Peter Heather

So, for instance, the Goths who are on the lower Danube opposite Romania and Bulgaria now, they can get better terms out of the Emperor Valens in the late 360s because the Persians are getting uppity in the east.

He needs to go and fight them.

So he's in the middle of a Gothic war, but the Persians are much more important.

So he'll do a deal.

that gives the Goths something of what they want in order to extract his armies and fight Persia.

So they need that kind of a thing to go in their favor.

There is nothing that we can see.

Yeah.

And this is, in fact, the colossal new data set that's become available since the 1980s.

Rural surveying.

You can date Roman pottery to within a decade, its progressions.

And we know from careful sampling how dense a concentration of surface pottery means a settlement underneath.

So rural surveying, very boring.

You go move forward a meter, collect everything, put it in a plastic bag.

No.

It's on the surface.

Yeah, tractors mean that modern playing techniques pull it all up.

So you just go forward, collect it all, look at it.

And what's emerged from that is not only where there are settlements in the Roman period, but when they're there, because you can date them to within a decade.

And the staggering, or it would have been to my older colleagues' fact, we're sort of getting used to it but still thinking about it now, that has emerged from all of that is that across the vast majority of this imperial landscape, the period of maximum rural population and rural productivity is the 4th century.