Discovered by de Robien Vers 1750 (≈ 1750)
First identification of aqueduct.
XIXe siècle
Study by de Closmadeuc
Study by de Closmadeuc XIXe siècle (≈ 1865)
In-depth analysis of the remains.
16 avril 2002
Protection of remains
Protection of remains 16 avril 2002 (≈ 2002)
Registration for Historic Monuments.
Aujourd'hui
Aujourd'hui
Aujourd'hui Aujourd'hui (≈ 2025)
Position de référence.
Heritage classified
The remains of the entire aqueduct, the fifteen piles and the masonry ramp, plus the ground (ground and basement) limited by a rectangle of the length of the work still visible, 106 metres, and extending in width up to two metres on either side of the piles of the work (Box C 589): inscription by order of 16 April 2002
Key figures
De Robien - Discoverer
Identifies the water supply around 1750.
De Closmadeuc - Researcher
Studyed in the 19th century.
Origin and history
The Gallo-Roman aqueduct from Rosnarho to Crac Its ruins, composed of an alignment of masonry piles descending to the Auray River, originally formed a 440-metre-long water bridge. The main objective of this construction was to supply water to the ancient agglomeration of Locmariaquer, located nearby. The main structure consisted of eleven arches resting on twelve piles installed in the river bed, now largely destroyed to expand the channel.
Downstream, the bridge extended about 160 metres before the pipeline sank into the ground. Of the 25 initial arches, spaced 2.50 metres apart, 15 remain, with heights ranging from 0.80 to 1.60 metres. The ramp marking the western end of the structure is also preserved. Since the 18th century, the remains have been gradually degraded, now reaching an advanced state of disrepair. This aqueduct is the only known example of Gallo-Roman aqueduct in Armorica, highlighting its historical and archaeological importance.
The aqueduct was studied extensively in the 19th century by de Closmadeuc, after its initial discovery by de Robien. In 2002, the remaining remains, including the fifteen batteries and the masonry ramp, as well as the surrounding terrain, were protected by a registration order under the Historic Monuments. The location of the work, although partially preserved, is estimated to be of poor accuracy (level 5 in 10), based on available data.
