and Eocene sediments from Trinidad, unpublished oil industry report held in Petrotrin library, Santa Flora. Erikson, J. P. & Pindell, J. L. 1998. Sequence stratigraphy and relative sea-level history of the Cretaceous to Eocene passive margin of northeastern Venezuela and the possible tectonic and eustatic causes of stratigraphic development. In: Pindell, J. L. & Drake, C. (eds) Paleogeographic Evolution and Non-glacial Eustacy: North America. SEPM (Society for Sedimentary Geology), Special Publication, 58, 261-281. Erlich, R. N. & Keens-Dumas, J. 2007. Late Cretaceous Palaeogeography of northeastern South America: implications for source and reservoir development. In: Proceedings of the 4th Geological Society of Trinidad and Tobago Geological Conference – "Caribbean Exploration – Planning for the next century", 17th–22nd June 2007, Port of Spain, Trinidad. World Wide Web Address: http://www.gstt.org Erlich, R. N., Farfan, P. F. & Hallock, P. 1993. Biostratigraphy, depositional environments, and diagenesis of the Tamana Formation, Trinidad: A tectonic marker horizon. Sedimentology, 40, (4), 743-768. Erlich, R. N., Villamil, T. & Keens-Dumas, J. 2003. Controls on the deposition of Upper Cretaceous organic carbon–rich rocks from Costa Rica to Suriname. In: Bartolini, C., Buffler, R. T. & Blickwede, J. (eds) The Circum-Gulf of Mexico and the Caribbean: Hydrocarbon habitats, basin formation, and plate tectonics. American Association of Petroleum Geologists, Memoir, 79, 1-45. Escalona, N. 1985. Relaciones estratigráficas con el método de minerales pesados, Área Machete, Faja Petrolífera del Orinoco. Memorias del VI Congreso Geologico Venezolano. Sociedad Venezolana de Geologia, Caracas, 1, 536-587. Feo-Codecido, G. 1955. Heavy mineral techniques and their application to Venezuelan stratigraphy. American Association of Petroleum Geologists Bulletin, 40, 984-1000. Foland, K. A., Speed, R. & Weber, J. 1992. Geochronologic studies of the hinterland of the Caribbean Mountains Orogen of Venezuela and Trinidad. Anonymous. Geological Society of America Abstracts with Programs, 24, (7), A148. Galehouse, J.S. 1971. Point counting. In: Carver, R.E. (ed.) Procedures in Sedimentary Petrology. Wiley Interscience. 385- 407. Gonzalez de Juana, C., Iturralde de Arozena, J. M. & Picard Cadillat, X. 1980. Geologia de Venezuela y de sus Cuencas Petroliferas. Ediciones FONINVES, Caracas. Hackley, P.C., Urbani, F., Karlsen, A.W. & Garrity, P. 2005. Geologic shaded relief map of Venezuela. United States Geological Survey Open File Report 2005-1038. Hedberg, H. D. 1950. Geology of the eastern Venezuela basin (Anzoátegui, Monagas, Sucre, eastern Guárico portion). Geological Society of America Bulletin, 61, 1173-1216. Hedberg, H. D., Sass, L. C. & Funkhouser, H. J. 1947. Oil fields of Greater Oficina area, central Anzoategui, Venezuela. American Association of Petroleum Geologists Bulletin, 31, 2089-2169. Higgs, R. 2006. Venezuela–Trinidad–NW Colombia geohistory revised: long-lived (Campanian–Quaternary) nappe-thrust belt and stacked foreland basins (Proto-Caribbean, Caribbean ), capped by supraorogen salt-dissolution basins (11–0 Ma), Extended abstract, XIII Congreso Venezolano de Geofisica, 22nd–25th October 2006, Caracas, Venezuela. Extended abstract available online at http://www.congresogeofisica-sovg.org/dyncat.cFormation?catid=4378 Higgs. R. 2009. Caribbean–South America oblique collision model revised (Campanian–Recent) on a Jurassic–Cretaceous passive margin. In: James, K., Lorente, M. A. & Pindell, J. L. (eds) The geology and evolution of the region between North and South America. Geological Society, London, Special Publications, this volume. Hung, E. J. 1997. Foredeep and thrust belt interpretation of the Maturin sub-basin, Eastern Venezuelan basin. Master's thesis, Rice University. Illing, V. C. 1928. Geology of the Naparima region of Trinidad. Geological Society of London Quarterly Journal, 84, 1-56. James, K. 2006. Arguments for and against the Pacific origin of the Caribbean Plate: discussion, finding for an inter- American origin. Geologica Acta, 4, 279-302. Jenkins, D. G. 1964. Panama and Trinidad Oligocene rocks. Journal of Palaeontology, 38, 606. Kamen-Kaye, M. 1937. Reconnaissance Geology in State of Anzoategui, Venezuela, South America. American Association of Petroleum Geologists Bulletin, 21, 233-245.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 57 Kamen-Kaye, M. 1942. "Ortiz Sandstone" and "Guarumen Sandstone Group" of North-Central Venezuela. American Association of Petroleum Geologists Bulletin, 26, 126-133. Kasper, D.A. & Larue, D.K. 1986. Palaeogene quartzose sandstones of Barbados and palaeogeography of northern South America. Tectonics, 5, 837-854. Kennan, L. & Pindell, J. L. 2009. Dextral shear, terrane accretion and basin formation in the Northern Andes: best explained by interaction with a Pacific-derived Caribbean Plate. In: James, K., Lorente, M. A. & Pindell, J. L. (eds) The geology and evolution of the region between North and South America. Geological Society, London, Special Publications, this volume. Kugler, H. G. 1953. Jurassic to recent sedimentary environments in Trinidad. Bulletin de l’Association Suisse des Géologues et l’Ingénieurs du Petrole, 20, 27-60. Kugler, H. G. 1996, Treatise on the Geology of Trinidad, detailed geological maps and sections, Natural History Museum of Basel. Kugler, H. G. 2001, Treatise on the Geology of Trinidad, part 4: The Paleocene to Holocene Formations of Trinidad, Knappertsbusch, M and Bolli, H. M. (eds), Natural History Museum of Basel. Lamb, J. L. & Sulek, J. A. 1968. Miocene turbidites in the Carapita Formation of eastern Venezuela. In: Saunders, J. B. (ed.) Transactions of the Fourth Caribbean Geological Conference, 28th March–12th April 1965, Port-of-Spain, Trinidad and Tobago. Caribbean Printers, Arima, Trinidad, 111-119. Larue, D. K. 1995. Non-conventional methods of determining clastic reservoir orientation using paleocurrent/paleoslope studies in oriented core–application to subsurface of Trinidad, In: Proceedings of the 3rd Geological Conference of the Geological Society of Trinidad and Tobago, 777-813. Lebron, M. C. & Perfit, M. R. 1993. Stratigraphic and petrochemical data support subduction polarity reversal of the Cretaceous Caribbean island arc. Journal of Geology, 101, 389-396. Locke, B. & Garver, J. 2005. Thermal evolution of the eastern Serranía del Interior foreland fold and thrust belt, northeastern Venezuela, based on apatite fission track analyses. In: Avé Lallemant, H.G., Sisson, V.B. (eds) Caribbean–South American plate interactions, Venezuela. Geological Society of America, Special Paper, 394, 315-328. Macsotay, O., Vivas, V. & Moticska, P. 1995. Lithostratigraphy of the Piemontine nappe of north central Venezuela: Senonian to Eocene gravitational sedimentation. Boletin de Geológia del Minsterio de Energia y Minas, Publicación Especial, 10, 114-123. Mange, M. A. & Maurer, H. F. W. 1992. Heavy Minerals in Colour. Chapman and Hall, London. Maresch, W. V., Kluge, R., Baumann, A., Krückhans-Lueder, G., Pindell, J., Stanek, K. & Stöckhert, B. 2009. The occurrence and timing of high-pressure metamorphism on Margarita Island, Venezuela: A constraint on Caribbean–South America interaction. In: James, K., Lorente, M. A. & Pindell, J. L. (eds) The geology and evolution of the region between North and South America. Geological Society, London, Special Publications, this volume. Mascle, A., Moore, J. C, Taylor, E., Alvarez, F., Andreieff, P., Barnes, R., Beck, C., Behrmann, J., Blanc, G., Brown, K., Clark, M., Dolan, J., Fisher, A., Gieskes, J., Hounslow, M., McLellan, P., Moran, K., Ogawa, Y., Sakai, T., Schoonmaker, J., Vrolijk, P. J., Wilkens, R. & Williams, C. 1988a. Site 671 – Barbados Ridge. Proceedings of the Ocean Drilling Program, Initial Report, 110, 67-204. Mascle, A., Moore, J. C, Taylor, E., Alvarez, F., Andreieff, P., Barnes, R., Beck, C., Behrmann, J., Blanc, G., Brown, K., Clark, M., Dolan, J., Fisher, A., Gieskes, J., Hounslow, M., McLellan, P., Moran, K., Ogawa, Y., Sakai, T., Schoonmaker, J., Vrolijk, P. J., Wilkens, R. & Williams, C. 1988b. Site 672 – Barbados Ridge. Proceedings of the Ocean Drilling Program, Initial Report, 110, 205-310. Maya-Sánchez, M. 2001. Distribución, Facies y Edad de las Rocas Metamórficas en Colombia, Memoria Explicitiva: Mapa metamorfico de Colombia. Instituto Colombiano de Minería y Geologîa (INGEOMINAS), Bogota. World Wide Web Address: http://productos.ingeominas.gov.co/productos/MEMORIA/Memoria%20MMC.pdf Maya-Sánchez, M. & Vásquez-Arroyave, E. 2001. Mapa metamorfico de Colombia (Scale: 1:2 000 000). Instituto Colombiano de Minería y Geologîa (INGEOMINAS), Bogota. World Wide Web Address: http://productos.ingeominas.gov.co/productos/OFICIAL/georecon/geologia/escmilln/pdf/Metamorfico.pdf Means, W. D. 1989. Stretching faults. Geology, 17, 893-896. Müller, R. D., Royer, J.-Y., Cande, S. C., Roest, W. R. & Maschenkov, S. 1999. New constraints on the Late Cretaceous/Tertiary plate tectonic evolution of the Caribbean. In: Mann, P. (ed.) Caribbean Basins. Elsevier Science, Amsterdam. Sedimentary Basins of the World, 4, 33-59. Oxburgh, E. R. 1966. Geology and metamorphism of Cretaceous rocks in eastern Carabobo state, Venezuelan coast ranges. Caribbean Geological Investigations. Geological Society of America, Memoir, 98, 241-310. PDVSA 2005. Codigo Estratigráfico de las Cuencas Petroleras de Venezuela. World Wide Web Address: http://www.pdvsa.com/lexico/index.html Peirson, A.L., Salvador, A. & Stainforth, R. M. 1966. The Guárico formation of north-central Venezuela. Boletin Informativo – Asociacion Venezolana de Geologia, Mineria y Petroleo, 9, 181-224. Perez de Armas, J. 2005. Tectonic and thermal history of the western Serranía del Interior foreland fold and thrust belt and Guárico Basin, north-central Venezuela: implications of new apatite fission track analysis and seismic interpretation. In: Avé Lallemant, H.G., Sisson, V.B. (eds) Caribbean–South American plate interactions, Venezuela. Geological Society of America, Special Paper, 394, 271-314.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 58 Pindell, J. L. 1985. Alleghanian reconstruction and subsequent evolution of the Gulf of Mexico, Bahamas, and Proto- Caribbean. Tectonics, 4, 1-39. Pindell, J. L. 1993. Regional synopsis of Gulf of Mexico and Caribbean evolution. In: Pindell, J. L. & Perkins, R. F. (eds). Transactions of the 13th Annual GCSSEPM Research Conference: Mesozoic and Early Cenozoic Development of the Gulf of Mexico and Caribbean Region. 251-274. Pindell, J. L. & Barrett, S. F. 1990. Geological evolution of the Caribbean region; a plate tectonic perspective. In: Dengo, G. & Case, J. E. (eds) The Caribbean Region. Geological Society of America, The Geology of North America, H, 405-432. Pindell, J. L. & Kennan, L. 2001. Processes and events in the terrane assembly of Trinidad and eastern Venezuela. In: Fillon, R. H., Rosen, N. C., Weimer, P., Lowrie, A., Pettingill, H., Phair, R. H., Roberts, H. H. & Van Hoorn, B. (eds) Transactions of the 21st GCSSEPM Annual Bob F. Perkins Research Conference: Petroleum Systems of Deep-Water Basins, 159-192. Pindell, J. L. & Frampton, J. 2007. Geology and Tectonic Evolution, Scotland District, Barbados, field trip guide. Proceedings of the 4th Geological Society of Trinidad and Tobago Geological Conference – "Caribbean Exploration – Planning for the next century", 17th–22nd June 2007, Port of Spain. Pindell, J. & Kennan, L. 2007a. Cenozoic kinematics and dynamics of oblique collision between two convergent plate margins: the Caribbean–South America collision in eastern Venezuela, Trinidad, and Barbados. In: Kennan, L., Pindell, J. L. & Rosen, N. C. (eds) Transactions of the 27th GCSSEPM Annual Bob F. Perkins Research Conference: The Paleogene of the Gulf of Mexico and Caribbean Basins: Processes, Events and Petroleum Systems, 458-553. Pindell, J. L. & Kennan, L. 2007b. Bow-wave model for deformation and foredeep development since 10 Ma, eastern Venezuela and Trinidad. Proceedings of the 4th Geological Society of Trinidad and Tobago Geological Conference – "Caribbean Exploration – Planning for the next century", 17th–22nd June 2007, Port of Spain. Pindell, J. L. & Kennan, L. 2009. Tectonic evolution of the Gulf of Mexico, Caribbean and northern South America in the mantle reference frame: an update. In: James, K., Lorente, M. A. & Pindell, J. (eds) The geology and evolution of the region between North and South America. Geological Society of London, Special Publications, this volume. Pindell, J. L., Cande, S. C., Pitman, W. C., Rowley, D. B., Dewey, J. F., Labrecque, J. & Haxby, W. 1988. A plate- kinematics framework for models of Caribbean evolution. Tectonophysics, 155, 121-138. Pindell, J. L., Erikson, J. P. & Algar, S. T. 1991. The relationship between plate motions and the sedimentary basin development in northern South America: from a Mesozoic passive margin to a Cenozoic eastwardly-progressive transpressional orogen. In: Transactions of the Second Geological Conference of the Geological Society of Trinidad and Tobago, 191-202. Pindell, J. L., Higgs, R. & Dewey, J. F. 1998. Cenozoic palinspastic reconstruction, paleogeographic evolution, and hydrocarbon setting of the northern margin of South America. In: Pindell, J. L. & Drake, C. L. (eds) Paleogeographic Evolution and Non-glacial Eustasy, northern South America. SEPM (Society for Sedimentary Geology), Special Publication, 58, 45-86. Pindell, J., Kennan, L. & Barrett, S. 2000. Kinematics: A key to unlocking plays. Part 2 of a series: "Regional Plate Kinematics: Arm Waving, or Underutilized Exploration Tool". American Association of Petroleum Geologists Explorer, July. World Wide Web Address: http://www.aapg.org/explorer/geophysical_corner/2000/gpc07.html Pindell, J. L., Kennan, L., Maresch, W. V., Stanek, K. P., Draper, G. & Higgs, R. 2005. Plate-kinematics and crustal dynamics of circum-Caribbean arc-continent interactions, and tectonic controls on basin development in Proto-Caribbean margins. In: Avé-Lallemant, H. G. & Sisson, V. B. (eds) Caribbean–South American plate interactions, Venezuela. Geological Society of America, Special Paper, 394, 7-52. Pindell, J. L., Kennan, L., Stanek, K. P., Maresch, W. V. & Draper, G. 2006. Foundations of Gulf of Mexico and Caribbean evolution: Eight controversies resolved. Geologica Acta, 4, 89-128. Pindell, J. L., Kennan, L. & Dewey, J. F. 2009. Tectonic evolution of the Gulf of Mexico, Caribbean and northern South America in the mantle reference frame: an update. In: James, K., Lorente, M. A., and Pindell, J. (eds) The geology and evolution of the region between North and South America, Geological Society of London, Special Publications, this volume. Potter, H.C. 1973. The overturned anticline of the Northern Range of Trinidad near Port of Spain. Journal of the Geological Society of London, 129, 133-138. Reyment, R. 1972. Some Lower Turonian Ammonites from Trinidad and Colombia. Geologiska Föreningens i Stockholm Förhandlingar, 94, 357-368. Roure, F., Carnevali, J.O., Subieta, T. & Gou, Y. 1994. Geometry and kinematics of the north Monagas thrust belt (Venezuela). Marine and Petroluem Geology, 11, 347-362. Roure, F., Bordas-Lefloch, N., Toro, J., Aubourg, C., Guilhaumou, N., Hernandez, E., Lecornec-Lance, S., Rivero, C., Robion, P. & Sassi, W. 2003. Petroleum systems and reservoir appraisal in the sub-Andean basins (eastern Venezuela and eastern Colombian foothills). In: Bartolini, C., Buffler, R. T. & Blickwede, J. (eds) The Circum-Gulf of Mexico and the Caribbean: Hydrocarbon habitats, basin formation, and plate tectonics. American Association of Petroleum Geologists, Memoir, 79, 750–775. Ruiz, G. M. H., Seward, D. & Winkler, W. 2004. Detrital thermochronology – a new perspective on hinterland tectonics, an example from the Andean Amazon Basin, Ecuador. Basin Research, 16, 413-430.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 59 Ruiz, G. M. H., Seward, D. & Winkler, W. 2007. Evolution of the Amazon Basin in Ecuador with special reference to hinterland tectonics: data from zircon fission-track and heavy mineral analysis. Developments in Sedimentology, 58, 907- 934. Sageman, B. B. & Speed, R. C. 2003. Upper Eocene limestones, associated sequence boundary, and proposed Eocene tectonics in eastern Venezuela. In: Bartolini, C., Buffler, R. T. & Blickwede, J. F. (eds), The circum-Gulf of Mexico and the Caribbean; hydrocarbon habitats, basin formation, and plate tectonics. American Association of Petroleum Geologists, Memoir, 79, 874-890. Sandwell, D. T. & Smith, W. H. F. 1997. Marine gravity anomaly from Geosat and ERS 1 satellite altimetry. Journal of Geophysical Research, 102B, 10039-10054. Salvador, A. & Stainforth, R. M. 1968. Clues in Venezuela to the geology of Trinidad, and vice versa. In: Saunders, J. B. (ed.) Transactions of the Fourth Caribbean Geological Conference, 28th March–12th April 1965, Port-of-Spain, Trinidad and Tobago. Caribbean Printers, Arima, Trinidad, 31-40. Saunders, J. B., Roberts, C. L., Ali, W. M. & Eggertson, E. B. (eds) 1998. Geological Map of Trinidad and Tobago. Ministry of Energy and Energy Industries, Trinidad. Schruben, P. G., Wynn, J. C., Gray, F. & Cox, D. P. 1997. Geology and Resource Assessment of the Venezuelan Guayana Shield at 1:500,000 scale – A Digital Representation of Maps published by the U.S. Geological Survey. USGS Digital Data Series, DDS-46. Schubert, C. 1971. Metamorphic rocks of the Araya Peninsula. Geologische Rundschau, 60, (4), 1571-1600. Senn, A. 1940. Palaeogene of Barbados and its bearing on history and structure of Antillean–Caribbean region. American Association of Petroleum Geologists Bulletin, 24, 1548-1610. Sidder, G. B. & Mendoza, V. 1995. Geology of the Venezuelan Guayana Shield and its relation to the entire Guayana Shield. US Geological Survey Bulletin, 2124, B1-B41. Sisson, V. B., Ertan, I. E. & Avé Lallemant, H. G. 1997. High-pressure (~2000 MPa) kyanite- and glaucophane-bearing pelitic schist and eclogite from the Cordillera de la Costa belt, Venezuela. Journal of Petrology, 38, 65-83. Sisson, V. B., Avé-Lallemant, H. G., Ostos, M., Blythe, A. E., Snee, L. W., Copeland, P., Wright, J. E., Donelick, R. A. & Guth, L. R. 2005. Overview of radiometric ages in three allochthonous belts of northern Venezuela: Old ones, new ones, and their impact on regional geology. In: Avé-Lallemant, H. G. & Sisson, V. B. (eds), Caribbean–South American plate interactions, Venezuela. Geological Society of America, Special Paper, 394, 91-117. Smith, C. A., Sisson, V. B., Avé Lallemant, H. G. A. & Copeland, P. 1999. Two contrasting pressure-temperature-time paths in the Villa de Cura blueschist belt,Venezuela: Possible evidence for Late Cretaceous initiation of subduction in the Caribbean. Geological Society of America Bulletin, 111, 831–848. Snoke, A. W., Rowe, D. W., Yule, J. D. & Wadge, G. 2001. Petrologic and structural history of Tobago, West Indies: A fragment of the accreted Mesozoic oceanic-arc of the southern Caribbean. Geological Society of America, Special Paper, 354. Socas, M. 1991. Estudio Sedimentologico de la Formacion Naricual, Estado Anzoategui. Geological Engineering Thesis, Universidad Central de Venezuela. Speed, R. C. 2002. Field trip to the Scotland District: Exposed example of an accretionary prism. In: Transactions of the 16th Caribbean Geological Conference, Barbados, 16th–23rd June, 2002. Speed, R.C., Westbrook, G.K., Biju-Duval, B., Ladd, J.W., Mascle, A., Moore, J.C., Saunders, J.B., Schoonmaker, J.E. & Stein, S. 1984. Atlas 10, Ocean Margin Drilling Program, Lesser Antilles and Adjacent Ocean Floor. Marine Science Institute Publications, Woods Hole, Massachussets. Speed, R. C., Sharp, W. D. & Foland, K. A. 1997. Late Paleozoic Granitoid Gneisses of Northeastern Venezuela and the North America–Gondwana Collision Zone. Journal of Geology, 105, 457-470. Sprague, A. R., Larue, D. K., Faulkner, B. L., Sykes, M. & Aden, L. 1996. Sequence stratigraphy, facies architecture and reservoir distribution, Cretaceous lowstand fan reservoirs, Southern Basin, Onshore Trinidad, Association Round Table. American Association of Petroleum Geologists Bulletin, 80, 1338. Stainforth, R. M. 1948. Description, correlation and paleoecology of Tertiary Cipero Marl Formation, Trinidad. American Association of Petroleum Geologists Bulletin, 32, 1292-1330. Stöckhert, B., Maresch, W. V., Brix, M., Kaiser, C., Toetz, A., Kluge, R. & Krückhans-Lueder, G. 1995. Crustal history of Margarita Island (Venezuela) in detail: Constraint on the Caribbean plate-tectonic scenario. Geology, 23, 787-790. Suter, H. H. 1960. The General and Economic Geology of Trinidad. British West Indies, Overseas Geological Surveys, Mineral Resources Division, London, UK. Van Andel, T. H. 1958. Origin and classification of Cretaceous, Paleocene, and Eocene sandstones of western Venezuela. American Association of Petroleum Geologists Bulletin, 42, 734-763. Van der Hilst, R. D. 1990. Tomography with P, PP and pP delay-time data and the three-dimensional mantle structure below the Caribbean region. PhD thesis, University of Utrecht. Villamil, T. 1999. Campanian–Miocene tectonostratigraphy, depocenter evolution and basin development of Colombia and western Venezuela. Palaeogeography, Palaeoclimatology, Palaeoecology, 153, 239-275. Villamil, T. & Pindell, J. L. 1998. Mesozoic paleogeographic evolution of northern South America: Foundations for sequence stratigraphic studies in passive margin strata deposited during non-glacial times. In: Pindell, J. L. & Drake, C.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 60 (eds) Paleogeographic Evolution and Non-glacial Eustacy: Northern South America. SEPM (Society for Sedimentary Geology), Special Publication, 58, 283-318. Vivas, V. & Macsotay, O. 1989. Miembro El Pilar de la Formación Quiamare: Ejemplo de molasa orogénica Neogena de Venezuela nororiental. GEOS, 29, 108-125. Vivas, V. & Macsotay, O. 1995. Formacion Tememure: Unidad olistostromica Eoceno Superior–Oligoceno Inferior en el frente meridional de la Napa Piemontina. Boletin de Geológia del Minsterio de Energia y Minas, Publicación Especial, 10, 95-113. Vivas, V. & Macsotay, O. 1997. Reinterpretación de la cobertura sedimentaria Cretácico–Palaeocena de la napa de Villa de Cura, Venezuela nor-central. Memorias del VIII Congreso Geológico Venezolano. Sociedad Venezolana de Geologia, Caracas, 2, 516-525. Weber, J., Ferrill, D. A. & Roden-Tice, M. K. 2001. Calcite and quartz microstructural geothermometry of low-grade metasedimentary rocks, Northern Range, Trinidad. Journal of Structural Geology, 32, 93-112. Wilkinson, B. H. & McElroy, B.J. 2007. The impact of humans on continental erosion and sedimentation. Geological Society of America Bulletin, 119, 140–156. Wilson, B. 2003. Foraminifera and paleodepths in a section of the Early to Middle Miocene Brasso Formation, central Trinidad. Caribbean Journal of Science, 39, (2), 209-214. Wilson, B. 2004. Benthonic Foraminiferal Palaeoecology Across a Transgressive-Regressive Cycle in the Brasso Formation (Early–Middle Miocene) of central Trinidad. Caribbean Journal of Science, 40, 126-138. Wood, L. J. 2000. Chronostratigraphy and tectonostratigraphy of the Columbus Basin, eastern offshore Trinidad. American Association of Petroleum Geologists Bulletin, 84, 1905-1928. Workman, W. 2000. Guyana Basin, a new exploration focus. World Oil, 221, 55-60. World Wide Web Address: http://www.worldoil.com/magazine/magazine_detail.asp?ART_ID=53 Wynn, J. C. 1993. Placer Titanium and other heavy minerals. US Geological Survey Bulletin, 2062, 89-90. Ysaccis, R. 1997. Tertiary evolution of the northeastern Venezuela offshore. PhD thesis, Rice University.

Pindell et al., in press 2009, PREPRINT Trinidad, Venezuela. Barbados heavy minerals and palaeogeography Page 61 -61° 55' -61° 50' -61° 45' -61° 40' -61° 35' -61° 30' -61° 25' -61° 20' -61° 15' -61° 10' -61° 05' -61° 00' -60° 55' -60° 50' 10° 55' 10° 55' Supplementary Figure 1. Location of Trinidad samples, with key geological, morphological, geographical features. Pindell, J. L., Kennan, L., Wright, D. T., & Erikson, J., Clastic domains of sandstones in central/eastern Venezuela, Trinidad, and Barbados: Toco Galera 10° 50' 10° 50' heavy mineral and tectonic constraints on provenance and palaeogeography

10° 45' 10° 45' NORTHERN RANGE DRAGON’S MOUTH 10° 40' 10° 40' NORTH ARIMA FAULT BASIN PORT OF SPAIN

10° 35' 10° 35' CARONI BASIN CARONI-FISHING POND FAULT

Cunapo Southern Road Mamon Brasso GULF 10° 30' 10° 30' East Mt. Harris OF Mamoral Rd Talparo Chaudiere River West Mt. Harris CENTRAL RANGE FAULT PARIA Cuche River Mamoral Rd Manzanilla Plum Mitan Road Cocos Bay 10° 25' 10° 25' Point Lisas Charuma Navet Dam Carolina Sand CENTRAL RANGE NARIVA Mapapire Quarry Area of Eocene Charuma outcrops Brake Factory “Coldon” Quarry SWAMP Piparo Gorge NARIVA FOLD- Point Radix Plaisance Piparo Mud Volcano 10° 20' 10° 20' Fabien Rd. Corbeau Hill THRUST BELT Area of Eocene Pt.-a-Pierre outcrops Williamsville Quarry SAN FERNANDO Sst Trace Balata Central (BC) well

ABM well POINT RADIX FAULT 10° 15' 10° 15' La Brea Catshill (CA) well

Point Fortin Barrackpore (BP) wells SOUTHERN 10° 10' LOS BAJOS FAULT BASIN 10° 10' Galeota Morne Diablo (MD) well Moruga East (ME) well Guayaguayare SOUTHERN “RANGE” 10° 05' 10° 05' SW Peninsula Galfa Point Moruga COLUMBUS CHANNEL

10° 00' 10° 00' -62° 00' -61°55' -61°50' -61° 45' -61° 40' -61° 35' -61° 30' -61° 25' -61° 20' -61° 15' -61° 10' -61° 05' -61° 00' -60° 55' Supplementary Figure 2. Location of Venezuela samples.

-65¡00' -64¡30' -64¡00' -63¡30' Caribbean allochthons/prism 25 km (metaigneous, sedimentary rocks)

ARAYA PENINSULA (metamorphosed slope, rise strata)

10¡30' EL PILAR FAULT 10¡30' Cumana V05-24A

V05-14 San Juan Graben V05-28 SERRANIA DEL INTERIOR ORIENTAL V05-30 V05-32 V05-04 GULF OF (folded, thrusted Cretaceous, Palaeogene platform strata) BARCELONA Pto. La Cruz 10/5-12 V05-13 10/5-2,4,5 10/5-11 V05-06 V05-19 Location of Figure 3 San Francisco Fault 10/5-10 stratigraphic sections V05-18 V05-08 V05-05 10¡00' V05-03 10¡00' V05-15 Onlap edge of piggyback basin URICA FAULT Pirital Thrust (buried)

Geology from Hackley et al., 2005 Urica MATURIN BASIN

-65¡00' -64¡30' -64¡00' -63¡30'

Pindell, J. L., Kennan, L., Wright, D. T., & Erikson, J., Clastic domains of sandstones in central/eastern Venezuela, Trinidad, and Barbados: heavy mineral and tectonic constraints on provenance and palaeogeography Supplementary Figure 3. Location of Barbados samples. -59¡36' -59¡30'

Geology from D.O.S., 1983 Pindell, J. L., Kennan, L., Wright, D. T., & Erikson, J., Geologic Map of Barbados Clastic domains of sandstones in central/eastern Venezuela, Trinidad, and Barbados: heavy mineral and tectonic constraints on provenance and palaeogeography

13¡18' 13¡18'

Scotland/Oceanics outcrops

Morgan Lewis E1,2 06/04 Scotland Gp. outcrop

St. Andrews D1,2,5,6 06/03 Walkers Beds 06/06 G2 06/05 06/02 Windy Hill Bridgetown Belle Hill 06/09 C2,4 B1,2 Chalky Mt. 10 km A1,2,3 Mount All

06/01 06/08a,b Bathsheba Bissex Hill 06/07

13¡12' 13¡12'

Landward edge of Scot- land Gp. outcrop (buried by Pleist. lst terraces) Oceanics Gp. outcrop

Ragged Point

Pleistocene coral terrace cover H1

5000 m

-59¡36' UTM Zone 20

-59¡36' -59¡30' Pindell, J. L., Kennan, L., Wright, D. T., and Erikson, J. Clastic domains of sandstones in central/eastern Venezuela, Trinidad, and Barbados Supplementary Table 1, Trinidad Heavy Minerals

Collector Year Number "DEPTH" UTM X UTM Y Location Formation Heavy Minerals Count Grain count breakdown Percentages (rounded) Others (raw counts) CUCHE Total Zircon Tourm. Apatite Rutile Others StauroliteEpidote GarnetGp Zircon Tourm. Apatite Rutile Others Staurolite Epidote Gp Garnet Clinozoisite Kyanite Chloritoid Others Total LK 2005 50 73 704744 1159961 MT HARRIS 1 Well Cuche 200 124 67 2 7 0 0 0 0 62 34 1 4 0 0 0 0 0 0 0 0 0 LK 2005 51 72 704744 1159961 MT HARRIS 1 Well Cuche 200 160 36 1 1 2 0 0 0 80 18 1 1 1 0 0 0 0 0 0 2 2 Other = unknown LK 2006 22 71 704228 1156660 Cuche River Cuche 193 172 20 0 0 1 0 0 0 89 10 0 0 1 0 0 0 0 0 0 1 1 Other = monazite LK 2006 24 70 704248 1156427 Cuche River Cuche 100 78 21 0 1 0 0 0 0 78 21 0 1 0 0 0 0 0 0 0 0 0 LK 2006 26 69 704041 1156260 Cuche River Cuche 200 166 32 1 1 0 0 1 0 82 16 0 1 0 0 1 0 0 0 0 0 0 LK 2005 63 68 723975 1198225 Toco, Andre Pt Toco/Cuche? 176 84 65 5 0 22 0 0 0 47 37 3 0 13 0 0 0 22 22 Others = brookite, some rdd, detrital LK 2005 65 67 723975 1198225 Toco, Andre Pt Toco/Cuche? 200 115 35 1 4 45 0 0 0 56 18 1 2 23 0 0 0 45 45 Others = brookite, some rdd, detrital

TOCO LK 2005 68 61 720213 1198397 Snake River beach Toco 118 71 4 39 2 0 0 0 2 60 3 33 2 0 0 0 2 0 0 0 0 0

GAUTIER/NAP HILL LK 2005 54 60 695995 1119797 ME15 Well 8875' Gautier 198 123 30 24 0 1 0 0 20 62 15 12 0 1 0 0 10 0 0 0 1 1 Other = spinel? (1) LK 2005 52 59 672918 1119879 MD34 Well 13550' Lr. Naparima Hill 200 103 22 75 0 0 0 0 0 52 11 38 0 0 0 0 0 0 0 0 0 0

GALERA LK 2005 24 58 728375 1198275 Galera Lighthouse Galera 131 102 28 0 1 0 0 0 0 78 21 0 1 0 0 0 0 0 0 0 0 0

WEST PAP LK 2005 29 57 672234 1142474 San Fabien Road Pointe-a-Pierre (PaP) 200 156 43 0 1 0 0 0 0 78 22 0 1 0 0 0 0 0 0 0 0 0 LK 2005 45+46 56 676800 1143050 Tormos Brake Factory Pointe-a-Pierre? 200 157 43 0 0 0 0 0 0 79 22 0 0 0 0 0 0 0 0 0 0 0

CENTRAL POINTE-A-PIERRE (PAP), NAVET DAM, PIPARO AREA LK 2003 TTM2 55 690805 1150250 Navet Dam Gully Pointe-a-Pierre? 200 165 34 0 0 1 0 0 0 82 17 0 0 1 0 0 0 0 0 0 1 1 Other = monazite (1) JP/RH 2003 T45-1 54 681788 1144821 Piparo Gorge Pointe-a-Pierre? 200 174 26 0 0 0 0 0 0 87 13 0 0 0 0 0 0 0 0 0 0 0

EAST MT. HARRIS LK 2005 20 53 707580 1159672 East Mt. Harris Chaudière-PaP? 200 125 70 3 2 0 0 0 0 63 35 2 1 0 0 0 0 0 0 0 0 0

WEST MT HARRIS LK 2005 33 52 705441 1158397 West Mt. Harris Chaudière-PaP? 117 103 14 0 0 0 0 0 0 88 12 0 0 0 0 0 0 0 0 0 0 0 LK 2007 20 51 705005 1158358 South of Mt. Harris Bridge Chaudière-PaP? 200 155 40 0 3 2 0 0 0 77 20 0 2 1 0 0 0 0 2? 0 0 2 Badly weathered, tentative

CHAUDIERE RIVER LK 2006 19 50 705846 1158756 Mt Harris Chaudière-PaP? 200 172 28 0 0 0 0 0 0 86 14 0 0 0 0 0 0 0 0 0 0 0 LK 2006 16 49 705326 1159290 Chaudière River Chaudière-PaP? 200 172 25 0 1 0 0 2 0 86 13 0 1 0 0 1 0 0 0 0 0 0 LK 2006 6 48 705262 1159351 Chaudière River Chaudière-PaP? 200 158 40 0 0 0 0 2 0 79 20 0 0 0 0 1 0 0 0 0 0 0 LK 2005 48 47 704833 1159979 Chaudiere River Chaudière-PaP? 200 185 13 0 1 0 0 1 0 93 7 0 1 0 0 1 0 0 0 0 0 0

PICNIC SITE + WMH GRIT LK 2005 14 46 706928 1162487 Mt. Harris Picnic Site Chaudière-PaP? 200 152 44 0 4 0 0 0 0 76 22 0 2 0 0 0 0 0 0 0 0 0 LK 2005 32 45 705772 1158391 West Mt. Harris Chaudière-PaP? 200 117 83 0 0 0 0 0 0 59 42 0 0 0 0 0 0 0 0 0 0 0 LK 2005 37 44 704925 1158425 West Mt. Harris Chaudière-PaP? 200 175 25 0 0 0 0 0 0 88 13 0 0 0 0 0 0 0 0 0 0 0

SAN FERNANDO, PLAISANCE LK 2005 43 43 670350 1143109 Plaisance Quarry Plaisance Mb. (Sfdo) 150 145 5 0 0 0 0 0 0 97 3 0 0 0 0 0 0 0 0 0 0 0 LK 2005 44 42 670469 1143148 Plaisance Pipeline Plaisance Mb. (Sfdo) 200 158 37 0 3 0 0 2 0 79 19 0 2 0 0 1 0 0 0 0 0 0

CHARUMA SILT AREA JP 2003 ANT1 41 701900 1151925 Charuma (sand phacoid) Charuma Silt Mb. (PaP) type locality 100 60 20 4 0 10 0 0 6 60 20 4 0 10 0 0 6 1 8 1 other = sphene LK 2003 LKSUN2 40 701900 1151925 Charuma (sand phacoid) Charuma Silt Mb. (PaP) type locality 199 139 35 2 4 6 0 0 13 69 18 1 2 3 0 0 7 2 other = sphene LK 2007 2 39 701900 1151925 Charuma (silty matrix) Charuma Silt Mb. (PaP) type locality 161 91 33 0 9 12 0 0 16 57 20 0 6 7 0 0 10 0 2 9 1 12 Other = monazite LK 2003 AUG03/06 (a) 38 706140 1155970 Plum Mitan Road Charuma Silt Mb. (PaP) 127 109 16 0 1 1 0 0 0 85 13 0 1 0 0 1 0 0 0 0 0

NARIVA LK 2006 4 37 704901 1159403 Chaudière River Nariva (our reinterpretation) 200 124 38 0 1 1 0 5 31 62 19 0 1 1 0 3 16 0 1 0 0 1 LK 2006 3 36 704901 1159403 Chaudière River Nariva (our reinterpretation) 200 122 64 0 2 4 0 0 8 61 32 0 1 2 0 0 4 0 0 4 0 4 LK 2006 2 35 704901 1159403 Chaudière River Nariva (our reinterpretation) 200 97 68 0 7 8 0 0 20 49 34 0 4 4 0 0 10 0 0 8 0 8 LK 2005 55 34 649654 1134789 ABM54 well Nariva 198 43 25 7 2 18 3 0 100 22 13 4 1 9 2 0 51 0 0 18 0 18 LK 2005 1 33 680373 1142799 Corbeau Hill Nariva 200 123 62 1 5 9 0 0 0 62 31 1 3 5 0 0 0 0 0 7 2 9 Others = corundum (2) LK 2005 12 32 677269 1140842 Williamsville Quarry Nariva 250 129 25 1 3 14 28 0 50 52 10 0 1 6 11 0 20 0 2 9 3 14 Others = blue amphibole (3) LK 2005 11 31 679361 1143439 Guaracara Road (Coldon Qy.) Nariva 200 79 114 0 2 5 0 0 0 40 57 0 1 3 0 0 0 0 0 5 0 5 LK 2005 6 30 676646 1142429 Sandstone Trace Nariva 201 103 51 0 6 40 0 0 1 51 25 0 3 20 0 0 0 0 0 40 0 40 Also ?corundum (1); ?spinel (1) LK 2005 7 29 676646 1142429 Sandstone Trace Nariva 200 112 31 0 15 38 0 0 4 56 16 0 8 19 0 0 2 0 0 27 11 38 Other = monazite (1); corundum (6); ?epidote (4) LK 2005 18 28 680953 1143165 Piparo Mud Volcano Nariva (probable) 201 95 11 14 5 10 1 0 65 47 5 7 2 5 0 0 32 0 1(?) 10 0 10 V. tentative identification of kyanite (1)

BRASSO LK 2003 AUG03/01 27 706674 1161841 Cunapo Sthn. Road Brasso? (previously mapped as PaP 200 104 82 0 6 8 0 0 0 52 41 0 3 4 0 0 0 0 0 8 0 8 Also 2 blue tourm. LK 2003 BR1 26 701150 1162750 Nestor-Mamon Rd Brasso 200 97 94 0 4 5 0 0 0 48 47 0 2 3 0 0 0 0 0 5 0 5 1 kyanite? Also 2 blue tourm.

HERRERA LK 2005 74 (b) 25 643521 1124226 Galfa Point Herrera? 32 20 4 2 1 1 0 0 4 63 13 6 3 3 0 0 12 0 0 0 1 1 Other = sphene LK 2006 30 (c) 643521 1124226 Galfa Point 200 2 2 0 0 196 0 0 0 1 1 0 0 98 0 0 0 0 0 0 197 197 Others = authigenic barytes only JKD 2005 BC1/1 24 626251 1114095 BC1 well Herrera 200 158 20 3 1 3 1 3 11 79 10 2 1 2 1 2 6 1 0 1 1 3 Others = hbl (1) JKD 2005 CA40/2 23 691705 1129621 CA40 well Herrera 200 148 41 3 3 3 0 1 1 74 21 2 2 2 0 1 1 0 0 0 3 3 Others = cassiterite (1); sphene (2); also anatase (1) JKD 2005 CA40/1 22 691705 1129621 CA40 well Herrera 200 125 31 7 2 7 0 1 27 63 16 4 1 4 0 1 14 0 0 7 0 7 Also anatase (7) JKD 2005 BP344/1 21 673828 1124559 BP344 well Herrera 200 153 31 2 1 1 0 0 12 77 16 1 1 1 0 0 6 0 0 0 1 1 Others = monazite (1) JKD 2005 BC1/2 20 626251 1114095 BC1 well Herrera 147 119 8 0 4 4 0 0 12 81 5 0 3 3 0 0 8 0 0 4 0 4 JKD 2005 BP362/1 19 675896 1126995 BP362 well Herrera 115 95 7 0 0 1 0 2 10 83 6 0 0 1 0 2 9 0 0 0 1 1 Others = sphene (1); also anatase d (5) JKD 2005 BC1/3 18 626251 1114095 BC1 well Herrera 200 135 30 11 1 7 0 0 16 68 15 6 1 4 0 0 8 0 0 7 0 7 Also anatase (2) JKD 2005 BP344/2 17 673828 1124559 BP344 well Herrera 200 143 34 2 1 2 0 4 14 72 17 1 1 1 0 2 7 0 0 2 0 2 Also anatase (5) JKD 2005 BP362/2 16 675896 1126995 BP362 well Herrera 177 134 20 4 1 1 0 3 14 76 11 2 1 1 0 2 8 0 0 1 0 1 JKD 2005 BP362/3 15 675896 1126995 BP362 well Herrera 196 142 25 6 2 4 2 4 11 72 13 3 1 2 1 2 6 0 0 2 2 4 Others = sphene (2) JKD 2005 BC1/4 14 626251 1114095 BC1 well Herrera 93 74 8 1 0 1 0 0 9 80 9 1 0 1 0 0 10 1 1 JKD 2005 BP347/1 13 674433 1124535 BP347 well Herrera 160 113 20 7 2 6 2 2 8 71 13 4 1 4 1 1 5 0 0 2 4 6 Others = monazite (1); sphene (3), also anatase (4) JKD 2005 BP347/2 12 674433 1124535 BP347 well Herrera 167 116 25 4 2 5 0 3 12 69 15 2 1 3 0 2 7 0 0 4 1 5 Others = sphene (1); also anatase (2) JKD 2005 BP347/3 11 674433 1124535 BP347 well Herrera 164 120 27 4 0 5 2 0 6 73 16 2 0 3 1 0 4 0 0 4 1 5 Others = monazite 1 LK 2005 53 10 672918 1119879 MD34 well Herrera 209 138 40 13 5 5 1 3 4 66 19 6 2 2 0 1 2 0 0 5 0 5

LOWER MANZANILLA LK 2003 LKSUN2c 9 706679 1168240 Cunapo Southern Road "Cunapo" (mismapped?) 200 149 45 0 6 0 0 0 0 74 23 0 3 0 0 0 0 0 0 0 0 0 LK 2005 13 8 706538 1163096 Cunapo Southern Road Manzanilla (San Jose Mbr) 200 190 8 0 2 0 0 0 0 95 4 0 1 0 0 0 0 0 0 0 0 0 LK 2005 16 7 707625 1156170 Plum Mitan Road citrus grove Unknown Lt Mio or yngr 200 150 44 0 6 0 0 0 0 75 22 0 3 0 0 0 0 0 0 0 0 0

UPPER MANZANILLA LK 2006 8 6 687543 1155438 Mamoral Road Manzanilla (Telemaque) 200 31 14 4 0 46 0 100 5 16 7 2 0 23 0 50 3 40 1 5 0 46

CRUSE LK 2005 70 5 626546 1113374 Galfa Point Cruse 199 145 17 0 5 12 11 0 9 70 10 0 3 6 6 0 5 0 5 2 5 12 Others = sphalerite? LK 2005 41 4 626546 1113374 Galfa Point Cruse 196 68 41 47 4 18 0 0 18 35 21 24 2 9 0 0 9 4 7 7 0 18

SPRINGVALE, TALPARO? LK 2005 9 3 678432 1143565 Mapapire Road Unknown Lt Mio or yngr 200 135 56 0 6 1 2 0 0 68 28 0 3 1 1 0 0 0 0 0 1 1 Other = spinel?

TALPARO LK 2006 7 2 689395 1159779 Mamoral Road Talparo 200 157 24 0 2 10 7 0 0 79 12 0 1 5 4 0 0 0 2 8 10 Others = sphene (2); xenotime (1); unknown (5) LK 2003 LK/TALP/SUM 1 672750 1148500 Carolina sand quarry Talparo 107 66 17 0 4 11 7 2 0 61 16 0 4 10 7 2 0 3 4 0 4 11 Others = sillimanit (3); glaucophane (1)

COMMENTS (a) Many coarse zircons >300 microns, some resorbed and embayed; some coarse euhedral clear zircons (b) Low grain count, overwhelmed by opaques, with carbonates, white mica and chlorite. 1 spherical apatite and 1 metamict zircon (c) Too little recovered other than authigenic barytes to be meaningful Pindell, J. L., Kennan, L., Wright, D. T., and Erikson, J. Clastic domains of sandstones in central/eastern Venezuela, Trinidad, and Barbados Supplementary Table 2, Venezuela Heavy Minerals

Collector Year Number "DEPTH" UTM X UTM Y Location Formation Heavy Minerals Count Grain count breakdown Percentages (rounded) Others (raw counts) BARRANQUIN Total Zircon Tourm. Apatite Rutile Others Staurolite Epidote Gp Garnet Zircon Tourm. Apatite Rutile Others Staurolite Epidote Gp Garnet Clinozois Kyanite Chloritoid Others Total JE 2005 V05-04 23 352242 1141866 Mochima Barranquin 200 120 80 0 0 0 0 0 0 60 40 0 0 0 0 0 0 0 0 0 0 0 JE 2005 V05-14 22 362839 1147821 Cumana Barranquin 200 180 13 0 2 2 3 0 0 90 7 0 1 1 2 0 0 0 2 0 0 2 JE 2005 V05-28 21 321141 1139591 Chimana Grande Barranquin 200 131 65 0 2 0 1 1 0 66 33 0 1 0 1 1 0 0 0 0 0 0 JE 2005 V05-30 20 321141 1139591 Chimana Grande Barranquin 200 151 39 0 3 1 1 5 0 76 20 0 2 1 1 3 0 0 0 0 1 1 Other = sphene

EL CANTIL JE 2005 V05-32 19 323936 1138288 Chimana Segunda El Cantil 148 111 36 1 0 0 0 0 0 75 24 1 0 0 0 0 0 0 0 0 0 0

SAN JUAN JE 2005 V05-13 18 422772 1123371 Triste San Juan 200 137 60 0 2 1 0 0 0 69 30 0 1 1 0 0 0 0 0 0 1 1 Other = xenotime JE 2005 V05-18 17 349292 1105727 Rio Querecual San Juan 200 162 38 0 0 0 0 0 0 81 19 0 0 0 0 0 0 0 0 0 0 0 JE 2005 V05-24A 16 469528 1155350 E of Casanay San Juan 151 114 36 0 1 0 0 0 0 75 24 0 1 0 0 0 0 0 0 0 0 0 JP 2005 10/05-11A 15 326819 1119925 San Juan 200 163 36 0 0 1 0 0 0 81 18 0 0 1 0 0 0 0 0 0 1 1 Other = sphene

CARATAS JE 2005 V05-05 14 428808 1113272 Rio Guarapiche Caratas 200 167 32 0 1 0 0 0 0 84 16 0 1 0 0 0 0 0 0 0 0 0 JE 2005 V05-06 13 428808 1113272 Rio Guarapiche Caratas 200 158 38 1 2 0 0 1 0 79 19 1 1 0 0 1 0 0 0 0 0 0 JP 2005 10/05-2A 12 319663 1123919 Caratas 200 168 31 0 1 0 0 0 0 83 16 0 1 0 0 0 0 0 0 0 0 0 JP 2005 10/05-2B 11 319663 1123919 Caratas 200 159 41 0 0 0 0 0 0 79 21 0 0 0 0 0 0 0 0 0 0 0

PEÑAS BLANCAS JE 2005 V05-03 10 256029 1110365 NW of Clarines Peñas Blancas 200 185 15 0 0 0 0 0 0 93 8 0 0 0 0 0 0 0 0 0 0 0

LOS JABILLOS JE 2005 V05-15 9 349292 1105542 Rio Querecual Los Jabillos 192 155 36 0 1 0 0 0 0 81 19 0 1 0 0 0 0 0 0 0 0 0 JP 2005 10/05-4A 8 319508 1123409 Los Jabillos 200 174 26 0 0 0 0 0 0 87 13 0 0 0 0 0 0 0 0 0 0 0

LECHERÍAS JP/MN 2005 10/5-12A 7 314487 1125273 Lecherías Lecherías 200 141 58 0 1 0 0 0 0 70 29 0 1 0 0 0 0 0 0 0 0 0

AREO-NARICUAL JP 2005 10/05-5A 6 319327 1123129 Lower Areo 200 185 12 0 2 1 0 0 0 92 6 0 1 1 0 0 0 0 1 0 0 1 Also 6 anatase JP 2005 10/05-5B 5 319327 1123129 Upper Areo 200 162 35 0 1 2 0 0 0 81 18 0 1 1 0 0 0 0 0 0 2 2 Others = sphene JE 2005 V05-08 4 416533 1110116 Arbolitos Areo-Naricual 200 167 33 0 0 0 0 0 0 84 17 0 0 0 0 0 0 0 0 0 0 0 JP/MN 2005 10/05-10D 3 320330 1114389 Naricual 174 134 36 0 2 0 0 0 2 77 21 0 1 0 0 0 1 0 0 0 0 0 JP/MN 2005 10/05-10F 2 320330 1114389 Naricual 200 120 19 45 6 0 0 0 10 59 10 23 3 0 0 0 5 0 0 0 0 0

QUEBRADON JE 2005 V05-19 1 256616 1112151 Clarines Quebradon? 199 98 37 11 5 0 1 0 47 49 19 6 3 0 1 0 24 0 0 0 0 0 Pindell, J. L., Kennan, L., Wright, D. T., and Erikson, J. Clastic domains of sandstones in central/eastern Venezuela, Trinidad, and Barbados Supplementary Table 3, Barbados heavy minerals

Collector Year Number "ZONE" "AGE" UTM X UTM Y Location Formation Heavy Minerals ZONE 20 Count Grain count breakdown Percentages (rounded) Others (raw grain counts) Total Zircon Tourm. Apatite Rutile Others Staurolite Epidote Gp Garnet Zircon Tourm. Apatite Rutile Others Staurolite Epidote Gp Garnet Clinozois Kyanite Chloritoid Others Total JP 2005 E1 2 1 221877 1468417 The Chase Morgan Lewis Beds 200 13 9 2 0 110 0 42 24 7 5 1 0 55 0 21 12 89 4 2 15 110 JP 2005 E2 2 1 221877 1468417 The Chase Morgan Lewis Beds 165 55 11 9 1 54 2 16 17 33 7 5 1 33 1 10 10 37 4 5 8 54 JP 2006 JP 12/06-03 2 1 221809 1466603 St. Andrews Walkers Beds 200 2 26 0 0 147 0 25 0 1 13 0 0 73 0 13 0 147 0 0 0 147 NB noted on further examination: glaucophane and lawsonite JP 2006 JP 12/06-05 2 1 220207 1465819 Bawdens Walkers Beds 200 6 36 0 0 135 0 21 2 3 18 0 0 67 0 11 1 135 0 0 0 135 JP 2005 E4 2 1 221820 1468598 The Chase Morgan Lewis Beds 200 15 17 6 1 56 0 88 16 8 9 3 1 28 0 43 8 47 1 5 3 56 brookite; unknown JP 2006 JP 12/06-04 2 1 220318 1468110 Behind Windmill Morgan Lewis Beds 200 22 16 5 2 134 0 19 2 11 8 3 1 67 0 10 1 125 0 0 9 134 others = glaucophane (1), ?tremolite (5) monazite (2) ?pumpellyite (1) JP 2005 D1 2 2 220449 1466307 Oil Quarry Walkers Beds 214 71 25 19 3 60 2 20 14 33 12 9 1 28 1 9 7 46 6 5 3 60 sphene; amphibole JP 2005 D2 2 2 220449 1466307 Oil Quarry Walkers Beds 200 122 15 3 0 24 1 4 31 61 8 2 0 12 1 2 16 14 5 1 4 24 sphene; xenotime JP 2005 D5 2 2 220449 1466307 Oil Quarry Walkers Beds 201 57 20 19 2 73 1 14 15 28 10 9 1 36 0 7 7 44 17 2 10 73 sphene; corundum? JP 2005 D6 2 2 220449 1466307 Oil Quarry Walkers Beds 200 10 42 10 3 135 0 0 0 5 21 5 2 68 0 0 0 135 0 0 0 135 JP 2006 JP 12/06-09 2 2 221704 1465118 Sth. Belle Hill Lower Scotland 23 1 1 0 0 21 0 0 0 (a) 17 3 1 0 21 JP 2005 A1 1 3 224630 1463800 Cambridge Chalky Mt. Mbr. 200 81 54 3 0 32 3 7 20 41 27 2 0 16 2 4 10 17 6 7 2 32 zoisite JP 2005 A2 1 3 224630 1463800 Cambridge Chalky Mt. Mbr. 200 116 41 0 2 15 2 2 22 58 21 0 1 8 1 1 11 0 4 10 1 15 amphibole JP 2005 A3 1 3 224630 1463800 Cambridge Chalky Mt. Mbr. 208 34 120 0 2 20 1 1 30 16 58 0 1 10 0 0 14 0 2 14 4 20 amphibole JP 2005 B1 1 3 224281 1464661 Barclays Pk Chalky Mt. Mbr. 200 76 58 11 1 20 2 6 26 38 29 6 1 10 1 3 13 0 11 8 1 20 amphibole JP 2005 B2 1 3 224281 1464661 Barclays Pk Chalky Mt. Mbr. 200 156 26 0 1 5 2 8 2 78 13 0 1 3 1 4 1 0 2 3 0 5 JP 2005 C2 1 3 223612 1464750 Chalky Mt. Potteries Chalky Mt. Mbr. 200 110 54 0 2 14 12 8 0 55 27 0 1 7 6 4 0 0 3 2 9 14 unknown; sphene JP 2005 C4 1 3 223612 1464750 Chalky Mt. Potteries Chalky Mt. Mbr. 200 151 40 1 1 7 0 0 0 76 20 1 1 4 0 0 0 1 4 1 1 7 unknown JP 2006 JP 12/06-06 2 3 223378 1465878 Windy Hill Windy Hill Mbr. 196 68 40 1 4 56 7 10 10 35 20 0 2 29 4 5 5 50 0 5 1 56 glaucophane; NB noted on further examination: chrome spinel, glaucophane, monazite, tremolite JP 2006 JP 12/06-01 2 4 220888 1463083 Mt. All Mt. All Mbr. 200 150 28 0 4 10 6 2 0 75 14 0 2 5 3 1 0 0 10 0 0 10 JP 2006 JP 12/06-08a 2 4 221121 1463165 Mt. All Mt. All Mbr. 200 110 68 0 4 18 0 0 0 55 34 0 2 9 0 0 0 9 4 4 1 18 unknown; 1 garnet outside count JP 2006 JP 12/06-08b 2 4 221121 1463165 Mt. All Mt. All Mbr. 201 137 41 0 1 19 0 0 2 67 21 0 1 10 0 0 1 14 1 4 0 19 JP 2005 G2 2 5 221433 1465255 Belle Hill Belle Hill 180 157 15 0 4 2 0 2 0 87 8 0 2 1 0 1 0 2 0 0 0 2 JP 2005 H1 2 5 236754 1456840 Ragged Point Belle Hill Mbr. 200 169 12 0 2 9 3 0 5 85 6 0 1 5 2 0 3 8 1 0 0 9 JP 2006 JP 12/06-02 2 5 221831 1445277 Belle Hill E. Belle Hill Mbr. 200 78 33 0 2 58 0 5 24 39 17 0 1 29 0 3 12 44 0 11 3 58 others = glaucophane (1); ?tremolite (3) JP 2006 JP 12/06-07 2 6 226917 1461801 Bathsheba Post-Scotland? 200 10 20 0 3 156 0 11 0 5 10 0 2 77 0 6 0 156 0 0 0 156

COMMENT (a) v poor sample; only v few, fine-grained heavy minerals, incl. tiny prisms of clinozoisite, kyanite and chloritoid; 1 zircon, 1 tourmaline; overwhelmed by opaque minerals