Paleogeography of the Aves Ridge and its potential role as a bio-colonization pathway linking South America and the Greater Antilles in the mid-Cenozoic

The Aves Ridge in the eastern Caribbean Sea is thought by many to have operated as a short-lived, mid-Cenozoic, dry-land path that was used by the ancestors of a number of South America-derived faunal groups to colonize the Greater Antilles. The idea is, however, contentious and has triggered much debate. Here, we review the various models, notably GAARlandia, focusing on mechanisms to account for the emergence of the Aves Ridge causeway, the time window over which it operated, where along northern South America its southern end abutted, and its portrayal in written descriptions and on maps. In response to perceived weaknesses, a new paleogeographical scenario is presented, which draws upon a sizeable body of literature including several recent publications. We suggest that from the start of the Late Middle Eocene (41.2 Mya) through to the late Early Miocene (c. 17 Mya) the Aves Ridge's crest was topped with scattered carbonate platforms and coral atolls, but these were separated from one another by 30- to 70-km-wide gaps. During this time period we find no support for high-elevation islands, with their accompanying diverse habitats. Instead, suggested analogues for what we term the 'Aves Ridge Reef Belt' are flat, atoll-like islands with limited ecospaces such as those currently in the southeast Bahama Banks (western North Atlantic) and the Spratly Islands (southern South China Sea). We also address the likelihood of 'stepping-stone' colonization whereby taxa are thought to advance from a source continent to a remote landmass or archipelago via a suite of closely-spaced islands. We argue that the mechanism has a much lower probability compared with direct (over-water) dispersal, and that it played no appreciable part in seeding the Greater Antillean land-vertebrate assemblages, either via the Aves Ridge or the nearby Lesser Antilles chain. We show that phylogenies of living species support direct dispersal rather than stepping-stone transfer.