Montserrat makes the beads and St Martin makes the axes

Some islands in the Lesser Antilles worked as factories. At Trants on Montserrat, people made carnelian beads on a scale far beyond what one settlement could wear, and the carnelian itself was not from Montserrat. It came from Antigua, brought in by boat. At Hope Estate on St Martin the specialism was stone celts, cut from greenstone found on the island and sent out to other islands finished. The way to tell a workshop from a customer is the rubbish. Sites like Trants and Hope Estate are full of production waste, the flakes and broken blanks that come off making something. Settlements of the same date elsewhere in the region hold the finished object and none of the debris. Put those two patterns beside each other and a working economy appears: raw material moving one way, finished goods moving another, and communities specialising because they could count on getting what they did not make. All of it was running in the first millennium.

Fighting resumed in 651 and an Arab force of five thousand under Abd Allah b. Sa'd marched on Old Dongola, the capital of Makuria, battering the cathedral with catapults. The peace that followed was set down as a covenant binding Nubians from the frontier of Aswan to the border of Alwa. Neither side would raid the other. Each could pass through the other's territory without settling in it. The Nubians would send back runaway slaves and Muslim fugitives, and would sweep, light and protect the mosque built in the courtyard of their capital. Every year they were to hand over 360 people to the governor at Aswan, male and female, the finest of the country, none of them elderly and none below puberty. Food, cloth and wine came back the other way. A scribe dated the text to Ramadan of the year thirty one. The arrangement held for roughly five hundred years.

By about 900 the Meillacoid tradition had become the dominant pottery of the northern Caribbean, spreading across communities that had previously made quite different things. The vessels themselves are plain: unpolished, decorated with narrow incisions, applied pieces and punched marks, made to be used. What travelled with the style was a way of farming. Meillacoid communities intensified swidden cultivation, clearing and burning plots and working them in rotation, probably with new crops arriving from South America. That shift moved people off the coasts and mangrove edges and up onto hilltops, because the priority had become access to workable soil rather than access to the shore. The sea did not stop mattering. Marine food stayed a large part of the diet, and the villages on the ridges kept sending people down to fish. The picture is of societies converging on a shared set of practices across island boundaries, which is what transculturation looks like in the ground.

At the Canas site in Puerto Rico the Saladoid and Ostiones layers sit one above the other with a clean line between them, and the early conclusion drawn from that line was that two separate cultures had occupied the site in turn. The difficulty with that reading is that nobody could find anywhere for the second culture to have come from. No source region, no migration route. So the material was examined again, and the current view is that Ostionoid pottery grew out of the Saladoid tradition in the islands themselves, with recent argument pointing to the eastern Dominican Republic and the exchanges taking place there. Other evidence supports continuity. Stone toolmaking runs straight through the transition with no equivalent break. What changed was the pottery, and pottery is the loudest thing in an excavation. The change marks the beginning of a period in which the Greater Antilles broke into regions, each developing a look of its own.