When it comes to maintaining a population, the primary factor is food. I was lucky enough to go on a trip to Rome last week, and during that time visited the port city of Ostia. Much of what I'm writing about is direct from the site, but the book Ancient Ostia: Past and Present (Giuseppina Lauro) is a great substitute for understanding the history of the city.

Ostia was the primary trade link by sea for Rome, situated a short distance away on the coast. The primary trade in terms of food was Egyptian grain, with luxuries alongside. This was an ultimately huge level of importation that saw Ostia itself grow from a small seaside town in 400BC to the huge trade port it was in 150AD. Somewhat luckily, the port was simply abandoned after the fall of Rome and the collapse of its Mediterranean Sea trade empire. The harbour silted up, preserving the site beautifully for us today.

The roads that have been uncovered show deep ruts in the stone from significant cart traffic, as well as a large portion of the city built around being given over to warehousing built in the 1st Century AD. Combined with the large apartments owned by the wealthy merchants in the city, it shows a roaring trade coming through the ports. The Square of the Guilds near the theater housed the headquarters of several trade corporations, with mosaic logos showing various foodstuffs.

I'm still trying to find records of the exact quantities of grain moved through Ostia, but its sheer size and status as Rome's primary port points to a significant amount.

If you ever find yourself making the trip to Rome, I would recommend Ostia if history is your passion. There are few other sites where one can walk freely through almost an entire Ancient Roman city that is so well preserved.

The real answer here is immigration. Yes, the city's location was particularly well suited to growth, being on the convergence of the north-south land route and the east-west river route in central Italy. Yes, as with every major city at any time, the city rapidly outgrew the carrying capacity of its immediate hinterland, and so growth was only possible through the mass importation of grain from overseas sources, which eventually became a stable supply in the early Principate. But what made the city's population sustainable was immigration. Even the most conservative estimates of urban mortality at Rome are staggering. It's quite clear that the city's population was not self-replicating, and that more people were dying within the city than being born within it. The population of the city remained steady and sustainable--indeed, even probably continued to grow during the first centuries BC and AD--because people from Italy and, eventually, the provinces were entering the city in enormous numbers. Not all of these immigrants were brought freely. While slaves remained a minority in the city, their numbers were not demographically insignificant, and freedmen made up a significant chunk of the free population, though we do not have clear numbers. Suffice to say that foreign populations, particularly slaves and freedmen, who are often surprisingly visible in our evidence (and certainly over-emphasized in epigraphy) were prevalent enough as to be constantly mentioned by ancient authors, with varying degrees of acceptance.

Various factors contributed to the growth and establishment of a steady population in the city, but at the end of the day none of that really matters very much for the sustainability of the urban population. The fact of the matter is that location, grain, and building programs notwithstanding, the urban population was not sustainable. Mortality was simply too high. As was typical in premodern megalopoleis, the only way the city could sustain its population, which was regularly bleeding to death, was by migration of outside populations into the city.

Before tackling the supply of such a metropolis I'd query the million: Glenn R Storey argues persuasively against the likelihood of a number over half as great: The population of ancient Rome, Antiquity 71:274 (Dec 1997) 966-78. Such a low count in turn halves the challenge of provisioning the city, which as Scheidel observes (Roman population size: the logic of the debate, here) drew upon the produce of coastal regions of much of the Mediterranean. Seen in the context of an Empire-wide urban population that Scheidel elsewhere puts at eight million or more, Rome sits at the apex of a hierarchy of tributary and trading relationships that channelled the surplus output of provinces like Egypt, north Africa or Sicily to the cities, towns and grain-deficit regions: the extent of that surplus is all the more remarkable when we realise that Rome's own consumption (even according to the higher count) must have represented only a fraction of it.