01Executive summary
Four milestones from the civilisation that industrialised construction: a cement that sets under water, gravity water supply at metropolitan scale, roads specified to their duty, and water power applied to production.
Greek building is exquisite and structurally conservative. Roman building is comparatively unrefined and structurally transformative, because the Romans had a material the Greeks did not: a pozzolanic cement that hardens into artificial stone, sets under water, and can be cast into any shape a form will hold.
02Pozzolanic concrete: a material that changes what is buildable
Volcanic ash from the region around Pozzuoli, mixed with lime, produces a cement that sets by chemical reaction rather than by drying, hardens under water, and continues to gain strength over long periods. Vitruvius praises it in the first century BC, and the Romans used it to build the great vaults and domes — the Pantheon, the baths of Diocletian and Caracalla, the basilica of Constantine.
Geometry stops being modular
Masonry is assembled from units, so form is constrained by what can be cut and stacked. Cast material takes the shape of its formwork, which makes continuous curved surfaces and monolithic domes possible for the first time.
Skill moves from the mason to the organiser
Placing concrete needs less individual craft than cutting stone and far more logistics, formwork and sequencing. This is the same shift toward the production system that the American System repeats eighteen centuries later.
Underwater construction becomes possible
A cement that sets under water permits harbour works, bridge piers and foundations in places where nothing could previously be built.
Formwork is expensive
This is why the great aqueduct arcades are stone rather than concrete. Casting requires elaborate temporary works; laying stone arches over a long arcade does not repay them.
The last card is the interesting one, and it answers a question that puzzles people looking at Roman remains: if concrete was so good, why are the aqueducts stone? Because the choice between casting and laying is not a contest between a better and worse material. It is a judgement about temporary works, repetition and access, and the Romans made it correctly in both directions.
In the largest domes the aggregate is varied with height — heavier stone low down where stress is greatest, lighter material such as pumice toward the crown where the priority is reducing weight. Varying a material's properties through a structure according to local demand is a sophisticated idea, and it would not reappear as a general design method until functionally graded materials and composite laminates in the twentieth century.
03The aqueducts: an economy that is usually missed
The image of a Roman aqueduct is an arcade marching across a plain. That image is accurate for about an eighth of the total length and wrong for the rest. The aqueducts of Rome followed contours for most of their route, tunnelling through a few ridges and bridging ravines, and came within about 16 kilometres of the city before being carried up onto high structures across the Campagna.
An arcade is enormously expensive per kilometre. A conduit following a contour at a shallow constant gradient is cheap. The Roman achievement is therefore principally a surveying and route-selection achievement: holding a very small, consistent fall over tens of kilometres of irregular ground so that arcades are needed only where genuinely unavoidable. Where a newer aqueduct was built, it was often placed on the substructure of an older one to avoid building a second arcade at all. The monumental part is the part they were trying to minimise.
The same reasoning explains why the Romans made comparatively little use of the inverted siphon that the Hellenistic engineers built at Pergamon. A siphon saves structure and costs pressure, and pressure requires pipe that can contain it. Given ample labour, ample stone and a preference for durability, running at atmospheric pressure along a contour was the better engineering decision for their circumstances.
04Roads specified to their duty
The Appian Way, begun by Appius Claudius in 312 BC and eventually carried to Brindisi, is the first of the Roman arterial roads. What is instructive is not its construction but the fact that it is not typical.
The Romans did not build every road to the standard of the Appian Way. They classified roads according to the traffic expected and built accordingly — which is exactly what a modern road authority does. A mountain path did not receive the effort and expenditure of an arterial highway. Presenting Roman road building as a single uniform specification, applied everywhere, misrepresents a system that was deliberately differentiated.
| Element | Function | Modern equivalent |
|---|---|---|
| Prepared, drained formation | Removes soft material and keeps water out of the subgrade | Subgrade preparation and subsoil drainage |
| Graded layers of decreasing coarseness | Spreads wheel load over a widening area with depth | Subbase and basecourse |
| Wearing surface | Takes abrasion and sheds water | Surface course |
| Camber | Sheds water laterally rather than letting it stand | Crossfall |
| Side drains | Carries water away from the structure | Table drains and kerbed drainage |
Every element on that list is about water. A road fails when water reaches and softens the material beneath it, and load spreading only works if the layers below stay stiff. This is the single most durable piece of knowledge in road engineering, it was understood in the third century BC, and it was substantially forgotten in Europe for a thousand years before Tresaguet, Telford and McAdam recovered it in the eighteenth century.
05Water power, and why it did not go further
The vertical water wheel with a horizontal shaft, geared to a millstone, was introduced across Europe by the Romans in the later centuries of the empire. It ground grain and sawed stone, and the overshot arrangement — using the head of a fall rather than the velocity of a current — was known to them.
So the Romans had mechanical power, and did comparatively little with it. The reason usually given is that an economy resting on slave labour has no incentive to mechanise, and there is real force in that: where muscle is cheap and abundant, a mill is a capital cost solving a problem that is not expensive. It should be said that this explanation is a simplification of a genuinely debated question, and that Roman water milling was more extensive than the older literature allowed. But the broad shape holds. What changed in the Middle Ages was not the invention of the water wheel but the disappearance of the conditions that made it optional.
06Takeaways for current practice
- A castable material changes what geometry is available. The shift from assembling units to forming a monolith is a bigger change than any property improvement.
- Judge casting against laying on temporary works, not on material quality. The Romans chose stone for arcades for a good reason.
- Minimise the monumental part. The aqueducts are a route-selection achievement first and a structural one second.
- Specify to the expected duty. Roman roads were classified by traffic, and a single universal standard would have been waste.
- Everything about a pavement is water. Drainage, camber and a dry subgrade are what make load spreading work.
The modern descendants sit in AS 3600 for concrete, Austroads pavement design guidance and AS/NZS 3500 for water services. Cited by number for orientation only — verify currency.