Earliest oral grave lining from 13,70011,700-y-old Natuan burials at Raqefet Cave, Mt.

Carmel, Israel
Dani Nadela,1, Avinoam Daninb, Robert C. Powerc, Arlene M. Rosend, Fanny Bocquentine, Alexander Tsatskina, Danny Rosenberga, Reuven Yeshuruna,f, Lior Weissbroda, Noemi R. Rebollog, Omry Barzilaig, and Elisabetta Boarettog
a The Zinman Institute of Archaeology, University of Haifa, Israel, 91905; bThe Hebrew University of Jerusalem, Jerusalem, Israel, 91904; cPlant Foods and Hominin Dietary Ecology Research Group, Max Planck Institute for Evolutionary Anthropology, 04103 Leipzig, Germany; dDepartment of Anthropology, University of Texas, Austin, TX 78712; eCentre de Recherche Franais Jrusalem USR 3132, Centre National de la Recherche Scientique, 75794 Paris, France; fProgram in Human Ecology and Archaeobiology, National Museum of Natural History, Smithsonian Institution, Washington, DC; and gWeizmann InstituteMax Planck Center for Integrative Archaeology, Dangoor Research Accelerator Mass Spectrometer Radiocarbon Laboratory, Weizmann Institute of Science, Rehovot, Israel, 76100

Edited by Ofer Bar-Yosef, Harvard University, Cambridge, MA, and approved May 23, 2013 (received for review February 7, 2013)

Flowering plants possess mechanisms that stimulate positive emotional and social responses in humans. It is difcult to establish when people started to use owers in public and ceremonial events because of the scarcity of relevant evidence in the archaeological record. We report on uniquely preserved 13,70011,700-y-old grave linings made of owers, suggesting that such use began much earlier than previously thought. The only potentially older instance is the questionable use of owers in the Shanidar IV Neanderthal grave. The earliest cemeteries (ca. 15,000 11,500 y ago) in the Levant are known from Natuan sites in northern Israel, where dozens of burials reect a wide range of inhumation practices. The newly discovered ower linings were found in four Natuan graves at the burial site of Raqefet Cave, Mt. Carmel, Israel. Large identied plant impressions in the graves include stems of sage and other Lamiaceae (Labiatae; mint family) or Scrophulariaceae (gwort family) species; accompanied by a plethora of phytoliths, they provide the earliest direct evidence now known for such preparation and decoration of graves. Some of the plant species attest to spring burials with a strong emphasis on colorful and aromatic owers. Cave oor chiseling to accommodate the desired grave location and depth is also evident at the site. Thus, grave preparation was a sophisticated planned process, embedded with social and spiritual meanings reecting a complex preagricultural society undergoing profound changes at the end of the Pleistocene.
burial customs

full-length postures, patterned orientations, head and body decoration with beads, removal of the skull after body decomposition, use of ochre-based pigments, provisioning with grave goods and offerings, and possibly an association with funereal feasts (13, 19). Materials and Methods
Here we report on the discovery of grave linings made of owering plants in four radiocarbon-dated (13,70011,700 Cal. B.P.) Natuan burials excavated at Raqefet Cave, Mt. Carmel, Israel (Figs. 1 A and B, 2, and 3; Figs. S1 S2, S3 and S4; and Table S1). Excavations at the site exposed 29 skeletons, all but one clustered in a small area (ca. 15 m2) (Fig. 1C). Although not suitable for all Natuan contexts, the term cemetery seems justied here, because frequent, repetitive interments took place at a specically dedicated location, probably for at least several generations. The retrieved skeletons include infants, children, and adults (14, 15). Most of the burials were single interments, although four were double, in which two bodies were interred together in the same pit. In the double burials, deaths were certainly contemporaneous or almost so, because the burials are simultaneous. However, the cause of death (epidemic, accident, violence) is unknown, and in most such cases is never identied (10). The current study focuses on the analyses of plant impressions in four graves (Table S1). To document the complex formation processes in the Natuan graveyard, to rene the microstratigraphy, and to assess depositional and postdepositional processes, we collected sediment samples from all contexts in and around the graveyard for micromorphological studies (SI Text, section S1). We also sampled a wide array of deposits across the site for phytolith analysis (SI Text, section S2). These samples include 16 samples from eight burials, 12 samples from four bedrock mortars, and seven samples from other contexts and from naturally accumulating soils located away from the caves entrance (Table S2).
ANTHROPOLOGY

| preburial preparation | radiocarbon dates

n the Mediterranean Levant, the earliest known burials involved disposal of the dead in sporadic and isolated pits dug in caves and their terraces. Such burial sites are known at the caves of Qafzeh, Skhul, Tabun, Amud, and Kebara, all dated to the latter half of the Middle Paleolithic, ca. 120,000 55,000 y ago (15). The bodies were usually interred in exed positions, sometimes with selected animal parts placed on them. In all instances only a few individuals were buried at each site, likely reecting intermittent burials. These sites were not cemeteries in the modern sense, where frequent, repetitive interment and memorial events take place at a specically dedicated location, usually for generations. Natuan [ca. 15,00011,500 calibrated years (Cal.) B.P.] sites used for burial reect a different pattern. To date, more than 450 skeletons have been unearthed in sites such as el-Wad Cave and Terrace, Eynan, Hayonim Cave and Terrace, Hilazon Tachtit Cave, Nahal Oren, and Raqefet Cave; these sites have several characteristics conceptually similar to modern cemeteries (6 18). At each, at least several dozen burials were found in a delineated and densely used area. The unprecedented density of graves and the great variety of inhumation practices in these sites represent some elaboration of earlier traditions and a wide variety of innovations. The new forms of burial include the combination of individual and multiple graves, exed and
www.pnas.org/cgi/doi/10.1073/pnas.1302277110

Results The graveyard deposits vary in thickness from 0.8 m in the concentration between Homo 9 and Homo 15 (Locus 1) to 0.6 m in the concentration of Homo 31 to Homo 28 (Fig. 1C and Fig. S3). In thin sections the grave lls appear as heterogenic calcareous ash deposits mixed with comminuted charcoal, bones, local soil, gravel, and snail shells (SI Text, section S1 and Fig. S5 A and B). The uppermost part of the anthropogenic deposits was strongly affected by bioturbation, manifested as a loose consistency, high porosity and crumbly structure (Fig. S5 C and D). Signicantly, bioturbation acted mainly on a millimetric scale and decreased considerably with depth.

Author contributions: D.N. designed research; D.N., A.D., R.C.P., A.M.R., F.B., A.T., D.R., R.Y., L.W., N.R.R., O.B., and E.B. performed research; R.C.P. analyzed data; and D.N. wrote the paper. The authors declare no conict of interest. This article is a PNAS Direct Submission.
1

To whom correspondence should be addressed. E-mail: dnadel@research.haifa.ac.il.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1302277110/-/DCSupplemental.

PNAS Early Edition | 1 of 5

Fig. 1. (A) Location map of Mt. Carmel. (B) Major Natuan sites in Mt. Carmel. (C ) Plan of the Natuan burial area in the rst chamber, Raqefet cave. See Fig. S3 for a section through graves Homo 28Homo 31.

Several graves rest on hard limestone bedrock with a distinct upper layer altered by weathering into a porous, friable crust, 23 cm thick, that does not effervesce with diluted hydrochloric

acid, upon which a mud veneer, 2 3 mm thick, is present. Micromorphological and scanning electron microscope (SEM)/ energy dispersive X-ray spectroscopy (EDS) analyses of the crust studied in the grave of Homo 25 and Homo 28 (SI Text, section S1 and Fig. S5 EH) provide data regarding the microstructure, elemental composition, and postdepositional diagenesis of the initial calcite in the limestone bedrock. SEM observations show that the crust has an amorphous microstructure with fractures (Fig. S5 EG). EDS measurements show that the crust is composed of a calcium-phosphorus mineral phase with 14% phosphorus by weight, indicating that initial calcite of the limestone rock has been largely replaced by hydroxyapatite. On a microscale level, the crust shows variations in chemical composition which appear as different shades of gray (Fig. S5G). However, EDS measurements in both brighter and darker areas are similar and apparently reect the presence of two different mineralogical phases; one is a phosphate mineral phase, and the other is an alumosilicate phase revealed by peaks of silicon, aluminum, and potassium (Fig. S5H). The presence of alumosilicates is probably the result of a mud veneer on the phosphatic crust. Preservation conditions hamper full characterization of the veneer and its postulated plaster-like utilization. A mud plaster was found lining the inside of a Late Natuan grave at Hilazon Tachtit (12, 13). The veneer at Raqefet Cave contains impressions of stems, leaves, and fruits and hence must have been damp and in a plastic state during the burial event or immediately thereafter, allowing soft delicate plant tissues to leave their precise impressions. The largest number of preserved plant impressions was discovered in the double burial of Homo 25 and Homo 28 (Figs. 1C, 2, and 3; Figs. S1 BD and S3; and Table S1). The two skeletons were found lying on their backs, parallel to each other with their elbows juxtaposed. One was a 12- to 15-y-old adolescent (Homo 28) placed with the knees folded to the left. The skull was ritually removed from the grave at a later time, after esh decomposition. Homo 28 has been directly radiocarbon dated to 12,55011,720 Cal. B.P. (Table S1 and Fig. S6). Homo 25 was an individual over 30 y old placed with his knees folded to his right. This individual had a stone slab set vertically behind the head, which was facing upwards. The head and slab were naturally dislocated in the grave and fell on their side, after body decomposition (Fig. 2).

Fig. 2. (A) Field photograph of skeletons Homo 25 (adult, on left) and Homo 28 (adolescent, on right) during excavation. Note the almost vertical slab behind the skull of Homo 25 and the missing skull of H28. Photograph reproduced with permission from E. Gernstein. (Scale bar: 20 cm.) (B) A reconstruction of the double burial at the time of inhumation. The skull of Homo 25 was displaced in the grave long after burial (A), but originally the head was facing upwards. The skull of Homo 28 was ritually removed months or years after burial. Note the bright veneer inside the grave on the right, partially covered by green plants.

2 of 5 | www.pnas.org/cgi/doi/10.1073/pnas.1302277110

Nadel et al.

rest of the bedrock in the cave is usually undulating and has a different color (Fig. S4 A and B). Only after the modication of the rock were the veneer and plant lining set inside, as indicated by the impressions of reeds (Phragmites?) and other species found on the thin vertical veneer (Fig. S4C). We identied another dense set of impressions at the base of the Homo 1 grave; it was exposed 40 y ago, although at the time the impressions were not observed (14). Here, there are several examples of square stems crossing each other at right angles and in regular intervals, seemingly in a loose net-like arrangement of the lining (Fig. 3B). Impressions of stems with round crosssections were exposed immediately under the skeleton of a child (Homo 31) (Fig. 1C and Figs. S1A and S3). Apparently, the practice of lining graves with owers was not age-related, according to the Raqefet burials. This nding is not surprising; when children were not excluded from the rest of the dead community (usually they were), their treatment did not substantially differ from that of the adults (10). The plant impressions were restricted to the burial area. No similar remains were observed elsewhere on the cave oor or cave walls. Furthermore, no int, stone, or bone impressions were found in the graves despite the presence of thousands of these hard and durable artifacts within the grave lls. This nding suggests that the green lining was thick and continuous, covering the entire grave oor and sides, preventing other objects from leaving impressions on the mud veneer. The abundance of phytoliths recovered from eight graves provides additional evidence for the habitual use of plants in the

Fig. 3. Photographs of plant impressions found at the bottom of graves. (A) Impressions under the skeletons of Homo 25 and Homo 28. The long impressions are of square stems. (B) Impressions under the skeleton of Homo 1. Note the short, parallel impressions of square stems crossing long stems (at the top center) and the Y-like branching stems (at the left). Photographs reproduced with permission from E. Gernstein.

At the base of the grave we recorded more than 30 impressions; we identied 13 stem impressions, 315 cm long and 0.52 cm wide, with rectangular cross-sections and visible longitudinal veins preserving stem vascular bundles (Fig. 3A and Figs. S1 BD, S2, and S3). One set of impressions was identied as the owering stems of Salvia judaica Boiss. (Judean sage) (Fig. 4A) based on the size and angles between the branching stems (Fig. 4B). Other rectangular stems belong to sage and closely related species of the mint family (Labiatae) or to the gwort family (Scrophulariaceae). Species of these families currently grow on the terrace and slopes below the cave. Local plants of these families ower in spring; most have a strong, aromatic fragrance, and some possess medicinal properties. The stems become stiff in April May and begin to deteriorate after desiccation during the summer. Thus, the impressions on the veneer were formed before midsummer. Additional impressions in this grave include stems with a round crosssection; three wide leaves, 5 cm in length, likely of trees; a row of ve small impressions with a close morphological afnity to seeds of Cercis silliquastrum L. (Judas tree) and small (<3 mm), round impressions of unidentied seeds. The double grave of Homo 18 and Homo 19 is unique because it preserves data regarding a preparation phase preceding the oral lining. Here, a bedrock bulge on the cave oor was chiseled to serve as an inner wall of the grave. The chiseled surface is ca. 40 40 cm in area, creating an almost vertical plane, whereas the
Nadel et al.

Fig. 4. (A) Impressions of owering stems in the grave of Homo 25 and Homo 28, marked by dashed lines. Photograph reproduced with permission from E. Gernstein. (B) Flowering stems of Salvia judaica, presented in the same scale and orientation as the impressions in the grave. Photograph reproduced with permission from A. Danin. (C and D) Jigsaw-puzzle phytoliths indicating tissue of a dicot plant. These phytoliths are associated with the graves and are also found in Salvia species growing near the cave today. Photographs reproduced with permission from R. Power.

PNAS Early Edition | 3 of 5

ANTHROPOLOGY

Raqefet Cave burials (Fig. 4 C and D). Phytoliths were identied in 35 contexts in the cave and on its terrace, including control samples. They consist of morphotypes from grasses, dicotyledon leaves, reeds, and sedges. The average density of samples from the eight graves is 61,199 phytoliths per gram of sediment, whereas the average for the off-site control samples is only 27,231 per gram. The eight grave samples produced 91% of all jigsaw dicotyledon phytoliths even though these contexts produced only 38% of all phytoliths (SI Text, section S1, Tables S2S4, and Fig. S7). The density of dicot phytoliths found in graves at Raqefet Cave is unique for phytolith concentrations at Natuan sites and supports the suggested use of large quantities of shrubs in the burials. Our recent study (summer 2012) revealed that phytoliths of local Lamiaceae species (including Salvia fruticosa) fall morphologically within the group of dicotyledonous phytolith morphotypes and thus may be one of the taxa used as grave lining at the Raqefet Cave burials (Fig. 4 C and D). Direct 14C dates of bone collagen (extracted from three human skeletons) are now available for two double graves. The Homo 18 and Homo 19 skeletons produced a range of 13,700 13,000 Cal. B.P. The Homo 28 skeleton was dated to 12,550 11,720 Cal. B.P. (Fig. S7). The dates indicate that the site was used for burial, repeatedly, in the same conned area and with the same customs and grave preparation, likely for as long as about 2,000 y. The characteristics of the burial customs (6, 14), the int assemblage (15), and the bedrock mortars accord with these dates. Indeed, the site was used as a graveyard for many generations, during which inhumation practices that included plant lining did not change. The burial pits contained abundant material remains. In many burials worked and natural stone objects were set on edge, perhaps as markers or symbols. Flints and butchered animal bones were very common. The low frequency of trampling, burning, and carnivore gnawing marks on the animal bones suggest instant burial of food remains during interment, as would be expected in funerary feasting. Abundant bedrock mortars and cup marks were hewn in the cave oor adjacent to the graves (15), and one human skeleton was found resting on the rim of a large bedrock mortar. This nding suggests that at least some mortars functioned in conjunction with the burials. Discussion The Raqefet Cave plant impressions and phytoliths provide evidence for the earliest known use of plant lining in grave preparation. Earlier examples of plant lining are limited to utilitarian grass beddings found at Middle Stone Age/Middle Paleolithic Sibudu Cave (20), Misliya Cave (21), and Tor Faraj rockshelter (22), where the lining was identied as thin laminar layers containing abundant phytoliths and other microscopic plant remains. The oldest macroscopic lining remains yet identied were found on brush hut oors at Ohalo II (23,000 y ago), where grass bedding composed of large bundles of Puccinellia convoluta was exposed and analyzed (23). The nds from Raqefet Cave indicate that 13,70011,700 y ago the Natuans lined graves with a soft mud veneer and then placed on the veneer a thick cover of fresh owering plants, thereby providing color and aromatic fragrance. Claims for earlier use of owers in the Shanidar IV Neanderthal burial (Iraq) were based on concentrations of microscopic pollen grains found adjacent to the skeleton (24, 25). However, the presence of rodent burrows, the abundant remains of jird (Meriones crassus) bones in the same layer, and this rodents habit of storing seeds and ower heads in its burrows cast serious doubts on this interpretation (26). Natuan grave preparation at Raqefet Cave included three basic patterns. The rst was the creation of a pit in either Middle Paleolithic or Natuan deposits. In several cases Natuan pits were dug successively in the same location, with the later ones

disturbing earlier ones (Fig. 1C and Fig. S3). These pits were dug in loose deposits and then were lled with sediments from the immediate surroundings (Fig. S5 AD). This pre-Natuan tradition was practiced for millennia and was very common at Raqefet Cave and many other Natuan burial sites (1, 2, 6, 9, 10, 16, 17). The second pattern was the chiseling of bedrock to accommodate the desired grave. The Natuan skills of high-quality rock and stone carving are well demonstrated by bedrock and stone mortars (2729) and by aesthetic and symbolic objects such as gurines and decorative designs (16, 30, 31). At Raqefet Cave there is direct evidence for bedrock chiseling in the graveyard as part of grave preparation. In one case the result was a vertical surface against which the foot of Homo 19 rested (Fig. S4). Rather than simply being a part of the physical setting of the grave pit, this chiseled surface may have had a symbolic meaning. Stones set vertically are very common in burials (e.g., Homo 25; see also Fig. S3) and even occur in bedrock mortars at Raqefet Cave (27). The third pattern was the lining of the pit with a thick layer of green plants, including owering species renowned for their aromatic fragrance and bright colors. This lining may have been the practice for all burials at the site, although remains were preserved only in some of the graves resting directly on bedrock. The lining practice appears to have taken place regardless of age and sex, in single and double burials. It may have been common in other Natuan sites, although such remains could be preserved for millennia only in particular conditions. Naturally, all three patterns of grave preparation could have been combined, as was the case for the Homo 18Homo 19 grave. Experimental studies have demonstrated owers signicant role as external sources of emotional stimuli, with measurable positive impacts on human social function (32). Flowers can be used to express sympathy, pride, and joy (33). They also are used to express religious feelings; in some religions owers are considered the direct route for spiritual communication (34). These relationships may benet tness in both humans and owers and have been linked to the domestication of certain species of owering plants more than 5,000 y ago (32, 35). At a signicantly earlier period, the use of owers in social events (funerals) may have served as yet another means for enhancing group identity and solidarity. The Natuan development of group-specic burials and related practices likely reduced social tensions and improved group cohesion in a period of uctuating environmental conditions, increasing population density, and growing social conicts (611, 36, 37). The emergence of Natuan cemeteries, such as those at Raqefet Cave and Hilazon Tachtit Cave, also may represent new and complex social organizations which could have included the establishment or strengthening of special interest groups, inheritance of corporate property, territorial ownerships, and aspects of social organization (611, 12, 13, 19). The careful preparation of the graves and the common use of owers add yet another perspective to Natuan funerary rites and their impact on the participants in a way that is familiar among many modern cultures.
ACKNOWLEDGMENTS. We thank G. Bar-Oz, A. Belfer-Cohen, L. Conyers, S. Filin, I. Hershkovitz, D. Kaufman, G. Lengyel, M. Weinstein-Evron, A. Weisskopf, and D. Bruggeman for their support, comments and advice; R. Shar and T.R. Sevi for laboratory and eld assistance; E. Mintz for help in sample preparation; and A. Lambert and G. Bosset, who also assisted in eld work. Digital gures were prepared by A. Regev. Photographs were taken by E. Bartov (Fig. S4B), A. Danin (Fig. 4B), R. Power (Fig. 4 C and D), M. Eisenberg (Fig. S2 C and D), and E. Gerstein (Figs. 2A, 3AB. 4A, S1 A, B, C, and D, S2 A and B, S3 A and C). Field work was conducted under permits from the Israel Antiquities Authority and the Israel Nature and Parks Authority. This project was supported by Grant 8915-11 from the National Geographic Society, Grant 7481 from the Wenner-Gren Foundation, and the Irene Levi-Sala CARE Foundation. Radiocarbon dating was funded by Grant 475/10 from the Israel Science Foundation (to E. B.).

4 of 5 | www.pnas.org/cgi/doi/10.1073/pnas.1302277110

Nadel et al.

1. Vandermeersch B (1981) Les Hommes Fossiles de Qafzeh (Israel) [The Fossil Humans of Qafzeh (Israel)] (Centre National de la Recherche Scietique, Paris). French. 2. Garrod DAE, Bate DMA (1937) The Stone Age of Mount Carmel (Clarendon, Oxford, UK). 3. Bar-Yosef O, Callander J (1999) The woman from Tabun: Garrods doubts in historical perspective. J Hum Evol 37(6):879885. 4. Hovers E, Kimbel WH, Rak Y (2000) The Amud 7 skeletonstill a burial. Response to Gargett. J Hum Evol 39(2):253260. 5. Bar-Yosef O, et al. (1992) The excavations in Kebara cave, Mt. Carmel. Curr Anthropol 33:497550. 6. Bocquentin F, et al. (2010) De la rcurrence la norme: Interprter les pratiques funraires en prhistoire [From repetition to norm: Interpreting prehistoric funerary practices]. Bull Mem Soc Anthropol Paris 22:157 171. French. 7. Bar-Yosef O (1998) The Natuan culture in the Levant, threshold to the origins of agriculture. Evol Anthropol 6:159177. 8. Byrd FB, Monahan CM (1995) Death, mortuary ritual, and Natuan social structure. J Anthropol Archaeol 14:251287. 9. Weinstein-Evron M (2009) Archaeology in the Archives. Unveiling the Natuan Culture of Mount Carmel (Brill, Boston). 10. Bocquentin F (2003) Pratiques funraires, paramters biologiques et identitis culturelles au Natouen: une analyse archo-anthropologique [Burial Practices, Biological Factors and Cultural Identities During the Natuan Period: A Bio-Archaeological Perspective]. PhD thesis. (Universit Bordeaux 1, Talence, France). French. Available at http://grenet. drimm.u-bordeaux1.fr/pdf/2003/BOCQUENTIN_FANNY_2003.pdf. 11. Grosman L (2003) Preserving cultural traditions in a period of instability: the Late Natuan of the hilly Mediterranean zone. Curr Anthropol 44:571 580. 12. Grosman L, Munro ND, Belfer-Cohen A (2008) A 12,000-year-old Shaman burial from the southern Levant (Israel). Proc Natl Acad Sci USA 105(46):1766517669. 13. Munro ND, Grosman L (2010) Early evidence (ca. 12,000 B.P.) for feasting at a burial cave in Israel. Proc Natl Acad Sci USA 107(35):1536215366. 14. Lengyel G, Bocquentin F (2005) Burials of Raqefet Cave in the context of the Late Natuan. J Isr Prehist Soc 35:271284. 15. Nadel D, et al. (2009) The Raqefet Cave 2008 excavation season. J Isr Prehist Soc 39: 2161. 16. Perrot J (1966) Le Gisement Natouen de Mallaha (Eynan), Israel. Anthropologie 70: 437484. 17. Belfer-Cohen A (1988) The Natuan graveyard in Hayonim Cave. Palorient 14:297308. 18. Stekelis M, Yizraely T (1963) Excavations at Nahal Oren. Isr Explor J 13:112. 19. Hayden B (2011) Guess Whos Coming to Dinner. Feasting Rituals in the Prehistoric Societies of Europe and the Near East, eds Jimnez GA, Montn-Subas S, SnzhezRomero M (Oxbow, Oxford, UK), pp 3063.

20. Goldberg P, et al. (2009) Bedding, hearths, and site maintenance in the Middle Stone Age site of Sibudu cave, Kwazulu-Natal, South Africa. Archaeol Anthropol Sci 1:95 122. 21. Weinstein-Evron M, et al. (2012) A Window into Early Middle Paleolithic human occupational layers: Misliya Cave, Mount Carmel, Israel. Paleoanthropology 2012: 202 228. 22. Rosen AM (2003) Neanderthals in the Levant: Behavioral Organization and the Beginnings of Human Modernity, ed Henry DO (Continuum, London), pp 156171. 23. Nadel D, et al. (2004) Stone Age hut in Israel yields world s oldest evidence of bedding. Proc Natl Acad Sci USA 101(17):6821 6826. 24. Leroi-Gourhan A (1975) The owers found with Shanidar IV, a Neanderthal burial in Iraq. Science 190:562 564. 25. Solecki RF (1971) Shanidar the First Flower People (Knopff, New York). 26. Sommer JD (1999) The Shanidar IV Flower Burial: a reevaluation of Neanderthal burial ritual. Camb Archaeol J 9:127 137. 27. Nadel D, Lengyel G (2009) Human-made bedrock holes (mortars and cupmarks) as a Late Natuan social phenomenon. Archaeology, Anthropology and Ethnology of Eurasia 37(2):3748. 28. Nadel D, Rosenberg D, Yeshurun R (2009) The deep and the shallow: The role of Natuan bedrock features at Rosh Zin, Central Negev, Israel. Bull Am Schools Orient Res 355:129. 29. Nadel D, Rosenberg D (2010) New insights into Late Natuan bedrock features (mortars and cupmarks). Eurasian Prehistory 7(1):65 87. 30. Weinstein-Evron M, Belfer-Cohen A (1993) Natuan gurines from the new excavations of the El-Wad cave, Mt. Carmel, Israel. Rock Art Research 10(2):102106. 31. Edwards PC (1991) in The Natuan Culture in the Levant, eds Bar-Yosef O, Valla FR, (International Monographs in Prehistory, Archaeological Series 1, Ann Arbor), pp 123148. 32. Haviland-Jones J, Rosario HH, Wilson P, McGuire TR (2005) An environmental approach to positive emotion: Flowers. Evol Psychol 3:104132. 33. Heilmeyer M (2001) The Language of Flowers: Symbols and Myths (Prestel, New York). 34. Stenta N (1930) From other lands: the use of owers in the spirit of the liturgy. Orate Fratres 4(11):462469. 35. Pollan M (2002) The Botany of Desire (Random, New York). 36. Bar-Yosef O, Belfer-Cohen A (1989) The origins of sedentism and farming communities in the Levant. J World Prehist 3:447498. 37. Bar-Yosef O (2012) Biodiversity in Agriculture: Domestication, Evolution and Sustainability, eds Gepts P, et al. (Cambridge Univ Press, Cambridge, UK), pp 5791.

Nadel et al.

PNAS Early Edition | 5 of 5

ANTHROPOLOGY

Supporting Information
Nadel et al. 10.1073/pnas.1302277110
SI Text S1. Micromorphological and Scanning Electron Microscope/ Energy Dispersive X-Ray Spectroscopy Analyses Materials and Methods. To characterize depositional and postdepositional processes at the site in general and the graveyard in particular, we retrieved 74 samples for micromorphological analyses. These samples were collected from all types of loci. So far we have analyzed 29 samples: 14 samples from the graveyard in the rst chamber (some Natuan burial pits were dug into earlier Natuan graves and sediments, several were dug into Middle Paleolithic sediments, and the bottom ones are located directly on limestone bedrock), 11 samples from bedrock mortars in the rst chamber and on the terrace, and four samples from tufas and rock crusts (1, 2). The 29 samples were collected as undisturbed blocks. All samples were processed in the Laboratory of Geology at the University of Haifa. They were impregnated with a polyester resin/ styrene mixture under vacuum, dried, and cut into small cubes. Then petrographic thin sections (30 m thick) were sliced off in the petrography laboratory of the Department of Maritime Civilizations at the University of Haifa. The thin sections were examined with Olympus BH-2 petrographic microscope in planepolarized (PPL) and cross-polarized (XPL) light using magnications of 40 to 400. They are described according to Stoops method (3), applying concepts accepted in micromorphology of archaeologically related sediments (4, 5). The microstructure and chemical composition studies were performed on a Quanta scanning electron microscope (SEM), Oxford Instruments FEA, linked with an energy-dispersive spectrometer (EDS) INCA 200, in the Materials and Environmental Engineering Department of the Technion-Israel Institute of Technology. The SEM/EDS samples were analyzed uncoated, and all SEM microphotographs were taken in the mode of back-scattered (BS) electrons.
Results. The burial pits were lled during the inhumation process

with ash-rich deposits mixed with local soil materials (Fig. S5). The thin-section analyses indicate that bioturbation was common in the upper deposits of the graveyard and decreased substantially downward. Moreover, the articulation of the skeletons and the associated in situ artifacts (e.g., stones set on edge, int cores near the skulls, selected animal bones in particular locations) clearly reect contextual preservation of macroscopic remains. Combined with the microscopic studies, the Raqefet Cave burials serve as excellent examples where not only the skeletons and their associated durable material remains were preserved intact but even perishable materialsthe impressions of plant liningswere preserved at the bottom of graves. The high phytolith densities in burials and bedrock mortars indicate good preservation conditions as well as distinct and common use of plants in both (see below). Overall, the variety of remains and observations, combined with thin-section analyses, establish that bioturbation and other postdepositional processes had little impact on the location of anthropogenic artifacts within graves or bedrock mortars. S2. Phytolith Analysis
Materials and Methods. Because macroscopic botanical remains are extremely scarce in the southern Levantine Natuan sites (6, 7), phytolith analyses provide one of the few possible direct ways to reconstruct past plant-related activities. To study plant use through the spatial distribution of phytoliths at Raqefet Cave, we analyzed 35 sediment samples retrieved from all types of in
Nadel et al. www.pnas.org/cgi/content/short/1302277110

situ Natuan features. These samples include 16 samples from eight inhumations, 12 samples from four bedrock mortars, two samples from a deep bedrock basin above which there were three burials, and two samples from Natuan tufa; we also studied a pre-Natuan (Kebaran) tufa and two control samples taken from naturally accumulating soils located 40 m and 100 m from the caves entrance (Table S2). The sediment samples were analyzed by R.C.P. at the Institute of Archaeology (University College London, London, United Kingdom) and the Plant Foods Research Group (Max Planck Institute for Evolutionary Anthropology, Liepzig, Germany). Phytoliths were extracted following a conventional protocol using gravity sedimentation, dry-ashing, and heavy liquid oatation (8). These phytoliths were treated with 15 mL of 10% HCl to remove any calcium carbonates. HCl then was removed by centrifuging. Clays were dispersed by adding 20 mL of a 5% solution of sodium hexametaphosphate and distilled water in a cylinder. Water then was added to a height of 8 cm. After 70 min the water was poured off, and the cylinder was relled with water, which was left to settle for 60 min and then was poured again. This process was repeated until the suspension was clear. The organic matter in the residue was removed by heating to 500 C in a mufe furnace for 2 h. The samples were transferred into 15-mL centrifuge tubes. Then 3 mL of sodium polytungstate solution calibrated to 2.3 specic gravity was added to each tube. The sample was centrifuged at 115120 g for 10 min. The tubes were removed, and the suspension containing the phytoliths was poured into clean 15-mL tubes. Distilled water was added to the tubes to reduce the specic gravity, and then the tubes were centrifuged at 721 752 g for 5 min. After two such treatments, the clean phytoliths were transferred into beakers, dried, and weighed. Entellan (Merck) or Permount (Fisher) was used to mount a weighed aliquot of 23 mg of residue from each sample. A wide variety of plant tissue and taxa were identied (Table S3). Phytolith counting was conducted at 400 magnication. Single-cell phytolith morphotypes were counted to 300 or more individuals. Because multicell phytoliths are composed of varying numbers of cells, they were tallied separately. Multicell forms were counted up to 100. The number of phytoliths on the slide was calculated using the following algorithm: n phytoliths per slide = n counted=n slide fields counted total n fields on slide: This value was used to derive a comparable unit of quantication, the number of phytoliths per 1 g of sediment. This value was calculated with the following formula: n phytoliths=g = n phytoliths per slide= total amount of sediment mounted mg total phytolith amount mg= total initial sediment mg 1; 000: We used this formula and an Excel spreadsheet (Microsoft) to calculate the number of each phytolith type on each slide and the number of each type per gram (<0.25 mm) of dry sediment.
Results. Control samples taken from outside the cave and in various contexts inside the cave indicate that jigsaw-puzzle phytoliths were rare outside the burial contexts (Table S4). Controls show low levels of all phytoliths and extremely low levels of dicot leaf
1 of 11

phytoliths. This nding veries that the presence of dicot phytoliths is the result of cultural practices associated with burials and not the result of natural processes or contamination. The high correlation between the burials and dicot phytoliths is apparent in the
1. Nadel D, et al. (2008) The Late Natuan at Raqefet Cave: The 2006 excavation season. J Isr Prehist Soc 38:59131. 2. Nadel D, et al. (2012) The 2010 and 2011 seasons of excavation at Raqefet Cave. J Isr Prehist Soc 42:3573. 3. Stoops G (2003) Guidelines for Analysis and Description of Soil and Regolith Thin Sections (Soil Science Society of America, Madison, WI). 4. Courty M-A, Goldberg P, Macphail RI (1989) Soils and Micromorphology in Archaeology. Cambridge Manuals in Archaeology (Cambridge Univ Press, Cambridge, UK). 5. Goldberg P, Macphail RI (2006) Practical and Theoretical Geoarchaeology (Blackwell Publishing, Oxford, UK).

burials of Homo 18 and Homo 19, where a set of well-preserved plant impressions was preserved on a vertical chiseled surface (Fig. S4). This nding implies an association between dicot leaf phytoliths and plant impressions in certain burial contexts.
6. Rosen AM (2010) Natuan plant exploitation: Managing risk and stability in an environment of change. Eurasian Prehistory 7:117 131. 7. Rosen AM, Rivera-Collazo I (2012) Climate change, adaptive cycles, and the persistence of foraging economies during the late Pleistocene/Holocene transition in the Levant. Proc Natl Acad Sci USA 109(10):36403645. 8. Rosen AM (1999) in The Practical Impact of Science on Aegean and Near Eastern Archaeology, eds Pike S, Gitin S, Wiener Laboratory Publication 3 (Archetype, London), pp 86 92.

Fig. S1. Field photographs of stem impressions. All scales are in centimeters. (A) Bones of Homo 31, a 6- to 8-y-old child, directly above impressions of round stems. (BD) Close-up views of rectangular stem impressions in the grave of Homo 25 and Homo 28. Photographs reproduced with permission from E. Gernstein.

Nadel et al. www.pnas.org/cgi/content/short/1302277110

2 of 11

Fig. S2. (A and B) Field photographs of stem impressions in the graves of Homo 25 and Homo 28. The impressions are rectangular in cross-section, as marked by the red line. Photographs reproduced with permission from E. Gernstein. (C ) Photographs of a stem segment and its rectangular cross-section (Salvia judaica). (D) Photographs of a stem segment and its rectangular cross-section (Salvia fruticosa). Both Salvia species currently grow near Raqefet Cave. Photographs in C and D reproduced with permission from M. Eisenberg.

Nadel et al. www.pnas.org/cgi/content/short/1302277110

3 of 11

Fig. S3. A section through selected burials discussed in text; see Fig. 1C for location. The presence of plant impressions on a mud veneer is indicated by the green line at the bottom of two graves. Note stones set on edge on top of Homo 28, Homo 25, and Homo 31 (diagonal hatching). The slab on the chest of Homo 25 is in addition to the one set near its skull (to the west of this section, not presented here; see Fig. 2). The shallow pit for the double burial on the left was dug through a Middle Paleolithic layer, down to bedrock. Homo 27 and Homo 30 are only partially preserved.

Nadel et al. www.pnas.org/cgi/content/short/1302277110

4 of 11

Fig. S4. The double burial of Homo 18 and Homo 19 (see location in Fig. 1C ). (A) Homo 19 during excavation (skull on right). Note the vertically chiseled bedrock surface at the left with foot bones resting on it. This surface was covered by more than 10 plant impressions. Photograph reproduced with permission from E. Gernstein. (B) An opposite view of the chiseled surface, after removal of the skeleton. Photograph reproduced with permission from E. Bartov. (C ) A close-up view of plant impressions found on the vertical chiseled surface. All scales are in centimeters. Photograph reproduced with permission from E. Gernstein.

Nadel et al. www.pnas.org/cgi/content/short/1302277110

5 of 11

Fig. S5. (A) Sample 3, the top of Locus 1: partially fragmented charcoal specimen from a hearth embedded in the clayey anthropogenic deposit; PPL, width of frame 2.4 mm. (B) Sample 4, the top of Locus 1: heterogenic ash-derived deposit encompassing a land snail shell fragment (1) and a yellow chip of bone (2); PPL, width of frame 2.4 mm. (C ) Sample 5, Locus 1: ll of the burial of Homo 9, a loose anthropogenic deposit comprising pulverized small, rounded organo-mineral aggregates; PPL, width of frame 0.98 mm. (D) Sample 6, Locus 1, ll of the burial of Homo 9, a channel lled with comminuted organo-mineral aggregates and an elongated phytolith (1); PPL, width of frame 0.98 mm. (E ) Sample Raq-111, oor of burial Homo 25 and Homo 28: back-scattered image of fractured (arrows), mainly amorphous mass of a crust. (F ) Close-up view of E. Square denotes area measured by EDS, arrows indicate rounded pores, asterisk shows tip of Legend continued on following page

Nadel et al. www.pnas.org/cgi/content/short/1302277110

6 of 11

idiomorphic hydroxyapatite crystal probably grown on parent calcite. (G) Sample Raq-111, oor of burial Homo 25 and Homo 28: Back-scattered image of heterogeneous crust surface dissected with fracture (arrow); note bright area (square 1, phosphate mineral phase) vs. darker area (square 2, alumosilicate phase). (H) EDS graphs of elemental composition in squares 1 and 2 in G. Major peaks of calcium and phosphorus are related to a phosphate phase, whereas silicon, aluminum, potassium, and iron are related to a silicate phase. Parts AB and EH reproduced with permission from The Israel Prehistoric Society.

Fig. S6. Three 14C dates obtained from collagen extracted from skeletons Homo 18 (RTK-6607), Homo 19 (RTK-6480-6540), and Homo 28 (RTK-6638) (following protocol of ref. 1).
1. Yizhaq M, et al. (2005) Quality controlled radiocarbon dating of bones and charcoal from the early Pre-Pottery Neolithic B (PPNB) of Motza (Israel). Radiocarbon 47:193206.

H11

H15

H17

H18

H19

H20H22 H31

C-XVI

C-XLIV

C-XXIII LXXIV

Tufa, controls & basin sed.

50,000 45,000 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5,000 RQ-11-26 RQ-11-4 RQ-11-10 RQ-11-9 RQ-11-12 RQ-11-13 RQ-11-20 RQ-11-21 RQ-11-22 RQ-11-17 RQ-11-18 RQ-11-19 RQ-11-23 RQ-11-24 RQ-11-56 RQ-11-57 RQ-11-27 RQ-11-28 RQ-11-49 RQ-11-50 RA-07-01 RQ-11-29 RQ-11-30 RQ-11-32 RQ-11-1 RA-07-02 RQ-11-47 RQ-11-48 RA-07-04 RA-07-05 RA-07-06 RQ-11-52 RQ-11-53 RQ-11-45 RQ-11-46 0

Fig. S7. The frequencies of single-cell jigsaw-puzzle phytoliths from dicot leaves. Basin sed., basin sediments; C, bedrock mortar; H, Homo.

Nadel et al. www.pnas.org/cgi/content/short/1302277110

7 of 11

Table S1. Data for four Natuan burials (six human skeletons) with plant impressions at Raqefet Cave
14 14

Nadel et al. www.pnas.org/cgi/content/short/1302277110

Skeleton n.m. n.m. Bottom

Posture and orientation Lab no. RTK

14

Age/sex

C age 1 years B.P.

C date for 1 calibrated years B.P. Location of impressions in grave

C date for 2 calibrated years B.P.

No. of impressions >3 cm >30

Type of impressions Square stems, seeds

Homo 1

Unclear details, head to north? 6607 6480, 6540 11,725 125 11,540 120 Average 11,630 87 n.m. 6638 10,320 110 12,40011,980 n.m. 12,55011,720 13,60013,370 13,71013,300 11,405 120 13,38013,160 13,58013,000

Mature female

Homo 18

Combined for both: >10 Vertical bedrock wall Bottom Bottom, side Combined for both: >30 Reeds? unidentied

Homo 19

On side, head to north On side, head to north

Mature female Young adolescent

Homo 25

Adult

Homo 28

On back, head to west; On back, head removed

Adolescent

Salvia, square stems, round stems, seeds Square stems, round stems, wide leaves, seeds n.m. Bottom, side >6 Round stems

Homo 31

On side, head to north

Child (810 y old)

n.m.

n.m., not measured; , standard deviation.

8 of 11

Table S2. Provenance of sediment samples analyzed for phytolith remains

Location Burials RQ-1126 RQ-114 RQ-1110 RQ-119 RQ-1112 RQ-1113 RQ-1120 RQ-1121 RQ-1122 RQ-1117 RQ-1118 RQ-1119 RQ-1123 RQ-1124 RQ-1256 RQ-1257 Bedrock features/ mortars RQ-1127 RQ-1128 RQ-1249 RQ-1250 RA-0701 RQ-1129 RQ-1130 RQ-1132 RQ-111 RA-0702 RQ-1247 RQ-1248 Tufa, basin, and controls RA-0704 RA-0705 RA-0706 RQ-1252 RQ-1253 RQ-1245 RQ-1246 K7 K8 L14a D14c E14a 4 4 5 3 3 2.15/2.27 2.20 0.60/0.70 2.48/2.55 2.70/2.75 Tufa Tufa Tufa Bedrock basin Bedrock basin Control Control Late Natuan matrix Late Natuan matrix Kebaran matrix Loose sediment Loose sediment 40 m northwest of cave 100 m northwest of cave F11d F11d F11d F11d F11d D12a D12b D12b F12 F12 E14b E14b 1 1 1 4650 cm below rim 5053 cm below rim 5355 cm below rim 5556 cm below rim Base, 5556 cm below rim 2.73 2.69 2.65/2.70 5 cm above base Base of feature 2.50 2.50 Mortar C-XVI Mortar C-XVI Mortar C-XVI Mortar C-XVI Mortar C-XVI Mortar C-XLIV Mortar C-XLIV Mortar C-XLIV Mortar C-XXIII Mortar C-XXIII Clay mortar LXXIV Clay mortar LXXIV C14c B12a B12a C12 C12ad C12 D16 D16 D16 D16a D16c D16c D14d/D15c E15c E14b E14b 3 1 1 1 1 1 3 3 3 3 3 3 3 3 3 3 2.06/2.15 2.49 2.49 2.87/2.88 2.85 2.80/2.82 2.38 2.25 2.27 2.20 2.27/2.28 2.20 2.06/2.15 2.16 254/259 243/248 Homo Homo Homo Homo Home Homo Homo Homo Homo Homo Homo Homo Homo Homo Homo Homo 11 15 15 17 17 17 18 18 18 19 19 19 20 22 31 31 Infant burial Near tibia Near tibias Under thorax Under slab B4 Left humerus/radius Under stone 15 at skull and hands Under sacrum Under proximal ulna Under stone 47 at abdomen Under stone 64 at proximal ulna Under stone 37 at upper chest Under skull Under thorax Chest area Above/near hand Sample Grid Locus Depth Feature Description

Sealed North wall, under Homo 9 North rim Hard internal sediment Hard sediment Hard sediment Negative of mortar, fragment 4 Negative of mortar, fragment 6

Nadel et al. www.pnas.org/cgi/content/short/1302277110

9 of 11

Table S3. Counted phytolith morphotypes and their highest level of taxonomic identication
Phytolith morphotype Single-cell phytolith morphotype Psilate long-cell Sinuate long-cell Dendriform/echinate long-cell Elongate verrucate Papillae Acicular hair Hairs Trichome Bulliform Oval Crenate Short-cell bilobe Polylobate Short-cell rondel Short-cell saddle Cones Stoma Achene type Platey Indeterminate echinate type Elongate Tracheid Two-tiered Oblong Elongate dicot wood/bark Block Smooth spheroid Scalloped Jigsaw-puzzle-shape Multicell phytolith morphotype Grass cells Leaf/Stem Indeterminate husk Triticum sp. husk Hordeum husk Cf. Hordeum Avena husk Small seeded grass husk Phragmites Sedge stem/leaf Sedge husk Mesophyll spheroid Stoma sheet Indeterminate dicot wood/bark Polyhedron Cf. Quercus polyhedron Polyhedron hair base Scirpus husk Multitier: square tier Multitier: (Quercus) Jigsaw Taxonomic attribution Grass stem Grass leaves Grass seed husk/inorescence Grasses Grass inorescences Monocots Monocots Grass leaves and inorescences Monocot leaves Monocot stem/leaves Grass leaves Panicoid grass stem/leaves Panicoid grass stems/leaves Pooide grass stem/leaves/oral Chloridoideae grass stem/leaves Cyperaceae stem/leaves Monocot Cyperaceae inorescence Dicot wood, stem and leaves Unknown taxonomical origin Indeterminate long cell/hair category Dicot/monocot Dicot leaves Dicot leaves/stem Dicot leaves/stem Dicot leaves/stem Dicot Dicot Dicot leaves/fruit Monocots leaves/stems Indeterminate grass inorescence Wheat orescence Hordeum sp. barley husk Confer barley inorescence Cf. Avena (oat) inorescence Small seeded grass inorescence of none above Phragmites (reed) leaf and stems Sedge (Cyperaceae) leaf/stem Sedge (Cyperaceae) inorescence Dicot and grass leaves Monocot (commonly found in Phragmites leaves) Dicot wood/bark Dicot leaves Dicot leaves Dicot leaves Scirpus inorescence Dicot leaves Dicot oak (cf. Quercus) Dicot leaves/fruit

Nadel et al. www.pnas.org/cgi/content/short/1302277110

10 of 11

Table S4. Density (number per gram of sediment) of select phytolith morphotypes organized by provenance
Provenance Burials RQ-1126 RO-114 RQ-1110 RQ-119 RQ-1112 RQ-1113 RQ-1117 RQ-1119 RQ-1124 RQ-1118 RQ-1120 RQ-1121 RQ-1122 RQ-1123 RQ-1256 RQ-1257 Total % of morphotype in grouping Average Bedrock features RQ-111 RQ-1127 RQ-1128 RQ-1249 RQ-1250 RQ-1133 RQ-1129 RQ-1130 RQ-111 RQ-1134 RQ-1247 RQ-1248 Total % of morphotype in grouping Average Tufa and controls RQ-1252 RQ-1253 RQ-1136 RQ-1137 RQ-1138 RQ-1245 RQ-1246 Total % of morphotype in grouping Average 42,160 56,558 11,489 11,412 33,718 42,495 99,932 179,066 107,613 43,324 21,337 97,504 58,343 113,231 52,115 8,893 979,190 61,199 2,710 24,550 154,612 286,040 305,843 333,281 30,884 16,447 2,710 1,917 65,676 85,008 1,309,678 109,104 111,699 65,374 7,073 30,464 10,378 22,228 32,233 279,449 39,921 113 0 0 0 0 0 1,579 2,283 1,517 1,636 464 9,087 3,286 1,134 756 881 22,736 52.7 1,421 6 183 1,151 2,732 4,386 7,281 366 180 6 53 464 665 17,473 40.5 1,456 1,479 954 23 217 167 100 32 2,972 6.9 424 562 0 7 0 12 21 4,372 2,344 2,743 894 2,398 26,823 13,988 5,599 965 0 60,728 91.1 3,795 170 2 73 0 1,384 0 0 14 170 0 347 1,750 3,901 5.9 3,910 1,172 568 0 0 254 0 0 1,994 2.6 285 Sample Total phytoliths Multicell phytoliths Jigsaw puzzle phytoliths

Nadel et al. www.pnas.org/cgi/content/short/1302277110

11 of 11