1A number of cave infillings in the Mediterranean area, settled by man during Neolithic to Bronze and/or Iron Ages, shows a typical thin-layered stratigraphic succession alternating blackish and whitish centimetre-thick strata (“zebra”-like or “layer-cake” facies). This peculiar stratigraphy has been tout court interpreted up to now as the result of agro-pastoral practices: repeated recovering of sheep and goats with following burning of stable waste (mainly herbivore dung) in order to diminish its volume and consequently amend the site for occupation. Such an interpretation is based on micromorphological observations coupled with experimental archaeology (e.g., Boschian, 1997; Macphail et al., 1997, 2004; Canti, 1999; Boschian and Montagnari-Kokelj, 2000). This paper shows evidence of a higher stratigraphic variability observed within the above thinly layered facies of the Santuario della Madonna Cave, located along the northern Tyrrhenian coast of Calabria (south Italy), close to Praia a Mare village: the main specific features, investigated through a multidisciplinary and multi-scale (macro- and micro-) approach, suggest that the effects of animal and human occupation are strongly superimposed by or alternated with the imprint of natural processes.
2The Santuario della Madonna Cave is one of a series of karstic caves located along the steeply westward-dipping scarp, hanging Early Pleistocene marine terraces and approximately parallel to the present-day coastline (Carobene and Dai Pra, 1990); this paleosea-cliff is carved across Meso-Cenozoic dolomite-marly limestone. In particular, the studied cave consists of a single, large room (over 2000 m2 wide and 15 m high) and three openings (fig.1A) that develop between 40 and 70 m a.s.l. with clear evidence of marine re-shaping; in fact, at the northernmost cave opening, sea-level indicators (lithodomus holes) and sea-cliff toe deposits, consisting of rounded, coarse conglomerates, occur up to the cave ceiling. They have been correlated with Middle Pleistocene terraces located to the North and to the South of the study area (Carobeneand Dai Pra, 1990). Outside of the southernmost cave opening, where the main Chapel is located, it is worthy to note a cave-collapse sinkhole consisting of two overlapping caves, as well as the floor of the highest cave onto which stratified slope deposits occur. Slope deposits, produced by the scarp facing the opening cave at issue (fig.1A), are fairly steep and in-cave dipping, where they underlie a Late Pleistocene-Holocene succession; at the end of the fifties of the last century, it was affected by early exploratory excavations that highlighted the archaeological-prehistoric interest of the site (Blanc and Cardini, 1961).
Fig. 1 - (A) Overview of cave openings in Mesozoic dolomite-limestone.
1, 2, 3: Santuario della Madonna Cave openings; 4: Cardini or Fumarolo Cave
Fig.1 - (B) Holocene stratigraphic succession excavated in the Grotta della Madonna Cave (see text for details).
3Excavations carried out in the sixties (Cardini, 1970) revealed an impressive stratigraphy, with several layers interested by a repeated settling by man from the late Upper Palaeolithic (around 12 ka BP) to the Bronze Age (III millennium BC).
4Starting from 2002, a new excavation area was opened next to the oldest one, close to the WNW wall of the cave; the new excavations, about 5m‑deep, have better pointed out the human occupation of the cave, from the earlier phases of the Middle Neolithic Age (6th-5th millennia BC) until the advanced phases of the Middle Bronze Age (2nd millennium BC) (Tinè, 2006).
5By means of biological and archaeological findings, four main units were recognised (fig.1B). The bottom early Middle Neolithic level (1st unit, 0.2 m thick), contains a great amount of typical “Red Stripes” pottery and the following 2nd unit (1 m thick) includes pottery shards suggestive of late Middle Neolithic age (Serra d’Alto style); both units consist of a well-structured soil, yellowish brown to greyish in colour, developed on silty clay to (clay) loam deposits; it consists of four argillic horizons (Bt1 to Bt4) with occasional clay coatings in pores and some hydromorphic features, such as small grey or yellowish red mottles. Some dated charcoal fragments indicate calibrated radiocarbon ages between 8735 ± 80 to 5555 ± 75 yr BP.
6The 2nd unit is buried by about 3 m of “zebra”-like layers (so called in the old excavation report), spanning the Late Neolithic to the Early Chalcolithic (4200 ± 65 yr cal BP); these layers occur in multi-storey sequences made up of centimetre-thick beds, locally lenses. They show a generally sheet-like geometry, locally wavy-based, and consist of alternating black, red, white (variable from clear white to “dirty” greyish) and brown loamy sand to sandy loam layers, not always in the same order. The little pottery recovered in the lower portion (3rd unit) suggests an age spanning from the Late Neolithic (Diana style) to the Early Chalcolithic (Piano Conte style), during which there was a low degree of human frequentation of the cave. The 4th unit, up to 1.2 m thick, starts with subtle levels with typical Laterza’s Culture pottery (Late Chalcolithic-Early Bronze Age), followed by the richest layers of the whole sequence (phases 1-2, the so called Proto-Apenninic cultural aspect and phase 3, Apenninic, 2920 ± 65 yr cal BP); in fact they show clear anthropogenic palaeo-surfaces of Middle Bronze Age, suggesting an intense human re-occupation of the cave during these phases. This is also evidenced by the finding of different systems of pole pits with irregular and partly overlapping spatial arrangement during the Eneolithic to Late Middle Bronze Age, and hearth remains spanning the Proto-Apenninic phases. The topmost portion of the succession is related to historical human fruition of the cave, addressed to cemetery purpose, as suggested by many burials revealed during excavations. In addition to ceramic artefacts, pole pits, hearths and charcoals, also human and animal bone fragments were found in the stratigraphic succession.
7The present cave floor is characterized by a very low, general westward gradient and by a typical wavy surface, with small subcircular concavities due to active dripping and small patches commonly filled with millimetric to centimetric carbonate concretions and soft concentrations.
8All the argillic horizons of the buried soil (Bt1 to Bt4, corresponding to the 1st and 2nd archaeological units) and one sample for each differently-coloured “zebra”-like layer (black, brown, red and white) of the 3rd and 4th units were collected for physical, chemical and mineralogical analyses. The main results are reported in tab.1.
Table 1. Main physical and chemical features of Bt horizons and “zebra” layers.
OM = organic matter; CEC = cation exchange capacity.
9In particular, the particle size distribution analysis shows the dominance of the silt and clay fractions in the buried soil, whereas the overlying layers are dominated by sand, always higher than 60%. Organic matter appears particularly high in the black (7.9%), red (6.35%) and brown (5.0%) beds of units 3 and 4, never exceeding 3.0% in the other samples. Soil reaction is always alkaline (pH mean values around 8.8). Total CaCO3 is particularly enriched in horizon Bt1 and in the layered succession, up to 21.4% in the white bed, and depleted in the lower Bt horizons (the lowest amount of 5.3% occurring in horizon Bt2). The major element chemical composition of all samples measured by X-ray fluorescence spectrometry (XRF) shows relatively high amounts of CaO (2.82‑15.08%) and MgO (2.85-13.79%), but much higher Al2O3 (5.25‑19.87%) and SiO2 (31.99-50.11%) content. Mean values of K2O and P2O5 only reach 2.23% and 2.60%, respectively, whereas range from 3.76% for Fe2O3 to 1.15% for TiO2; Na2O and MnO are always <0.5%.
10A micromorphological study was carried out in optical microscopy on thin sections prepared from selected undisturbed samples (Fitzpatrick, 1984) of units 3 and 4. The matrix in the “zebra” layers shows different fabric, from massive and compacted to highly porous, to fibrous, thinly laminated and lenticular. It is usually characterized by abundant micrite; in places, typical dense lenses, subcircular nodules or clusters of calcite crystals occur. Occasional wide zones made of brown, optically isotropic (between crossed polars) humified matrix, and frequent subrounded or thin elongated, calcite-cemented, fibre-like aggregates, respectively representing coprolites and fibrous dung, were identified, mainly in black and brown layers. Black charcoal fragments and subrounded to subangular, reddish clayey pedorelicts occasionally occur, with chaotic to laminated or imbricated fabrics. Abundant spherulites, 5-10 µm (and occasionally 20 µm) in diameter, occur within the fibrous calcitic dung, concentrated in lenses and in subvertical pores (fig.2A), or scattered in the matrix. They exhibit an anisotropic optical behaviour in crossed polarized light (XPL), indicated by a permanent, sharp extinction cross (Fig.2B). A radial and concentric pattern of calcite crystallites, with an organic-rich outer rim and an inner mineral core, were observed through SEM-EDS analysis. Electron microscopy permitted also to identify a variety of very small Ca-Mg carbonate concretions, ranging from a few to some tens microns in size, mainly in the grey and white layers: curtains, clusters of polygonal pisoliths (fig.2C) or coralloids, globular concretions, eccentric (often single-crystal) and druse-like, radial calcite crystals (fig.2D).
Fig. 2 - Microphotographs (in XPL).
(A) subvertical pore infilling enriched in faecal spherulites; (B) spherulite-rich layer showing sharp extinction crosses (B); SEM images of: (C) clusters of polygonal pisoliths; (D) druse-shaped floating calcite with radial crystals.
11Quartz, feldspars and micas, as well as calcite and dolomite, were detected by optical microscopy on thin sections and X-ray diffraction analysis on bulk samples both from the Bt horizons and the “zebra” beds. The mineralogical analysis of the clay fraction (XRD and FT-IR spectrometry) from the same layers shows the ubiquitous occurrence of calcite in all samples, the dominance of illite and kaolinite in the buried soil (coupled with minor smectite or chlorite) and in the red layer; black, brown and white layers are mainly characterized by organic (X-ray amorphous) substances, on which we focused using infrared spectroscopic analysis. In particular, treated (acetone-extracted) samples only reveal two main (but weak) bands at 2850 and 2820 cm‑1, indicating methyl (CH3) and methylene (CH2) groups, in the red and grey beds, whereas the black ones (both untreated and treated) exhibit more and higher peaks that can be related to organic substances (fig.3).
Fig. 3 - FT-IR spectra of untreated and treated (acetone-extracted) black samples.
-1-1-1-1-1-1-1-1-1
12On the basis of the main field and laboratory data, a characterization of the geoarchaeological context of the Santuario della Madonna Cave is proposed, with a special focus on the major natural and anthropogenic environmental changes that occurred from the Middle Neolithic to the Middle Bronze Age.
13The relatively high percentages of calcium and magnesium determined by XRF analysis in all samples are consistent with the dolomite-limestone bedrock of the cave; in contrast, the much higher amounts of silica and aluminium indicate the occurrence of allochtonous (Al-)silicate components, possibly transported by wind and/or reworked by surface runoff from outside the cave. Such results are confirmed by XRD and optical microscopy results, showing quartz, feldspar and mica minerals occurring together with calcite and dolomite: these mineral species represent the primary components in the soils developed on the surrounding Quaternary marine terraces (Scarciglia et al., 2006). The variable K, Na, Fe content (and some amount of Ca and Mg) presumably includes elements associated with both primary and pedogenetic aluminosilicates. The small amount of phosphorous can be related to the presence of apatite as a constituent of bone fragments or as a secondary product of their alteration, or to its release from animal dung. Organic matter showed higher values, as expected, in black and brown beds, being surprisingly high even in red layers. The amount of total CaCO3 is higher in the “zebra”-like succession, coherently with the occurrence of many secondary calcitic features (see below). On the other hand, its partial depletion in the underlying soil accounts for more intense leaching, enhanced by a relatively prolonged time of pedogenesis, allowing enough soil development and differentiation, also recorded by the kaolinite- and illite-dominated clay mineralogy. On the basis of the archaeological and radiometric stratigraphy (fig.1), major soil formation can be ascribed to the late early/middle Holocene climatic optimum, as also supported by the presence of clay coatings in argillic horizons (units 1 and 2) (cf. Cremaschi and Trombino, 1998; Scarciglia et al., 2006). In fact, this period was characterized by high moisture availability (higher than today) and enough seasonal contrast (Goodfriend, 1999; Di Donato et al., 2008) to allow clay translocation and its subsequent stacking within pores.
14The presence of humified organic matrix is observed in optical microscopy particularly in black layers from the “zebra”-like succession (units 3 and 4), although infrared spectroscopy did not allow clear identification of specific organic substances. Anyway, IR spectra showed more definite and abundant peaks of organic components in black samples, which appear fairly comparable with those diagnostic of humin and humic (and partly fulvic) acids (cf. Grasset and Amblès, 1998; González-Pérez et al., 2004; Fiorentino, 2005). No explicit evidence of burnt/unburnt organic matter has been detected from our data.
15The calcitic coprolites, fibrous dung and spherulites observed in the layered succession represent typical products of herbivore digestion (e.g., Boschian, 1997;Macphail et al., 1997, 2004; Canti, 1999). In particular, spherulite-rich layers are interpreted as a result of stabling practices by Neolithic people, who used caves as agro-pastoral recoveries for sheep and goats, periodically burning animal dung in order to reduce organic waste and restore a healthy cave environment: much vegetal tissues are degradated by fire, with a consequent enrichment in the spherulitic (mineral) components. This interpretation is consistent with the frequent occurrence of lenticular or clustered calcite aggregates, probably representing ash residues due to in situ fires. On the other hand, the abundance of spherulites in pores suggests that other mechanisms can be responsible for their concentration, such as reworking due to sheet wash with consequent pore filling. Surface runoff and colluvial processes are supported by the low gradient of the cave floor, the sedimentary macro- and micro-facies of some “zebra” layers, and the presence of occasional pedorelicts. The latter and the red beds themselves likely represent soil sediments reworked into the cave from the widespread red soils cropping out on the surrounding marine terraces, characterized by similar primary (see above) and neogenic mineralogical features (Scarcigliaet al., 2006). In particular, the presence of kaolinite among clay minerals indicates that whether they experienced fire, temperatures should have been lower than 500-550 °C, which represent the threshold for its stability. Moreover, the different carbonate concretions identified in white and grey beds by means of SEM-EDS, display shapes that are typical of cave environments, related to a variety of underground karst processes: eccentric calcite growths are interpreted as result of capillarity, curtains are emplaced by drop flowing, clustered polygonal pisoliths are due to spatial constraints, globular concretions to splash processes, coralloids to high evaporation rates, as well as druse-like radial calcite crystals, which represent typical floating calcite in pools of stagnant water (Forti, 2002). The lack of calcium oxalates in any sample proved by XRD and FT-IR results supports that a possible morphological convergence between karstic floating calcite and Ca-oxalate-pseudomorphous druses, calcitized after burning at T≥500 °C (Canti, 2003), should be excluded, as also confirmed by the above discussed kaolinite preservation upon heating. In addition, the occasional, small concretions and soft concentrations of calcite observed on the present cave floor, indicate early stages of formation due to an essentially active process of carbonate reprecipitation. On one hand, it is not surprising that in caves developed in carbonate rocks, secondary calcite features may occur and may have repeatedly formed in the past. On the other hand, the opposite, i.e. their lack, would be astonishing! In addition, water dripping is nowadays observed on the floor, varying in intensity according to seasons. The small hemispherical cavities thus produced on its surface, appear strictly conformable with the deep layer boundaries that characterize the “zebra”-like succession, suggesting the repetition through time of such a wavy paleotopography, with calcite filling the depressions and progressively forming, at least partly, white and grey beds. The rates of speleothem growth in Italian caves during the Holocene mainly range from 20-40 to 200-250 µm/yr (Sauro et al., 2003 and references therein), resulting 101 higher than those monitored in modern stalagmites throughout European sites (Genty et al., 2001). Even assuming (i) that white layers are only partly made of carbonate reprecipitation features (at least those with more whitish colour rather than the dusty greyish ones, which in turn probably consist of burnt calcitic ash), and (ii) that their growth rates are higher than those estimated in the literature (suggesting a buffering effect of organic matter on CaCO3 or a higher loss of pCO2 due to the open cave system), the hypothesis that such caves were possibly left abandoned for a couple of seasons (e.g., Boschian and Miracle, 2003) appears poorly supported. In fact, the time span to permit sufficient thickening of carbonate-concretioned layers is too short, as also constrained by the geochronological data of the stratigraphic succession.
16This work shows that an integrated multidisciplinary geoarchaeological study of “protected” sites such as karst caves involved by human occupation can contribute to highlight man’s behaviour and practices in the evolution of the natural landscape and climatic changes. As a whole, the pedosedimentary succession of the cave records important environmental changes during the Holocene. The buried soil depicts a relative (millennial-scale) geomorphological stability, under humid and seasonally contrasted conditions, that occurred during the Middle Neolithic climatic optimum. The overlying “zebra” layers record cyclical, intense and rapid (decadal to centennial) fluctuations in moisture regimes and morphodynamics: natural processes, such as water dripping, CaCO3 leaching and reprecipitation, humification, surface runoff and colluviation are intercalated with man’s activity (stabling, in situ burning, setting of poles) during the Late Neolithic up to the Bronze Age. These intensely alternating pedostratigraphic records are consistent with short-term climate changes during the middle Holocene, with increasing aridity and seasonality, as evidenced by several proxies (e.g., Ellwoodet al., 1997; Sauro et al., 2003; Riehl et al., 2008; Di Donato et al., 2008). However, the natural signal appears often overprinted (and perhaps partly controlled) by heavy human disturbance, as also highlighted by other authors (Courty and Vallverdu, 2001; Schuldenrein, 2001).