SHRIMP and conventional zircon dating place temporal constraints on the evolution of the Cretaceous Volcanic Arc system in central Cuba. The arc has a consistent stratigraphy across strike, with the oldest and deepest rocks in the south (in tectonic contact with the ~5-10-km-wide Mabujina Amphibolite Complex [MAC]) and younger rocks in the north. The MAC is thought to represent the deepest exposed section of the Cretaceous Volcanic Arc and its oceanic basement in Cuba. We undertook a single zircon geochronological study of five gneisses and two amphibolites from the MAC and seven rocks from the Manicaragua Batholith, which intrudes both the MAC and the Cretaceous Volcanic Arc. A SHRIMP zircon age of 132.9 ± 1.4 Ma for a trondhjemitic orthogneiss (MAC) from the Jicaya River dates the oldest phase of granitoid magmatism in this area and the entire Caribbean (Antillean) region. A tonalitic gneiss collected near the previous sample yielded an age of 123.9 ± 0.6 Ma, and a further tonalitic gneiss had an age of 112 ± 2.1 Ma, with one inherited zircon at 1045 ± 17 Ma. Two trondhjemitic orthogneisses from the central part of the MAC yielded ages of 93.8 ± 0.5 and 92.8 ± 0.7 Ma, whereas two amphibolites from the eastern part of the complex provided similar ages of ca. 93 Ma and zircon inheritance at 315, 471, 903, and 1059 Ma. Two weakly foliated Manicaragua granitoids from the eastern part of the massif provided ages of 89.3 ± 0.45 and 87.2 ± 1.2 Ma, whereas five unfoliated granitoid samples from the central and eastern part of the massif yielded ages of 88.7 ± 0.7, 87.4 ± 1.3, 87.0 ± 0.6, 84.2 ± 0.8, and 83.1 ± 0.8 Ma. Our age data support the view that the Mabujina Protholiths are exotic and formed somewhere NNW along strike of the nonmetamorphosed Cuban arc since pre-Middle Hau-terivian time (before ~133 Ma). The MAC became part of the Cuban Volcanic Arc during the Turonian (ca. 90-93 Ma), when it was intruded by plutonic rocks of the Manicaragua Batholith (Turonian-Campanian; ca. 89-83 Ma). The geology and geochronology of central Cuba do not support the idea of a polarity reversal event at any stage of the Cretaceous Arc-building process. Because most of our dated samples come from the narrow Mabujina Belt, the polarity reversal model would imply that the axis of a newly developing arc (with opposite polarity) would spatially coincide with the older arc, which appears unlikely. Inherited Precambrian and Palaeozoic zircons in the MAC granitic rocks (similar to inherited zircon populations in the Guerrero terrane from central-western Mexico) suggest a Neocomian proximal setting close to a cratonic area (probably SW Mexico/Maya Block) for the protolith of the MAC relative to the synchronous Primitive Island Arc of central Cuba.