In February 1988 the remains of a Roman building were identified in the municipality of Pollena Trocchia (on the north side of the Somma-Vesuvius, probably part of the territory of Naples in antiquity), in the locality of Masseria De Carolis. At the time the area was used for quarrying volcanic material, used for construction in the vicinity. Despite the fact that vast damage had been caused to the ancient walls by the quarrying machinery, it was possible to identify two vaulted structures, at the time interpreted as granaries belonging to a villa rustica of 2nd century A.D. date. Following this discovery the site was abandoned, partially fenced in and then buried by an illegal rubbish dump. From 2005 the “Apolline Project” has undertaken various research activities on the archaeological sites within the territory of Pollena Trocchia. During the 2005 and 2006 campaigns, the site in the locality of Masseria De Carolis was identified and cleaned of vegetation and the rubbish inside the fenced in area. The first excavation campaign began in 2007. A georadar survey was undertaken to the north and west of the vaulted structures. This showed the edge of the quarry created in 1988. Five trenches were excavated which documented the deposition sequence of the eruptive material. In fact, two thirds of the site had been buried by volcanoclastic material from the eruption of Vesuvius in 472 A.D. (known as the “di Pollena” eruption). A number of modest structures (including an oven) made from elements robbed from the walls that had not been buried post dated the eruptive material. These structures were then covered by the ash from another eruption, dated 505/512 A.D. Above this layer, in the eastern sector of the site, the collapse of a wall from a second construction level was found. This had been covered by a volcanic deposit that could not be precisely dated, but was probably medieval. A trench was dug inside one of the vaulted structures discovered in 1988. Below the volcanic material of 472 A.D. a semi “a cappuccina” burial of a boy about six years old was uncovered. The only object in the grave was a siliqua of Marcian (450-7 A.D.) emperor of the eastern empire. The burial overlay 5th century A.D. layers, which produced pottery, half of which coarse, mainly plain buff ware and cooking ware, and a lesser quantity of amphorae, colour coated ware and ARS. Worthy of note among the finds were fragments of glass and tegulae mammatae, which suggest that the vaulted structures were not used for agricultural purposes. There was no dating evidence for the walls, however they presented characteristics typical of the 2nd-3rd century A.D.

The 2008 campaign greatly increased knowledge of the site. In fact, five trenches were excavated in diverse parts of the site, which identified ten rooms. The two vaulted features (“g” and “f”), partially investigated in 1988, were completely excavated and identified as a praefurnium with a long furnace (probably for the combined use of metal boilers – aenea – and testudo alveolorum), steps leading to the furnace floor and three openings, two in the east wall (one providing access to room “f”) and one in the north wall. The praefurnium heated the calidarium “e”, where suspensurae were present (partially excavated). The vaulted feature “f” had a single access in the west wall and was also a praefurnium, with two small furnaces, one on the south side, the other to the east. The rooms “d” (partially excavated), “c” and “b” were situated to the south of the two praefurnia and had suspensurae. Moreover, room “c” also preserved the hypocaust floor, collapsed in the central part before the eruption, and the first row of flue tiles. Room “b” was probably the tepidarium, as it received heat in an indirect form, via an opening in the hypocaust connecting it to room “d”. It was also linked by a passageway to room “a”, which was probably in origin the frigidarium. In the north-east sector of the site there are three rooms yet to be investigated and of uncertain function. The phases of the vulcanoclastic fill identified in 2007 were confirmed and the directions of the lava flows were identified, the main one from the east (Somma-Vesuvio) towards west, the second from north towards the south, with the consequent thinning of the eruptive matter and the refinement of the material. All occupation layers date to the 5th century A.D., although in some cases the presence of residual 2nd and 3rd century material was noted. This probably corresponded to the removal of floors in many of the rooms, leading to the disturbance of the foundations. However, excavation below foundation level showed a thick ashy layer identified as that of the Vesuvian eruption of 79 A.D. The vulcanologists interpretation was confirmed by the presence of pottery and frescoes dating to the 1st century A.D., both above and below the ash layer. The terminus post quem provided by the ash of 79 A.D. further supports – together with the residual material in the 5th century contexts – the dating of the complex to the 2nd-3rd century A.D. It is still unclear whether the bath complex was autonomous or connected to a residential villa.

Several trenches were opened in various parts of the site. In the south-west sector two trenches were opened in order to gain a more precise plan of the calidaria “e” and “d”. However, this area had suffered severe disturbance caused by mechanical diggers in 1988. In fact, traces of a bucket from one of these machines were identified on the floor levels of the hypocausts and the walls were almost razed. Between them the north and east areas revealed the complete stratigraphic sequence. North of the praefurnium “g”, room “h” (probably the wood/coal store), was completely excavated and provided useful information regarding both the nature of the eruptive fill of 472 A.D. and the wood used for heating the rooms in the bath complex. Above room “h” there was a cistern “i”, also completely excavated. Following the 472 eruption, the cistern was emptied and used, together with other parts of the northern area of the site, and later buried by the ash from the 505/512 A.D. eruption. Numerous carbonised leaves were found within this layer which, together with the anthracological samples taken in previous years, provided a reconstruction of the late antique cultivations in the Vesuvian area. During this season mortars samples were also taken and a three-dimensional scan was made of the site.

The 2010 campaign concentrated on the southern side of the site where the heated rooms of the baths and the courtyard in front of them are situated. The excavation inside the _calidarium_ “e” defined almost its complete perimeter. The stratigraphy was heavily disturbed by building work undertaken at the time of the site’s discovery in 1988. Despite this the hypocaust floor was discovered, already robbed in antiquity. Some of the _sesquipedales_ were stamped “DVO DOM”, showing that they were produced in the kilns of the brothers _Domitius Lucanus_ and _Domitius Tullus_, active between 60 and 93/94 A.D. Considering that the entire construction was built above the 79 A.D. ashes it is possible to date the entire structure to between 79 and 93/94 A.D. This data is of particular importance for the understanding of resettlement in the Vesuvian area following the 79 A.D. eruption. The courtyard south of the baths, paved in _opus signinum_, was also excavated. Here the stratigraphy was less disturbed by the 1988 intervention and still preserved a thick volcanoclastic layer from the 472 A.D. eruption. Below this two infant burials, placed close to each other, in amphora (LRA1 with titulus pictus, Keay LII), were found. DNA analysis is being carried out in order to discover any relationship between the two.

The excavation concentrated on the northern slope, corresponding with the service sector of the baths, and on the south slope where the actual bath structures and the open area in front of them stood. The excavation of the area south of caldarium e (situated in the westernmost part of the excavation area) confirmed that the stratigraphy had been badly disturbed by work undertaken with a mechanical digger in the 1980s. In fact, the presence of an accumulation of ancient material, disturbed first with the remains of a burial and later again by modern work was documented. The burial reused a bipedales robbed from the structure. Mainly cranium fragments remained of the skeleton. A large bone pin was also found. The burial’s original position was probably not far from the two infant amphora burials found slightly further to the east, during the 2010 campaign. The guidelines for the construction of the calidarium walls were visible on the hypocaust floor. To the south, the position of these lines indicated that the wall delimiting this room on that side was not on the same axis as those of adjacent rooms d and b, but further forward. To the west of the baths was an area of _opus signinum_ paving, largely obliterated by rubbish dumped there in 1988. In the north-western corner, a well /cistern curb came to light, buried by the material from the eruption of 472 A.D. The well/cistern was internally and externally faced with opus signinum and had a cylindrical opening with an internal diameter of 1.16 m, which opened into a quadrangular space covered by a barrel vault. The well/cistern, only excavated to a limited depth, was filled with deposits from the 472 A.D. eruption. At the base of the well-curb, four small symmetrically arranged holes were identified in the _opus signinum_ floor, that probably served to sustain a structure, perhaps mobile, related to the well itself. The trenches dug in the northern part of the site, already investigated in earlier seasons, confirmed the absence of architectural structures and finds within the layer of eruptive material from 472 A.D. Therefore, it may be suggested that this was an external area leading towards the service rooms of the baths. The level of the 79 A.D. eruption was reached in one of the trenches excavated on the north-western edge. This level had been previously identified and partially investigated in a trench put into the praefurnium f. The trench north of f and east of the cistern i revealed part of what was probably a service room.

This season’s excavations were concentrated on three fronts: the western side, close to the well/cistern; along the entire northern side that is the area providing access to the service area; and in the eastern area, with the rooms leading into the baths. The well/cistern is situated west of the baths in an area paved in _opus_ _signinum_ and bordered to the north by a low wall, almost completely razed in the final occupation phase before the 472 A.D eruption. The south and east walls were wider and served to separate the service area (north) from the courtyard in front of the baths (south). These walls were largely destroyed by illegal activity with construction machinery in 1988. During the latest excavation campaign, work continued to empty the well/cistern of the volcanic deposits from the 472 A.D. eruption. The area north of the baths provided access to the service rooms. Here, in the entrance to the first _praefurnium_ g, excavation revealed the presence of a thick ashy fill from the 79 A.D. eruption underlying the foundations of the bath building. Finds datable to the 1st century A.D., in particular several fresco fragments that probably decorated another villa close to the bath building, were recovered from both above and below the layer of ash. A substantial accumulation of archaeological materials was found along the north wall of the service rooms. These were mainly pottery fragments from vessels used for the processing, storage and eating of food and for use in the baths. The removal of this late antique dump revealed a ramp that led up to one of the service rooms. Excavations took place on two fronts in the eastern area of the site. In the southernmost part, the modern deposits still covering the ancient walls were removed revealing the presence of a passageway to the _frigidarium_. Part of the volcanic material from the 472 A.D. eruption was removed from the _apodyterium_ in the central part of the east front. This revealed the crest of a room in the east section, which based on the vertical stratigraphy and construction technique could be one of the residential rooms of the villa annexed to the baths and predating them.

In 2013, trenches were opened on the north, east, and south fronts of the structures exposed to date. On the north front, excavation of the well continued, and this revealed that it was connected to the cistern to the west, but only volcaniclastic material from the 472 A.D. eruption was dug out of it. Close to the well, a trench was opened along the external north wall of the charcoal deposit. Here, a substantial deposit of material from the 472 A.D. eruption was noted that also ran below the original floor level of the area, as suggested by the level of the building’s foundation offset. The 472 material lay directly on the ashy deposit from the 79 A.D. eruption. Above this, there was a _dolium_ impression cutting into the ash later, a low wall and steps leading to the well area. A new room was identified along the eastern side of the site, perhaps the first belonging to the villa itself, covered by a layer of rubble, probably the result of the eruptive phase, over which lay the thick deposit of volcaniclastic material. The floor level was reached in this area: the black and white mosaic was very patchy and showed traces of postholes. On the south front several sections of wall were identified, of which one seemed to be of particular interest as it made it possible to establish a stratigraphic sequence in which all the walls relating to the villa seemed to be earlier than those of the bath building. Lastly, also on the south front, a terracing wall on a north-south alignment was identified, which probably divided the open-air area paved in _opus_ _signinum_ into two.

Most of the time and resources during the 2014 campaign were dedicated to the cleaning and restoration of the wall facings and finds recovered during this and previous campaigns. Particular attention was paid to a frescoed wall, found in a collapse, and stabilizing the area of the _apodyterium_. Excavation work took place in three areas: the _opus_ _signinum_ terrace south of the baths, the _tepidarium_, and the area at the edge of the villa in the south-eastern area of the site. The area of the _opus_ _signinum_ terrace south of the baths was badly disturbed in the 1980s by heavy machinery and in particular by the dumping of materials that transformed the entire sector into an illegal rubbish dump. This situation was documented in the earlier seasons, however it was deemed necessary to remove the modern fill still covering the _opus_ _signinum_ , in order not only to recover all the ancient pottery (although mixed with the modern material), but also to create drainage for rainwater, following the slope of the ancient floor. The excavation did not reveal any great surprises, as most of the layers were clearly modern, except for one ancient context in which there was an infant burial. This was in the area untouched by the mechanical digging in the 80s and where in 2010 two infant burials in amphorae were found. Excavation of the _tepidarium_ completed investigations undertaken in 2007-2008 during which only half the room was exposed. At the same time, the layers still overlying the floor were excavated, all relating to the final occupation phase (mid 5th century A.D.) before the 472 A.D. eruption. Two infant burials lay on the heated floor, which had collapsed at the centre at the time of the eruption due to the great weight of the volcanic material covering it. The burials were partially covered by tiles and no grave goods were present. In the third and final area investigated in 2014, in the south-east corner of the site, a collapsed wall overlying rubble and a corridor with a ramp leading to the second floor of the villa were uncovered, beyond which a low wall with an _opus_ _signinum_ facing delimited an area that was probably a garden.