CHRONOSTRATIGRAPHY OF THE PLEISTOCENE/HOLOCENE BOUNDARY: THE AZILIAN PROBLEM IN THE FRANCO-CANTABRIAN REGION

Lawrence Guy Straus*

ABSTRACT: The chrono-stratigraphic evidence for the Azilian of Vasco-Cantabrian Spain and France is reviewed and found to range from the Allerød to the Preboreal. A survey of the associated artifact and fauna assemblages indicates the transitional nature ofthe Azilian between the Magdalenian and the Mesolithic. The relationship between Azilian technology and supposedly abrupt adaptations to radically changed en­ vironmental conditions at the 10,000 B.P. boundary is not straightforward.

KEYWORDS: Azilian, Franco-Cantabria, Magdalenian, Upper Palaeolithic, Mesolithic, Pleistocene/Ho­ locene boundary, chrono-stratigraphy, radiocarbon, pollen, sedimentology.

1. INTRODUCTION Holocene transition in the Franco-Cantabrian re­ gion (fig. 1). This article attempts to clarify the question of the As a result of his excavations in the cave of El precise age of a culture-stratigraphic unit long Pendo and other sites in the Province of San tander recognized as transitional between the classic Up­ ('Cantabria'), J. Carballo (1922; 1960) argued that per Palaeolithic and Mesolithic: the Azilian. It is the Cantabrian Azilian was locally derived from the clear that normative characterization of prehistoric Upper Magdalenian and, consequently, that it was 'cultures' to 'date' archaeological deposits to pa­ an early manifestation of this 'culture'. Carballo leoclimatic ph ases is a procedure based on circular thus placed the 'origins' of the Azilian in Canta­ reasoning. It is also obvious that it cannot be brian Spain, a tendency followed to this day by assumed a priori that all Azilian-type assemblages certain other regional prehistorians, notably his or fossil directors must date to the same (brief) disciple, J. Gonzalez Echegaray. In recent years a period. Specifically, the assumption that all Azilian number of sites with Azilian deposits (unfortuna­ deposits should have precisely the same age as those tely often only impoverished remnants) have been of the classic regions of south western France is excavated in the region and in the adjacent French fa llacious. Such normative assumptions are based Pyrenees. Most of these excavations have incor­ on a view of culture as a simultaneous ethnic style, porated paleoclimatic analyses and radiocarbon rather than the extrasomatic means of adaptation. dating. Although few are as yet adequately pu­ It is the proposition of this article to investigate the blished, a corpus of data is beginning to accumulate chronology of the Azilian as the culture-strati­ concerning the absolute age ofthe Azilian along the graphic unit which corresponds to the Pleistocene­ northern flank ofthe mountain chain formed by the Holocene transit ion by means of various chrono­ Pyrenees and Cantabrian Cordillera. metric methods: radiocarbon, sedimentology and The palynological data for the sites in Asturias palynology. These methods are independent of the and Santander have all been produced as a result of artifact assemblages per se, although each is sus­ analyses by Arlette Leroi-Gourhan and her disci­ ceptible to errors of interpretation. No single datum ple, Anais Boyer-Klein, at the Musee de l'Homme and no single site can be considered as conclusive; it (Paris). Based primarily on Leroi-Gourhan 's (1967) is the ensemble of various chronological indicators interpretation ofthe pollen record from the cave of from various sites which can provide the basis for the establishment of a reliable chronostratigraphy. This article prop oses to contribute to the objective *Department or Anthropology, University or New Mexico, elaboration of such a scheme for the Pleistocene- Albuquerque, N.Mex. 87131, U.S.A.

89 90 L.G.STRAUS

At/antie Massif Oeean 22 Central

5 Cordille a 12 r 8 P 13 Yrel1 Sea j e es O 100 km

Figure I. Major Azilian sites discussed in text: I. La Paloma, Cueva Oscura de Ania; 2. Los Azules; 3. La Riera; 4. El Pendo, Cueva Morin; 5. Arenaza; 6. Santimamifie;7. Urtiaga, Ekain, Erralla; 8. Zatoya, Abauntz; 9. Isturitz; IO. Duruthy, Dufaure; II. Poeymau; 12. Mas d'Azil; 13. La Vache, Rhodes II; 14. La Tourasse; 15. Gazel 16. Chinchon, La Salpetriere; 17. Saut du Loup; 18. Tete du Chien; 19. Pegourie, Graves Pielago, Rascano, El Otero; 20. Borie del Rey; 21. Faurelie; 22. Pont d'Ambon; 23. Chez Jugie; 24. Bois Ragot; 25. Jean Pierre I; 26. Thoys; 27. Abri Gay; 28. Les Douattes, Blame-de-Thuy; 29. Varennes-Ies-Måcon; 30. Rochedane; 31. Mannlefelsen.

La Vache in the high Pyrenees of Ariege (South­ the Basque sites of Urtiaga, Arenaza, Ekain and central France), they invariably place Cantabrian Abauntz (as well as some possibie Azilian deposits Azilian deposits in the Allerød oscillation, which at Zatoya in Navarra) are all radiocarbon dated to would indeed give this culture-stratigraphic unit Dryas III and Preboreal. Palynological interpreta­ relative antiquity (c. 12,000-11,000 years B.P.). tions of deposits at Duruthy and Dufaure by M.M. Such an assignment has been given by these spe­ Paquereau (1978; n.d.) coincide with the sedimen­ cialists to Azilian deposits in Cueva Oscura de tological evidence and radiocarbon dates in placing Ania, La Riera, El Pendo, El Otero, Cueva Morin, the Azilian levels of these contiguous Gascon sites and Rascafio. This chronological assignment, ba­ in Dryas III-early Preboreal. sed on the interpretation of palynological evidence These data clearly point to the need for a re­ of sometimes relatively high AP (arboreal pollen) analysis of the age of the Vasco-Cantabrian Azi­ percentages including several thermophile taxa, is lian-is it indeed 'early' or 'Iate' vis il vis. the Azilian contradicted by radiocarbon and/o r sedimentolo­ elsewhere? Two recent doctoral theses, one on the gical evidence of a Dryas III age (c. 11,000-10,000 Azilian of Santander and Asturias by J.A. Fer­ B.P.) in the cases of La Riera, Rascafio, Morin and nandez-Tresguerres (1980) and the other on the El Pendo and a Preboreal age (c. 10,000-9,000 B.P.) Asturian ofthese two provinces by M.R. Gonzalez in that of Cueva Oscura de Ania. Reasonable cases Morales (1982), as well as a polemical article on for Dryas III age can be made at Pielago and for these two 'cultures' by L.G. Straus (1979), propose Preboreal age at Los Azules. The Azilian levels at the likelihood of a late date for the regional Azilian The Azilian problem in the Franco-Cantabrian region 91 and suggest its close temporal relationship (if not endscrapers an flakes and backed bladelets versus parti al overlap) with the specialized Asturian shell relatively low percentages af burins). Among the middens. other 'possible' Azilian deposits, the assemblage A systematic review af the question of Azilian from Poeymaii Level 'BS' is as yet undescribed and age should begin by a classification ofthe sites with is simply said to contain Azilian 'elements', so its chronometric information independent af subjec­ assignment is in same doubt. The assignment of tive judgements' af relative antiquity based an the Abauntz Level 'd' to the Azilian is based on the artifact assemblages themselves. (There are cer­ characteristics af its lithic assemblage and an its tainly toa few well-excavated Azilian sequences to stratigraphic position. La Riera Level 27 directly pro vide real bases for any pretence af 'seriation' ar underlies a level (28) with an Azilian harpoon, other relative chronological methods based on the overlies another (24) with a Magdalenian harpoon, artifacts (pace Garcia Guinea, 1975)). In the case af and contains 10.5% Azilian points and two ochre­ Vasco-Cantabrian Spain and the adjacent French stained cobbles, making its assignment to the Basque Country/Gascony, such a classification can Azilian 'possible' ar even 'probable', despite many be done as follows: similarities to underlying Magdalenian assembla­ l. Secure Azilian deposits dated directly by 14C: ges. These 'cases point out the inadequacy and Los Azules Level 3-upper; Cueva Oscura de Ania; circularity (but current inevitability) af normative Urtiaga Level C; Ekain Levels V, IV, III; Dufaure culture-stratigraphic schemes based an either fossil Level 3. director artifact types ar assemblage compositions 2. Secure Azilian deposits dated terminus ante described in terms af relative frequencies. and/or post quem by other levels dated by 14C: Los Same ofthe Vasco-Cantabrian deposits dated by Azules Level 3e & 2; La Riera Level 28; Morin Level 14C are als o dated geologicaIly and/or palynolo­ 1; Duruthy Level 2; Poeymau 'CPE'. gically, albeit with several cases of conflicting 3. PossibIe Azilian deposits dated directly by 14C: interpretations (as noted above), in which the La Riera Level 27; RascafioLevel 1;Arenaza Levels palynological assessment of age is invariably older III & II; Zatoya Levels 'III & II-Iower; Poeymaii than that of sedimentology and/or radiocarbon. 'BS'; Abauntz Level 'd'. Naturally, it should be noted that the data from all 4. Azilian deposits dated only geologicaIly and/or these sites are ofunequal quantity and quality. The palynologically: PieIago; El Pendo Level I; El Otero essential relevant facts from each site are presented Level 1(?). systematically in geographical arder, beginning in 5. Azilian deposit dated terminus post quem geo­ the west with the recently excavated Azilian sites of logicaIly or palynologically: La Palarna Level 2; Asturias and Santander. Isturitz.

Deposits considered to be 'secure' in their Azilian 2. TRE SITES OF ASTURIAS AND SANTAN­ attribution are those which contain flat-section DER (CANTABR IA) harpoons and no cylindrical-section anes. In same cases (El Pendo, Marin, Dufaure, Rascafio), Azi­ 2. 1. Cueva Oscura de Ania (Las Regueras, As- lian harpoons were found in old excavations but turias) not in the recent anes from which the independent chronological data are derived. In these cases This site is located near the confluence of the Rios correlations af variable security can be made be­ Anda1l6n and Nal6n in Central Asturias, about 9 tween the relevant strata uncovered in the old and km WNW of Oviedo. Excavated and summarily new excavations. In a few other cases (EI Otero, published by J.M. G6mez-Tabanera, M. Perez and Abauntz, Arenaza, Zatoya, Duruthy) no Azilian J. Cano (1975; Perez, 1977), Cueva Oscura de Ania harpoons (ar decorated cobbles) have (yet) been contains two Azilian leveis, overlying a sterile layer found, but an 'Azilian' attribution is likely ar at and an Upper Magdalenian stratum. Level I con­ least possible, due to the stratigraphic position af taineq a fragment of a flat-section harpoon and a the deposits in question (i. e. overlying the Upper lithic industry dominated by small endscrapers and Magdalenian and/ar underlying deposits with ce­ backed bladelets. There are at least two flat-section ramics) and/ar-more problematically-due to the unperforated harpoons in Level II, which appa­ composition af the lithic industries (L e. presence af rently has even higher percentages af small end­ many 'Azilian' points and high percentages af small scrap'ers and backed bladelets, and even fewer 92 L.G.STRAUS

burins than Level I. According to Arlette Leroi­ 2.3. Los Azules I (Cangas de Onis, Asturias) Gourhan and Renault-Miskovsky (1977), the pol­ This small cave, located in the mountainous in­ len, which are indicative of a warming trend, terior of eastern Asturias on a hillside dominating suggest an Allerød age for the Azilian deposits in the confluence of the Rios Guena and Sella, is at this cave. However, the samples are completely present the most important Azilian site in the isolated, making such a chronological assignment Cantabrian region due to its meticulous, extensive, high ly speculative. In fact, a radiocarbon date (on on-going excavation and publication by J.A. Fer­ unknown material) of about 9,400 B.P. for the nandez-Tresguerres (1980, with refs.). Los Azules I Azilian of Cueva Oscura has recently been made is especiaIly well-known for its Azilian burial with public by M. Perez (cited by Gonzalez Morales, associated grave goods. The stratigraphy is as 1982: p. 59). This would suggest a Preboreal age, follows: however, no specifics have yet been published Level l: sterile yellowish clay which reached the concerning this date and there are unfortunately no roof of the cave mouth. published data on fauna or sedimentology from this Level 2: compact reddish clay with a small site. Azilian industry and evidence of solifluction. Level 3a, b, c & d (=3-upper) plus 3e: alternating 2.2. La Paloma (Las Regueras, Asturias) reddish and blackish hearth lenses with abundant Located about 2 km north of Cueva Oscura de Ania residues of Azilian occupation. in the valley of the Rio Soto, the cave of La Paloma Level 3f:lens localized in the cave interior, as yet contained a long Magdalenian and Azilian cultural (1980) essentiaIly unexcavated. sequence excavated in 1914-15 by E. Hernandez­ Level 4: sterile yellowish clay. Pacheco, with the collaboration of J. Cabre and P. Level 5: reddish-brown sediments with Upper Wernert, as well as the Conde de la Vega del Sella. Magdalenian material. Recently a restudy of this important site was Level 6: black sediments with Magdalenian ma­ published by M. Hoyos et al. (1980). According to teria!. these authors, the levels variously labelled as '2', 'b' The burial was situated between lenses 3a and 3d, and 'c' in different parts of the cave correspond to which were themselves intact and undisturbed by the Azilian deposit described by Hernandez-Pac­ the grave excavation. Level 2 contains Azilian heco. They were able to locate and study 263 points and harpoons, despite the limited size of its retouched tools from this level. Over half (51 .7%) artifact assemblage (69 stone tools), which is do­ are endscrapers (mostly on small flakes)and 20.2% minated by typical small endscrapers on flakesand are backed bladelets (despite the early date of the backed bladelets. Levels 3a-d are artifactually ho­ excavation and subsequent vicissitudes of the col­ mogeneous. The 1819 retouched tools are domin­ lections). There are 4.6% Azilian points and only ated by backed bladelets (43.9%) and endscrapers 6.4% burins. The bone industry includes a flat­ (23.8%)-most of which are on small flakes-with section harpoon with a circular basal perforation. few burins (5. 1 %) and two atypical triangles. There The extant faunal collection from the Azilian are 8.0% Azilian points and 12 flat-section har­ level is dominated byCervus (MNI=ll), but also poons, several of which were in direct association includes Sus (MNI=I , the only boar in the whole with the burial, as were some ofthe painted cobbles 5-level sequence) and Capreolus (MNI=2), both found in Levels 3a-d. Level 3e yielded 761 re­ classic woodland species. La Paloma lacks radio­ touched tools: 40.6% backed bladelets, 23.7% end­ carbon dates and palynological analyses, and the scrapers (mostly on small flakes), 8.9% Azilian Azilian level could not be studied sedimentolo­ points, and only 3.0% burins. There are also flat­ gically, as it was entirely dug out, leaving no clear section harpoons and a few painted cobbles. witness section which could be sampled. Level 3, Levels 3a and 3d have been radiocarbon dated also now missing, was a thin, stony, nearly sterile (with wood charcoal samples). Undisturbed Level deposit with a possibie Magdalenian harpoon. 3a, overlying the burial, dates to 9,430±120 B.P. Level 4, with a high proportion of large angular (CSIC-216) and Level 3d, equally intact and under­ spalIs, and containing a 'Late Magdalenian' in­ lying,the burial, dates to 9,540±120 B.P. (CSIC- dustry with uniserial and biserial harpoons, was 260). Preliminary palynological results summarized forrned under cold, dry conditions. This rigorous by Fernandez-Tresguerres (1980: p. 128), include climatic episode could conceivably correspond to an AP of about 65% for the basal part of Level Dryas II or III. '3-upper', including Betula, Corylus , Pinus and The Azilian problem in the Franco-Cantabrian region 93

Ulmus. Arboreal pollen decrease somewhat in the tander (Clark, 1976: p. 343), and would hardly middle of Level '3-upper', while those of the qualify as a strong indicator of Pleistocene age for Gramineae increase. However, the AP increases the Los Azules burial. again in Level 3a. Fernandez-Tresguerres places 3-upper in the Preboreal and the solifluction of 2.4. La Riera (Posada de Lianes, Asturias) Level 2 in the la te Preboreal/early Boreal. The (as yet unquantified) fauna of Los Azules Level 3 is La Riera is a relatively small cave situated at the globally dominated by Cervus elaphus, but Sus base of the south face of the Llera ridge, 1.5 km scrofa is also abundant, while Capreolus capreolus from the present shore of the eastern sector of and other species which prefer wooded habitats Asturias. Discovered and first excavated by the (Felis sylvestris, Meles meles) are also present Conde de la Vega del Sella during the First World (Fernandez-Tresguerres, 1980: p. 41). All these data War, La Riera contained a nearly complete, intact clearly indicate a Post-Pleistocene age (probably stratigraphic sequence spanning the period from Preboreal) for the Upper Azilian levels at Los the Solutrean through the Asturian. Sandwiched Azules I. The total disappearance of Littorina between the Upper Magdalenian and the Asturian littorea from the molluscan fauna of Levels 3-upper shell midden, Vega del Sella (1930) found a thin and 2 may be an added indicator of Post-Pleisto­ Azilian layer with a typical flat-section harpoon cene age, since this species is relatively intolerant of (with buttonhole perforation) and a scanty lithic warmer ocean temperatures (Clark, 1976: p. 233). industry. Obermaier (1925) claimed that a painted Recently a series of bone collagen 14C dates has cobble was also found, but this piece was never been published for the lower part ofthe Los Azules figured, nor is it mentioned by Vega del Sella; it is at Azilian sequence: Level 3d/e: 1O,400±90 B.P. (BM- present missing from the old La Riera collections. 1879); Level 3e: 1O,330±190 B.P. (BM-1875), Subsequent test excavations by G.A. Clark and 10,700±190 B.P. (BM-1876), 11,190±350 B.P. by G6mez-Tabanera and M. Perez did not involve (BM-1877); Level 3f: 1O,720±280 B.P. (BM-1878) the Azilian stratum. However, the recent project (Burleigh et al., 1982; Fernandez-Tresguerres, pers. directed by Clark and L.G. Straus (Straus et al., comm.). These dates, which are coherent among 1981; Straus & Clark, 1986) revealed a series of themselves and vis il vis those of the upper levels Tardiglacial and early Holocene deposits, one or when standard deviations are taken in to consi­ more of which is attributable to this culture­ deration, indicate that the Los Azules Azilian stratigraphic unit. Level 24, with a cylindrical­ spanned the whole ofDryas III as well as Preboreal. section biserial harpoon fragment and rich lithic The statement by Burleigh et al. (1982) that these and bone industries typical of the Upper Magda­ are 'shell-midden' levels is inexact, as Los Azules lenian, is dated (on charcoal) to 1O,890±430 B.P. contains no true shell middens. Molluscs are pre­ (GaK-6982). (In Level 24, as in El Pendo Level II sent but rare. and a few other deposits assigned by the analysts to The Preboreal age of the Los Azules burial is the Upper Magdalenian, endscrapers outnumber doubted by NewelI et al. (1979: p. 160), who believe burins despite the norm to the contrary, and there it to be of Palaeolithic date, on the grounds that the are several Azilian points). This level presents associated buttonhole perforated harpoons are not geological evidence of cold, albeit humid climatic of Mesolithic type and that the associated mussel conditions, attributable to Dryas II, according to (Modiolus barbatus) is absent from Asturian (Pre­ the analysis of H. Laville. However, palynologi­ boreal/Boreal) shell middens in this region. an the cally it is considered to be Allerød in age; with a first count, buttonhole perforations are in fact far lowered AP (including birch) in the analysis of more common in Azilian harpoons than round Arlette Leroi-Gourhan, and, given the range of perforations in the Vasco-Cantabrian region (Car­ error in its radio carbon date, it could pertain to a ballo, 1960: pp. 164-173; Barandiaran, 1967: pp. cool episode within this oscillation. The Level 24 322-323; Fernandez-Tresguerres, 1980: pp. 145- fa una studied by J. Altuna includes a few remains 147). These types have indeed been found to lack of reindeer. The overlying Level 25 is a discon­ relative chronological value (Nougier & Robert, tinuous calcium carbonate crust with only two 1977). an the second count, a search of over a denticulates as its to ol assemblage. Level 26 is a dozen Vasco-Cantabrian Palaeolithic site reports thin, charcoal-stained lens with a small lithic in­ failed to reveal a single find of Mo diolus barbatus. It dustry which has 11.5% endscrapers, several back­ is found, however, at present in the Bay of San- ed bladelets, microgravette and Azilian points, and 94 L.G.STRAUS virtuaIly no burins. Sedimentologically, Levels 25 2.5. El Pendo (Escobedo de Camargo, Santander) and 26 indicate tempera te and very humid climatic The immense cavern of El Pendo is situated at the conditions attributable to the Allerød interstadial. base of a large dolina in hills which dominate the Sus is present and Capreolus is relatively abundant, Santander coastal plain. Explored since the earliest however, Littorina littorea is still fairly frequent in years of prehistoric investigation in the last quarter the small molluscan assemblage, though less so of the 19th century, El Pendo was sporadically than in Level 24. Monodonta lineata, absent in the exc vated f r four decades by J. Carballo (1960) latter level, makes a timid appearance in Level 26. � � dunng the fmt half of the 20th. Re found several There are no 14C dates for Levels 25-26. flat-section harpoons in different parts of the cave. Level 27 is a thick deposit, rich in mollusc shelIs These finds were supposedly made in situ in strati­ (still including Littorina), artifacts, bones and ebou­ graphic superposition above an Upper Magdale­ lis. The 1?5 stone tools include 17.5% endscrapers, nian deposit, and constituted the basis ofCarballo's 7.6% bunns, 29.8% backed bladelets sensu stricto theo ry of a gradual, early transition between the plus 10.5% Azilian points. There are 2 cobbles with Magdalenian and Azilian in Cantabrian Spain. traces of red ochre. The fauna includes Sus and In 1953-57, an international team, directed by J. Capreolus, but there are also remains ofthe Nordic Martinez Santa-Olalla in collaboration with Andre vole (Micro tus oeconomus). Level 28 is a very thi n Leroi-Gourhan and A. Cheynier, conducted a localized lens with only 32 lithic tools (17 of which major excavation in another sector of the El Pendo are back ed bladelets and 3 of which are Azilian entrance, results of which were only recently pu­ points). A nearly complete flat-section, uniserial . bhshed by J. Gonzalez Echegaray (1980). The harpoon with four barbs and a basal 'buttonhole' upper part ofthe stratigraphy ofthe immense Post­ perforation was, however, found in this level, along Palaeolithic, Upper and Middle Palaeolithic se­ with a cobble bearing red ochre traces. The few quence consists of Levels O (modern debris and fa unal remains include some of Sus and of Ca­ Bronze Age), Oa (stalagmitic layer), I (18 cm-thick preolus. Both Levels 27 and 28 are sedimentolo­ dark silt level-Azilian), la (sterile stalagmitic layer) gieally considered to be the products of co Ide r but and I� (50 cm-thick dark silt level-Terminal Mag­ still humid climatic conditions attributable to daleOlan). No harpoons were found in the 1953-57 Dryas III. Although the AP is still not very high, a excavation of Level I, however, the lithic industry broad di vers it y of thermophile taxa leads Arlette which includes 119 tools, seems to correspond t� Leroi-Gourhan (as usual) to assign these levels to the Azilian, with 25.2% endscrapers, 16.7% backed Allerød. Rowever, the bone collagen 14C date of bladelets, 5.0% Azilian points (despite curation 1O,635±120 B.P. (BM-1494) fo r the upper part of problems after the death of Santa-Olalla), and Level 27 corresponds to Dryas III. A second (bone) 18.5% burins (Gonzalez Echegaray, 1980). The old date from lower Level 27 of 12,270±400 B.P. p�llen nalysis by Arlette Leroi-Gourhan (pu­ (UCR-1275D) corresponds to Allerød even at -2 � bhshed In 1980) is, by her own admission rather s.d. (A third (charcoal) date of 14,760±400 B.P. ' unsatisfactory, as only a few samples were taken­ (GaK-6985) is fa r out of stratigraphic order and is discontinuously- throughout the long sequence. clearly far too old for a Terminal Wiirm deposit). Three samples from Level II are low in arboreal The contradictions among these 14C dates remain p llen, but include Pinus, Corylus and Juniperus, unresolved, although vertical movement of sample � wlth some Hippophae and abundant Ericaeeae items and interlaboratory errors are both possibie suggestive of the transition from Dryas II t� explanations. Level 28 is directly overlain by the Allerød according to Leroi-Gourhan. No sample Asturian conchero, the base of which dates (on was taken from Level I per se, but from the charcoal) to 8,650±300 B.P. and is assigned to the talagmitic layer overlying il. This sample, very rich late Preboreal or early Boreal. It is problematical to � . In pollen, IS overwhelmingly dominated by arboreal differentiate the large lith ic assemblages of Levels taxa, including over 50% Corylus, plus Fagus, 24 and 27, but Levels 24 and 28, with their Quercus, Fraxinus, and Pinus, as well as Hedera, distinctive harpoons, are Upper Magdalenian and Viburnum and numerous ferns. Leroi-Gourhan Azilian respectively, by definition. Probable or assigns this travertine layer to the Boreal. The certain Azilian Levels 25-28 would seem to span the travertine layer under/ying the Azilian Level I latter part of Allerød and Dryas III. I would be assignable to Allerød. A sedimentological analysis was more recently conducted by Butzer The Azilian problem in the Franco-Cantabrian region 95

(1980). The stalagmitic layer underlying Level Iper Allerød; however, the sedimentological analysis of se is considered evidence of a possibie temperate, Butzer (1973) indicates evidence of considerable humid episode, perhaps attributable to Allerød. contemporaneous frost weathering. These cold Level I itself, with a certain amount of frost­ conditions probably date to Dryas III. The traver­ weathering debris, was forrned under cold, albeit tine directly overlying Level l dates to 9,000±150 periodically humid conditions corresponding to B.P. (1-5 150) and would correspond therefore to the Dryas III and the beginning of the Holocene. The Preboreal. The small faunal assemblage from Level Level I fauna, heavily dominated by Cervus, in­ l, overwhelmingly dominated by Cervus, includes cludes several remains of Capreolus and a few of Sus one Capreolus bone and two of Sus. The Cueva (Fuentes, 1980). Naturally, given the huge size of Morin Azilian problem is like that of La Riera the cave, it is impossible to date precisely the Levels 27-28, where there is sedimentological evi­ Azilian harpoons found by Carballo in other parts dence of co Id climatic conditions attributed to the ofthe entrance by the results obtained in the Santa­ brief and relatively attenuated Dryas III episode, Olalla excavations, which uncovered the earlier but with a fa irly high AP and a few remains of discussed probable Azilian industry lacking har­ species of mammals which prefer woodland habi­ poons and dating, apparently, to Dryas III-Pre­ tats. boreal. 2.7. Rascafio(Mirones, Santander) 2.6. Cueva Morin (Villanueva de Villaescusa, San- Rascafio is a ca ve located in a cliffside dominating tander) the canyon of the Rio Miera in the fo othills of the This well-known Middle and Upper Palaeolithic Cantabrian Cordillera in Santander. This specia­ cave site, located near the edge of the coastal plain lized ibex-hun ting station was successively exca­ surrounding the present bay of Santander at the vated in the first two decades of this century by J. foot of a coastal mountain range, was excavated Carballo, L. Sierra and H. Obermaier, but was various times, notably by J. Carballo, the Conde de never extensively published. These prehistorians all la Vega del Sella, and, most recently, J. Gonzalez mention the discovery of an Azilian level or at least Echegaray and L. Freeman (1971). A thin Azilian 'typical' Azilian tools: microliths, small endscra­ deposit, localized near the cave mouth, was found pers (see summary in Straus, 1981). However, the in all these excavations sandwiched between a su­ only one to claim the presence of Azilian harpoons perficial travertine crust (in places containing mol­ is Carballo, who does so in his doetoral thesis luse shelIs) and an Upper Magdalenian level. Ty­ (1922) and in a general text (1924), but giving no pical flat-section Azilian harpoons were found by illustration thereof. No Azilian harpoons are pre­ both Carballo and Vega del Sella. Both early sent in the small remnant collections from these excavators speak of abundant small endscrapers early excavations. and backed bladelets, plus some geometric micro­ Rascafio was rediscovered by L.G. Straus and liths (see summaries in Femandez-Tresguerres, once again excavated in 1974 by J. Gonzalez 1980: pp. 96-99). The remnants of this ephemeral Echegaray, I. Barandiaran and him. Little re­ Azilian level (1) excavated in 1968-69 yielded 359 mained, particularly of the uppermost level (l) retouched stone tools, of which l3.6% are end­ which probably corresponds to the supposed Azi­ scrapers, 23.0% backed bladelets, and only 7.2% lian stratum of the earlier excavators. Level l burins. Although there are no Azilian points sensu contained only 13 stone tools and 4 bone artifacts stricto, microgravette points make up 5.0% of the (but no harpoons). Sedimentologically this level is assemblage, and there are indeed six geometric considered to be co Id and dry due to the presence of microliths (a circle segment and 5 triangles). No large quantities of frost-weathering produets; it is Azilian harpoons were found in the very limited assigned to Dryas III and is capped by a travertine recent excavation of this level. layer assigned to the early Holocene (Laville & There is a geological (and cultural) hiatus be­ Hoyos, 1981). In contrast, with an AP reaching tween Levels 2 (Upper Magdalenian) and l (Azi­ 18% (only) and including relatively many hazel lian). Level l contains abundant pollen of hazel and pollen but few of pine, alder or oak, together with pine, with small quantities of birch, alder and oak; many fem spores, Level l is assigned to Allerød by the Cichorieae are more numerous than the Gra­ Boyer-Klein (1981), explicitly folIowing the lead of mineae. Leroi-Gourhan (1971a) assigns this level to her mentor, Leroi-Gourhan. It should be noted that 96 L.G.STRAUS her two samples-both from the lower part of the tools) Level 2b collection, there is in reality no stratum (1.3) which is a loose, open rubble-are very strong reason for assigning it to the Terminal poor in pollen. Level 1.3 dates to 1O,485±90 B.P. Magdalenian (as do the excavators), rather than to (BM-1449) and 1.2 to 10,560±245 B.P. (BM-1448). the Azilian. The excavators consider it to be of The dates overlap at l S.d. and correspond to a possibie Allerød age, although they present no Dryas III age. They are preceeded by dates of independent evidence in support of such an hypo­ 12,280±165 B.P. (BM-1450) and 12,895±135 B.P. thesis. (BM-1451) for Upper Magdalenian levels 2. 1 and According to Arlette Leroi-Gourhan (1966), two 2.3 respectively. All are bone collagen dates. Level l pollen samples were taken from "the Epipalaeo­ is the only Rascafio stratum to contain remains lithic layers (Magdalenian and Azilian?) near the (albeit few) of Sus scrofa (Altuna, 1981). present surface". However, in her diagram the two isolated uppermost samples are both attributed to 2.8. Pielago (Mirones, Santander) Level l,with none for either Levels 2 or 3. The Level l samples have an AP of 3-4%, with a few pollen of The cave of Pielago is situated a short distance Quercus, luglans, Corylus, Alnus, Tilia, Buxus, as downstream of Rascafio on the same right bank of well as those of juniper, pine and birch and many the Rio Miera gorge. Excavated by M.A. Garcia spores of ferns. Level l is nonetheless considered to Guinea, this site remains unpublished except for a have had climatic conditions somewhat more tem­ brief reference (Garcia Guinea, 1975). The site is perate and humid than Magdalenian Level 4 (for said to have three Azilian levels (with flat-section which there are samples that lack thermophile harpoons) and an underlying 'Proto-Azilian' layer arboreal taxa and are poor in fern spores). In a (which lacks harpoons). All four levels are said to subsequent article, Leroi-Gourhan ( 1971 b) assigns contain Azilian points, as well as triangles and the El Otero 'Azilian' level (1) to Allerød and the other geometric microliths, which are rare in Can­ Terminal Magdalenian 'with Azilian influences' tabrian Spain. (i. e. Level 2b) to Dryas II. There are no radio carbon Fernandez-Tresguerres (1980) is skeptical about dates or sedimentological analyses from this site these as yet unillustrated geometric elements, al­ which might have helped to clarify the age of Levels though the collections displayed in the Museo de l and 2. The fauna of Level l includes red deer, ibex, Prehistoria of Santander do include some geome­ chamois and boar, while Level 2 contains all these trics, along with Azilian harpoons and points. As species, plus roe deer, horse and bovines (Mada­ at Rascafio, the fauna is said to be dominated by riaga, 1966). ibex-not surprising due to the surrounding steep, Fernandez-Tresguerres (1980) suggests an Al­ rocky cliffs-with some chamois and red deer. A lerød age for the stalagmitic layer (Level 2a) sedimentological analysis by Butzer (1973) points separating the El Otero Levels l and 2b. He notes to the presence of numerous cryoclastic elements in the existence of a similar stalagmitic layer between the Pielago Azilian (level unspecified), suggesting a the final Magdalenian and Azilian (with typical Dryas III age. harpoons) at the famous, but undated site of El Castillo in central Santander; he hypothesizes­ 2.9. El Otero (Secadura, Voto, Santander) albeit without addition al supporting evidence-an Allerød age for this travertine. It should be recalled The topmost level of this cave site, located near the that an analogous situation exists in the strati­ estuary of the Rio Ason in eas tern Santander , is graphy of nearby El Pendo cave and that there may very tentatively assigned to the Azilian. However, have been erosion in this same period between the there are only 4 lithic tools, 2 antier 'awls' and no deposition of the Upper Magdalenian and Azilian harpoons in the Level l collection (Gonzalez Eche­ levels at the also neighboring site of Cueva Morin. garay et al., 1966). This stratum overlies a supposed Similar erosion, specifically at the end of Allerød, is 'Magdalenian VI' level (2b), from which it is observed between the Magdalenian VI and Azilian separated by a stalagmitic layer (Level 2a). A at Duruthy on the edge of the French Basque solitary 'quadrangular'-section biserial harpoon tip Couqtry.2 fragment was found in Level 2b, while Level 3 contains several classic cylindrical-section uniserial 2.10. El Castillo (Puente Viesgo, Santander) harpoons, and is clearly Upper Magdalenian by definition. Although burins (20%) slightly out­ The vestibule of this vast cave was excavated in number endscrapers (16.5%) in the small (c. 60 191 1-14 under the direction of Hugo Obermaier, The Azilian problem in the Franeo-Cantabrian region 97 who uncovered a stratigraphy 16-18 m deep but (Apellaniz & Altuna, 1975: p. 153). The authors do never published a complete report of his work. not consider it to be Tardenoisian, it lacks geo­ According to Fernandez-Tresguerres (1980: p. 53) metric microliths. Azilian (and Magdalenian) har­ and Cabrera (1984: pp. 377-38 1) the extant artifact poons are thus far missing. The scanty fauna from assemblage from the thin but laterally extensive Bed III is heavily dominated by Cervus elaphus; Azilian deposit is very small (12 tools), but do es fo llowed by Capra pyrenaiea, Sus scrofa and Ca­ include three typical flat section harpoons, one of preolus. Interestingly, there is one remain of Canis which is bilaterally barbed (Breuil & Obermaier, familiaris, which, if in situ, would be one of the 1912: p. 10). Butzer (1981: pp. 164, 168, 174, 176) oldest known evidences of dog. Lacking further equates his level 18 with Obermaier's Azilian information, it is risky to affirm an Azilian at­ deposit and, because of an angular scree rubble tribution for these small Aranaza artifact assem­ content, concludes that it was forrnedunder con­ blages. There are as yet no paleoclimatic analyses ditions of co Id with slope instability which he would available for corroboration of a tentative Dryas III attribute to Dryas III. It should be noted that the and Preboreal dating of Beds III and II, respecti­ stalagmitic layer separating the Azilian and Upper vely. Magdalenian deposits may not have been ubiqui­ tous in the Castillo site (ef Breuil & Obermaier, 3.2. Santimamifie (Cortezubi, Vizcaya) 1912: p. 10 contra Obermaier, 1925: p. 176). This large cave, with well-known Upper Palaeo­ lithic paintings in interior galleries, lies in a moun­ tainside above Guernica and was excavated during 3. THE SITES OF V ASCONGADAS the first two decades of this century by T. Aranzadi (EUZKADI & NAV ARRA) and J.M. de Barandiaran (1935) and more recently re-excavated by Barandiaran (1962). The vestibule 3.1. Arenaza I (San Pedro de Galdames, Vizcaya) contained a long Upper Palaeolithic, Mesolithic There are a number of Azilian sites in the two and Neolithic/Eneolithic stratigraphic sequence, coastal Spanish Basque provinces (plus Navarra). which was unfortunately dug 'too early' and pu­ Unfortunately none of the recent excavations has blished in a rather confusing fashion. The totality of yet been adequately published, so the data are even its collections has never been studied systema­ scantier than those ofthe Azilian sites in Santander tically, although A. Cava (1975) has analysed the and Asturias to the west or the French Basque post-Azilian materiais. According to the excavators country and Pyrenees to the east. Arenaza I, a cave (Aranzadi & Barandiaran, 1935 and an unpu­ with Upper Palaeolithic paintings ofhinds done in a blished stratigraphic section on file in the Museo style very similar to that of paintings in caves in Historico de Vizcaya in Bilbao), it is Level 5, eastern and central Santander , is located in hill underlying a so-called Asturian shell midden and country not far inland of the city of Bilbao. It is overlying a Terminal Magdalenian level in the cave currently being excavated, and only summary re­ mouth, which is Azilian by definition, since it has a ports have been published to date (Apellaniz & flat-section harpoon. A radiocarbon date of Altuna, 1975). The stratigraphy in the vestibule is 9,470±400 B.P. (Gif-130), done on bones collected capped by Roman, Bronze Age and Neolithic in 1960 and attributed to 'Level 7', has been deposits, below which is a complex bed (II) com­ published by Delibrias et al. (1966: p. 88) who say posed of several levels with a scanty lithic industry the date corresponds better to Level 6. How the described as an 'atypical Tardenoisian'. Level D of level attribution of the sample was determined is this bed dates to 9,600±180 B.P. (CSIC-173), a quite unclear, as there was no Level 7 in the 1960 charcoal date which would make this stratum excavation, and the date is rejected by Barandia­ contemporaneous with Level 3-upper at Los Azu­ ran himself (Mariezkurrena, 1979: p. 242). 'Le­ les, although no Azilian harpoons have been found vel 7' in the original scheme of Aranzadi and at Arenaza. Barandiaran (1935) corresponds to the Solutrean. 1 Underlying Bed III is dated (on bone) to This 4C date has indeed sometimes been cited as 10,300±180 B.P. (CSIC-174). Its lithic industry pertaining to Level 6 (Final Magdalenian) (Moure, which notably includes back ed bladelets, small 1975: p. 30; Almagro Gorbea, 1970: p. l3). Gon­ endscrapers on flakes and nucleiform endscrapers, zalez Morales (1982: p. 183) has inexplicably sug­ is inexplicably (and incorrectly) labelled 'Asturian' gested that this date might in reality pertain to Level 98 L.G.STRAUS

5, the Azilian. Of the Level 5 fauna (Altuna, 1972), in his recent monograph on animal domestication we only know that it contained red deer, bovines, in the Basque country, convincingly argues that horse, ibex, roe deer, boar, chamois and several these remains are either intrusive (there is some carnivores, plus many birds and molluscs. Without evidence of badger disturbance, albeit not in the further studies of the industry and fauna, and sector where these bones were found) or mislabel­ lacking geological and palynological analyses, as led. Both the dog and goat bones are whole and well as reliable radio carbon dates, it is impossible to much better preserved than the rest of the bones in give a creditable estimate of the age of this deposit, Level C, which is repea ted ly assigned to the Azilian except to suggest that the Gif date might be based by Altuna (e. g. 1972; 1980; Altuna et al., 1982; see on a mislabelled or misinterpreted sample which in also Mariezkurrena, 1979). There is no reason to realit y came from Level 5. believe that the 14C determination dates intrusive shells, although this possibility cannot be totally excluded. If aquatic shells were those dated, the 3.3. Urtiaga (Itziar de Deba, Guipuzcoa) . 'real' date could in faet be even younger. The claim This important site, located in the rugged coastal made by Newell et al. (1979, p. 192) that Urtiaga hills of Guipuzcoa, was dug by J.M. de Baran­ Level C contained 'Asturian' artifacts is unfoun­ diaran in the late 1920's and 1930's but has never ded. I could find no reference to such pieces at been comprehensively published. The Azilian and Urtiaga, although they have been made for San ti­ Magdalenian lith ic assemblages have been studied mamifie, Lumentaxa and Mouligna (Barandiaran, at different times by J.M. de Barandiaran and D. de 1953). Sonneville-Bordes (1964), Laplace and Merino (1979) and Marsan (1979), and the bone industries 3.4. Ekain (Deba, Guipuzcoa) by Barandiaran (1967). Level C, which surmounts a long sequence of Middle and Upper Magdalenian Ekain is a major Upper Palaeolithic rupestral art deposits, contained a flat-section harpoon and a sanctuary with a small site in the cave entrance lithic industry heavily dominated by backed bla­ which includes Magdalenian and Azilian leveis. It is delets (26.2%) and Azilian points (c. 50%). Burins located in hill country not far from Urtiaga and and endscrapers-both rare-are in rough equili­ several other sites. A complete report on the recent brium (8.45% and 7.27% respectively), according to excavations by Altuna and Merino (1984) has just Barandiaran and De Sonneville-Bordes. been published. A preliminary report (Barandiaran Several years ago shells from Level C were & Altuna, 1977) c1early shows that at least levels radiocarbon dated to 8,700±170 B.P. (CSIC-63)­ III-Vare Azilian with flat-section harpoon frag­ the most recent date from a Spanish site definitely ments. Backed bladelets and backed ('Azilian'?) attributed to the Azilian by the presence of a classic points are very numerous. The lithic industry of harpoon. This date might be somewhat young Level n is said to have some Azilian characteristics, (although at +2 s.d. it exceeds 9,000 B.P.), but there but with elements (e. g. microburins) suggestive of is no reason for rejecting it a priori. (Likewise, shells the Sauveterrian; it lacks harpoons (Altuna et al., from underlying Terminal Magdalenian Level D 1982: p. 12). The end of the Magdalenian sequence have been dated to 1O,280±190 B.P. (CSIC-64). (base of Level VI/top of VII) is radiocarbon dated There are no sedimentological or palynological (on bone) to 12,050±190 B.P. (1-920). Overlying analyses for Urtiaga which could help date Level C Level V has been (charcoal) dated, however, to and either confirm or reject its apparent Preboreal 13,350±250 B.P. (CSIC-I 72) and Level III to age. The mammalian fauna includes Capreolus and 12,750±250 B.P. (CSIC-171). If correct, these dates Sus, along with 'alpine' caprids and abundant red would make the Azilian levels considerably older deer. Four dog remains (from 1 individual) found in than the underlying Upper Magdalenian. But there Level C may be intrusive (Altuna, 1980: p. 16). It are also bone dates for Levels IV & n of9,460± 185 should be noted that reindeer remains, which are B.P. (1-9239) and 9,540±21O (1-11666) which are unusually abundant for a Spanish site in the accepted by the excavators, who reject the CSIC Urtiaga Magdalenian leveis, are absent from Level dates for Levels nI and V (Barandiaran & Altuna, C (Altuna, 1972). Newell et al. (1979: p. 192) point 1977:'pp. 52-53; Mariezkurrena, 1979: pp. 241-242; out that sheep/goat remains were also found in Altuna, 1984: p. 43). These two dates for Levels IV Level C. In faet there are two domesticated goat and n overlap at l S.d. They would make the Ekain bones with Level C labels. Altuna (1980: pp. 16-18), Azilian contemporaneous with Los Azules Level The Azilian problem in the Franco-Cantabrian region 99

3-upper and Cueva Oscura de Ania iftheir dates are (Ly-l398). No Azilian harpoons have been fo und correct. Recent pollen results (Dupre, 1984) place and bone tools are virtuaIly absent, but the Level II the end of the Magdalenian sequence (Level VI) in Epipalaeolithic assemblage, unfortunately publi­ Allerød, followed by a brief cooling and then by the shed globally as a unit, contains many Azilian definitive return of warmer, more humid condi­ points, backed bladelets and endscrapers, but re­ tions, with an increase in (especiaIly thermophile) latively few burins. Level BIll from a pit dug in the trees (Preboreal). No chronological concJusions are cave interiOI', where ve ry few artifacts were found, drawn in the Ekain sediment analysis (Areso, 1984), has been dated (on bone) to 11,840±240 B.P. (Ly- although Levels IV-II are generally tempera te with 1400) and � 10,940 B.P. (Ly-1458), but the rela­ a brief cold snap in the course of Level III. The tionship between it and Level II in the cave entrance fauna from the Azilian levels includes dominant is unknown. It too is called 'Azilian' by the Cervus with 'alpine' caprids and Capreolus. Sus is excavator, Barandiaran (in Evin et al., 1979: p. present only in Levels III and II (Altuna et al. , 1982: 442). The dates for Zatoya Level BIll and II-Iower p. 61). would be the oldest ones for the Azilian from the Vasco-Cantabrian region (with the exception ofthe two Ekain'dates which are far out of stratigraphic 3.5. Erralla (Cestona, Guipuzcoa) order and therefore rejected by the excavators). The This small cave, located near Ekain, has recently date for Level II-uppet is, on the contrary, quite been dug by Altuna and his team from the Sociedad young. There is clearly an hiatus in sedimentation de Ciencias Aranzadi. Archaeologically sterile Le­ between the top of Level II and the lower part. The vel IV has been radio carbon dated (on bone) to Level II fauna incJudes red deer, ibex, chamois and 1O,580±270 B.P. (1-10803); it overlay a Middie boar; the former two species are represented by Magdalenian level. An early publication (Mariez­ especiaIly many young individuals (Barandiaran, kurrena 1979: 240) attributed the overlying deposit 1977), as is the case in the Azilian of La Riera. It to the Azilian, but the definitive site report (Altuna would seem likely that all or part of Level II is et al., 1985) attributes it to the Terminal Mag­ Azilian, but there are no other published data dalenian, based in part on a new date (on bone) of which might help resolve the problem of its age, 12,31O±190 B.P. (1- 13439) for Level III. There are which by its dates would span the period from no harpoons and the small lithic assemblage is Allerød to Boreal. banal, belying attempts to use it to date Level III to the Bølling or Allerød by the geologists and palyno­ 3.7. Abauntz (Arrais, Navarra) logist in question. Recently a singie 14C date (on bone) of 9,530±300 B.P. (Ly-1964) has been published for Level 'd' of 3.6. Zatoya (Abaurrea Alta, Navarra) Abauntz Cave, located at c. 700 m above sea level in This site, located at c. 900 m above sea levei on the northem Navarra near the Cantabro-Pyrenean edge of the Pyrenees near the French border, was watershed-like Zatoya in the Ebro basin (Evin et recently excavated by Barandiaran (1977; 1979) and al. , 1983: p. 113; Utrilla, 1982). The associated dated by Evin (1979; Evin et al. , 1979). The industry is described as 'Azilian (Epipalaeolithic) uppermost level (I) is Neolithic, bone dated to without geometries'. 6,320±280 B.P. (Ly-l 397), with ceramics but no Level 'd' is 10-25 cm thick. It contains Helix shells domesticated animais. It overlies a level (Ib) with a and 76 stone tools 14.4% endscrapers, 6.5% burins, pre-Neolithic stone industry charcoal dated to 2.6% perforators, 2.6% Azilian points, and 34.2% 8,260±550 B.P. (Ly-1457) and containing many backed bladelets. There are no harpoons. The geometric microliths attributed to the Sauveter­ fauna includes 2 MNI each of Cervus and Capra and rian(?). Level II in the cave entrance has been one each of Bovini and Rupicapra, plus a fish bone variously labelled as 'Azilian', 'epi-Azilian' or (Altuna & Mariezkurrena, 1982). The arboreal 'Azilian-like' (Barandiaran, 1977; 1979; Evin, 1979; pollen percentage ranges from c. 6O-c. 98%, with Gonzalez Morales, 1982). The lower part of this smali percentages of hazel, beech, alder and oak, level, which is c. 60 cm thick, has been radiocarbon together with dominant pine. High humidity is dated (on bone) to Il,480±270 B.P. (Ly-l 399), and indicated for this early Holocene level by a massive 11,620±360 B.P. (Ly-1 599), whereas the top of increase in fem spores vis il vis the underlying Level II has been charcoal dated to 8,150± 170 B.P. Magdalenian (Lopez, 1982).3 100 L.G.STRAUS

4. THE SITES OF GASCONY (py RENEES­ layer and atop the great Upper Magdalenian cobble A TLANTIQUES & LES LANDES) pavernentsin the rock shelter ofDuruthy. This vast site (like Dufaure, Grand Pastou and Petit Pastou) is located at the foot of the south-southwest facing 4. 1. Isturitz (Isturitz and Saint-Martin, Pyrenees- Pastou Cliff, overlooking the Gave d'Oloron near Atlantiques) its confluence with the Gave de Pau. The Basque Isturitz is a huge cave system forrned by the Pyrenees lie within view to the south. The early Arberoue river which has successively cut three excavators, who found a reindeer mandibIe in this tunnels through a ridge near the boundary of the deposit (their Level F'), noted the existence of a French Basque provinces of Labourd and Basse­ I layer of Helix shelIs between the Upper Magda­ Navarre. The uppermost tunnel, forrned by the lenian pavement and Level F' (Arambourou, 1978). Salles d'Isturitz (Grande Salle) and Saint-Martin, No mention is made of any flat-section harpoons. contains a long Middle and Upper Palaeolithic Beginning in 1957, R. Arambourou has been stratigraphic sequence excavated first by Passe­ excavating Duruthy. Only a small part of Level F' rnard and then by the Comte and Comtesse de (Arambourou's Couche 2) remained at the very Saint-perier. The Azilian is represented by two flat­ base ofthe cliff; it does not extend out into the slope section harpoons found by R. de Saint-Phier (1936: and lower terrace, as does the Magdalenian VI pp. 37-38) in the top of Level la in the Salle stratum. Arambourou was able to distinguish two d'Isturitz. Two cylindrical-section Upper Magda­ sublevels within Couche 2. F. Delpech (1983: pp. lenian harpoons were found in the lower part of this 112, 381) claims that the relative proportion of same pocket of jet black sediments, which meas u­ Rangifer decreased between the two sublevels, since red 5 m in horizontal extension and 20 cm in the lower contained 5 reindeer remains (out of a thickness and which locally overlay Level I. A third total of 15 ungulate remains), whereas the upper Azilian harpoon was found in a silt lens similarly contained 2 (out of a total of 14-including one each sandwiched between the flowstone and Level I in of boa r and roe deer). StatisticaI analysis does not the Grande Salle. confirm the significance ofthis observation. Aram­ Two pollen samples were taken from Level I: the bourou, like his predecessors, has found no flat­ uppermost one from right below and in contact section harpoons. The Stratum 2 lithic assemblage with the capping stalagmitic crust and the lower one includes 106 tools: 40.6% endscrapers, 18.9% bu­ from a few centimeters below (Leroi-Gourhan, rins, 8.5% backed bladelets and 5.6% Azilian 1959). Both samples are assigned to the 'Upper points. The sedimentological analysis of C. Thi­ Magdalenian' , but the stratigraphic relationship bault (1978) also divides Stratum 2 into two sub­ between them and the possibIe Azilian lens (which units: the older forrned under cool, humid con­ is not mentioned by Leroi-Gourhan) is not known. ditions corresponding to Dryas III and the more LogicaIly the uppermost pollen sample could cor­ recent under warmer, humid conditions correspon­ respond to the Azilian lens right below the flow­ ding to the beginning of the Post-Glacial. There is a stone, at the top of Level I, but this is not known. sedimentary hiatus between Strata 3 (Magdalenian Both samples have an AP of c. 1 % (with just a littIe VI) and 2; this was probably brief, since Stratum 3, pine) and high percentages of Cichorieae. Leroi­ with evidence of temperate, very humid conditions Gourhan stresses the great cold (albeit with hu­ (high AP, with thermophile and hygrophile taxa of midity) indicated by these pollen spectra, although both trees and herbs) is assigned to Allerød. The she does not suggest a chronological assignment for three phases of Allerød (the most recent of which­ Level I at Isturitz. Clearly either Dryas II or III are the thermal maximum-is truncated by erosion) are the possibilities, and the capping-stalagmitic crust represented by the pollen ofStratum 3, whereas the (post-Azilian) would consequently be either Al­ lower part of Stratum 2 shows a decline in the AP lerød or Preboreal. There are no radiocarbon dates (particularly in oak, hazel and alder), with a relative from Isturitz. increase in pine and birch. Hygrophiles are still present despite the colder climate attributed by Paquereau (1978) to Dryas III. Reforestation was 4.2. Abri Duruthy (Sorde-I'Abbaye, Landes) rapid 'in the upper part of Stratum 2, with a wide In 1874, before the discovery of the Azilian by variety of thermophile taxa. This is interpreted as Piette at Mas d' Azil, Lartet and Chaplain-Duparc the beginning of the Post-Glacial. These interpre­ uncovered a deposit lying below a Chalcolithic tations are supported by the geological evidence The Azilian problem in the Franco-Cantabrian region 101

(Thibault, 1978). It should be added that while itself (Ly-3 181) and 12,030±280 B.P. (Ly-3245), and the undated, Duruthy Stratum 2 overlies Stratum 3 the base dates to 12,260 ± 400 B.P. (Ly- 3182). Level 3 is top of which is radiocarbon dated (on bone) to dated to 9,600 ± 290 B.P. (Ly-4224) and 10,310 ± 270 1l,150±220 B.P. (Ly-919) and the base of which is B.P. (Ly-4223). Stratum 4 has yielded an abundant dated by thermoluminescence to 11,300 B.P. (c. lithicassemblage in which there are 12.8% burins. 10%) (BOR-6). These dates correspond in faet to 9.2% endscrapers and 47.8% backed bladelets the accepted range for Allerød. Stratum 2 is thus (including c. 1% Azilian points). There are a few considerably younger, especially given the strati­ geometric microliths and a Hamburgian point. graphic hiatus between the two levels noted above. There are als o over two dozen antler sagaie frag­ Several lines of evidence also point to the existence ments, a whole cylindrical section unilaterally bar­ of a Dryas III-Preboreal Azilian at the nearby site bed harpoon, two harpoon base fragments and a of Abri Dufaure. possible tip fragment. The fa una includes reindeer, horse, bovines and some red deer. Palynological analysis of ·Stratum 4 by M.M. Paquereau (n.d.) 4.3. Abri Dufaure (Sorde-l'Abbaye, Landes) reveals evidence of the three known phases of This site, located 230 m to the east ofDuruthy at the Allerød, the last ofwhich, with an AP ofup to 53%, base of the Pastou Cliff, consists of a small rock­ constituted a temperate, humid climatic maximum. shelter, a narrow ter race in front thereof, and a The lower part of Stratum 3, however, registers a steep talus slope which descends to the early Wiirm notable decline in the AP (28%), with pines and fluvial terrace of the Gave d'Oloron. In reality birches displacing the thermophile taxa present in Duruthy, Dufaure and the two sites which lie Stratum 4. The thermophiles increase once again between them, le Grand and le Petit Pastou, ean be in the upper part of Stratum 3. The colder, drier considered to be individual loci of one large Tar­ episode which caused the Stratum 3 rockfall pro­ diglacial station. The Dufaure rockshelter was bably corresponds to Dryas III, like Couche 2 at entirely excavated in one week in 1900 by H. Breuil Duruthy. Stratum 3 may contain at least two lens es and P. Dubalen. Despite some Post-Glacial disturb­ of residues from apparently ephemeral human anee of the surfaee deposits, they found materials occupations, with a possibie cobblestone pavement which would today be labelled as 'Azilian' in situ remnant in the squares closest to the abri. These amidst blocs and yellowish 'clays' (fo yer superieur). lenses seem to be separated by a rockfall episode, Among their finds was a flat-section harpoon while other such episodes within Stratum 3 overlie fragment and a couple each of possibly painted and the topmost Stratum 4 pavement and directly engraved cobbles. Small endscrapers were nume­ underlie the Stratum 2 silt. Nearly 9,000 items of rous, but burins fe w. They also found many backed knapping debris have been found in all of Stratum bladelets despite the early date and obvious haste of 3, together with only 272 retouched stone tools, the excavation. The jo yer superieur overlay a cob­ giving a debris/tools ratio of 32: l versus 19: 1 for

bl es tone pavement which capped the so-calledjo yer Stratum 4. Virtually no bone tools have been found, injerieur, with six cylindrical-section harpoon frag­ and although faunal remains are preserved, they are ments. This deposit also contained many more highly fragmented. Definite identifications by J. endscrapers than burins, as well as many backed Altuna and A. Spiess confirm the presence of rein­ bladelets, according to Breuil and Dubalen (1901). deer here as in Duruthy Couche 2. Small amounts As a consequence of the results obtained in test of Rangifer remains are distributed throughout trenches dug on the terrace in 1980 by L.G. Straus, all of Stratum 3. Endscrapers (14.7%) outnumber large-scale excavations were begun at Dufaure in burins (11.4%), while over halfthe tools (51 .5%) are 1981. Beneath a superficial layer of humus and old backed bladelets. There are 5.5% Azilian points, backdirt (Stratum l), an eboulis cone (Stratum 3) two worn ochre pebbles, and an engraved peb ble was uncovered on the terrace adjacent to the edge of like one found in Duruthy Couche 2. the Breuil/Dubalen excavation. This rockfall de­ A trench dug in 1981-84 connects the block posit, probably equivalent to the blocs and jo yer excavation on the terrace with the area dug in 1900 superieur ofBreuil and Dubaien, overlies a series of in the abri. A physical correlation between thejoyer up to 9 cobblestone pavement layers (Stratum 4 = superieur and Stratum 3 and between the jo yer jo yer injerieur). The top of Stratum 4 has been injerieur and Stratum 4 is apparent. Logically, radiocarbon dated (on bone) to 1O,9 1 0±220 B.P. therefore, the Stratum 3 assemblage should be (Ly-2666). The middle dates to 11,750±300 B.P. assigned by definition to the Azilian (due to the 102 L.G.STRAUS

1900 find of a flat-section harpoon) and Stratum 4 basis of fundamentally extrapolative palynological to the Upper Magdalenian (due to the finds of interpretations exaggerates its age by at least one or cylindrical-section harpoons in both the old and two millennia. Most of the existing radiocarbon new excavations), despite some general resem­ dates and sedimentological analyses point to a blances between the lithic assemblages of the two Dryas III age, with a few levels (Los Azules 3-upper levels aiready noted in 1901 by Breuil and Dubaien. and 2, Cueva Oscura de Ania, Abauntz d, Ekain Stratum 3 would date to Dryas III (Straus, 1983). IV, Dufaure 3-upper and possibly Arenaza II) pertaining to the Preboreal, and only Poeymaii 'CPE' clearly dating to Allerød. This evidence ca lis 4.4. Poeymaii (Arudy, Pyrenees-Atlantiques) for a re-interpretation of the significance of the Another Gascon site with Azilian deposits which palynological results, since the entire Tardiglacial has been excavated fairly recently is Poeymaii in period was quite humid in Cantabrian Spain, Arudy, in the Bearn sector ofPyrenees-Atlantiques. and the Dryas III cooling may have had a much Unfortunately the excavations of G. Laplace have more attenuated impact on the vegetation in this to date only been published in very preliminary southerly, hilly, coastal region than further north in fashion (Laplace, 1953). However, according to the flatter interior of France, where there would brief summaries given by Bordes (1972), Evin have been fewer possibilities for thermophile and (1979) and J.P. Rigaud(1980), the main Azilian level hygrophile taxa to find appropriate refugia during ('CPE' or couche il petits elements), with a 14C date this brief period of climatic deterioration.4 While on bone of Il,540±220 B.P. (Ly-1385) and con­ the 'facts' of the pollen analyses for Cantabrian sisting of graveis, sands and clays, is assigned to Spain (i.e. the presence of specific taxa) are not Allerød. It overlies the Terminal Magdalenian level called in to question (although the sample sizes are (BI), which dates on bone to 12,000±250 B.P. (Ly- often very small and the possibilities for percolation 1384) and consists of a rockfall attributed to 'older from overlying levels often great), it is the chrono­ Dryas'. Azilian or 'Azilo-Sauveterrian' elements logical significance attributed to often slight fluc­ were also found amidst blocs of an overlying tuations in percentages or to 'indicator' taxa which stratum (BS) dated on bone to 1O,420±230 B.P. must be treated with caution. In summary , the bulk (Ly-1386) and assigned to Dryas III. This level is in of the data from Vasco-Cantabrian Spain and turn overlain by a stalagmitic layer attributed to Gascony suggest that classic Azilian industries Preboreal, above which were deposited a long series there date to the period of transition between of so-called 'Aziloid' and 'Sauveterroid' (or 'Sau­ Pleistocene and Holocene, as generally defined by veterro-Arudian') Helix shell middens with geo­ the Dryas III-Preboreal boundary. metric microliths of early Holocene age, the lowest of which (foyers injerieurs il Helix) have been 14C dated to 9.960±210 B.P. (Ly-1379, charcoal), 9,470±220 B.P. (Ly-1380, charcoal), 9,430±21O 5. THE SITES OF THE CENTRAL PYRENEES & B.P. (Ly-1387, bone) and possibly 9,400±420 B.P. (ARIEGE HAUTE GARONNE) (Ly-1388, bone). There are also three dates on land snails for stratum FIH: 1O,250±240 B.P. (Ly-1 790), The Azilian chronology of the Vasco-Cantabrian 1O,700±290 B.P. (Ly-I789) and 10,300±250 B.P. region can be compared to the evidence from the (Ly-1788). Evin et al. (1980) judge these dates to be Central Pyrenees, the region in which the Azilian about 1000 years toa old. was first defined and where the site of La Vache, key The nearby site of Espalungue in Arudy yielded pollen sequence for the current standard chrono­ perforated flat-sectionred deer antler harpoons to logical assignment ofthe Azilian, is located. Dating E. Piette in his Level F, which overlies a series of means in the Pyrenees (and Languedoc) can be four Magdalenian strata. Level F is very tentative ly classified thus: attributed to Allerød by G. Marsan (1979) on Geological and/ol' palynological dating of secure typological grounds only. A possibie Azilian at Tute Azilian: Mas d'Azil, Rhodes II, La Tourasse, de Carrelore in Lurbe is als o essentially undated. Adaouste. While not entirely satisfactory, the corpus of data Tenninus post quem radio carbon dating of secure from Vasco-Cantabrian Spain and Gascony clearly Azilian: La Vache, La Salpetriere. suggests that the tendency to globally assign the Direct radiocarbon dating of possible Azilian: regional Azilian to the Allerød oscillation on the Gazel, Chinchon. The Azilian problem in the Franco-Cantabrian region 103

5.1. La Vache (Alliat, Ariege) however a lens within It In 'la Salle Garrigou' contained six flat-section harpoons, along with The cave of La Vache is located at an elevation of Azilian points and thumbnail scrapers (Nougier & about 550 m in the Vicdessos valley of the Ariege Robert, 1977). This pocket, clearly above the Pyrenees. It consists of a large vestibule ('la Salle 'Magdalenian' horizon, lacks cylindrical section Garrigou') first excavated in the latter half of the harpoons and has been labelled 'Proto-Azilian', 19th century and then again in the mid-20th although it is apparently more recent than the Salle century, and an interior chamber ('la Salle Moni­ Monique horizon containing both Azilian and que'), discovered and excavated by R. Robert Magdalenian harpoons and although its own har­ during the middle years of this century. While the poons are fully 'Azilian' in morphology. On�y results of the recent excavations have never been Upper Magdalenian-type harpoons were found In comprehensively published, there have been nu­ the cultural layer underlying the flowstone in 'la merous articles on the bone and lithic industries, Salle Garrigou'. fauna and pollen of La Vache in Bulletin de la Archaeological Levels 4-2 in 'La Salle Monique', Societe Prehistorique de f'Ariege over the past three with ve ry few arboreal pollen (mostly of pine) are decades. In both Sal/es a thick, extremely rich assigned to ·Bølling by Leroi-Gourhan (1967), while cultural deposit assigned to the Upper/Final Mag­ archaeological Level l, with a similarly low AP (c. dalenian is overlain by a potent flowstone layer; this 10%) but showing a replacement of Graminae by basic stratigraphic fact underlies all interpretations Cichorieae indicating drier conditions, is assigned of the La Vache sequence. La Vache is one of the to Dryas II. Consequently the overlying stalagmit ic key sites for understanding and dating the rela­ layer is palynologically attributed to Allerød. It tionship between the 'Upper Magdalenian' and the contains high albeit variable (c. 20-50%) arboreal 'Azilian' and its sequence has been used to 'date' pollen percentages, composed of a variety of ge­ Azilian levels throughout the Vasco-Cantabrian nera, notably Corylus, Tilia, Quercus, Betula, Ulmus region in particular (despite the physical distance and Alnus, as well as the ubiquitous Pinus. While and topographical differences involved). there is clear evidence that the flowstone represents There are three charcoal radiocarbon dates from a ve ry temperate, humid climatic phase, its dating the archaeological horizon (artificially divided into re lies partly on acceptance of the two 'older' four levels all assigned to the Upper/Final Mag­ Groningen dates and the rejection of the 'younger' dalenian) in 'la Salle Monique': 12,850±60 B.P. Lamont date (of Allerød age) for Level 2 of the (GrN-2026) for basal Level 4, 12,540±105 B.P. underlying archaeological horizon in 'la Salle Mo­ (GrN-2025) and 11,650±200 B.P. (L-336c) for nique'. Thus an alternate interpretation of the 14C Level 2. The latter date is rejected by Arlette Leroi­ dates per se could lead to an initial Holocene Gourhan (1967) 'because it does not fit with the assignment of the stalagmitic layer, with the under­ results of her palynological analysis' (Schmider, lying cultural horizon pertaining to Dryas III 1978: p. 15, fn . 5). Associated with abundant, (Level l) and Allerød (Levels 2-4). diversified Magdalenian lithic assemblages and The coexistence of Azilian and Magdalenian­ cylindrical harpoons (only uniserially barbed in type harpoons throughout the Salle Monique ar­ uppermost Level l and biseriaIly barbed in the chaeological horizon would indicate either signi­ lower leveIs) and sealed below the flowstone are ficant mixture of once separate assemblages or the eleven flat-section Azilian-type harpoons (Nougier arbitrariness (or even illegitimacy) of the classi­ & Robert, 1977). The combined Salle Monique ficatory distinction between these two culture­ lithic assemblage contai ns many backed bladelets stratigraphic units, which is the viewpoint defended but only low percentages of Azilian points (0.80%), by Nougier and Robert (1977). Taken at face value, thumbnail scrapers (0.67%) and geometric micro­ however, the La Vache evidence would seem to liths (0.25%), although (as in several other Pyre­ indicate an early appearance of Azilian-type har­ nean 'Upper Magdalenian' sites) burins do not poons, at leas t at this particular locality. The two outnumber endscrapers (Schmider, 1978). The harpoon types also appear together at a much later same is true of the Salle Garrigou assemblage from date:however, at Le Morin in Gironde (see below). the cultural horizon underlying the flowstone; supposed Azilian lithic types are present but rela­ 5.2. Rhodes II (Arignac, Ariege) tively scarce, as is true of many Upper Magdalenian assemblages. No cultural materials have been found The Rhodes II rockshelter is located less than 5 km within the stalagmitic layer in 'la Salle Monique', downstream from La Vache ab ove the west bank of 104 L.G.STRAUS the Ariege River. Like La Vache, Rhodes II con­ analyses were however recently conducted on sam­ tains a stratigraphic sequence which documents the ples which Piette had collected at the 'Left Bank' Magdalenian-Azilian 'transition' in the Central locus nearly a century ago and which were pre­ Py renees (Clottes & Simmonet, 1979). 'Foyers' 1-4 served at the Musee des Antiquites Nationales in of 'Couche' 2 are considered to be typical Upper Saint-Germain-en-Laye (Girard, Moser & Orliac, Magdalenian in assemblage composition, but the 1979). These samples, which could be correlated lithic assemblages of 'Foyers' 5 & 6 show changes in with Piette's published stratigraphy (still visible the Azilian 'direction' both in terms of particular today), suggest that the Azilian sequence at Mas types and lithic raw materials present. There are, d'Azil began at the end of Dryas III. The corre­ however, no harpoons. 'Foyer' 7 has a typical sponding sample is highly enriched in coarse sands Azilian assemblage, with several flat-section har­ and contains few arboreal pollen, almost none of poons. According to the palynological analysis of which are ofthermophile taxa. An open, steppelike M. Girard, the Allerød oscillation began with vegetation seems to have characterized the begin­ Foyer 4, being preceded by Dryas II in Foyer 3 and ning of Piette's Azilian ('Couche F'), as well as the Bølling in Foyers 1 and 2. The temperate, humid Terminal Magdalenian which underlay it on the Allerød conditions of Foyers 4-6 ended in Foyer 7 Left Bank. (Azilian 'Couche F' yielded the last times with a colder episode attributed to Dryas III reindeer remains in the Mas d' Azil sequence (Bahn, (Simmonet, 1983; 1976). A faunal change began in 1984: p. 398)). These conditions ended abruptly and Foyer 5, with increasing quantities ofred deer, roe were followed by an explosive increase in arboreal deer and boar, which replaced such 'coid' forms as vegetation (AP: c. 50-80%), with a wide variety of reindeer, Lagopus sp. and the Nordic vole. thermophile taxa indicative of the Preboreal, Bo­ The base of Foyer 6 was radiocarbon dated (on real and early Atlantic phases of the Holocene. charcoal): 12, 100± 150 B.P. (MC-997) and Fo­ Naturally these results should be interpreted with yer 5: 12,160±160 B.P. (Gif-2258), 12,250±200 B.P. caution, but they would seem to point to a late age (MC-1366) and 12,300±150 B.P. (MC-996). These for the Azilian at the type site. Piette's overlying dates are older than the age generally accepted fo r 'Arisian' levels with flat-section harpoons (= Later mid to late Allerød, although at minus 2 s.d. they Azilian, 'Couche G') are clearly Postglacial in age could fit into the traditional scheme. In addition and environmental characteristics. they are a bit toa young to pertain to Bølling. Ifthe culturally 'transitional' Foyers 5-6 at Rhodes II are 5.4. La Tourasse (Saint-Martory, Haute-Garonne) accepted as being Allerød in age, it could be asked why the same could not be true of the culturally Like Mas d'Azi l, the other Azilian type site, La 'transitional' stratum underlying the flowstone at Tourasse, has recently been re-analyzed. It is a cave La Vache with its similar 14C dates. The comparison and rockshelter located near the confluence of the ofthe neighboring La Vache and Rhodes II chrono­ Garonne and Salat rivers, 35 km west-northwest of stratigraphies raises the question ofthe identity and Mas d'Azil. First dug in the 1890's, La Tourasse synchrony ofTardiglacial climatic oscillations such yielded industries (which included flat-section har­ as the Allerød in Northwest Europe, West-central poons), labelled 'Tourassian' at the time, which like France and the Pyrenean region. the 'Azilian'j'Arisian' were thought to fill the 'hiatus' between the Magdalenian and the Neo­ lithic. The term 'Tourassian' was later dropped in 5.3. Mas d'Azil (Mas d'Azil, Ariege) favor of the 'Azilian'. The sequence at La Tourasse Littie hard evidence exists for the dating of the includes Terminal Magdalenian, Azilian, Meso­ Azilian at the type site ofle Mas d' Azil, a huge cave lithic and ceramic-bearing leveis. Excavations were in the Pyrenean fo othills ofwestern Ariege. Azilian started again in the 1970's at this classic site by E. deposits were excavated by Piette, Pequart and and M. Orliac (1973). While 14C dates are not yet others on both banks of the Arize River's passage available, preliminary sedimentological analyses through the cave, yielding vast quantities of the permit some initial hypotheses on the age of classic Azilian flat section harpoons, painted co b­ Tardiglacial and initial Postglacial deposits (Gi­ bles and lithic artifacts. They found extremely rich rard, Moser & Orliac, 1979). The Terminal Magda­ Magdalenian deposits as well. More recent excava­ lenian (with reindeer) levels (E3-1) and the first tions in this exceptionally complex site are as yet Azilian levels (D 1 1-6) were deposited under weakly unpublished. Sedimentological and palynological oscillating climatic conditions tentatively attribu- The Azilian problem in the Franco-Cantabrian region IOS ted to Dryas II and Allerød, whereas the bulk ofthe section harpoon and a painted pebble in a level levels reported by Girard et al. to be Azilian were immediately overlying an Upper Magdalenian se­ forrned under decidedly cold conditions rather quence (Sacchi, 1976). We as yet lack chronological speculatively attributed to Dryas III (Levels DS-2) data which would perrnit placement of these and under dramatically warmer conditions as­ Azilian levels in the Terminal Pleistocene/Initial signed to Preboreal and early Boreal (C7-1). Gi­ Holocene chronostratigraphic frarnework. How­ rard, Moser an'd Orliac (1979) believe, therefore, ever, recent excavations by D. Sacchi (1976) at the that the Later Azilian ('Arisian') at Mas d'Azil and cave of Gazei on another Aude tributary north of La Tourasse were contemporary, Postglacial phe­ Carcassonne have produced a Terminal Magdale­ nomena. (Another parallel ean be found in the nian level (6) 14C dated (on charcoal) to 10,760±190 'Arudian' levels (foyers injerieurs ilHelix) at Poey­ B.P. (Gif-26S4), fo llowed by a series of 'Epimagda­ mali). The Azilian at the two classic type sites would lenia n' levels (without Azilian harpoons or deco­ thus have spanned the Pleistocene/Holocene rated pebbles), the lowest of which (S) is dated (on boundary. R. Simrnonet(19 76: p. 1414) concludes charcoal) to 1O,080±190 B.P. (Gif-26S3)-right at his overview of the Pyrenean Epipalaeolithic by the Dryas III1Preboreal boundary. These levels at suggesting that the Azilian of this region may have Gazel have'high percentages of Azilian points, with been younger than once believed, extending through many backed bladelets, endscrapers (including and even beyond the Preboreal, although having thumbnail types), but relatively many geometrics begun as early as the Allerød (at least at La Vache). only in the uppermost Level 3. Other sites in this This is the conclusion he reaches on the basis ofthe region for which Azilian materials have been analyses of the stratigraphic sequences at Rhodes claimed include Balma Abeurador and La Grotte II, Mas d'Azil and La Tourasse. du Salpetre de Pompignan, the latter ofwhich has a The Central French Pyrenees contain a number date of 9,900±200 B.P. (MC-224 1) (Geddes et al. , of other sites with Azilian materiais. They include 1986). Le Trou Violet near le Mas d'Azil, which, like it, contained some reindeer remains in an Early Azi­ lian level (D) (Bahn, 1984: p. 396) and a possibly 7. THE SITES OF EASTERN LANGUEDOC Azilian burial with associated red-painted pebbles AND PROVENCE (Sawtell & Treat, 1927), reminiscent of the burial found at Los Azules. The recent excavation of La Still further eas t and far from the Pyrenees, in the Balma Margineda, at nearly 1000m in Andorra, has Vaucluse, there is an iso lated Azilian level (A) at uncovered several Azilian strata (8-S). Painted and Chinchon. This level, which contai ned a typical engraved pebbles have been found in Level 6 and flat-section harpoon, overlay a series of cryoclastic basal Level S. The two strata are placed palyno­ Upper Magdalenian levels (Livache, 1976: p. 1160) logicaIly in the Preboreal and early Boreal; basal and has been 14C dated to 8,980±8S0 B.P. (Ly-S98), Level 6 dates to 1O,640±260 B.P. (Ly-2843) and its although the humus sample is des eribed as 'prob­ top dates to 9,2S0±160 B.P. (Ly-2842). Levels 7-8 ably polluted' (Schvoerer et al. , 1979: p. 34). These may date to Dryas III. This was a specialized same authors state that the level in question re­ summer ibex-hunting site (Guilaine et al., 1985). presents a 'cold climatic phase' (Dryas III?), but Escalon de Fonton et al. (1979: 27S) place it within a humid phase they assign to Allerød. However the 6. THE SITES OF WESTERN LANGUEDOC Chinchon Magdalenian VI has been dated (on bone) to 12,000±420 B.P. (Ly-597). The Azilian The easternmost site along the northern flank ofthe sensu stricto is otherwise lacking in the Vaucluse, Pyrenees which ean be assigned to the Azilian is La although there are other post-Magdalenian lith ic Crouzade, located in l'Aude to the southeast of industries similar to that of Chinchon Level A at Narbonne, very near the Mediterranean. Here, sites like Soubeyras (i. e., IG IB, presence ofbacked Level (4), which overlay an Upper Magdalenian points and backed bladelets) (Livache, 1976: pp. deposit, con tai ned at leas t a dozen painted pebbles 1160-'161 1; Escalon de Fonton et al., 1979). (although it has yielded no harpoons) (Sacchi, The only other known true Azilian deposits in 1976). La Petite Grotte de Bize, on a tributary ofthe eas tern Languedoc and Provence were the ones Aude River north-northwest ofNarbonne, has also uncovered in old excavations at Adaouste (Bouches­ produced an Azilian assemblage with a typical flat- du-Rhone), Feraud and la Salpetriere (Gard) for 106 L.G.STRAUS which there are no absolute dates (Escalon de Coulonges and most recently by J.M. Le Tensorer Fonton, 1976a; 1976b; Escalon de Fonton et al., (1976). Coulonges rejected the term 'Azilian' and 1979). The Azilian layer at Adaouste apparently used that of 'Laborian'. Le Tensorer (1979), while originally overlay a 'Magdalenian V' level (12) maintaining this term, has qualified the industries dated (on bone) to 12,280±190 B.P. (Ly-541) from of Levels 3 and 4 at La Borie del Rey as Azilian in which it was separated by at least two (undated) general characteristics; Azilian points (sensu lato) 'Magdalenian VI' levels (11 and 10). The Adaouste are very abundant, but burins-while few-out­ 'Magdalenian VI' is believed to have been deposited number endscrapers. Level 3 has been dated (pro b­ during the Dryas II and eroded during Allerød ably on bone collagen) to 10,400±230 B.P. (Ly- (Escalon de Fonton & Onoratini, 1976; Escalon de 1386) and 10,350±340 B.P. (Ly-140 1) and Level 4 Fonton et al. , 1979). The Salpetriere Azilian was to 9,870±320 B.P. (Ly-1402) (Evin, 1979: p. 12). located in deposits with evidence of high humidity Thus, like La Tete du Chien in relative ly closeby also assigned to Allerød. It follows a Terminal northem Tam, the Borie del Rey Azilian straddles Magdalenian in cryoclastic sediments attributed to the Pleistocene/Holocene boundary. Le Tensorer Dryas II (Es cal o n de Fonton & Bazile, 1976) (1979) assigns it to Dryas III specifically, on the (despite a 14C date of 10,680±300 B.P. (Ly-937), basis of sedimentological evidence. which would place it in Dryas III (Schvoerer et al., 1979: p. 32)). At other sites in eastem Languedoc 8.3. Pegourie (Caniac-du-Causse, Lot) and in Provence the classic Azilian is replaced by the Valorguian (= Provenc;:al Romanellian), a re­ This cave has been under excavation for over 15 gional facies lacking harpoons and Azilian points years byM.R. Seronie-Vivien (Seronie-Vivien & Le and dated to the Allerød (Escalon de Fonton, Tensorer, 1979; Seronie-Vivien et al., 1981). It 1976a; 1976b; Escalon de Fonton et al. , 1979). contains three chronologically isolated cultural The Azilian record from Vasco-Cantabrian Spain units separated by occupational hiati: Early/Mid and extreme southem France can now be compared Bronze Age, Azilian and Initial Magdalenian. The with the chronological evidence for the Azilian in Magdalenian dates to about 17,400 B.P. and has the Perigord, Quercy and other regions further to faunal assemblages dominated by bovines and the north and east in France. Due to the large reindeer. Levels 7-4 are classified as Azilian, and number of Azilian sites in these regions, only those .indeed Level 4 contained three flat section har­ with radiocarbon dating will be discussed and poons. (An engraved pebble was also found, but evaluated at length. out of stratigraphic context.) The assemblages of stone tools contain high percentages of short end­ scrapers, few burins and regular backed bladelets, 8. SITES OF LE QUERCY (TARN, LOT AND and many Azilian points of a variety of sub-types. LOT-ET-GARONNE) Sedimentological ana lys is shows the Azilian levels to have been deposited under cold/humid, war­ 8.1. Abri de la Tete du Chien (Penne, Tam) mer/drier, temperate/very humid and finally co 1- This rockshelter is adjacent to the cave of La der/drier conditions very tentatively attributed as Madeleine, a final Magdalenian with famous re­ an ensemble to the Allerød. The palynological clining female bas relief sculptures. Excavations in analysis is based on four isolated samples from the early 1950's yielded an Azilian industry from a separate columns, the last of which has only a very deposit which has recently been radio carbon dated small pollen sample. There is an apparently pro­ (on bone) to 10,1 1O±440 B.P. (Ly-1175) (Clottes, gressive increase in the AP, first with birches and 1979: p. 666). This date clearly places the Tete du then more thermophile species (plus fems), al­ Chien Azilian in Dryas III or at the Dryas III/Pre­ though Cichorieae continue to outnumber Gra­ boreal limit. minae throughout the Azilian leveis. Levels 6-4 are assigned by Arlette Leroi-Gourhan to Allerød on the basis of the supposedly characteristic floral 8.2. La Borie del Rey (Sauveterre-la-Lemance, associations said to be reflected by the pollen Lot-et-Garonne) spectra. However, given the discontinuous nature This site is one of a group of Terminal Pleistocene of the sampling and the very long hiatus between loci investigated beginning in the 1920's by L. levels 8a and 7, this attribution may be at the least The Azi/ian problem in the Franco-Cantabrian region 107

quite risky. There are, however, 10 radiocarbon could be identified. Fragments of charcoal pro­ dates for the Azilian of Pegourie: vided a radiocarbon date of 9,450±190 B.P. (Gif- 1807) for the very base of Sauveterrian Level D-3 Level 4: 1l,390±320 B.P. (Ly-1390, bone) and a second Sauveterrian date of 9,060±190 B.P. 8,31O±220 B.P. (Ly- 1939, shell) (Gif-428 1). Attribution of this assemblage is rather Level 5: tenuous. 8,450±250 B.P. (Gif-2568, shell) 8,450±31O B.P. (Ly- 1837, shell) 8.4. Abri de Graves (Leobard, Lot) 11,680±330 B.P. (Ly-1391, bone) 11,870±290 B.P. (Ly-1832, bone) This single-component Azilian site, located WNW 11,850±280 B.P. (Ly-1833, bone) of Pegourie near the border of Dordogne, has been Level 5 base only briefly reported by J. Clottes (1979: pp. 640- 12,690±530 B.P. (Ly-1392, bone) 64 1; 1983: p. 489). To date the abundant lith ic Level 6: industry is heavily dominated by Azilian points (of 8,730±890 B.P. (Ly-1393, bone) which there are 138!) and endscrapers (130). There Level 7: are only a few other tools (including almost no 12,250±350 B.P. (Gif-2822, charcoal) burins), but debitage items are numerous. So far, two decorated cobbles have been found: one en­ The series of 'old' dates would support assignment graved and the other painted with ochre spots. No of Levels 6-4 to Allerød (and Level 7 to Dryas II). harpoons or other bone implements have been The three shell dates are, however, on terrestrial found, but there is a triturated fauna dominated by gasteropods (Helix), which are said to be good rabbits and red deer. Spit b of the 70 cm thick material for 14C dating (unlike aquatic molluscs) Azilian layer has been radiocarbon dated (on bone) (Evin, 1977: p. 136), so they, like the Level 6 bone to 9,900±180 B.P. (Gif-5318) (Clottes, 1983: p. collagen date must be explained away by mixing of 489), which would place this occupation in the the sediments (Seronie-Vivien et al., 1981: p. 253). Preboreal. Similarly , the few reindeer bones (15) in Level 7 are One final site in Lot may shed some light on the said to be intrusive. The Azilian faunas are do­ age of the Azilian (and on the ambiguity of its minated by Cervus elaphus, Sus scroja and Orycto­ temporal and compositional differences from the lagus cuniculus (rabbit), whose remains (the source Terminal Magdalenian): Sainte-Eulalie in Espag­ of 14C date Ly-1833) number in the thousands and nac (Lorblanchet, 1976a: p. 1193; Schvoerer et al., were systematically broken, charred and bea r cut 1979: p. 31). Level I in this cave contains classic, marks-all evidence of non-intrusiveness and defi­ rich 'Magdalenian VI' artifact assemblages, in­ nite human exploitation for food and/o r skins. The cluding several cylindrical-section, biseriaIly bar­ Pegourie Azilian faunal assemblages als o include bed harpoons. It has two radiocarbon dates on smaller percentages of roe deer, horse and bovine bone: 10,830±200 B.P. (Gif-1697) and 10,400±300 remains, as well as fish. B.P. (Gif-2193), would place it in Dryas III and The nearby collapsed cave of Les Fieux (Miers, make the Magdalenian here more recent than the Lot) contains a Mousterian and Early Upper Pa­ Azilian at nearby pegourie. While less abundant laeolithic (Aurignacian and Gravettian) sequence than in the underlying Middle Magdalenian leveis, and a Holocene sequence with Sauveterrian and reindeer still makes up nearly 59% of the faunal Neolithic levels (Champagne & Jaubert, 1981). remains in Level I. As along the flanks of the Between the two sequences lies a deposit (Level E) Pyrenees (Dufaure and Duruthy), Rangife r seems with an unusual lithic artifact assemblage labelled to have survived into the Dryas III to be hunted 'Epipalaeolithic' and said by the au thors to re­ here by people with Magdalenian equipment in sem ble the Azilian assemblage from La Borie del northern Quercy on the flanks of the Massif Rey, despite a lack of thumbnail endscrapers and Central. Lorblachet (1976b: p. 1398) hypothesizes Azilian points at Les Fieux. But this is a rather contemporaneity between Late Magdalenian and specialized indl1stry, with nearly 56% backed bla­ Azilian 'groups' (or, should we say, 'technolo­ delets (plus 26.55% burins and -only 2.6% end­ gies'). At the rock-shelter Coronzac-vers-Lot, in scrapers) among the 467 tools recovered to date. fact, he has found an assemblage combining Mag­ There are no harpoons. Bone is badly preserved, dalenian and Azilian traits (the later including a but a few remains of horse, bovines and reindeer flat-section harpoon bearing Magdalenian-type inci- 108 L.G.STRAUS sions along the shaft). Among the various types of (though both were modern excavations). Both sites undated Azilian assemblages in Lot is one without have biseriaIly barbed cylindrical-section harpoons, harpoons but containing painted and engraved but Morin has four flat-section ones as well in its peb bles, many endscrapers and Azilian points, and uppermost levels (A III-I) (along with 31 of the few burins or backed bladelets, at Abri Pages in 'Magdalenian VI' traditional type and Magdale­ Rocamadour, ve ry near Pegourie (Lorblanchet, nian harpoons alone in underlying Levels BIl, BI & 1976b: p. 1398). AIV) (Bordes & De Sonneville-Bordes, 1979). The top of the Morin sequence (All-I) is marked by sedimentological evidence for a brief, moderately 9. SITES IN THE PERIGORD (DORDOGNE) cold period, with a slight increase in the percentage of reindeer remains, while Rangifer dominates the 9.1. La Faurelie II (Mauzens-Miremont, Dor- whole Couze sequence, despite generally tempe­ dogne) rate, humid conditions (Bordes and De Sonneville­ While one of the classic Azilian sites, Villepin (with Bordes, 1979: p. 450). The Gare de Couze Terminal deposits horizontally contiguous and temporally Magdalenian (Levels G-B) is placed within 'Wiirm subsequent to the type Magdalenian IV-VI se­ IV, phase 8' (= Allerød) by Laville et al. (1980: p. quence at La Madeleine), is located in the Perigord, 312). Conceivably, the top of the Morin sequence this important prehistoric culture area has only one lies in the Dryas III as hypothesized by Delpech Azilian site with radiocarbon dates, Le Pont d' Am­ (1983: p. 294), and could theoretically have in­ bon, dug recently by G. Celerier (1976). The site of cluded an Azilian deposit perhaps mixed with the La Faurelie, near Les Eyzies, however, has a Terminal Magdalenian. Both these sites suggest sequence of U pper Magdalenian levels surmounted that the Magdalenian/ Azilian 'transition' took by an undated Azilian level (2), from which they place in Dryas III and tend to corroborate the are, in some parts of the site, separated by a sterile Preboreal date of the Faurelie Azilian. layer. The 'Terminal Magdalenian' of Level 5 (well below the Azilian) is dated to 11,780±180 B.P. 9.2. Le Pont d'Ambon (Bourdeilles, Dordogne) (Gif-364-on bone) (Schvoerer et al., 1979: p. 30; Lavilie et al. , 1980: p. 326). Sedimentological This is a small rockshelter in the northern ('White') analyses show La Faur€:lieLevel 5 (+ lower Level4) perigord. It has a series of Late Magdalenian leveis, to have been la id down under mild/humid con­ the uppermost ofwhich (4 & 3b) date to 12,840±220 ditions of 'Wiirm IV, phase 8' (= Allerød), upper B.P. (Gif-3369) and 12,130±160 B.P. (Gif-3739). Level 4 under colder/drier conditions of 'phase 9' Level 3b has large numbers of 'Azilian elements',

(= Dryas III) and Level 2 under temperate con­ according to its excavator, C€:lerier(19 76: p. 1429). ditions of early post-Pleistocene (= Preboreal) The first level classified as 'Azilian' (meaning here times (Laville et al., 1980: pp. 312, 326-29). In­ that there are many morphologically variable Azi­ terestingly, like at Pegourie, Level 2 contains huge lian points, short endscrapers, few burins and few quantities of rabbit remains, plus a few of reindeer backed bladelets or geometrics) (Level 3a), dates (Laville et al., 1980; Cel€:rier, 1976). (on bone) to 9,830±130 B.P. (Gif-2570). Level 3 There are two other sites from Guyenne which (above 3a) dates (on bone) to 9,990±250 B.P. (Gif- add some indirect information on the age of the 3561) and, on a sample described by the authors as Azilian in this region: Le Morin shelter (Gensac, 'arganic matter' to 1O,350±190 B.P. (Gif-3368) Gironde), and Gare de Couze (Lalinde, Dordogne). (Schvoerer et al., 1979: p. 30). (Taking standard 'Magdalenian VI' deposits (exact level unknown) deviations into account, the Level 3 dates overlap were dated (on bone) at Morin to 10,480±200 B.P. and the Level 3a date could actually be older than (Gif-2105) and at Couze to 10,900±230 B.P. (Lv- them, despite the apparent chronological inversion 391) (Schvoerer et al., 1979: p. 29). (Level C at of the means.) Level 2, the only ane at Pant Couze 'Magdalenian IV' has specifically been dated d'Ambon to contain a flat-section harpoon, dates to 11,750±31O B.P. (Ly-976) and Level H 'Mag­ (on bone) to 9,640±120 B.P. (Gif-3740). dalenian V-VI'). Both sites' assemblages are char­ On the basis of both sedimentological and paly­ acterized as being very final Magdalenian, with nological analyses, upper Level 4 and Level 3b are many short endscrapers and Azilian points, al­ said to be relatively cold and dry, Level 3a tem­ though Couze has very high microlithic indices perate/humid, Level 3 colder/drier and Level 2 (backed bladelets) and Morin has much lower ones temperate/humid (Celerier, 1976: p. 1431; Lavilie The Azilian problem in the Franco-Cantabrian region 109 et al. , 1980: pp. 335-37). Despite the evidence of 513). An Azilian industry is als o present in another significant environmental change, the fa unal as­ site in the Brive basin, Chez Bonny, but it is semblages are essentiaIly the same and 'temperate' undated. in character throughout. There are no reindeer. Red dee r make up about 80% ofthe ungulate remains in Levels 3a and 3, but only 37% in Level 2 (where 10.2. La Grotte du Bois Ragot (Gouex, Vienne) there are more 'borses 'and bovines). Roe deer and Further north, this site is located near the famous boar are also present, while there are large quan­ Magdalenian mobile art site of La Marche. The tities of rabbit and fish bones as well. The faunal Azilian sequence is underlain by a pair of final ana lys is (by F. Delpech) would place the whole Magdalenian levels (6 & 5b), from which it is sequence within the early Holocene (plus Dryas separated by sterile Levels 5a and 4c. These Mag­ III), which jibes with the radiocarbon dates. How­ dalenian levels ha ve cylindrical-section biseriaIly ever, the sedimentological and palynological ana­ barbed harpoons. Level 6 has been date d (on lyses, rejecting the dates from Levels 3 and 3a, charcoal) to 1O,180±160 B.P. (Gif-3579). This date attribute the latter level to 'Wiirm IV, phase 8' (= . is said to be contaminated by rootlets by Chollet et Allerød) and the former to 'phase 9' (= Dryas III). al. (1979: p. 369, note 4), although such contami­ In this scheme, at least Level 2 would be post­ nation is not noted by the radiocarbon analysts Pleistocene (Preboreal) (LaviIIe et al. , 1980: p. 312; themselves (Schvoerer et al., 1979: p. 29). Ac­ Celerier, 1976: p. 1431). Ifnot entirely Holocene in cording to Schvoerer et al. (1979: p. 29), Level 5b age, the Pont d'Ambon Azilian straddles the Plei­ dates (on charcoal) to 10,030±140 B.P. (Gif-2537), stocene/Holocene boundary. but according to Chollet et al. (1979: p. 369), the date should be 1l,030±140 B.P. Azilian Levels 4b and 3 date (on charcoal) to 10,990±160 B.P. (Gif- 10. SITES IN LIMOUSIN AND POITOU 3580) and 8800±220 B.P. (Gif-1588) respectively. (CORREZE AND VIENNE) These two levels are separated by a thick, ap­ parently archaeologically sterile layer (4a) com­ There are no dated Azilian sites in the Charentes posed of large rockfall blocks in a clay matrix. and only one each in Correze and Vienne, both The Level 4 lithic assemblage has high per­ departements which lie to the north of the Dor­ centages of Azilian points and short endscrapers, dogne. and there are far fewer backed bladelets than in the Upper Magdalenian assemblages. Lacking har­ poons, it has a curved bone fish hook. The Level 3 10.1. Chez Jugie (Cosnac, Correze) assemblage is poorer, but contains an even higher This rockshelter, right near the Dordogne border percentage (32.3%) of Azilian points, many short and not far from Le Pont d'Ambon, has been endscrapers and fewer burins. It contains two fish excavated recently by G. Maziere and J.P. Raynal. hook fragments, and indeed fish bones are abun­ Here an Azilian deposit (Level 5) lies between a dant and pertain to many species in the two Azilian sterile layer (6) and a series of Sauveterrian level s strata (much more so than in the Magdalenian (2-3), the latter dating to 7,000-8,000 B.P. Level 4, ones) (Chollet et al., 1979). There are also a few which lies between the Azilian and the earliest 'Azilian' engraved cobbles. Sauveterrian, is sterile (Maziere & Tixier, 1976: p. The ungulate faunas of Levels 5b and especia Ily 6 1441). Charcoal samples from Level 5 gave results are heavily dominated by Rangifer remains. The of 1 1,840±580 B.P. (Ly-1572) and 1 1,730±530 B.P. rodents are indicative of cold but increasingly (Ly-1601), dates which are considered satisfactory humid conditions. Snowy owl and arctic hare 'despite difficulties of measurement and risk of remains are also abundant in Level 5b and may be contamination' by 'roots, rootlets and burrows' in further evidence of cold, humid conditions (Chollet sandy sediments with high acidity (Evin et al., 1979: et al. , 1979: p. 374). However, the sediments of pp. 440-44 1). These dates would place the Chez Level� 6 and 5b are practically lacking in cryoclastic Jugie Azilian in the early Allerød. Its lithic as­ elements of the coarse and medium fractions. They semblage is 50% Azilian points, 33% endscrapers are mostly silty, clayey sands. In contrast, Level 4 (mostly short) and littIe else (only 1.5% burins). No has evidence of cooler, drier conditions, with an bone or bone tools are preserved (Maziere & Tixier, increase in frost-weathering. Reindeer remains are 1976: p. 1443; Maziere & Raynal, 1979: pp. 512- still present, but red deer is the dominant ungulate 110 L.G.STRAUS

in Level 4b. lronically, forest and waterside rodents Cantabrian and Pyrenean regions). Levels III and increase in Level 4b, and boar and roe deer, which II have sedimentological evidence of co Id condi­ appeared in Level 5b, are also present. The coarse tions, as does Level I (Combier, 1977: p. 576). Level granulometric fraction is again practically absent in III has three radiocarbon dates: 12,080±3 1O B.P. Level 3 and there are illuvial concretions. Reindeer (Ly-3 19-on humus), 11,750±300 B.P. (Ly-318-on is absent as is the tundra-dwelling arctic lemming charcoal) and 11,500±380 B.P. (Ly-320-on bone) (Dicrostonyx torquatus) which had been present (Schvoerer et al., 1979: p. 39). The latter two dates before. In general, the rodents are tempera te forest are considered more accurate. The faunas are and waterside species. Red deer dominates the poor dominated by rabbit remains, together with some ungulate fauna (Chollet et al. , 1979). Given the red deer, boar and ibex. Levels III and II, in problems of attempting broad climatic/chrono­ agreement with the radiocarbon dates, are believed stratigraphic correlations based on a site in a humid to belong to a cold phase within Allerød and Level I streamside microhabitat surrounded by upland pla­ to Dryas III. The lithic industries include many teaux, the authors do not propose phase assign­ Azilian points and small endscrapers (Combier , ments for their strata, although one hypothesis 1977: p. 576). could attribute Levels 6 and 5b to Allerød, Level 4 (Azilian) to Dryas III and Level 3 to Preboreal. This 11.2. Abri Jean Pierre l (Saint-Thibaud-de-Couz, hypothesis accomodates all the radiocarbon dates Savoie) (especiaIly when standard deviations are conside­ red) except the one for Level 6, which is rejected by This site (one of a pair of caves partly destroyed by a the excavator. Once again, then, the Azilian would quarry before excavation) is located at 500 m above straddle the boundary between Pleistocene and sea levei on the western slope of the Chartreuse Holocene. massifin very eastern France. It was dug in 1969-72 Two nearby Magdalenian deposits have radio­ by P. Bintz and yielded a series ofTerminal Magda­ carbon dates placing them in Dryas III. That ofLes lenian and Azilian assemblages, the latter equated Terriers in Lussac-les-Chateaux is 1O,450±250 B.P. by the excavator with the term 'Epipalaeolithic'. (Gif-1128), while the 'Magdalenian V' of the fa­ The radiocarbon dates for the Magdalenian range mous site of Le Roc-aux-Sorciers at Angles-sur­ in age from about 13,000 to about 12,000 B.P. I'Anglin dates to 1O,840±120 B.P. (GrN-2912). These levels (9 and 8) are assigned to Bølling and These dates help confirm the lateness of the Termi­ there is said to be a hiatus equivalent to Dryas II, nal Magdalenian and, consequently, Azilian in the just before the first occupations labelled as Azilian Vienne, as in many areas to the south. (Bintz, 1976; Bintz & Desbrosse, 1979). Conditions, based on sediments, macro- and microfauna, were Il. THE SITES OF THE RHONE-ALPES RE­ quite rigorous. The Jean Pierre Azilian sequence GION begins with Level 7, dated to 1l,900±360 B.P. (Ly- 429-on bone). This and Levels 6B and 6C (the latter There are a number of deposits which have been dated on charcoal to 10,750±300 B.P. (Ly-596)) are attributed to the Azilian in East-central France assigned to A:lerød, as there is some palynological which have been dated by radiocarbon. One of evidence ofwarmingplus greater evidence from the these, Le Saut du Loup, located in Ardeche, is rodents, birds and ungulates of increased foresta­ rather isolated, although actually closer to the Vau­ tion under more moderate climatic conditions, also cluse site of Chinchon than to the rest ofthe Azilian testified to by the sediments. However, while the sites in the Rhone-Alpes region discussed here. trend was generally toward more temperate, humid conditions, there were still strong, numerous episo­ des of freeze and thaw (Bintz & Desbrosse, 1979: p. 11.1. Le Saut du Loup (Bidon, Ardeche) 244). Level 6A (undated) shows a definite return to Also known as Abri Dumas, this site is one of a cold, steppic conditions, clearly apparent in the dense cluster of Pleistocene art and habitation loci pollen, sediments and molluscan fa una. The site along the gorge of the lower Ardeche river near its was little occupied at this time, presumably attribu­ confluence with the Rhone. Dug in the early 1970's table' to Dryas III. Level S had two Azilian occupa­ Saut du Loup has three cultural levels, all assigned tions dated to 1O,620±21O (Ly-l 190-on bone) (a to the Azilian, although no flat-section harpoons date for 5C which does not correspond exactly with were found (as is frequently the case outside of the the one given by Schvoerer et al. (1979: p. 38), and The Azilian problem in the Franco-Cantabrian region 111 which is believed by Bintz and Desbrosse (1979: p. 1O,220±650 B.P. (Ly-620-charcoal), 9,390±150 244) to be a bit too old for this level) and 9,050±260 B.P. (Ly-599-ca1cium carbonates) and 9,350±300 B.P. (Ly-428-on bone) for 5A, whose scanty indu­ B.P. (Ly-270 - charcoal) (Schvoerer et al., 1979: p. stry is tentatively called 'final Azilian' by the exca­ 37). vators. The pollen and sediments point to markedly warmer and more humid conditions for all af Level 11.4. Abri Gay (Poncin, Ain) 5, assignable to ·Preboreal. The large mammals are dominated by ibex and marrnot (understandable This is a huge rockshelter located near the famous given the site's steep rocky surroundings), plus Upper Palaeolithic site af la Colombiere an the chamois, red deer and moose. edge af the Ain River. Dug since 1970, it has pro­ The lithic assemblages ofLevel 7-assigned to the vided a 10ng stratigraphic sequence like Saint­ 'Early Azilian'- and those af Levels 6-5C ('Middle' Thibaud, running from the Late Magdalenian and 'Final Azilian') all have substantial percentages through the Azilian. The Azilian level, dated to of Azilian points and backed bladelets. Small end­ 11,660±240 B.P. (Ly-725-on bone), contains 7 scrapers and burins are, however, more numerous in painted pebbles and ane engraved ane together Level 7 (with a nearly identical burin index in with a lith ie industry almost entirely composed of 'Magdalenian' Level 9). There are no harpoons, Azilian points (nearly 90% to date) (Bintz, 1976: p. either af Magdalenian ar Azilian type. The upper­ 1407; Bintz & Desbrosse, 1979: p. 249). Sedimento­ most, 9,000-year old assemblage at Jean Pierre l is logi cal analyses indicate cold, humid conditions poor in tools and virtuaIly lacking in Azilian points, probably attributable to Allerød. Reindeer domi­ though it has a high percentage of small endscra­ nates the ungulate fauna (Bintz & Desbrosse, 1979: pers and relatively many backed bladelets (Bintz, pp. 244, 249), and there are arctic species of birds 1976: p. 1405; Bintz & Desbrosse, 1979: p. 253). Its (Mourer-Chauvire, 1979). attribution is c\early problematical. Interestingly, There is a Dryas III date for a 'Terminal Magda­ this Preboreal level has domesticated dog remains. lenian' deposit at the Abri des Douattes in Mussiege Industries partly similar to that af J ean Pierre Level (Haute-Savoie): 1O,680±450 B.P. (Ly-453-on bo­ 5A (Iabelled as 'Upper' ar 'Final Azilian') have ne). There is a second date of 12,480±260 B.P. been found at the adjacent caves af La Passagere (Ly-435), apparently from the same level. This is and Colomb in Meandre (Isere), to the west af overlain by a poor Azilian level (Thevenin et al. , Saint-Thibaud. These deposits are dated, respecti­ 1979: p. 226; Desbrosse & Girard, 1974). Reindeer vely, to 8,790±190 B.P. (Ly-1189-on bone) and and horse are present in both leveis. Desbrosse and 8,960±420 B.P. (Ly-420-on bone) (Bintz & Des­ Girard (1974) tend to reject both dates, and an the brosse, 1979: pp. 244, 253; Schvoerer et al., 1979: p. basis af two isolated pollen spectra, tentatively 39). These two sites also contain evidence af con­ place the Magdalenian level in early Allerød. The siderable concentration an marrnotand ibex procu­ Azilian would thus be more recent. Indeed, at the rem en t in early Post-Pleistocene times, although 'VieiIIe Eglise' rockshelter in La Balme-de Thuy other animaIs inc\uding a relict population af rein­ (Haute-Savoie) an Azilian deposit (Level 7 A) has deer (!) were also hunted by people in this region recently been dated an bone collagen to 9,820±200 flanking the Alps (Bintz & Desbrosse, 1979: p. 244). (Ly-2619) and 9,485±325 B.P. (CRG-41O) (Gine­ stet, 1984). The lithic industry inc\udes 77 retouc­ hed tooIs, af which 34.6% are endscrapers (mostly 11.3. Thoys I (Arbignieu, Ain) short forms), 3.8% burins, 22. 1 % backed bladelets, This small shelter, also known as Abri de la Touv­ 1l.7% Azilian points, and no perforatars ar geo­ iere, was dug from 1968-1971. It yielded a stratigra­ metric microliths. There are no harpoons pre­ phy particularly rich in recent deposits (Iran & served. The Azilian fauna in this alpine site is Bronze Ages, Chalcolithic and Neolithic) overlying dominated by ibex, with no reindeer, and only small a Sauveterrian series and a basal level which is quantities af red and roe deer. Overlying Level 6 characterized as 'Azilian-like' (Combier, 1977: p. contains lithic assemblages resembling that af Level 562; Bintz, 1976: p. 1407). This poor, problematical 7 in some respects, but has a Sauveterrian point, assemblage contains an Azilian point along with several geometric microliths, no Azilian points, and more 'Mesolithic' (i. e. Sauveterrian) elements. It a 14C date af 8,170±160 B.P. (Ly-1936). The has been firmly dated to the Preboreal/Dryas III Balme-de-Thuy Terminal Azilian thus seems firmly limit by three radiocarbon determinations: dated to Preboreal. 112 L.G.STRAUS

12. THE SITES OF BOURGOGNE bone) from Level Dl, whose sediments show evi­ dence of intense frost weathering activity. The 12. 1. Varennes-Ies-Måcon rodents are also cold species and reindeer is present. Given the 14C date, this level is assigned to the Sites were found during highway construction and Allerød period by the researchers (Thevenin et al., salvage excavations were conducted by J. Combier 1979: p. 227). Level Cl, which lies between Levels (1979). A hearth at site No. 10, characterized as Dl and 02, dates to 1I,090±260 B.P. (Ly-lI 92-on 'Azilian', has been charcoal dated to 8,080±280 bone) (Schvoerer et al., 1979: p. 40). This would B.P. (Ly-848) (Schvoerer et al., 1979: p. 40). This seem to confirm a late Allerød age for these depo­ date is said by the excavator to be apparently 'too sits. The Cl industry is similar to the D assemblages young' for an Azilian industry, and therefore must and also contained an engraved pebble. have been 'probably polluted by roots' (Evin et al., Level A4 has a similar industry, still without 1975: p. 26). large Azilian-type points, many short (thumbnail) Site No. 11 is a separate locus in the same bog as endscrapers, plus some backed bladelets, but very No. 10. Charcoal from it has been dated to few burins. Six more engraved pebbles were found 11,860±190 B.P. (Ly-849), a date which is judged in this level (and many more in old backdirt). The by Com bier (1979: p. 265) to be toa old. Pollen and base of Level B has been dated to 1O,730±190 B.P. fa una are poorly preserved. The only ungulates are (Ly-1 194-on bone), and on the basis of the mul­ bovines. The lithic industry is quite lamellar and is . . tidisciplinary studies, has been tentatively assIgned overwhelmingly dominated by Azilian points. to Dryas III (Thevenin et al., 1979: p. 2�7) . There are also specialized types found in the Termi­ . Level A4 has a similar industry, still wlthout nal northern Magdalenian (e.g. Hamburgian and geometric microliths, despite a decrease in the size Zonhoven points). Endscrapers outnumber burins. of the tools. There are a fa ir number of back ed Combier (1979: p. 263) sees similarities to the Azi­ bladelets and short, thick endscrapers are abun­ lian but also to the Terminal Magdalenian of the dant. This level has been dated to 9,21O± 120 B.P. Paris Basin at such sites as Pincevent (dated in­ (Gif-2530-on bone). Cold rodents are entirely cidentally, by a large number of radiocarbon de­ absent, although among the bird remains there are terminations ranging from about 12,000 to about . still some of Lagopus sp. The macrofauna mcludes 9,500 B.P.; Schvoerer et al., 1979: pp. 27-28). This large quantities of Cervus, Bos primigenius and Sus. site in Burgundy, lacking Azilian harpoons and Level A4 is assigned to Preboreal (Thevenin et al., cobbles, should thus best be qualified as 'Epimag­ 1979). It is followed by levels with geometric micro­ dalenian'; it is clearly on the frontier ofthe distribu­ liths of classic 'Mesolithic' appearance dating to tion of assemblages which can reasonably be as­ late Preboreal and Boreal (Thevenin, 1976a). signed to the Azilian. Such assemblages are absent In summary, Rochedane, far from the classic in North-central France. Other assemblages with Azilian area, but containing numerous decorated surer Azilian 'trademarks' have been found at Roc­ cobbles and Azilian points (despite its lack of har­ hedane in northeastern France (Franche-Comte), poons), has an Azilian stratigraphic sequence which although here there are als o similarities to the north fairly clearly spans the period from Allerød to Pre­ European Epimagdalenian. boreal, confirming the evidence from sites to the south. 13. THE SITES OF FRANCHE-COMTE AND ALSACE (DOUBS AND HAUT-RHIN) 13.2. Mannlefelsen (Oberlarg, Haut-Rhin) This review concludes with a site about 1050 km in a 13.1. Rochedane (Villars-sous-Danjoux, Doubs) straight northwest direction from La Palom�, the The Rochedane rockshelter, located near the bor­ rockshelter of Mannlefelsen, located near the Junc­ ders ofSwitzerland and Alsace, has been excavated, tion of the French, Swiss and German borders, not analyzed and extensively published by A. Thevenin far from Basel. Excavations underway since 1971 (1976a; Thevenin et al., 1979). Levels 02 and Dl have revealed 7 m of stratigraphy in this modest contain Azilian-type points and short endscrapers cavity, the basal level of which (T) was laid down . in large numbers, but few burins. They have also under very co Id conditions with intense cryoclastic yielded painted and engraved peb bles. There is a activity and an almost treeless vegetation attributed radio carbon date of 1 I,060±470 B.P. (Ly-1 193-on to Dryas III (Thevenin, 1976b: p. 1491). Level S The Azilian problem in the Franco-Cantabrian region 113 contains an Azilian assemblage, with Azilian points and both the pollen and sediment 'dates'. In only a and angular backed knives, plus small backed bla­ fe w more cases is 14C in agreement with either delets. This level dates to 1O,220±330 B.P. (Ny-2 1) pollen or sediments (Ekain II, Rascafio l, La Riera and has sedimentological evidence of decreased 27, Borie del Rey 3, Le Morin A, Abri Gay). Howe­ frost-weathering (smaller eboulis which are also ver even in these cases of apparent agreement there more rounded than in Level T). Pollen and macro­ are instances of equivocal 14C dates or problema­ botanical analyses show that pine was the main ticai Azilian attribution. There are many more arboreal taxon. Level S is either Late Dryas III or at instances where palynology and sedimentology are the limit with Preboreal (Thevenin, 1976b; Theve­ in apparent agreement but are either in disaccord nin et al., 1979: pp. 227-229). with 14C dates or lack radiocarbon controis. Whe­ Level R is cu Itu rally nearly sterile, but Level Q ther in cases of one-way dating, two-way or three­ contains an assemblage rich in geometric microliths way chronological concordance, as in disaccord, (including isosceles triangles, convex base micro­ the dates for assemblages assigned to the Azilian points, etc.). It has pollen from thermophile trees, culture-stratigraphic unit clearly do span the Alle­ which, given the northerly, continental location, is rød, Dryas III and Preboreal, with more problema­ good evidence of fairly warm, humid conditions. ticai taiIs to the distribution in the Dryas II and There are two radiocarbon dates for Level R: Boreal. Table 2 shows that frequently radiocarbon 9,41O±l1O B.P. (Lv-859) and 9,030±160 B.P. (Gif- dates are younger than palynological 'dates' , as 2387), both on charcoal samples. There is also a noted earlier particularly with regard to deposits in concordant date from overlying Level P: 8,230±300 northern Spa in palynologically attributed to Al­ B.P. (Ly- l297) (Thevenin, 1976a; Thevenin et al. , lerød. 1979:p. 229; Schvoerer et al, 1979: p. 40). A Prebo­ There is considerable evidence of temporal over­ real age for Level Q seems clear. The Level Q indu­ lap on the one hand with 'Magdalenian' assembla­ stry is characterized as 'Early Mesolithic' or 'Ger­ ges (including ones with cylindrical section har­ manic Epipalaeolithic'. poons) and, on the other hand, with 'Mesolithic' ones (e. g. Sauveterrian, Asturian). Of course part of this overlap may be due to the many uncertainties 14. CONCLUSIONS and errors which are inherent to the radiocarbon method (e. g. the variable ratio of atmospheric A summary of the likely ages of Azilian assembla­ 14Ci2c, contamination with both old and recent ges discussed in this review is presented in table l. In carbon, stratigraphic disturbance, differences be­ general, assemblages labelled as 'Azilian' ean be tween distinct sample materiais, differences among found in deposits dating from as old as 12,000 B.P. laboratories, etc.). However, as the Magdalenian, to as recent as 9,000 B.P. They ean span, as an Azilian and Mesolithic are all part of a cultural ensemble, the Allerød, Dryas III and Preboreal continuum, with fe w clear-cut, abruptly-marked pollen zones or periods, and thus straddle the Pieis­ differences in technology, real overlap is to be tocene/Holocene boundary as it is traditionally expected. Indeed cases of physical association of (albeit arbitrarily) defined. Recent synthesis of the such supposedly temporal diagnostics ('fossil direc­ oxygen isotope evidence of deep sea cores also pla­ tors') as cylindrical- and flat-section harpoons ces the stage 1-2 boundary at c. 10,000 B.P. (Hays et prove this point quite dramatically. al., 1976: p. 1131; see also Woillard and Mook Any review such as the present one is beset by the (1982) for correlation to the onset of Preboreal at fundamental problem of definition. There is no the Grande Pile bog at this date). such thing as a 'typical' Azilian assemblage. As in It is clear that there is considerable variability all Upper Palaeolithic industries, for example, there and disaccord among the different methods used to is considerable inter-assemblage variability both in date the Azilian deposits. Table 2 presents the Azi­ terms of'fossil director' types (harpoons, de corated lian levels as they have been dated by radiocarbon pebbles) and in terms of relative frequencies not and by interpretation of palynological and sedi­ only of Azilian points and short endscrapers, but mentological results. Site-by-site and level-by-Ievel also rriicroliths,endscrapers, burins and other ma­ there are significant differences among the dates jor to ol classes in general. There is admittedly a fair derived from the various methods. Only ve ry rare ly degree of circularity involved in the assignment of (Pont d' Ambon 2, Jean Pierre 5A and Pegourie 5-4) assemblages to the Azilian (or to any other culture­ is there agreement among the 14C determinations stratigraphic unit arbitrarily concocted by prehis- 114 L.G.STRAUS

Table I. Azilian chronology (X =Iikely age; ?=possible age).

o.. � � � ...... � � ...... co c;;" P:i :::: P:i '" P:i .... co co .� o '" o o '" o o co V') o c;; o c V') co V') � V') ..o V') V') N ;>, t- ., � .... t-, � ' o � � Sites (" o e o 'D, Cl :;;: Cl o: 00 co c-: �

Oscura de Ania X Los Azules X X La Riera X El Pendo X Cueva Morin X Rascano X Pielago ? Arenaza Santimamine Urtiaga X Ekain ? X Zatoya X Abauntz X Duruthy X Dufaure X Poeymati X ? La Vache X Rhodes II X Mas d'Azil Tourasse ? Margineda X X X Gazel X Pompignan X Chinchon Salpetriere ? Tete du Chien X Borie del Rey X X Pegourie X ? Graves X Faurelie ? Pont d'Ambon ? X X Chez Jugie X Bois Ragot X X Saut du Loup X X Jean Pierre I X X X Thoys X X Abri Gay X Varennes-Ies-Måcon ? Rochedane X X X Mannlefelsen X Les Douattes Balme-de-Thuy X

torians). Hopefully this revision has at leas t called 2. Decorated pebbles are concentrated in just a attention to the significant variability in composi­ few sites (i. e. Mas d' Azil, Rochedane and, to lesser tion and dating of assemblages commonly attribu­ extents, Abri Pages and Los Azules), but the styles ted to this classic archaeological construct. of decoration cross-cut the rat her rigid ly demarca­ Some broad patterns also seem to be revealed by ted distribution of harpoons (Thevenin, 1982; see the exercise: als o Fernandez-Tresguerres, 1980; Bahn & Cou­ l. Harpoons are far more common in the sites of raud, 1982). While absent in many (or most) Azilian the Vasco-Cantabrian, western and central Pyre­ deposits, these peb bles help confirm that the Azi­ nean regions than in the rest ofthe Azilian 'culture­ lian had some sort of reality or unit y (as yet to be area'. The reason is likely to be functional, such as fully defined or understood). the possibility of a concentration on salmon fishing 3. Dated Azilian deposits in the southern sector (to be proven) in rivers draining into the southern (Cantabria, Vascongada, Gascony, Foix, Roussil­ shore of the Bay of Biscay. Ion, Languedoc and Provence) seem as a wholp tn The Azilian problem in the Franco-Cantabrian region 115 cluster more tightly in the later time range (D ryas B.P. The nature ofthis relationship is, however, far III and Preboreal). While there are also many from clear. For one thing the fa unal change at the assemblages from more northerly sites which date Pleistocene/Holocene boundary was far from uni­ to these periods, many more there seem to clearly form. Reindeer was never a significant resource in date to Allerød. La Vache and the basal Azilian Vasco-Cantabrian Spain, where red deer and ibex deposit at Poeymaii are the sole strong exceptions were the key Upper Palaeolithic and Azilian ungu­ to the 'rule' in the southern sector. This tentative late game species. Nevertheless, the Azilian of this pattern presents an apparent paradox, as it would region is quite similar in artifact content to that of seem to contradiet expectations of earlier adaptive the Pyrenean region, where reindeer (and other change to meet logicaIly earlier major environmen­ open country grazers) were replaced in the early tal changes at the close of the Pleistocene in the Holocene by species favoring temperate, woodland southern area. In faet the latest sedimentological environments. In faet, within the territory which and paleontological evidence from Aquitaine seems now constitutes continental France, reindeer extir­ to suggest that 'Allerød' began and lasted longer pation cam'e at different times, depending on lati­ there than in northern Europe (Laville, 1979; Del­ tude and, particularly , on the presence or absence pech, 1979; pace Leroi-Gourhan & Girard, 1979). of mounta:in ranges, where relict populations of The relationships between Stone Age technologies Rangife r apparently survived up to or just past the and environmental changes are clearly not simple Pleistocene/Holocene boundary, and were exploi­ or direct, and this is one obvious case in point, as ted by 'Azilian' hunters. This seems to have been fu rther noted below. the case in parts of the Pyrenees, Massif Central and 4. A chronological synthesis such as the present Alps, although the data are still fairly scanty and study which pia ces primary emphasis on the overall controversial. Despite major faunal differences pattern ofthe radiocarbon dates, gives significantly among the various French regions (e. g. SE vs. SW different results than the interpretations of the vs. northern) and between these and Vasco-Canta­ pollen record from many of the same archaeolo­ brian Spain, there are both broad and detailed simi­ gical sites. As noted throughout, the conclusions larities in Azilian technology. based more heavily on radiocarbon are at particu­ For another thing, some of the changes in subsis­ lar variance with those of Arlette Leroi-Gourhan, tence strategies had been underway long before the especiaIly for Cantabrian Spain, tending to be con­ somewhat arbitrary date of 10,000 B.P., and seem sistently younger than hers (Leroi-Gourhan, 1971b; to have been rather independent of major environ­ Leroi-Gourhan & Girard, 1979; Leroi-Gourhan & mental changes. This is certainly the case with Can­ Renault-Miskovsky, 1977). There are problems tabria, where is substantial evidence of intensifica­ with dating based on radiocarbon, of course, but tion through specialized hunting and diversifica­ palynological dating contains an added measure of tion of the resource base (to inc!ude, notably, subjectivity, plus major components of sampling increasing amounts and kinds of aquatic foods) error (involving habitat differences, winds, over­ (Straus et al. , 1980; 1981; Clark & Straus, 1983). and under-representation of species in the pollen These changes occurred during the Solutrean and record, effects of small sample size, differential pre­ Magdalenian, and continued apace during the Azi­ servation, problems with the specificity of iden­ lianiAsturi an. Glacial/interglacial transitions had tifications) which have yet to be at all adequately occurred at leas t Il times beforeiduring the c. l my critiqued and controlled. Palynological dating is, in tenancy of Homo in Europe, but the adaptive shift short, several steps removed from 'reality' . Only to intensive, efficient, broad-spectrum subsistence further research ean help us decide upon the accu­ strategi es only occurred this last time. I suggest that racy and consistency of this method as a cross­ the added element which made a crucial difference dating device for archaeological sites, which is not was a denser human population. In physically con­ to deny palynology's tremendous value as an indi­ strained areas such as Cantabrian Spain, the shift reet indicator of past vegetations on the local and had to come somewhat earlier, during the Wiirm regional scales. Upper Pleniglacial. Here the Azilian adaptations 5. The Azilian is undeniably in some way related constituted a refinement, but were still a continua­ to the major readjustments in human adaptations tion of long-developing trends. In other regions which came about at the time of (and in part, due they may have represented more of a novelty and a to) the glacial/interglacial transition which occur­ transition to. true 'Mesolithic' -type adaptations. red during the period centered on the date of 10,000 The formation oftrue shell middens of considerable 116 L.G.STRAUS

Table 2. Dating of Azilian depos,ts.

Phases Direct 14C lndirect "C Pollen Sediments

7, 4S0 B.P.

Boreal Varennes 10(?) Faurelie 2 Pegourie S-4( )* Margineda S top Tourasse ( ) Zatoya IIup(?) Mas d'AziI'( ) Mas d'AziI3( )

8, 6S0 B.P. Los Azules 2 Los Azules 2

Preboreal Los Azules 3d-3a Ekain III, IF Oscura de Ania Salpetriere Mas d'Azil G3( ) Mas d'Azil G3( ) Arenaza I1(?) Ekain III Tourasse( ) Santimamine 5( ) Pont d'Ambon 2' Pont d'Ambon 2' Urtiaga C( ) Ekain IV, TF Jean Pierre S' Jean Pierre S' Abauntz d Chinchon A( ) Los Azules 3up Borie del Rey 4 Pegourie 6( ) Graves Bois Ragot 3 Dufaure 3' Pont d'Ambon 2' Jean Pierre SAt?)' Rochedane A4 Balme-de-Thuy Pompignan Margineda 6 top Margineda 6,S base'

IO, ISO B.P. Pont d'Ambon 3a, 3 Douattes(?) Mannlefelsen S' Gazel S(?) Tete du Chien Thoys(?) Mannlefelsen S'

Dryas III Los Azules 3f, 3e* Cueva Morin I El Pendo I( )3 La Riera 27', 28 Rascano l' Duruthy 2 La Riera 27*' Ekain IV( ) El Pendo Il Arenaza III(?) Adaouste( ) DlIruthy 23 Cueva Morin I Poeymaii BS Sallt du Loup F Castillo DlIfaure 3' La Riera 28 Borie del Rey 3' Rascano l' Le Morin(?)' Rhodes 7 Pielago Jean Pierre SC Mas d' Azilf3( ) Duruthy 23 Rochedane B Margineda 6 base Margineda 7,8 Pont d'Ambon 33 (Dufaure 3)l Jean Pierre 6A3 Mas d'Azilf3( ) Tourasse( ) Borie del Rey 3' Pont d'Ambon 33 Jean Pierre 6A3 Saut du Loup F Le Morin All, AF(?)

IO, 7S0 B.P. Los Azules 3f, 3e* Jean Pierre 6B Jean Pierre 6C La Riera 2S, 26( ) The Azilian problem in fhe Franco-Canfabrian region 117

Table 2 (continued).

Phases Direct I'C Indirect I'C Pollen Sediments

Allerød La Riera2 27* La Riera2 25,3 263, 272, 28 La Riera2 25, 26 Zatoya !I low, BIll Poeymali CPE . Pegollri'e 5-4( )* 1 Tourasse( ) Cueva Morin I Chez JlIgie Oscura de Ania( ) Pegourie 7-41 Bois Ragot 4b Otero I(?) Pont d'Ambon 3a Jean Pierre I 7, 6CI, 6B3 Saut de LOllp !IF Rascano I Abri Gai Abri Gay2 Rhodes 5, 6(?)2 Sallt dll Loup2 !IF, II Rochedane Dl, Cl La Vache stalag2 Pegourie 6-41 Pont d'Ambon 3a2 Jean Pierre 7, 6C, 6BI

11, 750 B.P. Rhodes 5, 6(?)2 Varennes II(?) Jean Pierre 71

Dryas II Pegourie 7

12, 250 B.P.

* = alternate I'C dates fo r same level 1 = three-way dating concordance 2 = two-way dating concordance with I'C 3 = concordance between pollen and sediment 'dates' only (?) = problem with 'Azalian' attriblltion ( ) = possibIe problem with dating

volume in Late Azilian times (the 'Asturian', the reflected by the standard sorts of data and analyses 'Arudian', and 'Arisian') represent the logical con­ archaeologists use in the study of the Azilian. The sequence of this trend toward intensification in the more we leam, the more we discover our ignorance efforts of human gro ups to survive against a back­ of the past. ground of both changed environments and increa­ sed populations. As I said before, the relationship between Azilian technology and adaptations is fa r 15. ACKNOWLEDGEMENTS from straightforward. We are only now beginning to perceive and pose the correct questions of the The idea for this review stemmed from correspon­ prehistoric archaeological record. dence with Christopher Meiklejohn (University of Azilian assemblages of artifacts and fa una do Winnipeg). It profited from his suggestions, as well indeed lie in between the Magdalenian and Mesoli­ as those ofM.R. Gonzalez Morales (Universidad de thic in composition and in their behavioral implica­ San tander), although the conclusions are entirely tions. The Azilian clearly belonged to two worlds my responsibility. Much of the bibliographic re­ and can genuinely be viewed as 'transitional'. One search and writing was done while on a sabbaticai can but only speculate that the apparently abrupt from the University of New Mexico. R.R. NewelI change in artistic activity coeva l with the Azilian, (Groningen) and T. Constandse-Westermann (Am­ might be a reflection of new forms of social organi­ sterdam) read an intermediate draft of the manus­ zation. Indeed, it is likely that the adaptation s to cript and provided very helpful comments, sugges­ post-Pleistocene conditions may have involved con­ tions and criticisms, many ofwhich led to clarifica­ siderable readjustments of social organization, in­ tions and other useful changes in the final version. cluding territorialism, at best only inadequately In other instances we have decided to respect our 118 L.G.STRAUS

disagreements and they have agreed to provide a ques, paleontologiques et palynologiques, bien commentary on the paper. I deeply appreciate their qu'on montre que les interpretations chronologi­ collegiality. The National Science Foundation ques. Il y a une tendance systematique vers un (USA) provided grants to G.A. Clark and myself vieillissement relatif des datations palynologiques, for the excavation and analysis of La Riera Cave surtout dans les Cantabres. and to myself for research at Abri Dufaure. To all L'ensemble des gisements aziliens embrasse la my sincere thanks. periode qui comprend Allerød, Dryas III et Prebo­ real, entre 12.000 et 9.000 ans B.P. environ. Donc, I'Azilien est å califourchon sur la limite Pleistod:ne/­ 16. SUMMARY Holod:ne. Cependant, les gisements meridionaux ont une tendance a etre plus strictement jeunes que Starting from the perspective of Vasco-Cantabrian ceux qui se trouvent plus au nord. Spain, this article reviews and evaluates the evi­ L' Azilien est aussi tres variable en ce gui con­ dence for the age of the Azilian there and in South, cerne la composition des ses assemblages indu­ Southwest, Southeast, East and Northeast France. striels et fauniques. La relation entre la technologie A relatively high degree of reliance is placed on azilienne et la supposee adaptation abrupte aux r-adiocarbon dating, which is used in conjunction environnements et ressources changes vers la limite with sedimentological, paleontological and palyno­ de 10.000 B.P. n'est pas simple. L'Azilien est en logical indicators, although interpretations of the effet une etape de 'transition' par le fait de ren­ pollen record are shown to often be at variance with fermer des elements industriels et fauniques trouves (and usually 'older' than) the radiocarbon dates. dans le Magdalenien et le Mesolithique, avec les­ The Azilian is shown as an ensemble to spa n the quels il existait un recouvrement tempore!. La Allerød, Dryas III and Preboreal periods, from c. nature des changements adaptifs, qui etaient I'apo­ 12,000 to c. 9,000 B.P., thus straddling the Pieis to­ gee des tendances mises en route bien avant 10.000 cene/Holocene boundary. However the southern B.P. dans des regions comme l'Espagne cantabri­ sites tend to be more strictly later than the more que, aurait consiste en fin de compte en des northerly ones in France for reasons which are modifications dans l'organization sociale des grou­ unc1ear. The Azilian is also highly variable in terms pes humains (ref1echies peut-etre dans les change­ of the composition of its artifact and faunal assem­ ments brusques dans le comportement artistique blages. The relationship between Azilian techno­ entre le Magdalenien et I'Azilien). logy and supposedly abrupt adaptation to radically changed environments/resources at the 10,000 B.P. boundary is not straightforward. The Azilian is 18. RESUMEN indeed 'transitional' in that it con tai ned artifactual and faunal elements found in both the Magdalenian Empezando con la perspectiva de la region canta­ and Mesolithic, with which it temporally overlap­ brica, este articulo pasa revista y evalua la evidencia ped. The nature of the adaptive change, the culmi­ para la edad del Aziliense aili como en el sur, sur­ nation of trends underway long before 10,000 B.P. oeste, sur-este, este y nor-este de Francia. Se da un in areas such as Cantabrian Spain, may ultimately grado relativamente alto de confianza a las fechas have lain in alterations in the social organization of del radiocarbono, que se usan en conjuci6n con los human groups (perhaps reflected in the brusque indicios sedimentol6gicos, paleontol6gicos y pali­ changes in artistic behavior between the Magdale­ nol6gicos, aunque las interpretaciones de los es­ nian and Azilian). pectros polinicos se mostran muchas veces desave­ nidas (y por Io general mas viejas) en relaci6n con 17. RESUME las fe chas del radiocarbono. Como conjunto, los yacimientos azilienses atra­ Prenant comme point de depart les donnees de la vesan el periodo de Allerød, Dryas III y Preboreal, region cantabrique, cet article analyse l'ensemble desde alrededor de 12.000 B.P. hasta 9.000 B.P., y existant des temoins pour l'age de I'Azilien lå et asi esta a horcajadas al limite entre el Pleistoceno y dans le Sud, Sud-Ouest, Sud-Est, Est et Nord-Est de el Hol'oceno. Sin embargo, los yacimientos meri­ la France. On met un haut degre de confiance dans dionales tienden a ser mas estrictamente tardios que les datations du radiocarbone, qui sont employees los yacimientos mas septentrionales. El Aziliense es ici en conjunction avec les indices sedimento10gi- tam bien muy variable en cuanto a la composici6n The Azilian problem in the Franco-Cantabrian region 119 de sus conjuntos industriaies y faunisticos. La rela­ The Aitzbitarte Azilian fa una, which is quite diversified, cian entre la tecnologia aziliense y la supuesta adap­ includes SI/S and Capreoll/s (as well as certainly intrusive remains of sheep/goat). The Ermittia Azilian level (II) also tacian abrupta a los medios ambientes y recursos en contains remains of Sus, but none of Capreolus (Altuna, torno al limite de 10.000 B.P. no es sencilla. El 1972). Aziliense es de hecho 'transicional' al tener unos 4. In fact, the rapid reforestation begun in Allerød and only temporarily affected by Dryas III is testified to by the sys te­ elementos industriaies y faunisticos hallados en el matic presence of roe deer (Cap reoll/s capreolus) and boar Magdaleniense y en el Mesolitico, con los cuales se (Sus sero/a ) in deposits pertaining to both phases. These sobrepone temporalmente. La naturaleza del cam­ species are true woodland creatures (unlike Cervus elaphl/s, which is far more flexible in its habits), so their presence in bio adaptivo, que era la culminacian de unas ten­ Dryas III deposits in Vasco-Cantabrian Spain is clear dencias en camino mucho antes de 10.000 B.P. en evidence of the attenuated nature of the cooling associated zonas como la region cantabrica, habria podido with this brief climatic phase in this region. Presence of roe deer and boar should not be taken as evidence necessarily consistir ultimamente en alteraciones de la organi­ contradictory of a Dryas III age determined on the basis of zacian social de los grupos humanos (quiza refleja­ radiocarbon and/or sedimentology at such southerly latitu­ das en los cambios bruscos en el comportamiento des. artistico entre el Magdaleniense y Aziliense).

:�o. REFERENCES 19. NOTES ALMAGRO GORBEA, M., 1970. Las fechas de C- 14 para la l. PossibIe Azilian stone tools were also found by the Conde de prehistoria y la arqueologia peninsular. Trabajos de Prehisto­ la Vega del Sella (1916; 1930) in the adjacent Uera sites of ria 27, pp. 9-42. Cueto de la Mina and Balmori, but in very uncertain stra­ ALTUNA, J., 1972. Fauna de mamiferos de los yacimientos tigraphic position. These pieces (small endscrapers and bac­ prehistoricos de Guipuzcoa. Munibe 24, pp. 1-464. ked bladelets-which are also very frequent in the regional ALTUNA, J., 1980. Historia de la domesticacion animal en el Upper Magdalenian) were not accompanied by flat-section Pais Vasco desdesus origenes hasta la romanizacion. Munibe harpoons. A possibIe painted cobble from the cave of El 32, pp. 1-163. Pindal, also in extreme eas tern Asturias, was found comple­ ALTUNA, J., 1981. Restos oseos del yacimiento prehistorico del tely out of stratigraphic context during construction done to Rascano. In: J. Gonzalez Echegaray & l. Barandiaran (eds.), provide protection and access to this Upper Palaeolithic El paleolftico superior de la Cueva del Rascano (= Centro de rupestral art sanctuary. No other possibIe Azilian materials Investigacion y Museo de Altamira, Monografias 3). San­ were found in subsequent test excavations (Fernandez­ tander, pp. 221-269. Tresguerres, 1980: p. 51). ALTUNA, J., 1984. Capitulo II. In: J. Altuna & J. Merino 2. There are a few other sites with Azilian harpoons in Asturias (eds.), El yacimiento prehistorico de la Cueva de Ekain. San and Santander. These include Cueva Oscura de Peran, La Sebastian, pp. 17-45. Meaza (an important,inadequately published site in western ALTUNA, J., A. BALDEON & K. MARIEZKURRENA, 1985. Santander, where the Azilian deposit with painted cobbles Cazadores magdalenienses en la Cueva de Erralla. MI/nibe apparently overIay an Upper Magdalenian level and possibly 37, pp. 1-206. underIay an Asturian conchero), El Salitre and El Valle. ALTUNA, J. & K. MARIEZKURRENA, 1982. Restos oseos There are no independent chronological indices available for del yacimiento prehistorico de Abauntz. Trabajos de Prehis­ these sites. There is a radiocarbon date (on charcoal) of toria Navarra 3, pp. 447-353. 1O,400±515 B.P. (GaK-2548) for a 'pre-Asturian' shell ALTUNA, J., K. MARIEZKURRENA, A. ARMENDARIZ, midden deposit at the top ofthe Upper Palaeolithic sequence L. DEL BARRIO, T. UGALDE & J. PENALVER, 1982. in El Cierro, a cave in eas tern Asturias (Clark, 1976). There Carta arqueologica de Guipuzcoa. Munibe 34, pp. 1-242. are as yet no published results from the new University of APELLANIZ, J.M. & J. ALTUNA, 1975. Excavaciones en la Salamanca excavations in this site which could clarify the Cueva de Arenaza L Noliciario Arqueologico Hisponico, Pre­ culture-stratigraphic attribution of the TerminalPleistocene historia 4, pp. 123- 154. deposit. ARAMBOUROU, R., 1978. Le giselllent prehistorique de Du­ 3. There are several other sites in the Spanish Basque Country ruthy il Sorde-I'Abbaye. (= Memoires de la Societe Prehisto- with Azilian harpoons but for which there are few other data rique Franyaise 13). • and no independent chronological indicators. These include ARANZADI, T. DE & J.M. DE BARANDIARAN, 1935. Balzola (Dima, Vizcaya), Goikolau (7), Lumentxa (Level Exploraciones en la Caverna de Santimamitie, 3a Memoria. 8-lower) in Vizcaya, Ermittia and Pikandita in Guipuzcoa. Bilbao. Other deposits contain no Azilian harpoons, but are assigned ARESO, P., 1984. Capitulo III. In: J. AItuna &J. Merino (eds.), by the authors to this culture-stratigraphic unit because of Elyacimiento prehistorico de la Cueva deEkail1. San Sebastian. their position above the Upper Magdalenian or below the pp. 47-60. Bronze Age/Eneolithic and because of the composition of BAHN, P., 1984. Pyrenean prehistory. Warminster. their lithic assemblages. These include Level III ofBerrober­ BAHN, P. & C. COURAUD, 1981. Les galets peints du Mas da in Navarra (Maluquer de Motes, 1964), Level I of Aitzbi­ d' Azil dans les collections britanniques. Bulletin de laSociete tarte IV and with less security, Silibranka, Bolinkoba, Mhidionale de Spe/eologie et de Prehistoire 21, pp. 9- 17. Atxurra in Vizcaya, Agarre,Kobatxo and AitzbeItz in Gui­ BARANDIARAN, L, 1967. El paleomesolftico del pireneo puzcoa (Altunaet al., 1982). Despite the importance of Aitz­ occidental. Zaragoza. bitarte IV, a major Upper Palaeolithic site near the Franco­ BARANDIARAN, L, 1977. El proces o de transicion Epipaleo­ Hispanic border in the hills above San Sebastian, and despite litico-Neolitico en la Cueva de Zatoya. Principe de Viana its fairly recent excavation, there are no paleoclimatic studies 146/147, pp. 5-46. and no radiocarbon dates for the Azilian or Magdalenian. BARANDIARAN, l., 1979. Azilien et post-Azilien dans le Pays 120 L.G.STRAUS

Basque meridional. In: D. de Sonneville-Bordes (ed.), Lafln CLOTTES, J., 1979. Circonscription de Midi-Pyrenees. Gallia des /emps glaciaires ell El/rope. Paris, pp. 721-732. Prlihis/oire 22, pp. 629-67 1. BARANDIARAN, J.M. DE, 1953. El hombre prehis/6rico ell el CLOTTES, J., 1983. Circonscription de Midi-Pyrenees. Gallia Pais Vasco. Buenos Aires. Prehistoire 26, pp. 465-510. BARANDIARAN, J.M. DE, 1962. Sall/iII/amine (= Excavacio­ CLOTTES, J. & R. SIMMONET, 1979. Le Paleolithique final nes arquelagicas en Espaiia). dans le bassin de Tarascon-sur-Ariege, d'apres les gisements BARANDIARAN, J.M. DE & J. ALTUNA, 1977. Excavacio­ des Eglises et de Rhodes II. In: D. de Sonneville-Bordes (ed.), nes en Ekain. Ml/Ilibe 29, pp. 3-58. La fllldes /ell/ps glaciaires ell Europe. Paris, pp. 647-658. BARANDIARAN, J.M. DE & D. DE SONNEVILLE-BOR­ COMBIER, J., 1977. Circonscription Rh6ne-Alpes. Gallia Pre­ DES, 1964. Magdalenien final et Azilien d'Urtiaga: Etude his/oire 20, pp. 561-668. statistique. In: E. Ripoll (ed.), Misceltillea ell Homellaje al COMBIER, J., 1979. Facies et chronologie du Paleolithique Aba/e Hellri Brel/il. Barcelona, pp. 163-171. final et de l'Azilien dans le sillon rhodanien. In: D. de BINTZ, P., 1976. Les civilisations de l'Epipaleolithique et du Sonneville-Bordes (ed.), Lafllldes /emps glaciaires ell El/rope. Mesolithique dans les Alpes et le Jura meridional. In: H. de Paris, pp. 259-264. Lumley (ed.), La prehis/oire/rall,aise. Paris, pp. 1405-1411. DELIBRIAS, G., M.T. GUILLIER &J. LABEYRIE, 1966. Gif BINTZ, P. & R. DESBROSSE, 1979. La fin des temps glaciaires natural radiocarbon measurements II. Radiocarboll 8, pp. dans les Alpes du nord et le Jura meri diona l. In: D. de 74-95. Son neville-Bordes (ed.), Lafllldes /ell/ps glaciaires ell El/rope. DELPECH, F., 1979. Les faunes de la fin des temps glaciaires Paris, pp. 239-255. dans le Sud-Ouest de la France. In: D. de Sonneville-Bordes BORDES, F., 1972. Circonscription d'Aquitaine. Gallia PreJzis­ (ed.), Lafllldes /ell/ps glaciaires ell El/rope. Paris, pp. 169-175. /oire 15, pp. 487-497. DELPECH, F., 1983. Les/al/Iles du pa/eoli/hique superiel/r dans BORDES, F. & D. DE SONNEVILLE-BORDES, 1979. l'Azi­ le Sud-Ol/es/ de la France. Paris. lianisation dans la vallee de la Dordogne. In: D. de Sonne­ DESBROSSE, R., & M. GIRARD, 1974. Azilien et Magdale­ ville-Bordes (ed.), Lafllldes /emps glaciaires ell El/rope. Paris, nien des Douattes. /'AIl/hropologie 78, pp. 48 1 -498. pp. 45 1-459. DUPRE, M., 1984. Capitulo IV. In: J. Altuna &J. Merino (eds.), BOYER-KLEIN, A., 1981. Analisis palinolagico del Rascaiio. El yacimiell/o prehis/6rico de la Cl/eva de Ekaill. San Sebas­ In: J. Gonzalez Echegaray & I. Barandiaran (eds.), El tian, pp. 61-63. paleolf/ico sl/perior de la Cl/eva del RascOllo (= Centro de ESCALON DE FONTON, M., 1976a. Les civilisations de l'Epi­ Investigacian y Museo de Altamira, Monografias 3). San­ paleolithique et du Mesolithique en Provence littorale. In: H. tander, pp. 215-220. de Lumley (ed.), La prehis/oire Frall,aise. Paris, pp. BREUIL, H. & P. DUBALEN, 1901. Fouilles d'un abri il Sordes 1367-1378. en 1900. ReVl/e de /'Ecole d'AIl/lzropologie de Paris 8, pp. ESCALON DE FONTON, M., 1976b. Les civilisations de l'Epi­ 251-268. paleolithique et du Mesolithique en Langedoc oriental. In: BREUIL, H. & H. OBERMAIER, 1912. Les premiers travaux H. de Lumley (ed.), La prehis/oire Frall,aise. Paris, pp. de l'Institut de Paleontologie humaine. /'AIl/hropologie 23, 1382-1389. pp. 1-27. ESCALON DE FONTON ,M. & F. BAZILE, 1976. Les civilisa­ BUTZER, K.W., 1973. Notas sobre la geomorfologia regional tions du Paleolithique superieur en Languedoc oriental. In: de la parte occidental de la provincia de Santander y la H. de Lumley (ed.), La prehis/oire Fralll;:aise. Paris, pp. estratigrafia de Cueva Morin. In: J. Gonzalez Echegaray & 1163-1173. L.G. Freeman (eds.), Cl/eva MorfIl: Excavaciolles 1969. ESCALON DE FONTON, M. & G. ONORATINI, 1976. Les Santander, pp. 267-276. civilisations du Paleolithique superieur en Provence littorale. BURLEIGH, R.,J. AMBERS &K. MATTHEWS, 1982. British In: H. de Lumley (Ed.), La prehis/oire Frall,aise. Paris, pp. Museum natural radiocarbon measurements XV. Radiocar­ 1145-1156. bOll 24, pp. 262-290. _ ESCALON DE FONTON, M., M.F. BONIFAY & G. ONO­ CABRERA, V., 1984. El yacimiell/o de la Cl/eva de El Cas/illo (= RATINI, 1979. Les industries de filiation magdalenienne Bibliotheca Praehistorica Hispana). Madrid. dans le Sud-Est de la France, leurs positions geo-chronolo­ CARBALLO, J., 1922. El paleolf/ico ell la Cos/a Call/librica. giques et les fa unes. In: D. de Sonneville-Bordes (ed.), Lafln Unpubl. doctoral thesis, Univ. Madrid. des /ell/ps glaciaires ell El/rope. Paris, pp. 269-286. CARBALLO, J., 1924. Prehis/oria gelleral y especial de EspaIla. EVIN, J., 1977. Criteres de choix des echantillons pour la data­ Madrid. tion par le radiocarbone. Bulle/in de la Socie/e Pre/zis/orique CARBALLO, J., 1960. lll ves/igaciolles prehis/6ricas Il. San­ Fran,aise 72, pp. 135-138. tander. EVIN, J., 1979. Reflexionsgenerales et donnees nouvelles sur la CAVA, A., 1975. La industria iitica de los niveles postazilienses chronologie absolue "C des industries de la fin du PaIeoli­ de Santimamine. Sal//l/ola I, pp. 53-73. thique et du debut du Mesolithique. In: D. de Sonneville­ CELERIER, G., 1976. Les civilisations de l'Epipaleolithique en Bordes (ed.), Laflll des /ell/ps glaciaires en Europe. Paris, pp. perigord. In: H. de Lum\ey (ed.), La prehis/oire /r all,aise. 5-13. Paris, pp. 1427-1432. EVIN, J., J. MARECHAL & G. MARIEN, 1983. Lyon natural CHAMPAGNE, F. &J. JAUBERT, 1981. La grotte des Fieux, radiocarbon measurements IX. Radiocarbon 25, pp. 59-128. il Miers. In: J. Clottes (ed.), La Prehis/oire dl/ Ql/ercy (= EVIN, J., J. MARECHAL, C. PACHIAUDI & J.J. PUISSE­ Congres Societe Prehistorique de France, XXI session). G UR, 1980. Conditions involved in dating terrestrial shelIs. Paris, pp. 85-104. Radiocarboll 22, pp. 545-555. CHOLLET, A., P. BOUTIN, A. DEBENATH, F. DELPECH EVIN, J., G. MARIEN &C. PACHIAUDI, 1975. Lyon natural & J.C. MARQUET, 1979. La grotte du Bois-Ragot ilGouex. radiocarbon measurements V. Radiocarbon 17, pp. 4-34. In: D. de Son neville-Bordes (ed.), Laflll des /ell/ps glaciaires EVIN, J., G. MARIEN & C. PACHIAUDI, 1979. Lyon natural ell El/rope. Paris, pp. 365-378. radiocarbon measurements VIII. Radiocarboll 21, pp. CLARK, G.A., 1976. El As/uriellse Call/librico (= Bibliotheca 405-452. Praehistorica Hispana). Madrid. FERNANDEZ-TRESGUERRES, J., 1980. El Aziliellse ell las CLARK, G.A. & L.G. STRAUS, 1983. Late Pleistocene hunter­ provillcias de As/urias y Sall/allder (= Centro de Investiga­ gatherer adaptations in Cantabrian Spain. In: G.N. Bailey cian y Museo de Altamira, Monografias 2). Santander. (ed.), Hl/Il/er-ga/herer ecollomy ill prehis/ory. Cambridge, pp. FUENTES, C., 1980. Estudio de la fa una de el Pendo. In: J. 131-148. Gonzalez Echegaray (ed.), El yacimiell/o de la Cl/eva' de el The Azilian problem in the Franco-Cantabrian region 121

Pelldo (= Bibliotheca Praehistorica Hispana). Madrid, pp. LEROI-GOURHAN, A., 197 1b. La fin du Tardiglaciaire et les 217-237. industries prehistoriques (Pyrenees-Cantabres). Munibe 23, GARCIA GUINEA, M.A., 1975. El Mesolitico en Cantabria. pp. 249-254. In: M.A. Garcia & M.A. Puente (eds.), La prehistoria ell la LEROI-GOURHAN, A., 1980. Analisis polinico de el Pendo. Comisa Canttibrica. Santander, pp. 177-197. In: J. Gonzalez Echegaray (ed.),EI yacimiento de la Cueva de GEDDES, D., M. BARBAZA, J. VAQUER & J. GUlLAINE, el Pelldo, pp. 265-266. 1986. Tardiglacial and Postglacial in the eastern Pyrenees LEROI-GOURHAN, A., 1986. The palynology of La Riera and western Languedoc. In: L.G. Straus (ed.), Th e el/d of the Cave. In: L.G. Straus & G.A. Clark (eds.), La Riera Cave (= Paleolithic ill the Old World (= British Archaeological Anthropological Research Papers 36). Reports). Oxford, pp. 63-80. LEROI-GOURHAN, A. & M. GIRARD, 1979. Chronologie GINESTET, J.-P., 1984. L'abri sous roche de la VieiIIe Eglise, la pollinique de quelques sites prehistoriques li la fin des temps Balme-de-Thuy (Haute Savoie): Premier resultats. Bulletill glaciaires. In: D. de Sonneville-Bordes (ed.), Lafill des temps de la Societe Prehistorique Franraise 81, pp. 320-342. glaciaires ell El/rope. Paris, pp. 49-52. GIRARD, M., F. MOSER & M. ORLIAC, 1979. Comparison LEROI-GOURHAN, A. &J. RENAULT-MISKOVSKY, 1977. des gisements de la Tourasse et du Mas d'Azi!. In: D. de La palynologie appliquee li I'archeologie. In: H. LaviIIe & J. Sonneville-Bordes (ed.), Lajill des temps glaciaires ell Europe. Renault-Miskovsky (eds.), Approche ecologique de /'homme Paris, pp. 637-643. fo ssile. Paris, pp. 35-49. GOMEZ-TABANERA, J.M., M. PEREZ & J. CANO, 1975. LE TENSORER, J.M., 1976. Les civilisations du Paleolithique Premiere prospection de Cueva Oscura de Ania dans le bas­ superieur dans le Lot-et-Garonne. In: H. de Lumley (ed.), La sin du Nalon et connaissance de ses vestiges d'art rupestre. prehistoire Francaise. Paris, pp. 1232-1 236. Bulletill de la Societe Prehistorique de /'Ariege 30, pp. 59-69. LE TENSORER, J.M., 1979. Donnees nouvelles sur le Magdale­ GONZALEZ ECHEGARAY, J., 1980. El yacimiellto de la nien final et I'Epipaleolithique en Lot-et-Garonne. In: D. de Cl/eva de el Pelldo (= Bibliotheca Praehistorica Hispana). Son neville-Bordes (ed.), Lajin des temps glaciaires ell Europe. Madrid. Paris, pp. 485-499. GONZALEZ ECHEGARAY, J., M.A. GARCIA GUINEA & LIV ACHE, M., 1976. Les civilisations de I'Epipaleolithique et A. BEGINES, 1966. Cueva del Otero (= Excavaciones du Mesolithique en Haute Provence et dans le Vaucluse. In: Arqueologicas en Espana 53). H. de Lumley (ed.), La prehistoire FraIIcaise. Paris, pp. GONZALEZ ECHEGARA Y, J. & L.G. FREEMAN, 1971. 1379-1381. Cueva MorfII: Excavaciolles 1966-1968. Santander. LOPEZ, P., Abauntz: analisis polinico. Trabajos de Prehistoria GONZALEZ MORALES, M.R., 1982. El Asturiense y otras NavO/Ta 3, pp. 355-358. CIIlturas loeales (= Centro de [nvestigacion y Museo de Alta­ LORBLANCHET, M., 1976a. Les civilisations du Paleolithique mira, Monografias 7). Santander. superieur dans le Haut Quercy. In: H. de Lumley (ed.), La GUllAINE, J. et al., 1985. La Balma Margineda: aux origines prehistoire Fram;aise. Paris, pp. 1189-1195. du peup1e andorran. Histoire et Archeologie 96, pp. 10-33. LORBLANCHET, M., 1976b. Les civilisations de I'Epipaleoli­ HAYES,J.D.,J. IMBRIE&N.J. SHACKLETON, 1976. Varia­ thique et du Mesolithique dans le Haut Quercy. In: H. de tions in the earths orbit: pacemaker of the Ice Ages. Sciellce Lumley (ed.), La prehistoire FraIIcaise. Paris, pp. 1398-1404. 194, pp. 1121-1131. MARIEZKURRENA, K., 1979. Dataciones de radiocarbono HOYOS, M., M. MARTNEZ, T. CHAPA, P. CASTAN OS & existentes para la prehistoria vasca. Mllllibe 31, pp. 237-255.

F.B. SANCHIZ, 1980. La Cueva de la Paloma (= Excavacio­ MARSAN , G., 1979. Les industries du Tardiglaciaire des nes Arqueologicas en Espana 116). Pyrenees-Atlantiques et du Guipuzcoa. In: D. de Son neville­ LAPLACE, G., 1953. Les couches li escargots des cavernes Bordes (ed.), Lajin des temps glaciaires ell Europe. Paris, pp. pyreneennes et le probleme de l' Arisien de Piette. Bulletin de . 667-692. la Societe Prehistorique Franraise 50, pp. 199-211. MAZIERE, G. & J. TIXIER, 1976. Les civilisations de l'Epipa­ LAPLACE, G. & J.M. MERINO, 1979. Application de latypo­ leolithique et du Mesolithique en Limousin. [n: H. de Lum­ logie analytique et structurale li l'et ude du processus d'azi­ ley (ed.), La prehistoire FraIIcaise. Paris, pp. 1441-1446. lianisation: la serie phyletique de la grotte d'Urtiaga en Pays MAZIERE, G. & J.-P. RA YNAL, 1979. La fin des temps gla­ Basque. In: D. de Son neville-Bordes (ed.), Lajill des temps ciaires en Limousin. In: D. de Son neville-Bordes (ed.), Lajin glaciaires ell Europe. Paris, pp. 693-709. des temps glaciaires ell Europe. Paris, pp. 1441-1446. LAVILLE, H., 1979. Chronostratigraphie des depots de la fin MOURE, J.A., 1975. Cronologia de las industrias tardiglaciares du Wiirm en perigord. In: D. de Son neville-Bordes (ed.), La en el Norte de Espafia. Trabajos de Prehistoria 32, pp. 21-34. jill des temps glaciaires ell Europe. Paris, pp. 159- 167. MOURER-CHAUVIRE, c., 1979. Les oiseaux de la fin des LAVILLE, H. & M. HOYOS, 1981. Estudio geologico de la temps glaciaires en France. In: D. de Sonneville-Bordes (ed.), Cueva del Rascano. In: J. Gonzalez Echegaray & I. Baran­ Lafi ll des temps glaciaires ell Europe. Paris, pp. 105-1 11. diaran (eds.), El PaleoUtico Superior de la Cueva del Rascallo NEWELL, R.R., T.S. CONSTANDSE-WESTERMANN & C. (= Centro de Investigacion y Museo de Altamira, Monogra­ MEIKLEJOHN, 1979. The skeletal remains of Mesolithic fias 3). Santander, pp. 191-210. man in Western Europe: an evaluative catalogue. Joumal of LAVILLE, H., J.P. RIGAUD &J. SACKETT, 1980. Rockshel­ Humall EvolutiolI 8, pp. 1-228. ters of the perigord. New York. NOUG[ER, L.-R. & R. ROBERT, 1977. Harpons 'aziliens' et LEROI-GOURHAN, A., 1959. Resultats de I'analyse pollinique harpons 'magdaleniens' de la grotte de la Vache li Alliat. de la grotte d'lsturitz. Bulletill de la Societe Prehistorique Bulletin de la Societe Prehistorique de /'Ariege 32, pp. 13-47. FraIIraise 56, pp. 619-624. OBERMAIER, H., 1925. El hombre f6 sil. (= Memorias de la LERO[-GOURHAN, A., 1966. Analisis polinico de la Cueva del Comision de Investigaciones Paleontologicas y Prehistoricas Otero. [n: J. Gonzalez Echegaray et al. (eds.), Cueva del 9). 2nd ed. Madrid. Otero (= Excavaciones Arqueologicas en Espana 53), pp. ORLIAC, E. & M. ORLIAC, 1973. La succession des industries 83-85. li la grotte de la Tourasse. Bulletin de la Societe Prehistorique LEROI-GOURHAN, A., 1967. Pollens et datation de la grotte FralIcaise 70, pp. 66-88. de la Vache (Ariege). Bulletill de la Societe Prehistorique de PAQUEREA U, M.-M., 1978. Analyses palynologiques de I'abri /'Ariege 22, pp. 115- 127. Duruthy li Sorde-I'Abbaye. In: R. Arambourou (ed.), Le LEROI-GOURHAN, A., 197 1a. Analisis polinico de Cueva GisemelIt Preh,istorique de DU/'llthy il Sorde-/'Abbaye, pp. Morin. In: J. Gonzalez Echegaray & L. Freeman (eds.), 96, 109. Cl/eva MorfII: Excavaciones 1966-1968, pp. 359-365. PAQUEREAU, M.-M., n.d. Preliminary report on the Abri 122 L.G.STRAUS

Dufaure pollen. Unpubl. manuscript. Journalof Field Archaeology JO, pp. 371-378. PEREZ, M., 1977. Presentaci6n de aIgu nos materiales proce­ STRAUS, L.G., G.A. CLARK, J. ALTUNA & J.A. ORTEA, dentes de Cueva Oscura de Ania, Las Regueras. Actas XIV 1980. Ice Age subsistence in northem Spain. Scielltific Ame­ Congreso Nacional de Arqueologfa. Zaragoza, pp. 179- 195. rican 242 (6), pp. 142- 1 52. RIGAUD, J.-P., 1980. Circonscription d'Aquitaine. Gallia pre­ STRAUS, L.G. & G.A. CLARK (eds.), 1986. La Riera Cave histoi!'e 23, pp. 39 1 -426. (= Anthropological Research Papers 36). Tempe. , SACCHI, D., 1976. Les civilisations de l'Epipaleolithique et du STRAUS, L.G., J. ALTUNA, G.A. CLARK, M. GONZALEZ Mesolithique en Languedoc occidental et en Roussillon. In: MORALES, H. LAVILLE, A. LEROI-GOURHAN, M. H. de Lumley (ed.), La prehistoire Fran(:aise. Paris, pp. MENENDEZ DE LA HOZ & J.A. ORTEA, 1981. Paleoe­ 1390-1397. cology at la Riera. Currellt Allthropology 22, pp. 655-682. SAINT-PERlER, R. DE, 1936. La Grotle d'lsturitz Il:Le Mag­ THEVENIN, A., 1976a. Les civilisations de l'Epipaleolithique dalenien de la Grallde Salle (= Archives de l'Institut de et du Mesolithique dans le Jura et en Franche-Comte. In: H. Paleontologie Humaine, Mern. 17). Paris. de Lumley (ed.), La prehistoire Frall(:aise. Paris, pp. SA WTELL, R.O. & I. TREAT, 1927. Primitive hearths ill the 1481-1485. Pyrellees. New York. THEVENIN, A., 1976b. Les civilisations de I'Epipaleolithique SCHMIDER, B., 1978. L'industrie lithique de la grotte de la et du Mesolithique en Lorraine et en Alsace. In: H. de Lum­ Vache, commune d'Alliat. Bulletill de la Societe Prehistorique ley (ed.), La prehistoire Frall(:aise. Paris, pp. 1491-1498. de I'Ariege 33, pp. 13-56. THEVENIN, A., 1982. A propos des galets graves aziliens. SCHVOERER, M., C. BORDIER, J. EVIN & G. DELlBRIAS, Bulletill de la Societe Prehistorique Frall(:aise79, pp. 267-269. 1979. Chronologie absolue de la fin des temps glaciaires: THEVENIN, A., M. CAMPY, F. GEISSERT, J. HElM, M. recensement et presentation des datations �e rapportant ildes HOFFERT, J.e. MARQUET, e. MOURER-CHAUVIRE, sites fran9ais. In: D. de Sonneville-Bordes (ed.), La fin des T. POULAIN-JOSlEN, J. SAINTY, O. SCHAAF, F. temps glaciaires ell Europe. Paris, pp. 21-4 1. SCHWEINGRUBER, e. VELASQUEZ&H. VOGT, 1979. SERONlE-VIVIEN, M.R. &J.M. LE TENSORER, 1979. Don­ Fondernentschronostratigraphiques des niveaux il industrie nees recentes sur l'Azilien du Lot: la grotte de Pegourie il Epipaleolithique de I'abri de Rochedane il Villars-sous­ Caniac-du-Causse. In: D. de Sonneville-Bordes (ed.), Lafill Dampjoux et de I'abri du Mannlefelsen I il Oberlarg. In: D. des temps glaciaires ell Europe. Paris, pp. 47 1-482. de Sonneville-Bordes (ed.), La fill des temps glaciaires ell SERONIE-VIVlEN, M.R., e. BENSCH, R. CAPOEGELLE, Europe. Paris, pp. 215-230. B. CHEVET, M. HEMINGWAY & A. LEROI-GOUR­ THIBAULT, e., 1978. Etude geologique du gisement de HAN, 1981. La grotte de Pegourie il Caniac-du Causse. In: J. Duruthy. In: R. Arambourou (ed.),Le gisemelIt prehistorique Clottes (ed.), La prehistoire du Quercy (= Congres Prehisto­ de Dumthy il Sorde-I'Abbaye, pp. 67-96. rique de France, XXI Session). Paris, pp. 250-265. UTRILLA, P., 1982. El yacimiento de la Cueva de Abauntz. SIMMONET, R., 1976. Les civilisations de I'Epipaleolithique et Trabajos de Prehistoria NavO/Ta 3, pp. 203-345. du Mesolithique dans les confins pyrenees de la Gascogne et VEGA DEL SELLA, R. DUQUE DE ESTRADA, CONDE DE du Languedoc. In: H. de Lumley (ed.), La prehistoire Frall­ LA, 1916. Paleolftico de Cueto de la Milla (= Memorias de la (:aise. Paris, pp. 1412-1419. Comisi6n de Investigaciones Paleontol6gicas y Prehist6ricas SIMMONET, R., 1983. Stratigraphie du Magdalenien final et de 13). Madrid. l'Azilien il Rhodes II, commune d'Arignac, Ariege. Bulletill VEGA DEL SELLA, R. DUQUE DE '<;TRADA,CONDE DE de la Sociere Prehistorique Frall(:aise 80, pp. 12-13. LA, 1930. Las O/evas de la Riera y Balmori (= Memorias de STRAUS, L.G., 1979. Mesolithic adaptations along the nor­ la Comisi6n de Investigaciones Paleontol6gicas y Prehist6- them coast of Spain. Quaternaria 21, pp. 305-327. ricas 38). Madrid. STRAUS, L.G., 1981. Las antiguas excavaciones en la Cueva del WOILLARD, G. & W. MOOK, 1982. Carbon-14 dates at Rascafio. In: J. Gonzalez Echegaray & l. Barandiaran (eds.), Grande Pile: correlation of land and sea chronologies. El Paleo/ftico Superior de la Cueva del Rascallo (= Centro de Science 215, pp. 159-161. Investigaci6n y Museo de Altamira, Monografias 3). San­ tander, pp. 15-24. STRAUS, L.G., 1983. The ab ri Dufaure prehistoric project: Paper submitted Nov. 1983, revised and updated Febr. 1986 and Preliminary report of the 1980, 1981 and 1982 field seasons. Sept. 1987.