The earliest dated Saladoid sites cluster in the northern islands

For a long time the Lesser Antilles were described as stepping stones, a chain the first farmers crossed island by island on a slow walk north. The radiocarbon dates complicate that picture. The earliest well dated Early Ceramic Age sites cluster in the northern part of the archipelago. Vieques, Puerto Rico, the Virgin Islands, Montserrat, Antigua and St Martin have all produced dates showing established colonies in the window from about 800 to 200 BCE, while many of the southern islands on the supposed route have produced nothing so early. One reading is that the first Saladoid migrants passed the southern islands by and settled in the north. Another is that archaeologists have simply dug harder in the places where the early dates keep turning up. Keegan and Hofman take the second possibility seriously and still conclude that the expansion through the chain went faster than anyone had previously proposed. Hispaniola sits at the far end of that fast run.

A Greek commercial handbook of about 340 BC, later wrongly credited to Skylax, placed the island of Cerne twelve days' sail from the Straits of Gibraltar and said no ship could get past it for shallows, mud and weed. Phoenician merchants moored there, pitched tents on the island and ferried goods to the mainland in small boats. They took away hides of deer, lion, leopard and cattle, and the hides and tusks of elephants. The people they dealt with used ivory ornaments and cups, put ivory anklets on their women and ivory decoration on their horses, drove chariots, and made a great deal of wine from vines, which the merchants also carried off. In return came perfume, Egyptian stone, Attic pottery and jars. The tents suggest that the permanent colony Hanno had planted on Cerne was already abandoned. Cerne has never been located.

Excavations at the Corinthia III site in Cuba turned up some of the earliest known Archaic Age pottery in the Caribbean. The find matters because the standard scheme awards pottery to the Ceramic Age migrations that arrived later out of South America, with Archaic peoples cast as foragers who did without it. The Cuban vessels show people working clay well before those migrations reached the island, alongside a move towards more settled living in the Late Archaic. Keegan and Hofman use the case against the period boundaries themselves. Archaeological ages are labels applied afterwards by researchers dividing up a continuum, and the material keeps crossing the lines drawn for it. What the Corinthia III pottery records is a wide range of practice across the archipelago, with communities inventing and adopting at their own pace. The people of the islands were doing their own experimenting, on their own schedule, for centuries before anyone brought them anything.

By about 250 BCE, on the radiocarbon evidence Rouse reviews, the Cedrosan Saladoid migration had reached the northern West Indies. The colonists were selective about where they stopped. They favoured high, mountainous islands, which hold rain, carry rivers and support the luxuriant forest their farming needed. Flat, dry islands were passed over. That preference explains a good deal of later Caribbean settlement, including on Hispaniola, whose mountains and well watered valleys made it one of the most heavily populated islands in the region by 1492. This is farming logic applied to an archipelago by people who had crossed open water to reach it. They knew which land would grow their crops and they chose accordingly, across a chain more than a thousand kilometres long. The indigenous history of Haiti begins with decisions of this kind, taken two thousand years before the Spanish arrived and recorded now in charcoal, potsherds and dates.