An alternative method is by looking at the archaeological extent of the city and its suburbs and their possible population density. From this it is possible to produce a simple estimate of the total area of settlement and to suggest the residential areas as opposed to the commercial, public or open areas of the settlement. In London, this could be done at three main periods - AD60/61, AD100/120 and about AD200.
When the town was totally destroyed in AD60/61 by Boudica, the settlement lay mainly on the eastern hill (now Cornhill) with some occupation on the western hill, the suburb to the east, along the waterfront and in Southwark.
This area has been calculated to be 44.85 hectares. With the removal of nearly 6 hectares of non-residential areas and areas of lower-density housing, the estimated area is 39 hectares.
In AD100/120, the town had grown with extensive residential occupation of both hills, a developed waterfront zone and areas in Southwark. From the later 1st and early 2nd century, the style of occupation can be shown as a rough grid of rectangular city blocks with long narrow strip buildings, back yards and town houses.
These residential and commercial blocks were divided by several large buildings complexes including the forum and basilica, amphitheatre, public baths and temples.
Major topographic features, such as the Walbrook valley and steep hillsides would have also influenced the scale and density of occupation while the waterfront area would have had reduced domestic occupation.
The town now covered an area of some 135 hectares including 34 hectares of supposed non-residential and lower density housing. With those removed, there would be an estimated area of almost 101 hectares.
The construction of the city wall in about AD200 delineated the extensive area of occupation within the walled area and in Southwark. This amounted to nearly 163 hectares. With public buildings and less densely occupied areas amounting to about 44 hectares, the estimated area of residential occupation was just over 118 hectares.
Pompeian examples have suggested 6-8 people per property or 9-12 people if there was an upper storey. Pompeian models have also calculated 1 person per 34-45 square metres producing an estimated population for Pompeii of about 10,000.
It is possible, therefore, to produce an estimate based upon calculations of the numbers of people per hectare within residential areas using the estimate of people per hectare of residential space multiplied by the estimate of total residential space.
Alternatively, an estimate could be based on a more general estimate of urban populations using the estimate of people per hectare multiplied by the estimate of total urban space.
A population figure could then be based on the average of these two estimates. A possible suggestion of a population in AD60/61, therefore, would be about 10,000; a population of about 25,000 – 30,000 in AD100/120 and, although the area in AD200 is calculated as being larger, property density was falling so a slightly smaller population of about 25,000 might be suggested for AD200.
This should only serve as a guide, however, as it will never be possible to produce a reliable and accurate population figure for Roman London.
For a more detailed explanation of these estimates, read Swain H & Williams T, ‘The Population of Roman London’ in Clark, Cotton, Hall, Swain & Sherris (Eds), Londinium & Beyond CBA Research Report 156 (2008), 33-40.