PREDORE – The baths of a large Roman villa

PREDORE - The baths of a large Roman villa

Dell’esistenza nel sottosuolo di Predore di un grosso edificio di età romana si sapeva da tempo grazie ad una serie di sporadici ritrovamenti risalenti al secolo scorso che avevano portato in luce porzioni di murature e parti di pavimentazioni musive. Solo in tempi più recenti però alcuni interventi eseguiti in concomitanza di attività edilizie hanno permesso di determinarne meglio la natura, l’estensione e l’orizzonte cronologico. Le indagini effettuate tra il 1998 e il 2001 presso l’Albergo dell’Angelo e in via Roma, pur nella loro limitata estensione, avevano già fornito preziose informazioni sulla storia dell’edificio e documentato la ricchezza del suo apparato decorativo.

Nel 2003 un grosso intervento edile volto alla riqualificazione della vasta area occupata dall’ex fabbrica Lanza Gomme ha permesso di effettuare, secondo metodologie archeologiche, una indagine estensiva ed esaustiva di un’area di circa 1.000 mq pertinente all’edificio romano che si confermava essere una villa di notevoli dimensioni, circa 15.000 mq allo stato attuale delle conoscenze. Infatti la prima importante informazione che è stato possibile ricavare, unendo i nuovi dati a quelli già in nostro possesso, è stata proprio la delimitazione abbastanza precisa dell’estensione dell’edificio; i limiti occidentale ed orientale sono dati da due corsi d’acqua: il torrente Rino ad ovest e quello che scende dalla valle Muradella ad est; a sud la villa era prospiciente il lago, la cui riva all’epoca era circa 70 m più vicina dell’attuale, mentre a nord terminava ai piedi del monte e, quindi, all’incirca poco oltre l’attuale strada provinciale.

The existence of a large Roman building under the town of Predore had been know for some time thanks to a series of chance discoveries during the 20th century that brought to light stretches of wall and areas of mosaic flooring. Only more recently have investigations conducted in advance of building work yielded a clearer picture of its nature, extension and chronology. Excavations carried out between 1998 and 2001 at the Albergo dell’Angelo and in Via Roma, although limited in size, furnished valuable information regarding the building’s history and revealed the opulence of its decoration.  

In 2003 a large-scale construction project to redevelop the area previously occupied by the Lanza Gomme factory allowed the careful open-area archaeological excavation of about 1000 m2 of the Roman building. This was a villa of considerable size: according to current estimates it covered around 15,000 m2. The first important information revealed by putting together the new results with what was already known was a fairly precise indication of the property’s limits; those to the west and east were determined by two watercourses, the Rino (to the west) and the stream from the Muradella Valley (to the east). To the south the villa faced the lake, the shore of which was about 70 m further inland than at present, and northwards it finished at the foot of the mountain (where today the main road passes).

The excavation unearthed the western portion of the residential complex; half the area was occupied by the baths, while in the other 500 m2 was an open area that contained unmistakeable evidence of building activities – such as the preparation of mortar and plaster – related to the construction and maintenance of the property. It is interesting to note that the area retained this function throughout the building’s lengthy and varied use.

To the west the physical boundary of the complex was a large wall, originally at least two metres high and made of split stone and mortar, that stood on the eastern bank of the River Rino. To the west of the stream no barrier was present, probably to allow it to freely overflow its western bank in times of high water volume. 

Four main building phases were distinguished, spanning from the 1st century BC until the 4th AD, some of which involved substantial modification of the layout and functions of this part of the villa. 

The first phase, dated to the 1st century BC by stratigraphic means and examination of the finds recovered, was only partially brought to light, since at times it had been incorporated within or concealed by later building. It is represented by a number of structures, in particular the western outer wall: notable for its lack of linearity, with several changes of direction, the northern part of this structure remained in use until the 4th century. Other Phase 1 walls (which unfortunately only survived at foundation level due to the presence of a factory basement) were found to the south-east, and demonstrated the existence of a wing with north-south/east-west orientation. The construction technique – as also in later centuries – involved the use of brick and tile as well as stone; it is not clear whether aesthetic or practical considerations determined the specific choices made. The limited information currently available does not permit a reliable interpretation of the function of this portion of the building, but it may have belonged to the baths complex. 

Similarly, few components could be assigned to the next phase, of 1st century AD date: just part of two north-south aligned rooms, the southernmost of which ends with a semicircular apse in which part of the floor preparation in grey mortar still survives.

Most of the remains may be assigned to Phase 3 (2nd-3rd century AD), to which the most substantial architectural interventions belong; this probably coincided with the decision to construct a baths suite in this part of the villa.

The centrally-heated rooms are followed by other components laid out in an L-shaped itinerary, which terminates to the south-west with a large pool, in accordance with the arrangement described in classical writings (and also seen in the Lake Garda villas quite nearby). 

The four hypocausts found were fed by a single praefurnium that was accessed from a vaulted underground room to the north. Prolonged use of these facilities is indicated by the layer of ash around 1 metre thick that was found inside the praefurnium; this was studied by the Como Museum Archaeobotanical Laboratory and yielded interesting information regarding the tree species present in the area at that time. The specialists have emphasized the importance of the traces of fig and olive that were found – particularly the latter, because this evidence from Predore is the earliest sign of cultivation of the olive in the pre-Alpine lake zone. The hypocausts, which still retain a good proportion of their regularly-spaced circular pilae, were associated with rooms of various sizes.

The most significant changes, though, concerned the south-western area, where the pre-existing rooms were demolished to make way for the tepidarium - frigidarium sequence consisting of two chambers followed by a large pool (natatio). The first was circular with a diameter of about 5 metres and served as a junction with the new wing, the axis of which was rotated eastwards by about 15°. The reasons for this choice – which created some architectural difficulty with regard to connection with the rest of the complex that followed the old alignment – are not clear. The second room was rectangular (4.5 x 2.5 m), with floor and walls covered with stone slabs: some are still in situ and traces of others may be seen in the hydraulic mortar (cocciopesto); the remains of lead pipes and an impression left on the floor surface indicate the presence of a fountain in the north-western corner. To the south this room opened onto a large pool (5 metres wide and more than 10 metres long) which extended to the south-south-west; it was about 1 metre deep and the walls and bottom were covered with large squared slabs of a white local limestone set in well-made hydraulic mortar.

The construction of the baths suite necessarily involved the creation of a series of water channels, generally made with brick/tile bases and covers, and stone-built side walls. One of these contained a brick with one side covered in graffiti, probably produced as a scholastic exercise. A study conducted by Dr. C. Molle revealed the existence of a sequence of letters, numbers and other signs of uncertain meaning, together with two inscriptions in cursive script which may be dated by  palaeography to the late 2nd – early 3rd century AD. Of particular interest is the more extensive piece, seven rows long and apparently part of a composition in which the god Neptune is referred to.


The building’s last period of use (Phase 4) dates to the 4th century, and corresponds to a general structural and functional continuity in this part of the villa; the most evident change is the substitution of the circular room by a rectangular one equipped with a fountain. The north-western hypocaust fell into disuse; the aperture leading to it was sealed and covered over with rubble that included pieces of polychrome mosaic and many stucco fragments. The ash deposit inside the praefurnium had in the meantime built up to such an extent that it blocked the access room; for this reason a new point of access was created in the north-east corner of the room, a new entrance with descending stairs made with re-used worked stone. It is not clear why so much fuel residue was allowed to accumulate and not simply removed; presumably this was due to structural modifications to the west of the hypocaust in the zone where construction of the factory basement has almost completely removed the ancient walls and stratigraphy.

Subsequently the villa – as demonstrated by the lack of layers created by widespread burning or demolition and the scarcity of  moveable finds – underwent a slow process of progressive abandonment. It is quite possible that (as is often seen in similar situations) parts of the building continued to be used in this following period.

Since April 2012 the site, enhanced by conservation treatment and development and equipped with a small museum, has been open to the public.


Angelo Ghiroldi, Archaeologist