1The archaeological excavations conducted by the Archaeological Superintendency prior to large-scale construction work in the Campanian plain during the last decade have yielded a wealth of information about the interaction between Phlegraean Fields and Somma Vesuvius volcanic eruptions and prehistoric and proto-historic human occupation (Marzocchella, 1998; Navaet al., 2007). In particular, the Early Bronze Age/18th century BC (Vanzetti1998) Vesuvian “Pomici di Avellino” (PdA) eruption (Sulpizioet al., 2008 and references therein), constitutes an important stratigraphic marker and provoked a significant hiatus in the occupation of the southern part of the Campanian plain.
2Numerous settlements in this region were destroyed by this eruption; the first found was close to the town of Palma Campania near Naples (Albore Livadie and d’Amore, 1980), which gave its name to the culture. The most spectacular example, several large houses belonging to a village, came to light at Croce di Papa, Nola. These sites all contain the same pottery, belonging to the Palma Campagna culture, which dates to a late phase of the Early Bronze Age, and are characterised by exceptional evidence of the organic materials once present, in the form of impressions in the volcanic deposits (Albore Livadie and Vecchio, 2005; Albore Livadie, 2007).
3The discovery of a village 14 km NW of Vesuvius, at Afragola (Province of Naples) (fig.1), during preliminary work in an area threatened by infrastructures associated with the building of the Rome-Naples high-velocity rail link (Laforgia et al., 2007), has alloweda varied and detailed reconstruction of the local effects of the PdA eruption. Furthermore, abundant information was obtained concerning the organization of the village, the nature of the buildings and other features, and their reaction to the arrival of the pyroclastic density currents (PDCs). The settlement was extensively investigated; it was inhabited at the moment the eruption struck and the structures it contained were surrounded by and buried in volcanic ash, creating the conditions for an outstanding degree of preservation. The results outlined below are the fruit of interdisciplinary studies which are still in progress.
Fig.1 - The Campanian plain showing Vesuvius, the Afragola Bronze Age village and the distribution of Pomici di Avellino deposits.
Isopachs in cm: EU2: white; EU3: black, EU5: dark grey (after Di Vito et al., 2008)
4According to M. Di Vitoet al.(2009), the eruption may be divided into three principal phases (opening, Plinian and phreatomagmatic) which produced five eruptive units (EU1‑5, fig.1). EU1 was produced by the opening phase of the eruption, which saw the formation of a short-lived, low and pulsating eruptive column. The relative deposits are distributed to the NE, at a distance of several tens of km from the volcano. The following EU2, 3 and 4 were deposited in the principal phase of the eruption by a sustained Plinian eruptive column, which reached a maximum height of about 30 km. The products of this phase consist of white, and then grey, pumice fallout deposits (EU2-3) again distributed towards the NE, which cover an area of more than 15,000 km2 (fig.1). Because these products were dispersed northeastwards from Vesuvius, none of them are found in the vicinity of the village, which only suffered from the effects of the final stages of the eruption. The last eruptive unit, EU5, generated by repeated phreatomagmatic explosions, is almost entirely composed of a succession of deposits formed by dilute, turbulent PDCs, distributed mainly to the WNW at a distance of up to at least 25 km from the volcano. Fine ash from these phases, elutriated by the currents, was dispersed and deposited by fallout at distances of several hundred km to the NW (Sulpizio, et al. 2008). On the basis of sedimentological characteristics, EU5 has been divided into 4 main depositional units (a - d), laid down by a corresponding number of PDCs (fig.4).
Fig.4 - Part of storehouse (Building 20), with partial internal collapse
The baulk shows the external eruption sequence of thickened deposits in section.
5Each of these units is topped by a massive layer of fine-ash, dispersed by the pyroclastic currents and deposited by fallout. After the eruption, large sectors of the Campanian plain were affected by repeated floods which emplaced variable thicknesses of finely-laminated sand deposits.
6In this section the results of careful and systematic excavation of the PdA deposits in the village area are presented. Beds deposited by the PDCs were removed one at a time, insofar as this was possible, in order to reveal the effects of the currents on the structures and vice versa. This procedure also allowed the recognition of traces of the passage of humans and animals at certain intervals between emplacements of the PDCs (fig.4). The first of the EU5 pyroclastic deposits covers a fine sandy/silty dark brown palaeosol with small pumice fragments, derived from underlying volcanic material, which shows widespread disturbance due to human activity. The pyroclastic deposit was identified over the entire excavation area, including the hut interiors, and is composed of two subunits. The lower of these (a1) is a massive fine-ash deposit with accretionary lapilli and abundant imprints of vegetation; it exhibits notable variations in thickness in correspondence to obstacles and irregularities in the underlying surface. At the Afragola site subunit a1 penetrated inside buildings (see below), thickened against obstacles and completely filled the spaces between the reeds in the huts’ thatching. During this phase some buildings suffered damage, in particular to their more elevated parts. The upper subunit (a2) is a laminated medium- to fine-ash deposit of more uniform thickness, which tends to mantle obstacles and huts. As palaeotemperature measurements have confirmed (Di Vitoet al., 2009), after the deposition of this subunit there was a lull of several hours which permitted the deposits to cool down sufficiently to allow the passage of men and animals who left large, deep imprints on a2, filled in by material of the unit above. EU5b is the coarsest of the sequence and composed of two subunits (b1 and b2). EU5b1 is strongly laminated, composed of a succession of coarse-, medium- and fine-ash layers, and tends to mantle buildings and increase in thickness close to them. This unit is again plastically deformed by footprints. EU5b2 consists of a sequence of massive fine- to medium-ash layers with accretionary lapilli. Above this there is no evidence of human activity, but only some chaotic redeposited material derived from slippage of ash build-ups on buildings which accumulated in highly irregular fashion. Numerous larger buildings collapsed during this phase, probably due to the destabilizing effects of heavy ash loads on the weakened structures. EU5c has similar characteristics to b, but is less coarse and exhibits a rapid lateral transition to fine-ash deposits. At the top of this unit, redeposited material is found once again, due again to slippage from the roofs of buildings; at least one large building collapsed at this point. The stratigraphic sequence terminates with EU5d, composed of weakly laminated to massive ash layers, grey to pinkish in colour and fining upwards from medium- to fine-grained, which was deposited by the last PDC. Most of this unit is the result of the slow sedimentation of fine-ash particles elutriated by the pyroclastic currents. The surface of the unit (d) was abundantly deformed by animal hoof-prints and human footprints formed at a time when the deposit was still highly plastic, with a mud-like consistency.
7Subsequent deposits were composed of layers and lenses of fine- to medium-grained ash which almost completely cover the irregularities in the surface of the eruption deposits; they were formed during periods of flooding and waterlogging which strongly affected the area after the eruption.
8It is interesting to note that the PDCs did not cause the immediate collapse of many small buildings and some larger ones (see following section) and did not blow over fences or trees. Buildings generally fell down due to the weight of volcanic ash accumulated on roofs and in some cases it is likely that they were subsequently kept in an upright position by the consolidating volcanic ash which surrounded the lower parts of supporting posts; this latter effect seems to have applied to all the fences.
9A synthesis of the information obtained from the excavation and recording of consolidated ash deposits and the underlying soil level, with the establishment of the archaeological stratigraphic sequence, allowed the timescale and mode of collapse of each construction to be determined and reconstructions (of varying degrees of completeness) of standing buildings to be made. It was thus possible to distinguish between structural traces found in situ and in a more or less original position, vertical elements inclined (either during or shortly after the eruption) into positions of partial or complete collapse (but the lower extremities of which remained roughly in place), and other structural remains which had been subjected to three-dimensional translation (and perhaps rotation and distortion) with respect to their (presumed) original locations (generally on first floors or roofs).
10Excavation at the site covered an area of around 4,500m2, and led to the identification of the village boundary to the north and south, whereas the eastern and western limits lay beyond the area investigated. To the north and south the settlement was surrounded by an enclosure, outside of which there lay an extensive wooded zone, in which oak predominated. Internally, the village contained a mainly residential area containing huts and associated small outbuildings and, to the SE, a concentration of storage buildings.
11The dwellings were aligned NW-SE; in plan they were basically rectangular but curved on one of the short sides (fig.2).
Fig.2 - U-shaped residential hut (Building 9) with collapsed internal fixture and pots.
- 1 The archaeobotanical aspect of the construction materials is currently under study by E. Castiglion (...)
12The steeply-pitched sloping roof, made of bundles of reeds tied to wooden stakes1 reached down to the ground on either side. The wall on the straight short side, generally to the SE, was triangular in elevation and made of slender juxtaposed stakes; this wall contained the doorway, and was protected by an overhanging roof. The buildings were 5.5 to 9 m long, 4 to 5 m wide and were supported by internal posts. A dividing wall partitioned the interior into two rooms; the main rectangular room housed an oven and/or fireplace for cooking, whilst a roughly semicircular room in the back contained vessels for the storage and consumption of foodstuffs, in part arranged on shelves or hung on the walls. Their height, based on the dimensions of the collapsed front wall of Building 1 (one of the smaller examples), was at least 3.6 m.
13The houses were often accompanied by nearby smaller buildings: some of these were circular in plan (diameter 1‑1.5 m) with conical roofs and tiny slightly sunken fireplaces inside; they were probably for smoking or drying foodstuffs. Others were oval (c. 4x2 m), with the entrance on a short side and thick reed walls (fig.3); perhaps they were used as general storage facilities or sheds (in one case apparently a wood store).
Fig.3 - Small oval outbuilding with baulk (Building 15).
The surface of EU5a1 is exposed. A surrounding fence, together with the steeply sloping hut walls and a central supporting post, may be seen in section.
14The larger storerooms were oriented perpendicularly to the dwelling huts and rectangular in plan (c.5x2.5 m); robust posts supported a first floor enclosed by walls and covered by a straw-thatched pitched roof, but the ground floor was open.
15External areas were divided in irregular fashion by fences, curved or straight, which at times were joined together and enclosed ample areas containing more than one building. They were generally composed of a continuous band of lightweight material (in one case horizontal planking) held up by a line of posts.
16The settlement seems to have been established not very long before the eruption, although some signs of expansion in an intermediate period were evident.
17The rapid partial burial of the village led to the preservation of detailed moulds of seeds, cereal ears, fruits, and seeds, as well as structural building components. Imprints of almonds (Amygdalus communis), previously found nearby in Nola (Province of Naples, Albore Livadie and Vecchio, 2005), confirm the presence of the species in the Mediterranean during this period and its alimentary importance, whilst those of berries and acorns are an indication that wild or semi-cultivated fruits supplemented agricultural production. Within the village, at the margin of what seemed to be the residential area, there were numerous bramble bushes (Rubus sp.), and a possible fruit imprint was found elsewhere. The volcanic ash deposits contain abundant evidence of endemic plants; leaf imprints are often chaotic and difficult to identify, but preliminary work indicates the presence of common oak (Quercus robur), downy oak (Quercus pubescens), hop-hornbeam (Ostrya carpinifolia), bracken (Pteridium aquilinum) and dogwood (Cornus mas L.). Inside the remains of a presumed granary, carbonized kernels and spikelets of einkorn (Triticum monococcum), emmer (Triticum dicoccum) and barley (Hordeum vulgare) were found.
18Archaeobotanical work currently under way will lead to an improved understanding of the ecosystem which surrounded the village and shed light on agricultural practices, the use of plant resources and alimentary habits. The picture which emerges is of a settlement bordered by mixed woodland, probably containing fruit trees, open grazing areas and cereal cultivation. Knowledge of the environment will also supply information concerning the inhabitants’ socio-economic status. The presence of a granary bears witness to the importance of cereal cultivation and preparation; a large area would have been devoted to grain threshing.
19During the excavation, few traces of domestic animals were found, although fenced enclosures were present. Only one individual appears to have died as a result of the eruption, a calf that was found inside a dwelling hut, into which it perhaps bolted in fright and then became trapped.
20The animal bones recovered were the remains of butchering and meals. Preliminary work has indicated the presence of three domesticated species, cattle (Bos taurus), sheep/goats (Ovis/Capra) and pigs (Sus domesticus); the first two of these predominate slightly with respect to the third. There would seem to have been a diversification of animal resources which would have favoured economic stability. The age at death, young to young adult for both Sus and Bos, indicates that these species were raised in order to provide meat. Ovocaprines, on the other hand, would have been used for wool and milk in addition to meat. The discovery of a calf which had been butchered at about six months of age furnished useful information regarding butchering practices. Remains of dog (Canis familiaris) were found, as well as those of a cat (Felis): it is not yet clear whether this was wild or domesticated. Wild animals include ground vole (Arvicola terrestris) and unidentified bird species.
21The village was apparently well organised with regard to its use of environmental resources, and its inhabitants – of whom no concrete biological trace was found – were probably relatively well nourished.
- 2 A Cyrax 2500 scanner was used for general scans of the excavation area at a precision of ± 2 mm and (...)
22Thousands of human footprints and animal hoof prints were found on ash deposits inside the village and in an extensive zone surrounding the settlement, at distances of up to 1 km. The prints were found on the surface of depositional units EU5a2, EU5b1 and EU5d, and often deformed the underlying ash beds which were still highly plastic. Numerous imprints were also found on the surface of many of the succeeding thinly bedded flood deposits which covered a sizeable area of the Campanian plain. In order to obtain accurate field surveys of the imprints, 3D laser scanner2 and GIS records of the phenomenon were prepared for palaeoanthropological and archaeozoological analysis.
23Inside the village, a plethora of imprints, mostly human footprints, were found above on the surface of EUa2 and b1, passing in a fairly chaotic fashion around the buildings and other obstacles present (fig.5).
Footprints made during the eruption and traces of 3 storehouses (with baulks), bramble bushes (top right) and a large tree (bottom right).The regularly shaped cuts, 2 circular and 1 rectangular, are irrelevant later features (Top c. NW).
24In places these consisted of readily legible single tracks, but were often confused due to the overlaying of multiple print sequences which stirred up the ash deposits. Sometimes there were signs of slipping and sliding, of turning back or jumping over fences. Externally, “pathways” extended beyond the excavation area in various directions, roughly northwards and southwards.
25With regard to the human footprints, the length, width and depth of the individual imprints was measured in a sample area external to the village. This allowed the height of individuals to be determined (Giles and Vallandigham, 1991; Jasuia et al., 1991), although estimating sex was more problematic; the sample was divided into three groups, using the presumed height of three individuals of known sex found in PdA eruption deposits at San Paolo Belsito (Province of Naples) as points of reference (Albore Livadie and Vecchio, 2002). 13 individuals of presumed height not in excess of 140 cm were considered as subadults of unknown sex, whilst 57 individuals of presumed height 140-170 cm, together with 9 in excess of 170 cm, were considered adults of unknown sex.
26Numerous tracks consisting of a sequence of 3 imprints with similar dimensions and morphological and directional characteristics were identified; the distance between corresponding points (i.e. big toes and/or heels) in pairs of consecutive prints was measured so as to calculate speed of travel using the method of Charteris (Charteris et al., 1982). This study was performed on infra-eruptive and post-eruptive tracks, yielding walking speeds of between 2.5 and 5.4 kmh‑1 in 89.6% of cases: not particularly rapid velocities.
27During the process of abandonment of the zone, which became marshy after the rapid accumulation of volcanic products disrupted the natural drainage pattern, the hoof prints of animals and imprints left by human feet and cartwheels are found repeatedly in the sequence of thin alluvial laminations. These post-eruptive tracks show an exodus towards the NW, N and NE; only sporadic and isolated prints were found towards the E, S and W. Tracks of animal prints running parallel to those of human beings strongly suggest the movement of domestic herds. The most commonly recorded animals are cattle (Bos taurus), followed by ovocaprines, equids (both Equus and Asinus) and carnivores. Dog and cat prints are generally not associated with those of humans. Wild animals are represented by deer tracks, together with scattered prints of rodents, porcupines and squirrels.
28A careful integration of archaeological and geological research, supported by other specialist studies, has allowed a deeper understanding of the PdA eruption, and has produced a wealth of new archaeological information regarding village life in Campania in this late phase of the Early Bronze Age.
29The archaeologically exceptional nature of the site is, of course, due to the fact that the accumulation and subsequent hardening of volcanic deposits preserved negative impressions of the wood and other plant materials from which buildings in the Afragola village were constructed, allowing the classification of buildings into various standard types and revealing fascinating details concerning construction techniques and materials. An essential contribution proved to be the analysis of the mode of collapse or partial disintegration of the various structures.
30The palaeo-environmental evidence is also important, deriving from the study of both the impressions of fruits, seeds and leaves etc. present in the volcanic ash and of carbonized cereals and other plant fragments sealed by the eruption. It has shed light on the woodland vegetation present outside the village and species employed in agriculture, together with the inhabitants’ use of wild (or semi-wild) plants and fruits, and therefore on their Early Bronze Age diet. Similarly, the numerous animal bones recovered have yielded data concerning domesticated animal species and butchering practices, to which information from animal prints must be added. In addition to significant anthropological measurements, a particularly vivid vignette of the exodus of human populations from the area hit by the PDCs has resulted from the recording and study of human footprints.
31From a geological perspective, the Afragola excavation confirmed that the area of the Campanian plain where the village was situated, to the north of Vesuvius (fig.1), was only affected by the final stages of the eruption. Detailed examination and laboratory study of the PDC deposits within the village allowed the reconstruction of the sequence of volcanic and non-volcanic deposits, a definition of their lithology and sedimentological characteristics, flow directions and deposition temperatures. Archaeological evidence also contributed to some of the geological conclusions reached (see Di Vito et al., 2009). Details of structural collapse demonstrate, unambiguously, that the impact of the pyroclastic currents was not violent enough to destroy buildings; some fell towards the end of the eruption under the weight of accumulated sediments. The presence of human and some animal tracks indicates that two of the pauses between PDCs (EU5a2-b1 and b1-b2) were sufficiently long to allow cooling to temperatures low enough for human passage; anthropological study shows that gait was generally normal and unhurried. Furthermore, PDC temperatures were clearly not high enough to cause the combustion of wood or other more delicate plant materials such as bushes and fallen leaves (which were all completely eliminated by subsequent decay); temperature determinations of about 280°C were made on potsherds (as well as lithic inclusions).
32The abundant deposits produced by the PdA eruption interfered with natural drainage mechanisms over a wide area and seem to have provoked a marked interruption in the human occupation of this sector of the Campanian plain. The characteristics and quantities of the eruption products were not conducive to a rapid return to well-drained soil conditions. It was only in the Late Bronze Age (13th‑12th century BC), that settlement returned to this area, in the form of an extensive occupied zone (c. 15 ha) – perhaps composed of several separate villages – which has yielded pottery showing regular contacts with the Aegean (Nava et al., 2007).
We are indebted to the archaeologists who took part in the Afragola excavations, with the coordination of the Soprintendenza per i Beni Archeologici di Napoli e Caserta. Our thanks to the anonymous for their most helpful comments. The work was funded by Condotte d’Acqua SpA on behalf of TAV SpA as part of the high-speed rail project and the INGV-DPC 2004-2006 Project (RU Di Vito).