Areas that became densely set with trace italienne fortresses [Wiki] – particularly the towns of the Low Countries – became almost impossible to conquer. The Army of Flanders [Wiki], then arguably the finest in Europe, tried for eighty years to subdue the Dutch and largely failed: the cost of endless sieges of trace italienne fortified towns made the task effectively hopeless. Once again, I don’t want to imply that this is the only factor (it isn’t, by any means), but after the long series of remarkably decisive wars from 1450 to 1550, the trace italienne contributed to the frustratingly inconclusive (but expensive and bloody) wars of the following century.

At the same time, this shift back towards defensive stalemate didn’t lead towards more fragmentation – as the castle had – because trace italienne fortresses were too expensive for any individual aristocrat to build merely to protect his house. Not merely because of the massive fortifications themselves, but the large garrisons they required, either of full-time soldiers or town militia, which were simply beyond the resources of what was left of the old medieval aristocracy. Indeed, warfare in this period (both offensive and defensive) largely proved to be beyond the resources of Europe’s newly centralizing states; the Spanish crown, for instance, went bankrupt, primarily from military expenses, in 1557, 1560, 1575, 1596, 1607, 1627, 1647 and 1653.

It is part and parcel of the era that as much as building and defending a trace italienne fortress was an exercise in mathematics, so was attacking one. Because the entire point of the fortress is to project firepower – often at great distances (with cannon) – the attacker cannot simply form up outside or just set up their artillery in an open field and begin firing; they’d be cut up by counter-battery fire before they had gotten very far. So the attacker had to set up their own earthworks, which naturally being based on the same principles and weapons as the defender’s, would resemble them.

First, the attacking army would generally set up its own fortified camp, outside of cannon-shot. Because that camp needed to be resistant to enemy attack either from the garrison of a relieving army, it was likely to be built in the same style as a trace italienne fortification, albeit with earthworks and gabions (wicker baskets filled with earth) in place of the heavier stonework of the fort. The attacker then has to isolate their target from help, preventing the defenders from leaving, or supplies or reinforcements from arriving. The act of enclosing a besieged settlement in a wall is called circumvallation (recall our word vallum there; this is “walling around”). The act of constructing a line of outward facing defenses to defend that line from attack by a relieving army from behind is called contravallation. Armies in this period would do both.

The first such line of defenses was often started from the fortified camps and was built effectively out of range of the fort; this was the “first parallel”. Now, by the time this project was well along, both the defenders and the attackers could do some basic calculations. They both might have a sense of how quickly reinforcements could arrive and if they were likely to be large enough to give battle or lift the siege. Both sides also know how long their supplies will last and can probably have a decent guess of how long their opponent’s supplies will last. And at this point, they can both calculate fairly well how many meters of earthwork and trench the attackers can dig per day and how many they need to dig to complete their siege operations.

But – and this is important – no one wants this siege to come down to its conclusive, final assault. The attackers don’t want this: every day their siege army sits out here, it is eating money such that each day wasted besieging this town limits what this army can accomplish overall before supplies and money run out. Moreover, the actual assault is likely to incur very high losses on the attackers, because even with a breach, they have to cross all of that open, fire-swept ground and then force the breach in close combat – and the defenders will know exactly where they are going days in advance. On the other hand, if the defenders make the attackers go through all of that effort, danger and death and lose – and they must assume the attackers wouldn’t have laid siege if they weren’t confident they had enough men to take the breach when it comes to it – then the city would be looted, its populace raped or massacred as the attackers vent their rage on the city. This was, at the time, considered the normal result of holding out, to the point that it seems to have been general practice that it was appropriate to wait something like three days before beginning the process of getting control of an army that had breached a city in this fashion.

So once the attacker has completed that first parallel, he is going to send a message to the garrison, announcing that he has done so and offering them the chance to surrender. The standard terms for such surrender was called the “honors of war” – normally the defenders would be permitted to march out, with its flags flying, bayonets fixed, matches (for their matchlock muskets) lot on both ends and “ball in mouth” (that is, a musketball held in their mouth ready to be swiftly loaded as protection against treachery). The defeated army was generally required to turn over its arms, but generally allowed to march back to their own territory. And finally, a town that surrenders like this might have to pay a ransom, but ought to be immune from pillage. To modern readers, these sorts of rituals seem quaint, like having a battle break for tea in the midafternoon (a thing that, to be clear, did not happen), but in fact these were fairly hard-nosed considerations: generous surrender terms aimed to induce a garrison or town to surrender and so spare the attacker both the time but also the blood of storming the place, because after all the goal here isn’t to destroy the enemy garrison but to get control of the town.

One that first parallel is completed, assuming the defender, upon doing the math, doesn’t decide the matter is hopeless and just give up, the attacker now proceeds to begin digging his works forward to set up advanced firing positions (most of the digging was done at night when enemy artillery couldn’t accurately fire at the work parties); there were typically three parallels, the first two providing protection to the digging works and the third parallel, dug into the glacis, providing the firing position directly into the structure of the fort. Once the cannon were emplaced, the attacker might again report that he was prepared to begin firing but if the defenders would just surrender, we could all skip that nasty business. If not, the attacker’s cannon – now secured in their own trace italienne-style firing positions – would start working on the enemy fortifications in a gunnery duel (since the defender’s guns are firing back). Progress here would mean both demolishing ravelins and other outworks which shielded the main curtain wall, as well as storming assaults on various outworks to take them and deny them to the enemy (Americans may note effectively all of these steps in the Siege of Yorktown (1781), albeit on fast-forward since this was a siege of a fairly small town defended merely by field fortifications rather than a large trace italienne fortress complex). The final option to surrender came once the curtain wall had been breached (there was little use in trying to “rush” the breach since it would take days to produce so everyone knew where the final fight would be). The defender then had the terrible choice of either surrendering or risking the terrible slaughter that would follow if the breach was forced (as it was very likely to be).

This whole process could take a really long time and it involved a lot of digging, but it was almost mechanistic in its push towards success. In a sense, those demands forced European states to develop the sort of state capacity that had been the norm under Rome or in China in order to run this new version of the “big army” siege playbook which demanded such tremendous amounts of work. As I’ve noted elsewhere, sieges of these sorts against large trace italienne complexes could last a long time; Parker (Military Revolution, 13) notes the siege of Breda in 1624 lasted nine months and was fairly short while the siege at Ostend in Flanders in 1601 lasted three years and was fairly long, to give a sense of the range.

Bret Devereaux, “Collections: Fortification, Part IV: French Guns and Italian Lines”, A Collection of Unmitigated Pedantry, 2021-12-17.