Coral files and conch tips fill the tool kit at Krum Bay

The Krum Bay complex on St Thomas gives the fullest picture available of Archaic Age material life in the Virgin Islands. The deposits hold flaked igneous stone, hammer stones made from pebbles, beads cut from shell, and red ochre, which had a symbolic use or a practical one and possibly both. The marine component is the striking part. Files made from coral, and tips taken from the columella of Lobatus, the central pillar of the conch shell, show people turning the reef itself into a set of tools for working other materials. Archaeologists reading the variation between Krum Bay and neighbouring complexes suspect they are looking at one subsistence and settlement system with different tasks carried out in different places. Early maritime hunters and gatherers in these islands worked several island niches at once, with a tool kit assembled for each, two thousand years before the ceramic villages arrived.

African potters commonly fire their work in pits dug into the ground. Across the woodland savanna that runs from West Africa to the eastern edge of the Central African Republic, the ground is lateritic and heavy with iron oxide. The archaeologist Merrick Posnansky proposed that firing at high temperature in such pits would have produced small pieces of iron out of the soil itself, repeatedly, until somebody noticed. That is a plausible route to the discovery, and it puts the origin of the technology inside a craft that women had run for thousands of years. What happened next is a familiar sequence. Men took up smelting, built the rituals and symbols that assigned the work to them, and made the arrangement look ancient and proper. Women kept essential roles regardless, mining the ore, and in southern Africa and probably elsewhere shaping the ceramic tuyeres that carried air into the furnaces and the crucibles used for copper. The foundation was women's. The credit went elsewhere.

Around 1500 BCE a distinct ceramic tradition, which archaeologists call the Barrancoid series, developed in the middle and lower stretches of the Orinoco valley in what is now Venezuela. Over time its makers moved downriver toward the coast. The people already holding that ground, the communities Rouse files under the Saladero heading, were pushed out of their positions along the river, and out of that disturbance came the subseries he named Cedrosan Saladoid. This is upstream bookkeeping that matters a great deal for Haiti. The Cedrosan Saladoid are the people who a thousand years later carried farming, pottery and village life out of South America and into the Antilles, island by island, until they reached Hispaniola. The population Columbus met in 1492 descended from them. So the chain that ends in the Taino villages of the north coast begins with a pottery style shifting position on a South American river three thousand years earlier.

Jolly Beach sits on the west coast of Antigua, looking out over reefs and a productive stretch of shallow sea. Around 2000 BCE the people there were making sophisticated flint blades, which sets the site apart from the general run of early stone working in the islands. Their diet was wide. They took eight major species of mollusc, fish from shallow water, sea turtles and manatees. They also took far more land animals than the other coastal sites on Antigua from the same period, and that is the second thing marking Jolly Beach out. Better stone technology and a broader food base together suggest a place people used for longer stretches, or for more purposes, than the camps around it. Two settlements on one small island, a short walk apart, could run on different economies. Archaic populations in the Caribbean read the ground in front of them and adapted site by site.