This report summarizes the results of paleontological and geological investigations carried out during the 1990s at Domo de Zaza, a late Early Miocene vertebrate locality in south-central Cuba. Paleontologically, the most important result of fieldwork at Zaza was the first discovery of terrestrial mammals of Tertiary age in Cuba. Three terrestrial mammal taxa are now known from this locality—a megalonychid sloth (Imagocnus zazae), an isolobodontine capromyid rodent (Zazamys veronicae), and a platyrrhine primate (Paralouatta marianae, new species). In addition to these finds, a number of selachian, chelonian, crocodylian, cetacean, and sirenian remains have been recovered.

Domo de Zaza is a low hill transected by a large artificial channel, the Canal de Zaza, whose walls provide an extensive exposure of Miocene sediments attributable to the Lagunitas Formation (Fm). This formation is laterally and vertically complex, showing evidence of at least four different depositional regimes. However, the sedimentary sequence indicates that all depositional phases took place within a broader episode of transgression. Estimated Burdigalian age (16.1–21.5 Ma) for Lagunitas Fm is based on the presence of marine invertebrate taxa corresponding to the late Early Miocene Miogypsina–Soritiidae zone. The overall transgressive aspect of Lagunitas suggests rising sea level, possibly in correlation with a global onlap event. Within Burdigalian time, the most likely correlate is the eustatic rise centered on 17.5–18.5 Ma.

Most of the vertebrate fossils were recovered from lagoonal and alluvial beds; those from lagoonal beds are exceptionally well preserved. The terrestrial facies displays evidence of paleosol formation, subaerial erosion, and plant life in the form of grass and palm pollen. Other evidence indicates that most of the present-day highlands of Cuba, including the Cordillera del Escambray near Zaza, have been continuously subaerial since the latter part of the Late Eocene. Although no land vertebrate fossils of this age are known from Cuba, recent discoveries elsewhere in the Greater Antilles indicate that land vertebrates could have colonized landmasses in the Caribbean Basin as early as 33–36 Ma.

Recently, marine geological data have been interpreted as showing that (1) the Mona Passage began to form in the Early Oligocene, and (2) the Puerto Rico/Virgin Island block was entirely transgressed by shallow marine environments during the period between the Late Oligocene and the Early Pliocene. However, the seismic reflection profile evidence for an Early Oligocene opening of the Passage is ambiguous. Even if the separation of Puerto Rico and eastern Hispaniola occurred relatively early, it remains more probable than not that this happened in the medial Oligocene or even somewhat later (i.e., ≤30 Ma). On the other hand, the evidence is not at all ambiguous concerning the hypothesized mid-Cenozoic inundation of Puerto Rico: it did not happen. When available land and marine indicators are adequately compared, apparent contradictions in datasets can be evaluated and resolved. When examined in this way, the preponderance of evidence supports the contention that Puerto Rico has been an emergent landmass and has supported terrestrial environments continuously since the latest Eocene.