Could the Dacian fortresses of the Orăștie Mountains see each other?

What is said

A system of fortresses that warned one another#

Dacia was a kingdom north of the Danube, in what is now Romania; Rome conquered it in two wars, in 101–102 and 105–106, under the emperor Trajan, against the last Dacian king, Decebalus. Around its capital, Sarmizegetusa Regia, in the Orăștie Mountains of Transylvania, stand five stone fortresses: Costești-Cetățuie, Costești-Blidaru, Piatra Roșie, Bănița and Căpâlna. The six fortresses have been on the UNESCO World Heritage list together since 1999; before their inscription, the name "the Dacian fortresses of the Orăștie Mountains" referred to four of them, and the more distant Bănița and Căpâlna joined the group when the six were inscribed (Aurora Pețan, 2018). Romania's National Heritage Institute describes them as one defensive system, in which the fortresses were linked by a branching system of guard towers, including between Costești-Cetățuie and the capital.

Archaeologists write cautiously about signals. Of the towers around the Blidaru fortress, their excavators write that their purpose "did not consist only in the observation and signalization of a potential danger" (Pescaru and colleagues, 2014). None of the studies cited here describes the remains of a fire lit as a signal.

Part of the question can be answered by calculation: can one fortress be seen from the walls of the other?

How it is checked

Two neighbours: they see each other#

On a terrain model with 30-metre cells, a straight line is drawn from one fortress to the other, from 6 m above the ground to 6 m above the ground: the height Ioana Oltean and João Fonte use for Dacian walls. If the ground between stays under the line, the two places see each other. The calculation allows for the curve of the Earth and for the way air bends light.

Costești-Cetățuie and Blidaru are 1.6 km apart. The line passes 8.6 m above the bare ground. At that distance 30 m cells are too coarse for a firm answer: when the two ends are moved by a few tens of metres, the line stays clear in 71 of 100 tries.

A finer measurement settles it. On laser-scanned relief, with one-metre cells, Edmond Nogyi (2025) finds that both the Costești-Cetățuie fortress and its towers are visible from Blidaru at the height of a standing person.

From the gate to the capital

The line to Sarmizegetusa runs into the mountain#

Costești-Cetățuie stands at the mouth of the Grădiște valley, the road to the capital. From here to Sarmizegetusa Regia it is 13.6 km in a straight line. The line leaves the walls and after a few kilometres enters the ridges between the valleys.

From Blidaru the ground rises 120 m above the line to the capital, from Piatra Roșie 114 m. No tower is tall enough to make that up, and forest can only add to it.

Every pair

Of 15 pairs, one#

The six fortresses make 15 pairs. In 14 the line is blocked on both surveys of the relief, with or without forest. Even with 12 m towers at both ends, set on any 30 m cell inside the site boundaries recorded in the National Archaeological Record, the ground still rises at least 100 m above the line. That leaves Costești-Cetățuie–Blidaru. Among the four fortresses of the older group (the capital, the two at Costești and Piatra Roșie), five of the six pairs are blocked.

Together with Vârful lui Hulpe and Cucuiș-Dealul Golu, two more fortified places with known positions, that makes eight places. In 27 pairs the two places are more than 3 km apart: none has a clear line, and the least blocked line passes 81 m under the ground.

What is seen

Each fortress as a lamp#

A lamp set on the walls would light exactly the ground that can be seen from there. The lit areas on the map are that ground: every lit cell has a clear line to the fortress.

Blidaru sees 96 km² within 25 km: the valley of the Apa Orașului, where the road comes in, and the hills towards the river Mureș; the archaeologists who dig here also describe a good view over the Grădiște valley and on to the Mureș valley. To the south, towards the capital, the light stops at the first ridge.

The road

The two fortresses at the entrance overlook the lower valley#

The National Heritage Institute also names the main way to the capital: from the Mureș valley, upstream along the river below Costești. The ancient road is not mapped, so the present line of the river stands in for it here: 46 km from the Mureș to the foot of the capital's hill. On bare ground, a person standing on that line could be seen from the walls of Blidaru along about 26 km of it and from those of Costești-Cetățuie along about 22 km: most of the lower valley, from the Mureș to the two fortresses.

Here the two do better than their neighbours. Blidaru sees less ground than most of the hilltops around it, but more of the river than 13 of the 14 hilltops it was compared with; Costești-Cetățuie, more than eight of 12.

Above Blidaru the valley leaves the light. The river runs another 17 km to the foot of the capital, and along 13 to 15 of them, depending on how the line is shifted by a few tens of metres, it cannot be seen from any of the eight fortified places; on today's road, along 12 km of 16. The calculation does not show whether the upper valley was watched in antiquity: it leaves out the towers without an exact position and the mountain road by which, according to the archaeologists working at Blidaru, the capital can be reached from the tower at La Vămi.

The capital

Sarmizegetusa sees only its own valley#

The capital's lamp lights almost nothing but the slope across the valley. The fortress stands on a spur at about a thousand metres, below higher ridges on every side. Paul Pupeză wrote in 2012 that it is "dominated by all the surrounding heights" and that it was probably placed so "not to see the landscape nearby but to be seen from the landscape nearby". Using laser scans of the terrain, Oltean and Fonte found that only part of the surroundings could be seen from it. The lit slopes on the map are both those visible from the walls and those from which the walls could be seen.

The fair comparison is with the neighbours. Set beside 12 similar hilltops around it, the capital's hill ranks third from the bottom by the area of ground it sees. The rest are alike: each of the eight fortified places sees less ground than half of the hilltops it was compared with.

The comparison with other hills

Hills picked at random can lack lines too#

Perhaps these mountains let nobody see far. For comparison, the place of each fortress is replaced, at random, by another hilltop of similar height within a few kilometres, and the lines are computed again, 4000 times.

The networks made this way have on average 2.0 pairs in sight out of the 27, and 16 in 100 have none, like the real one. The absence of lines between the fortresses is therefore not unusual for the relief of these places. Nothing in the calculation shows the sites were chosen to see one another.

One pattern does recur at all eight: each sees less ground than at least half of the hilltops it was compared with. Pupeză had already observed that, although the fortresses stand high, the landforms around them dominate them. The count of ground seen is a separate result, and it points the same way. The calculation does not show why these sites were chosen.

A chain of signals

One post would have been enough. Where would it stand?#

Nobody claims a fire was seen directly across 13 km of mountains; a signalling system means posts in between. How many again depends on the relief. With posts set on the hilltops of these mountains, 6 m above the bare ground, one post is enough between Blidaru and the capital if a signal can be seen 17 km away, and two if it can be seen 8 km away. The two distances are only markers: they come from a trial in Finland with smoke from burning tyres.

The ground from which both Blidaru and Sarmizegetusa Regia are in sight adds up to 5.19 km², in patches on a few ridges: the lit areas on the map. Of the 27 places in the list used here, one touches these patches: Muncel, a fortification that Oltean and Fonte read as a base of the Roman army during the siege. Even there, the line between Muncel and the capital stays uncertain.

None of the Dacian fortifications in the list touches these patches. This does not prove that no post stood there: the list includes only places with published positions, and no exact positions were found for the towers the National Heritage Institute mentions between Costești and the capital. The National Archaeological Record lists La Vămi as a Dacian fortress and Șesul Ciorii as a Dacian tower, both without an exact location. The 17 towers around the Blidaru fortress appear on a plan published by Pescaru and colleagues; on Nogyi's maps they stand less than a kilometre from the fortress. A chain with more posts along another route remains possible; the calculation gives the smallest number of posts for the positions and signal ranges tried.

The other army

Between the Roman camps, the lines are there#

During Trajan's wars against Dacia the Roman army camped on the ridges around the capital: on Comărnicel, at almost 1,900 m, on Vârful lui Pătru, on Jigoru Mare, on Muncel. Among nine such camps (Comărnicel hill, with three camps, is counted through Comărnicel I), 14 of 36 pairs have a clear line from 6 m on bare ground, and 13 from 2.8 m, the height Oltean and Fonte use for Roman camps. Six pass every check: three terrain models, forest, moved positions.

The comparison shows no plan here either: hilltops taken at random in the same places give 9.9 pairs with a clear line on average. The camps stand high, on the crests; the fortresses on lower hills. The calculation records the difference without being able to give it a cause.

The siege

From most camps the capital could not be seen#

Of 11 camps (here the three on Comărnicel are counted one by one), eight have no line to the fortress on Grădiște Hill. From the Comărnicel I camp, 10.2 km away, the ground rises 71 m above the line; Oltean and Fonte reached the same answer on laser-scanned relief.

From the siege camps beside it, Șesului, 1.9 km away, and Muncel, 2.4 km, the model's answer is uncertain; for Șesului, Marcu and Mateescu write that the fortification and the sanctuaries could be observed from it. More unexpected is the camp at Târsa on the Luncani plateau: at 11.7 km the line stays clear on all three models, by at least 5 m. So small a difference needs checking on the ground.

Where the map misleads

Up close, 30-metre cells are too coarse#

Around the Blidaru fortress stand 17 isolated towers. For most of them no positions are published precisely enough, and a few hundred metres apart a model with 30 m cells often leaves the answer uncertain. The map has lines for only one, Poiana Perții, whose position is approximate.

For those there is Nogyi's study on laser-scanned relief: at the height of a person, 10 of the 17 are visible from one point or another of the fortress; from about 6.5 m all would be, if the towers were at least as tall. At that scale the fortress and its towers could see each other. Whether they also signalled, the calculation cannot say.

The forest on the map is the one the radar recorded in recent decades, not that of two thousand years ago; nobody knows how bare the ridges were then, so every line is computed on bare ground and with forest. And a clear line says only that a signal could have been seen.

A blocked line does not say everything either. The calculation assumes walls and towers no more than 12 m high; that height was chosen for the calculation and has not been established at every site. Smoke rises higher: a column over the capital would have to stand between 280 and 340 m tall to show above the ridges from Costești-Cetățuie, Blidaru or Piatra Roșie.