Feeding Rome

Somewhere in the second century, a clerk on a quay beside the Tiber picked up a brush and wrote on the shoulder of a pottery jar. He wrote the weight of the empty vessel, the weight of the oil inside it, the name of the merchant whose oil it was, where in southern Spain it had come from, and the name of the inspector who had checked it, with a date.

Then the oil was decanted, and the jar — too greasy inside ever to reuse, too heavy to be worth shipping back empty — was carried a short distance and thrown on a heap.

Do that for a hundred and fifty years and the heap becomes a hill. Monte Testaccio stands about thirty-five meters high and runs roughly seven hundred meters around, and it is made of nothing but broken jars: on the usual reconstruction, something on the order of fifty-three million of them. The sherds were not dumped at random. They were stacked in courses by workers who arranged them for stability and dusted the layers with lime to keep the smell down, which means the hill is a piece of civil engineering as well as a rubbish tip. And because so many of them still carry that clerk's brushwork, it is also an archive: a stratified, dated, named record of who shipped what to Rome, and when.

That is the kind of evidence this subject runs on, and it is why the book exists.

A Million People, No Farmland, Every Day

Rome at its height held somewhere near a million people. The figure is contested — competent scholars put it anywhere from well under that to somewhat above, and the book spends its first chapter showing you the five different methods they use and why the methods disagree. But whatever the number, the surrounding countryside never came close to feeding it, and nobody in Rome expected it to.

So the food came by sea. Grain from Sicily first, then Africa, then Egypt, in ships that ran from modest coasters up to Alexandrian carriers of three or four hundred tons, with a few giants that may have topped a thousand. Working from the population and the ration, the annual requirement comes out somewhere around two hundred thousand metric tons — and once you divide that by what a hull could carry and by the two hundred and thirty days a year the Mediterranean was open to large-scale sailing, you get the number that makes the whole thing real:

Two or three loaded grain ships arriving, every day, all season, for four centuries.

And that is only the sea leg. At the coast everything had to come out of the seagoing hulls and into river barges, because Rome's own river mouth could not take a laden merchantman. The barges were then hauled up the Tiber against the current by men and oxen walking a towpath — the most laborious step in the entire chain, and one that no amount of imperial engineering ever solved. At the Emporium the cargo was unloaded, weighed, sampled and inspected again, and put into warehouses whose raised floors and ventilation slots were designed to keep grain alive in storage.

Two hundred thousand citizens then collected a free monthly ration, five modii of wheat, on a schedule, at numbered gates, on the strength of a token that could be inherited and sold. The whole apparatus had a name — the annona — and the official who ran it held one of the most dangerous posts in the empire, because when it failed the failure was visible to a million people on the same morning.

Why I Wrote This

I went looking for a book about how Rome actually ate and could not find one. The subject lives in academic economic history and excavation reports, and in the popular Roman books it gets a paragraph about bread and circuses — usually a slightly sneering one, as though a free grain ration were a decadence rather than the largest logistical undertaking in the pre-modern world.

What changed it for me was Monte Testaccio. A hill of labeled containers is a strange kind of monument: nobody built it on purpose, nobody meant it to survive, and it records the operation more honestly than any triumphal arch does. Once I understood that the hill is what four centuries of routine looks like when it piles up, the routine became the subject.

The other thing that kept me going was the arithmetic. Most of what we say about the ancient world rests on judgment about texts. This subject can be counted — not precisely, and the book is careful about how imprecisely, but counted. You can start with a population estimate and a ration and work forward to a number of ships on a given day, and then go and look at a harbor built to receive exactly that many. That is a rare and satisfying thing in ancient history and I wanted to show a reader how it is done.

Frequently Asked Questions

Do I need to know Roman history to read this?

No. The empire, the provinces, the offices and the sources are explained where they first matter. The book assumes you know Rome was large and are curious how it worked.

How much of this is numbers?

Enough to make the argument, presented as pictures rather than tables. The house rule throughout is ships per day rather than tonnes per annum, loaves per oven rather than throughput. There is a full appendix carrying the working — every figure with the assumption chain that produces it — so the chapters do not have to stop and show it twice.

Is Rome's population really a million?

Probably close, and honestly contested. Chapter 1 gives five independent estimation methods, what each one actually measures, and why they disagree. The book uses a working figure of about a million and says plainly that it is a reconstruction, not a census.

Does the book draw modern comparisons?

No, deliberately. Readers who work in operations or logistics tend to enjoy this subject, and the temptation to flatter them with a modern parallel is constant. The book refuses it. The Roman operation is interesting on its own terms, and the comparisons a reader makes for themselves are worth more than any I could supply.

What's the source policy?

Every ancient author is paraphrased rather than quoted, including Lucian's famous description of the grain ship Isis, which is the single most-lifted passage in this whole field. The Roman legal texts on shipping liability are summarized in plain English. No papyrus, inscription or jar label is reproduced as quoted text — the appendix that teaches you to read an amphora describes the fields instead.

How long is it?

Around 78,000 words: an introduction, fourteen chapters in four parts, a conclusion, and three appendices — the full numerical working, a guide to reading an amphora, and the grain year month by month. The paperback runs 251 pages.

If You Liked This, You Might Like

• The Ancient World's Greatest Engineers — The direct companion. Same approach, same register, different system: that book is about what the ancient world built, this one is about what it ran.
• The Height of Empire — The Rome this operation was feeding, at its fullest extent, described from the political side.
• The Fall of Rome — Where the last chapter of this book leads. The aqueducts cut, the population gone, and a city living inside the plant of a much larger one.

A hill of fifty-three million labeled jars, two or three ships a day for four hundred years, and a city that was always exactly as large as its supply lines allowed.