Eurasian Prehistory, 5 (1): 85- 94.

THE RIO SECCO CAVE, A NEW FINAL MIDDLE PALEOLITHIC SITE IN NORTH-EASTERN ITALY

Marco Peresani and Fabio Gurioli

University ofFerrara , Dipartimento delle Risorse Naturali e Culturali, Corso Ercole I d 'Este 32 I-44100 Ferrara, Italy; [email protected] [email protected]

Abstract This article describes Rio Secco Cave, a newly discovered Middle Paleolithic site in the eastern Italian Pre-Alps. Sedimentary succession, faunal remains, lithic assemblages and one 14C date define a chronological range from OIS 3 to the Holocene with evidence of human presence at the end of the Middle Paleolithic. This site shows for the first time the presence of the last Mousterians in the central northern Adriatic region between the Venetian Alps and Dalmatia.

INTRODUCTION some cases can be related to tool production tasks, The final phase of the Middle Paleolithic in due to their very close proximity to lithic raw ma­ northeastern Italy is documented through numer­ terial sources. In other cases these short-term ous sheltered sites and open-air settlements that camps were simply used as waypoints in a logisti­ show evidence of short-term occupations or re­ cal system of mobility. Segmented tool produc­ peated use for complex tasks mostly aimed at ex­ tion sequences like those recorded in the lithic as­ ploiting mineral, non-mineral and food resources. semblages are the most useful indicators for Large amounts of lithic raw material as well as the predicting human behavior and variability in the physical-geographical and ecological variability way these items circulated (Peresani and Porraz, at the belt between the upper Venetian-Friulian 2004). The settlement system spans from the Ve­ alluvial plain and the Pre-Alps depict the context netian Pre-Alps to the neighboring western and in which Neanderthal groups lived, occupied ter­ eastern regions, where sites are exclusively ritories and migrated according to seasonal ephemeral and reveal limited and specific evi­ rhythms with low residential mobility (Fiore et dence that can infer Neanderthal exploitation at a/. , 2004; Peresani, 200 I; Porraz, 2005). Accord­ the edge of the mountain context (i.e., Caverna ing to their topographic position, some of the Generosa, Bona et al., in press) or to specific main caves in the Venetian Pre-Alps (Grotta di situations in availability of mineral resources. Fumane, Riparo Tagliente, Riparo Mezzena) can To this backdrop a recently discovered site be considered reference sites for sites that have stands out: Rio Secco Cave on the Pradis plateau, been subjected to frequent, complex and intense with its rich scientific evidence that shows human human occupation where the lithic productions mobility indexes, territory occupation, and collec­ were intimately integrated with the acquisition, tion and exploitation of mineral and non-mineral processing and consumption of animal resources resources in the key region situated in the plain­ (Fiore et al. , 2004; Peresani, 2001 ; Thun Hohen­ alpine contact region. Including the neighboring stein and Peretto, 2005). These contexts were as­ Piedmontese transect, the Pradis plateau is located sociated with various ephemeral camps that in between distinct morphological and ecological 86 M. Peresani & F. Gurioli

Fig. 1. Northeastern Italy and the position of the Pradis Plateau in the Carnic Pre-Alps

areas: on the west the Venetian area rich in caves, shelters and open-air sites, some of which are relevant to Mousterian culture; at the east the Giulian Alps with Divje Babe I cave (Turk, Fig. 2. Sketch map of the Pradis Plateau showing the 1997), the Trieste Karst with several sites mostly main morphological features and the position of Rio still undated (Tozzi, 1994) or reported to the Early Secco Cave and Verdi caves: 1- very stepped slopes, 2 Wiirm (Grotta degli Orsi - Boschian, 2003), - stepped slopes, gentle slopes, 4 - sinkholes, 5 - main northwestern Croatia and Dalmatian coast where stream cuts, 6 - Paleolithic cave well-known sites and a few recently discovered archives are reported (Karavanic, 2001).

steep slopes that are sometimes inaccessible. PRESENTATION OF THE SITE The plateau is enclosed by the following Rio Secco cave was discovered in 2002 dur­ mountains: the Pala Mount (1 ,231 m) to the east, ing a survey aimed to explore archaeological evi­ the Rossa Mount ridge (1,369 m) to the north, the dence and to reconstruct the earliest human his­ Ciaurlec Mount (1,148 m) to the west (Fig. 2). To tory of the western Friulian region. It lies at an the south it terminates above the Cosa stream cut altitude of 580 m on the Pradis Plateau in the east­ which runs southward dissecting the hilly land­ em part of the Carnic Pre-Alps (Fig. 1), an oro­ scape, a system formed by WSW- ENE oriented graphic unit 850 km2 bounded to the north by the crests between 250 m and 500 m with progres­ high water-course of the Tagliamento River, to sively decreasing elevation until they reach a clear the south by the high western Friulian plain, to the contact to the upper Pordenone alluvial plain. To east by the eastern bank of the Cavazzo Lake and the southwest, the Col Palis (374 m}--Col Vaita to the west approximately by the meridian passing (367 m) ridge rises from the Meduna fan and from through the Claut Village. Altitudes range from the Ciaurlec Mount and Castelnuovo hills, and 310 m in the Tagliamento valley at the extreme bounds to the north the Sequals marsh in proxim­ northeastern sector of the area to 2,479 mat Mt. ity of which a few Mousterian artifacts were re­ Pramaggiore. Several N- S and W- E oriented val­ covered from the surface. leys with 400 to 800 m elevations and high crests Almost 6 km2 in area, the·plateau elevation (2,000- 2,300 m) between them dissect the region rises from 530 m to 590 m. From its northeastern and make the landscape extremely complex with side through a threshold formed by reliefs 1,000 The Rio Secco Cave, Italy 87 m higher than the plateau it is possible to enter the Corai, 1980; Peresani, in prep). The Grotte Verdi narrow Arzino Stream valley, a tributary of the caves were almost totally emptied by uncon­ Tagliamento running southward from the inner trolled excavations during the 1960s and were in­ Pre-Alps to the plain. To the west, the Chiarza vestigated by research teams in 1970-1971 that stream valley connects with the Meduna Stream recovered several Mousterian lithic artifacts from in the Tramontina Valley, crossing the Pre-Alps the lowermost deposit in Riparo I (sections 7- 13): up to the upper course of the Tagliamento. Due to few sidescrapers, one sidescraper on a thinned its geographic setting between the upper Friulian blank, various flakes partly affected by pseudo­ plain and the Carnic Pre-Alps, the Pradis Plateau retouches. From the same shelter, just a few side­ stands at a strategic position, which may have fa­ scrapers and flakes damaged by pseudo-retouches cilitated human penetration into the alpine region were recovered in sections 3, 4 and 5. Additional and the upper Tagliamento basin. Middle Paleolithic artifacts were found in sec­ The Pradis Plateau has a gentle undulating tions 4 and 5 at Riparo II and in a reworked de­ landscape deriving mainly from the lightly sloped posit. carbonatic formations (Rudist Limestones and Scaglia Rossa - Cretaceous) and partly from the Flysch (Upper Paleocene- Middle Eocene) over LITHIC RESOURCES one third of the total surface (De Nardo, 1999. The stratigraphic series spanning from the The limestone bedrocks were affected by karst Upper Triassic to the Miocene in the Carnic Pre­ degradation processes producing an uneven mi­ Alps is complex due to the various sedimentary crotopography with isolated blocks, brattices and settings (platform, basins, etc.) and as a result of large dolines lined up along the main fractures or intense tectonic activity in this area since the tectonic discontinuities. The bedrock is perme­ Mesozoic. Rocks are mostly carbonatic - dolo­ ated by a dense system of more than 200 explored mites and micritic, oolitic or bioclastic lime­ cavities, some of which penetrate some kilome­ stones, marls and marly limestones - overlain by ters deep and vary in altitude by a few dozen me­ the Claut and Clauzetto Flysch respectively and ters (Cucchi and Finocchiaro, 1981). The few wa­ by the molassic succession (Carulli et al., 2000). terways dissecting the plateau run through the These formations have variable chert content: the bottom of deep and narrow gorges with several dolomites of the Monticello Formation, with rare shelters and caves opening on the walls. nodules and/or black flint lenses; the Dolomia di The Flysch outcrops on the northwestern side Forni, the most important and extended basin and connects to the karst plateau to the Pala unit, with frequent flint levels in the lower portion Mount. Due to its impermeability, a surface hy­ of the sequence; the Chiampomano Limestone, drographic system developed draining toward the with dark flint nodules and beds sporadically karst sector where it is almost completely ab­ present and becoming more frequent at the top; sorbed in the substrate. The landscape assumes the Soverzene Formation in carbonatic facies and typical fluvial features with_ several valleys, ter­ in dolomitic facies, in which dark/black flint nod­ raced surfaces and thin alluvial sheets. The Cosa ules and beds abound, flint sometimes has a yel­ and Rio Secco streams run in gorges almost 1,000 lowish color in the dolomitic facies; the Verzeg­ m deep originating from a combination of tectonic nis Encrinites Mount, containing flint beds in the uplift with karst and alluvial erosion processes encrinitic calcarenites lower portion; the Igne and which divide the plateau from the western Formation, in which dark flint beds abound dis­ slope of the Ciaurlec Mount. continuously; the Vajont Limestone, with rare Some caves opening at the base of rockwalls brownish or reddish flint nodules and flint beds within large collapsed dolines or in the gorges cut mostly in the lower and upper portions of the for­ by the Cosa and its tributaries have been explored mation; the Fonzaso Formation, rich in brownish for the presence of Pleistocene fills. Of these or reddish flint nodules and flint beds, with local caves, the Grotte Verdi and the Grotta del Clusan­ predominance of the silicified portion over the tin were used during the Mousterian and late­ carbonatic one; the Rosso Amrnonitico Superiore, glacial Epigravettian (Bartolomei et al., 1977; presenting in its nodular facies large red flint nod- 88 M. Peresani & F. Gurioli

0

us 1

so

tOO

US4

0 CJO ~ 150 0 ~ us 5 ..

I em ... Fig. 3. Above left, the site viewed from the right side of Rio Secco section with position of the GRS I sounding; right, representation of the sedimentary succession with the anthropically modified bone recovered in unit 5 and used for radiocarbon dating

ules and beds; the Biancone micritic limestones, flint sometimes in large nodules; the Scaglia where flint varies from light (white) to dark in the Rossa and its red nodular flint locally present. form of nodules and beds; the Soccher Limestone Nevertheless, the extensions covered by these and its bioclastic carbonatic sandstones with flint formations are restricted: partly they are distrib­ becoming more frequent in the western zone than uted in W- E strips according to the main tectonic in the easternmost where it dramatically decreases lines corresponding to the overthrusts as in the or disappears; the Scaglia Variegata, with dark case of the Dolomia di Forni; partly they consti- The Rio Secco Cave, Italy 89 tute the upper portions of some main massifs like bounded from the outside through a large boulder the Valcalda, Verzegnis and Piombada mounts, ridge positioned north- south on which two dry­ 13- 14 km away from the site to the N- NE. stone walls lean forming the presumed remains of Looking north, behind the important struc­ a shelter used until a few dozen years ago. The tural element known as the Alto Tagliamento Line cave opens at the center of the rockshelter wall and partially merging with the present-day river and continues as a gallery heading for 12m until course, the micritic and bioclastic limestones of the sediments completely fill this old karst sys­ the Livinallongo Formation outcrop 20 km from tem. Outside the fill forms a slope-waste deposit the site. These limestones are partially or com­ thickening along the present-day drip-line where pletely silicified and sometimes include flint beds the large boulders define the vast original roof ex­ and nodules typically colored green. Their geo­ tension. graphic distribution is restricted to the upper During the summer of 2002 the fill was ex­ Tagliamento valley, the Lumiei Stream valley and plored through two test pits named GRS I and the middle Chiarza stream valley from which it is GRS II: GRS I was opened at the left side of the rapidly dismantled and transported by the river cave-mouth on a 2 x 2m square, then was re­ flows. The richest exploitable outcrops occur in stricted to 0.5 x 0.5 min the NE comer following the Bivera Mount area, at the Creta Forata Massif the lower boundary of unit 4; GRS II was dug at foot, over Ampezzo and along the Lumiei Stream. the foot of the rock-wall facing GRS I but several In spite of such relative abundance and their meters from it. This second test pit proved to be suitability for flaking, these fine-textured pre­ completely sterile and for this reason was discon­ alpine flints were not intensely exploited for two tinued after l m excavation depth. reasons: l) poor accessibility to the primary expo­ GRS I exposed a l . 7 m thick sedimentary sures, which were often scattered on the highest succession including four distinct units at the NE mountain ridges and far from the main rivers; and comer of the test pit that were observed and de­ 2) intense fissuration of the flint beds and nodules scribed (Fig. 3). resulting from tectonic activity. Forming a dense Unit 1: reworked deposits resulting from un­ discontinuity net, this second factor dramatically controlled excavations; bones, fragments of pot­ affects the possibility of using these raw materials tery and a few flint flakes including one pseudo­ except when re-cementation processes periodi­ Levallois point have been recovered. cally ameliorated the lost isotropic qualities. Unit 2: layer of charcoal visible on the west­ Secondary deposits are also varied and are em section incorporated into the upper part of unit sources of pebbles and sub-rounded flint cobbles. 4 and partially removed by the uncontrolled exca­ Well rounded pebbles are available in the poly­ vations. It contains a few lithic artifacts covered genic breccia of the lower Preplans sandstone, an by a thin carbonatic concretion and late Neo­ Oligocene-Miocene clastic unit in the western­ lithic-Bronze Age pottery. It includes unit 3, a central part of the region that has been interpreted small cache with fragmented pottery. as pertaining to the molassic sequence. Coarser Unit 4: clast-supported coarse breccia with materials can be collected· both on river and discontinuous incoherent yellowish-brown dolo­ stream gravel plains and on glacial and fluviogla­ mitic sand infills. Stones and boulders have a cha­ cial deposits at various points along the Taglia­ otic orientation and are covered by thin blooming mento basin. concretions on the lower surfaces. Locally these are visible levels made of thin breccia with sub­ rounded to rounded coarse grains locally ce­ STRATIGRAPHY mented. Archeologically sterile, this unit yielded Rio Secco Cave is a large sheltered cave rare bones. opening on the left slope of the homonymous Unit 5: this unit includes several levels made stream, ca. 20 m above the present-day bed. Fac­ of incoherent matrix-supported breccia. Medium­ ing south, the great shelter has a large, flat roof small sized carbonatic stones tl).at are occasionally derived from the collapse of large slabs driven by larger, sub-angular to sub-rounded; dark yellow­ the stratified limestone. The sheltered area is ish-brown clayey-silty matrix; reddish-brown 90 M . Peresani & F. Gurioli clayey-silty coatings, continuous on the upper worked and thus it was impossible for us to infer stone surfaces. It contains fragmented charred information about the termination of this sedi­ wood, small bone fragments and lithic artifacts. mentary cycle. The lower boundary was not reached. This succession documents two main sedi­ FAUNAL REMAINS mentary events controlled by freezing-thawing From GRS I, dozens of faunal remains were processes and rock collapses that acted to degrade recovered. Due to the small size of this assem­ the walls and the roof of the sheltered cave. Ow­ blage some very preliminary observations will be ing to the restricted size of the explored area we advanced about the agents that acted to produce it cannot evaluate whether the large collapses at the and about the human and non-human occupations drip line or the stream deep cut facing the site re­ at the mouth of the cave. At the moment no data fer to this sedimentary cycle or to other cycles. can be provided to support paleoenvironmental Unit 5 records a depositional phase at the cave­ reconstructions. mouth due to colluvial processes from the outer­ Unit l contains 43 pieces with variable de­ most zone and/or of non-carbonatic fine sedi­ grees of preservation: strongly weathered remains ments originating from the degradation of the ~re assoc.iated with very fresh present-day depos­ Flysch and reworked in the inner karst system. tts. Identified pieces consist of one badger mandi­ Reworked deposits may occur frequently as it has ble (Meles meles), one phalanx and one incisor been revealed elsewhere in other inactive cavities. probably belonging to wild boar (Sus scrofa), and In most cases these are occluded or nearly filled two Caprinae teeth. In unit 2, among the 33 re­ up. The deposition of these fine-textured silicatic covered faunal remains, there are a marmot clavi­ sediments maybe also relate to the outermost cle (Marmota marmota) and one indeterminate stream activity, although neither sedimentary fragment showing a series of short cut-marks. The structures were observed or siliciclastic rock out­ unit 5 bone assemblage counts 92 remains includ­ crops throughout the Rio Secco basin upstream ing one meta podium and three cave bear ( Ursus from the cave. spelaeus) deciduous canines and one marmot A fragmented diaphysis (species undetermin­ (Marmota marmota) metacarpus. A small frag­ able) recovered in unit 5 displays cut-marks that mented herbivore rib has preserved subcircular were made with a lithic tool (Fig. 3) was radiocar­ a~eas (pits) produced by the pressure of large car­ bon dated to 37,790 ± 360 B.P. (LTL429A). As mvore teeth (Ursus?). Anthropic activity on ani­ discussed by Housley eta/. ( 1997) and Zilhao and mal bones in this unit is evident by a dozen burnt D'Errico (2003), direct dating of human-modified bones and by the diaphysis with cut-marks which animal bones gives the most accurate proof of hu­ was sacrificed for 14C dating. Bone preservation is man occurrence in caves used by animals and hu­ good- the surfaces are fresh with sharp edges and mans, while unmodified animal bones - often contrast with ca. 12 pieces bearing blunt or roun­ use d m. 14c ana I ysts. - can be totally unrelated to ded edges presumably altered by water transport any human settlement. ,. or more likely trampling. In sum, this faunal as­ The successive sedimentary phase records a semblage is the result of both large carnivores us­ conspicuous cave-fill aggradation coincidental ing the cave as a den and from human carcass with the deposition of unit 4, which led to an al­ processing. Butchering activities are also re­ most complete obstruction at the cave-mouth. In corded in unit 2. Remains collected from unit 1 relation with the sedimentological features and ar­ define a non-homogeneous assemblage produced ch.aeological content observed at the top of unit 4, by the reworking of both old and modem items. thts event can likely be temporally placed at an in­ terval extending from the late Middle Weichse­ lian to the Late-glacial. The variable sedimenta­ LITHIC ARTIFACTS tion rate throughout the succession is also discer­ Very few fictile remains and lithic artifacts nable in the opposing degrees of anthropization were recovered in a good state of preservation recorded in both units. The uppermost 50 em of from the GRS I sounding. Except for unit 1 with the stratigraphy at this stage had been recently re- its reworked content, only two units provided The Rio Secco Cave, Italy 91 evidence of human presence and these consist of very few items, presumably due to the limited ex­ tent of the excavation. Although unit 5 was ex­ plored on a more reduced area (0.5 x 0.5 m), it re­ veals the highest concentration of lithic and ~ \ faunal remains along the overall stratigraphy. Unit 1: Some pieces were recovered from re­ w worked sediments: fragmented Bronze Age pot­ tery and four flint artifacts, two small cores and two flakes in a good preservation state unaffected by patina or other alteration. Both of the cores dif­ 2 ferentiate from flakes by the presence of thin car­ bonatic films on their surfaces: the first one was found on a rounded block of Soverzene flint col­ lected in the Meduna stream plain, the second one on a greenish very fine cretacious flint nodule of unknown provenance. Flakes are made in lime­ stone or in Jurassic flint. The latter is a pseudo­ Levallois point struck from a cobble bearing 3 traces of stream/fluvial transport. The thin edge opposed to the back displays modification pre­ Fig. 4. Artifacts recovered in unit 2: I - bladelet flake-core) and unit 5; 2- pseudo-Levallois point; 3 - sumably due to utilization. Because these items Levallois core. Drawings by G. Almerigogna and S. are indistinguishable in their taphonomic features, Muratori they can be grouped according to their techno­ logical features into two assemblages: 1) Late Neolithic-Bronze Age (the cores); and Mouste­ rian (the pseudo-Levallois point). trimmed striking platform covering the overall Unit 2: This unit contains one fragmented perimeter and centripetal recurrent scars on the bladelet with cortical back and one bladelet flake­ extraction surface with repeatedly struck hinges core. The flint is gray and vitreous, lightly pati­ leading to its discard. Among the five entire and nated, presumably collected from cretaceous for­ indeterminate fragmented flakes we describe one mations (Biancone?) in the Venetian Pre-Alps be­ flake with centripetal scars affected by pseudo­ hind the Alpago Basin. The core was obtained on retouches and one small pseudo-Levallois point a thin plunged flake struck from flint filled with produced with discoid technology (Fig. 4: 2) cemented fissures. Flaking occurred along the made from coarse textured flint similar to the Ver­ main axis following the distal thickening (Fig. 4: zegnis Encrinites. 1). As listed below, the following traces are visi­ The unit 5 group defines a homogeneous as­ ble: remains of lamellar scars, a stop notch, the semblage containing Levallois technology and last failed or hinged detachments ending the se­ presumably discoid technology as suggested by quence. Further operations aimed to shape and the core and the pseudo-Levallois point, which is regulate the volume can be observed on the lower a typical primary product of this method. As a face. rule, the technological features (dihedral butt with Unit 5. From this unit one Levallois core and convex surfaces, the operative sequence) exclude five artifacts were recovered. The core is in any doubt that this item is a by-product of the Le­ Livinallongo flint and the lower surface has abra­ vallois procedure. Indeterminable fragments and ded ridges and features typically derived from flu­ other flaking waste products indicate on-site flake vial transport (Fig. 4: 3). Similar provisionable production. However, less certainty is possible for raw materials are likely to be found in the fluvial framing the artifacts recovered in unit 2 either or fluvio-glacial alluvial gravels spread through­ chronologically or culturally. Technological fea­ out the Tagliamento river basin. The core has a tures like to place them in the Upper Paleolithic 92 M. Peresani & F. Gurioli and presumably the Epigravettian, but they clash another well-known site in the Lessini Mountains, with the co-association of fragmented pottery. the Tagliente shelter, which has recently provided Moreover, no carbonatic film has been deposited evidence of technological variability at the very on these potsherds, contrary to the lithics. An as­ end of the Mousterian or at least during OIS 3 sociation resulting from reworked non-coeval (Arzarello and Peretto 2005). Rio Secco should items cannot thus be excluded a priori. also be contemporary with levels from B to El and El/E2 in Mujina Pecina, Dalmatia, a site dated to 45- 34 ka BP (Rink eta/., 2002). The ar­ CONSIDERATIONS chaeological content counts Levallois items in Our estimates using chronometric, lithologi­ lower levels (E- D) and small tools like notches cal and paleontological data seem to support the and denticulates in the uppermost ones (C- B). interpretation that the explored stratigraphy of the Rio Secco cave records for the first time the Rio Secco cave was rapidly deposited as a conse­ final Middle Paleolithic in Friuli and provides a quence of roof collapse and other events in a few numeric age at the MP- EUP boundary, a period tens of thousands of years. At the moment no pa­ which has until now been better investigated in leoecological data are available to infer whether the neighboring regions. Given its particular posi­ the final Middle Paleolithic at Rio Secco is em­ tion at the crossroads of the plains-mountains ar­ bedded in sediments formed under interstadial or eas and the total preservation of the cave fill, stadia! conditions, or to confirm the chronological many new topics could be approached by future position of the human occupation recorded in unit research on stratigraphy and paleoecological set­ 5. Recent improvements from the Azzano X pol­ ting, chronology, site-function and site dynamics, len record, a core from the Friulian plain, indicate as viewed through techno-economical and ar­ continuous oscillations in the percentage values chaeozoological analyses of lithic and non-lithic of Pinus, Picea, Betula, Graminae and xero­ relics recovered in this human context. phytes, possibly resulting from alternations be­ tween steppe/taiga environments. According to Acknowledgments Pini et a/. , (in press) the climate never reached Archaeological surveys on the Pradis Plateau and temperatures supporting broad-leaved forests. research on the Rio Secco Cave have been carried out A handful of sites with Mousterian industries through concession by the Ministero dei Beni Culturali 4 fall by different numeric ages ec, U/Th, TL, rac­ and the Soprintendenza per i Beni Archeologici of emization) in the radiocarbon interval of 40-30 ka Friuli-Venezia Giulia (Dr. S. Vitri). The authors are BP in northeastern Italy and Croatia. At Fumane very grateful to Dr. P. Visentini, Director of the Museo Cave in the Veneto Pre-alps, the group of units A della Grotta at Pradis for his invitation to commence reveals variable crioergic activity and wooded and conduct field research and surveys; Clauzetto Mu­ landscape in units from All to AlO that persisted nicipality for financial support; Mr. L. Colledani for in­ until unit A3, when conditions shifted toward formation about the site location; geologist A. Riva for details concerning the geology of the area and the flint more open contexts and a cool-dry climate. The sources; geomorphologist R. Avigliano for text revi­ Rio Secco date covers the dispersed radiocarbon sion; students and collaborators that took part in the set depicted by units A 11, A l 0, A9 and A6, which field-work; N. Uomini for revision of the english text. contain Levallois (All, AlO and A6) and discoid The following study is part of the Research Project (A9) lithic assemblages (Peresani eta/., in press). granted by the University of Ferrara and entitled "Evo­ Levallois products are mostly long and thin and lution of techno-functional systems and subsistence were obtained using the unidirectional recurrent strategies from the Palaeolithic to the Mesolithic in dif­ method, whereas the centripetal recurrent was ap­ ferentiated ecological contextes". For this specific plied during the final steps of the reduction se­ work F. Gurioli carried out the analysis of faunal re­ mains. quence (Broglio eta!., 2003). An abrupt techno­ logical change observed in unit A9 is due to the appearance of complete and exclusive discoid re­ REFERENCES duction sequences (Peresani, 1998). To the same ARZARELLO M., PERETTO C. 2005. Donnees interval belong the undated upper sections from nouvelles sur les caracteristiques et I' evolution The Rio Secco Cave, Italy 93

techno-economique de l' industrie mousterienne du HOUSLEY R. A., GAMBLE C. S. , STREET M. , PET­ Riparo Tagliente (Verone, ltalie). Actes du CoHo­ TITT P., 1997. Radiocarbon evidence of the Late­ que International Donnees recents sur les modalites glacial human recolonisation of northern Europe. de peuplement et sur le cadre chronostratigra­ Proceedings of the Prehistoric Society, 63, 25-54 phique, geologique et paleogeographique des in­ KARA V ANIC I. (200 I). The Middle Paleolithic settle­ dustries du Paleolithique inferieur et moyen en Eu­ ment of Croatia. In Conard N. J. (ed.), Settlement rope, BAR Int. S. , Sl364, 281-289. Dynamics of the Middle Palaeolithic and Middle BARTOLOMEI G., BROGLIO A. , PALMA DI CBS­ Stone Age. Tiibingen Publications in Prehistory, In­ NOLA A. 1977. Chronostratigraphie et ecologie de troductory Volume, 185-200. l'Epigravettien en Italie. In AA. VV., La fin des temps PERES ANI M. I 998. La variabilite du debitage dis­ glaciaires en Europe - Chronostratigraphie et co'ide dans Ia Grotte de Fumane (Italie du nord). ecologie des cultures du Paleolithique final, Actes Paleo, 10, 123- 146. Colloque International C.N.R.S., 271,297-234. PERES ANI M. 200 I . An overview of the Middle Pa­ BONA F., PERESANI M., TINTOR! A. in press. Les laeolithic settlement system in North-Eastern Italy. grottes aours avec indices de frequentation humaine In Conard N. J. (ed.), Settlement Dynamics of the au Paleolithique moyen final. L'exemple de Ia Middle Palaeolithic and Middle Stone Age. Tiibin­ Caverna Generosa dans les Prealpes lombardes, gen Publications in Prehistory, Introductory Vol­ ltalie. L 'Anthropologie. ume, 485-506. BOSCHIAN G. 2003. Proceedings International Round PERESANI M. 2003. An initial overview of the middle Table The Alps: environment and mobility, Preisto­ Palaeolithic discoid industries in Central-Northern ria Alpina 39,91-102. Italy. In Peresani M. (Ed.) Discoid Lithic Technol­ BROGLIO A., LEMORINI C., PERESANI M. , ogy. Advances and implications. British Archaeo­ ROSSETTI P. 2003 . Modifications culturelles et logical Reports, International Series, 1120, 209- comportamentales entre Mousterien et Aurignacien 223. au sud des Alpes. In J. Bruzek, B. Vandermeersch, PERESANI M., CREMASCHI M., FERRARO F., M. D. Garralda (eds.) Changements biologiques et FALGuERES Ch., BAHAIN J.-J., GRUPPIONI G., culturels en Europe de Ia fin du Paleolithique SIBILIA E., QUARTA G. , CALCAGNILE L. , moyen au Neolithique. Actes du Colloque Interna­ DOLO J.-M. (in press). Age of the final Middle tional, Laboratoire d' Anthropologie des Populations Palaeolithic and Uluzzian levels at Fumane Cave, 14 234 30 du Passe. Universite Bordeaux I, 39-59. Northern Italy, using C, ESR, UP Th and CARULLI G. B. , COZZI A., LONGO SALVADOR thermoluminescence methods. Journal of Archaeo­ G., PERNARCJC E., PODDA F., PONTON M. logical Science. 2000. Geologia delle Prealpi Carniche. Museo Friu­ PERESANI M. , PORRAZ G. 2004. Re-interpretation lano di Storia Naturale, 44. et mise en valeur des niveaux mousteriens de Ia CORAl P. 1980. Le piu antiche culture preistoriche Grotte du Broion (Monti Berici,Venetie). Etude della Ladinia (Paleolitico e Mesolitico). Ladinia, techno-economique des industries lithiques. Rivista IV, 183-2 18 . di Scienze Preistoriche, LIV, 181-247. CREMASCHI M., FALGUERES C., GRUPPIONI G., PIN! R. , DONEGAN A M., RA VAZZI C., A VIGLI­ PERESANI M., QUART A G., SIBILIA E. submit­ ANO R., CALDERON! G. in press. Late Pleisto­ ted. Age of the Late Mousterian levels at Fumane cene pollen stratigraphy and vegetation history at Cave, Northern Italy, using 1 ~, electron spin reso­ Azzano Decimo, southeastern Alpine foreland. 234 30 nance, UP Th and thermoluminescence methods. PORRAZ G. 2005. En marge du milieu alpin. Dy­ Journal ofArchaeological Science. namiques de formation des ensembles lithiques et CUCCHI F., FINOCCHIARO F. 1981. Note sui car­ modes d 'occupation des territoires au Paleolithique sismo ipogeo dell'altopiano di Gerchia (PN). Attie Moyen. Ph.D. dissertation, Universite de Provence, Memorie Commissione Grotte "E. Boegan", Tri­ Formation Prehistoire, Archeologie, Histoire et este, 21, 19-29. Civilisations de I' Antiquite et du Moyen-Age. DE NARDO M. T. 1999. Rilevamento geologico RINK W. J., KARAVANIC I. , PETTITT P. B. , VAN dell 'area di Pradis di sotto (Prealpi carniche). Atti DER PLICHT J. , SMITH F. H. , BARSTOLL J. VIii Convegno Regionale di Speleologia del Friuli­ 2002. ESR and AMS-based 14C dating of Mouste­ Venezia Giulia, 131 - 134. rian levels at Mujina Pecina, Dalmatia, Croatia. FIORE I. , GALA M., T AGLIACOZZO A., 2004, Ecol­ Journal ofArchaeological Science, 29, 943-952. ogy and subsistence strategies in the Eastern Italian TOZZI C. 1994. II Paleolitico inferiore e medio del Alps during the Middle Palaeolithic. International Friuli - Venezia Giulia. In Atti XXIX Riunione Sci­ Journal ofOsteoarchaeology, 14, 273- 286. entifica lstituto Italiano Preistoria e Protostoria, 94 M. Peresani & F. Gurioli

19- 36. geographique des industries du Paleolithique TURK I. (Ed.). 1997. Mousterian "Bone Flute" and inferieur et moyen en Europe, Actes du Colloque In­ other finds from Divje Babe I cave Site in Slovenia. ternational, British Archaeological Reports, Int. S. , Opera Instituti Archaeologici Sloveniae (2). 1364,261-268. Znanstvenoraziskovalni Center SAZU, 119- 156. ZlLHAO, J., D'ERRlCO F. 2003. An Aurignacian THUN HOHENSTEIN U. , PERETTO C. 2005. Faunal "Garden of Eden in Southern Germany"? An alter­ exploitation in the Middle Palaeolithic: Evidences native interpretation of the Geissenkliisterle and a from Riparo Tagliente (Verona, Italy). In: Donnees critique of the Kulturpumpe model. Paleo 15, recents sur les modalites de peuplement et sur le 69- 86. cadre chronostratigraphique, geologique et paleo-

...