The contact relationships between metabasalts (eclogites, glaucophanites, prasinites, etc.) and the enclosing mica schists in the Venezuelan Coast Ranges favour a common petrological history for both. Mineralogical disequilibria, such as replacement textures and mineral zoning in the metabasic assemblages, can all be related to a single metamorphic cycle. Relict deuteric/late magmatic hornblende epitaxially overgrown by the barroisitic amphibole of this metamorphic event shows that the latter has been the only regional metamorphic episode to have affected these rocks. The high-pressure character of the metamorphism is a logical consequence of overthrusting related to collision of the Aruba-Blanquilla island arc with the South American continental margin. The well-defined stratigraphy and detailed radiometric dating of intrusive rocks in the Netherlands Antilles indicate that this collision took place in the Coniacian/Campanian interval. The igneous rocks of the Netherlands Antilles, which are considered to form part of the colliding arc, consist largely of submarine volcanics, as well as a tonalite/gabbro batholith. These rocks range in age from middle Albian to Coniacian. The volcanics of Curaçao and Aruba are composed of basalts with a MORB chemistry and are oceanic in origin. Nevertheless, these sequences differ from "normal" oceanic crust by their thickness, chemical homogeneity and the non-depleted nature of the source, thus suggesting that they were fed by a prolific chondritic mantle plume. The volcanics of Bonaire range from basalt to rhyolite in composition and are chemically related to the primitive island arc series. An important characteristic is the high initial water content of these magmas, as shown by the geometry and mineralogy of the flows. The tonalite/gabbro batholith on Aruba is of calc-alkaline composition. The similarity in chemistry between the volcanics of Bonaire and the Villa de Cura Group of the Venezuelan mainland supports the view that the latter is an overthrust remnant of the colliding arc. Comparison with the metabasalts of the La Rinconada Group of Margarita Island is equivocal. Existing paleomagnetic evidence indicates that the arc underwent a 90° north-south to east-west rotation shortly before collision, and that the colliding arc extended via the Aves Ridge into the Greater Antilles. The age of the oldest volcanics of this arc, and thus the age of its origin, is uncertain. Most data favour formation in the Early Cretaceous, but a Late Jurassic age is also possible. Consequently, two alternative models for the evolution of the arc are proposed: one in which the arc forms as a lengthening Central American "isthmus" in response to opening of the Caribbean in the Late Jurassic, and a second in which the arc originates in the Pacific in the Early Cretaceous. In order to collide with the northern as well as southern margin of the Caribbean, the arc must have lengthened to more than twice the width of the gap. Lengthening probably occurred at a trench-trench-ridge triple junction, thus explaining the large volume of MORB-volcanics in the arc. The high water content of the primitive island arc series is attributed to a high rate of subduction due to the worldwide surge in spreading rates in the mid-Cretaceous. In the second model, a strong causal connection between the mid-Cretaceous thermal event in the Pacific as well as the origin and evolution of the arc/trench system is postulated.