This Article Appeared in a Journal Published by Elsevier. the Attached

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy

Journal of Human Evolution 60 (2011) 618e636

Contents lists available at ScienceDirect

Journal of Human Evolution

journal homepage: www.elsevier.com/locate/jhevol

On the evolution of diet and landscape during the Upper Paleolithic through Mesolithic at Franchthi Cave (Peloponnese, Greece)

Mary C. Stiner a,*, Natalie D. Munro b a School of Anthropology, P.O. Box 210030, University of Arizona, Tucson, AZ 85721-0030, USA b Department of Anthropology, Unit 2176, 354 Mansﬁeld Rd., University of Connecticut, Storrs, CT 06269, USA article info abstract

Article history: Franchthi Cave in southern Greece preserves one of the most remarkable records of socioeconomic change Received 1 July 2010 of the Late Pleistocene through early Holocene. Located on the southern end of the Argolid Peninsula, the Accepted 17 December 2010 area around the site was greatly affected by climate variation and marine transgression. This study examines the complex interplay of site formation processes (material deposition rates), climate-driven Keywords: landscape change, and human hunting systems during the Upper Paleolithic through Mesolithic at Zooarchaeology Franchthi Cave based on the H1B faunal series. Building on earlier work, we establish the full spectrum of Vertebrate taphonomy the meat diet using taphonomic evidence, and we analyze these data for trends in socioeconomic reor- Site formation processes “ ” Diet breadth ganization. Foraging patterns during the Aurignacian and Gravettoid occupations at Franchthi were fi Prey choice model terrestrial and already rather diversi ed in comparison to Middle Paleolithic diets in southern Greece. Marine foraging Hunting shifted abruptly to a mixed marineeterrestrial pattern during the Final Paleolithic, and fishing Hunting activities intensified though the Mesolithic. The zooarchaeological data indicate two consecutive trends of Fishing increasing dietary breadth, the first within an exclusively terrestrial context, and the second as marine habitats came into use through the end of the Mesolithic. The intensity of the human occupations at this site increased in tandem with intensified use of animal and plants. Comparison to the inland site of Klissoura Cave 1 indicates that the trend toward broader diets was regional as well as local. Ó 2010 Elsevier Ltd. All rights reserved.

Introduction groups from subsistence risk (e.g., Davis, 1982; Belfer-Cohen, 1991; Tchernov, 1992, 1993, 1994; Stiner et al., 2000; Stiner, 2001; Munro, Franchthi Cave on the Argolid Peninsula of the Peloponnese was 2004, 2009). Comparatively little is known about human subsis- visited by humans over a span of more than 30,000 years. The tence in southern Greece during the Late Pleistocene and earliest massive archaeological sequence in this cave includes two of the Holocene. The situation is improving with recent excavation projects most important cultural transitions in prehistorydthe reorgani- on the Peloponnese (e.g., Koumouzelis et al.,1996, 2001, 2004; Tomek zation of foraging societies associated with social and economic and Bochenski, 2002; Panagopoulou et al., 2002e2004; Starkovich, intensification during the Late Pleistocene (Paleolithic through 2009) and revisiting older collections. Mesolithic), and the shift from a foraging to a farmingeherding The Franchthi Cave sequence presents a fascinating range of (Neolithic) way of life. subsistence and environmental extremes through time. Because This study examines the complex interplay of landscape change, the small Argolid Peninsula projects into the Aegean Sea, the material deposition rates, and the evolution of local human hunting locality was directly affected by marine transgression towards the systems at Franchthi Cave from the Upper Paleolithic through end of the Pleistocene. With global warming, shorelines were Mesolithic. In the eastern Mediterranean Basin, this interval saw reconfigured, water tables rose, and habitat structures and biotic significant expansions of human populations, greater site perma- diversity altered (e.g., van Andel and Lianos, 1983). Early reports on nence, and increasingly complex cultural mechanisms for insulating the Franchthi faunas by Payne (1975, 1982) suggested that human foraging strategies underwent a dynamic evolution as well. Knowledge of Late PleistoceneeHolocene shoreline dynamics in the Franchthi vicinity has led Jacobsen (1976) and others (e.g., * Corresponding author. Payne, 1975; Lambeck, 1996) to ask whether the appearance of the E-mail addresses: [email protected] (M.C. Stiner), natalie.munro@ uconn.edu (N.D. Munro). first marine resources in the archaeological sequence at about

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 619

12e10 ka BP, and the exploitation of larger deepwater fish by about (Aurignacian, “Gravettoid” phases), “Epigravettian,” Final Paleo- 9.2 ka BP, were explained foremost by coastal evolution rather than lithic, and multiple Mesolithic and Neolithic phases (Jacobsen, human economic evolution. Unfortunately, no further publications 1981; Perlès, 1987; Farrand, 2000). have appeared on the faunal remains from Franchthi Cave, with the The excavated deposits in the cave span roughly 39,000 to 5000 notable exceptions of the marine mollusks studied by Shackleton years ago. An additional 5 m or more of unexcavated sediments (1988) and fish from the FAS trench studied by Rose (1995). (Stratum P and below) lie below the water table, equivalent to This paper presents the first results from a larger, on-going modern sea level, and could not be excavated for this reason (Perlès, study of the Franchthi Cave faunas. Using taphonomic criteria, we 1999). The transitions between the major cultural phases within identify the full range of prey animals in the diet. Our findings the Paleolithic deposits are relatively clear-cut and generally follow confirm some important trends in taxonomic representation first the lithostratigraphic divisions. Cultural phase changes are ill- identified by Payne (1975, 1982), but the taphonomic observations de fined in the Mesolithic by contrast, probably because of the clarify the true boundaries of the prey spectrum. We then take up strongly anthropogenic character of these deposits (Farrand, 2000; Jacobsen’s challenge through a detailed analysis of the trends in see cross-sections in Jacobsen and Farrand, 1987), high degree of hunting economics from the Upper Paleolithic through late Meso- lateral variation within layers, and low frequencies of diagnostic lithic periods. We want to know how humans responded to well- artifacts (C. Perlès, pers. comm.). For the Mesolithic, we follow documented changes in the local environment, and which of these Perlès’ (n.d.) recent revisions of the cultural phasing and its relation responses represented temporary adjustments to local conditions to the lithostratigraphy (see Table 1). as opposed to more fundamental reorganization(s) of technology The earliest dated event in the Franchthi stratigraphic sequence and social contracts. The null model in this study therefore overall is a wind-blown volcanic ash (Campanian tephra) repre- proposes that foragers simply adjusted to known changes in habitat sented by Stratum Q. Originating from the Naples area of Italy, this structure, without substantive changes in the energetic founda- tephra is dated to 39.28 0.11 ka by 40Ar/39Ar (De Vivo et al., 2001). tions of the economy. Alternatively, we hypothesize that prey A depositional hiatus of perhaps 2000e2500 years may separate choice was evolving in response to impoverishment of terrestrial the ashfall and the first Aurignacian occupations (Stratum R; food supplies and strong selection to intensify foraging efforts on Farrand, 2000), but this is not certain. Another hiatus in occupation land and sea. In the latter situation, following the predictions of certainly separates the Aurignacian from the “Gravettoid” occupa- a prey choice model (Stephens and Krebs, 1986), there should be tions of Stratum S, which bear general similarities to Gravettian evidence of increasing use of costly resources alongside the more industries elsewhere in Europe but are sufficiently distinctive to economical (high-return) types. merit caution in their classification. These “Gravettoid” occupations continued until about 20 ka based on evidence from multiple Background to Franchthi Cave trenches. The cave was used little if at all from 20 ka until about 15 ka. Farrand (2000) identified three major pulses of occupation Franchthi Cave is a very large, elongate chamber in a Lower after 15 ka, punctuated by brief hiatuses: the Final Paleolithic of Cretaceous limestone headland. It sits on the southwestern shore of Stratum T (T1e3); the Mesolithic of Stratum W (W1e3) that coin- the Argolid Peninsula (Fig. 1) near the modern town of Kiladha. At cides with a major increase in plant macrofossils (Hansen, 1991); the top of the same peninsula, ca. 50e60 km from Franchthi, are the and the Neolithic of Strata X and Y, which are not considered in this cave sites of the Klissoura Gorge (Koumouzelis et al., 1996, 2001, presentation. 2004; Karkanas et al., 2004), whose rich Paleolithic and Meso- The function of the site and/or the intensity of the occupations lithic deposits only partly overlap the cultural chronology of are believed to have shifted with time, with generally light use prior Franchthi (see Table 1). to the Last Glacial Maximum (LGM) and much heavier use after it. Franchthi Cave was excavated in the 1960s and 1970s by a large According to Farrand (2000), geogenic processes dominated the team of archaeologists and geologists from Indiana University and formation of the Paleolithic sedimentary layers, because cultural the University of Pennsylvania (USA) and the American School of inputs were low in comparison to sedimentary inputs. Anthropo- Classical Studies at Athens (Jacobsen, 2000). The excavations yiel- genic processes strongly influenced the formation of the later ded an enormous collection of well-preserved bone specimens, layers, particularly during the Mesolithic. Perlès (1999) has shells, artifacts, and other materials dating to the Upper Paleolithic proposed from multiple lines of evidence that the site became

Figure 1. Locations of Franchthi Cave (F) and Klissoura Cave (K) on the Argolid Peninsula of the Peloponnese, Greece. Left panel shows modern shorelines. Right panel shows estimated shorelines of the LGM and distribution of paleolakes (black ﬁll) in the interior during this period; adapted from Petit-Maire et al. (2005). Author's personal copy

620 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636

Table 1 Summary of dates, stratigraphy, and occupation phases for the H1B trench in Franchthi Cave.

Lithostratigraphic Lithostratigraphic Calibrated Occupation Occupation phase Cultural layer unit ranges radiocarbon phases unit ranges attribution years BP (general range) X2, Y, Z Mixed in H1B 8,500 and younger e Mixed in H1B Neolithic, mixed X1 Mixed in H1B Undated 9 Mixed in H1B Final Mesolithic/Neolithic, mixed (X1) W3 106e116 10,000e9,500 8 106e115 Upper Mesolithic W3 117e127 10,900e10,000 7/8 117e120 Mixed Mesolithic phases W1e2 128e147 10,900e10,000 7 121e147 Lower Mesolithic (intense occupations; snail middens; several graves) V 148e151 Undated 6.2 Cross-cutting Mixed Paleolithic/Mesolithic levels in H1B U 151/152e154 11,700e12,500 6.1 151e153/154 Final Paleolithic (small microliths with microburins and geometrics; few snails) T 155e163 13,300e12,700 5 155e160/161 Epigravettian (double-backed bladelets, microburins; snail midden) S2 164e172 14,000e15,000 4 162e171/172 Epigravettian (with microburins and La Mouillah points) S1 173e180 Undated 3 173e180 “Gravettoid” (single-backed points and backed bladelets) R 181e212 26,000e27,000 2 181e207 “Gravettoid” (double-backed points and backed bladelets) R 181e212 Undated 1.2 208e212 Aurignacian (carenated end-scrapers, twisted bladelets) Q (tephra) 213 39,000 1.1 213 Aurignacian (carenated end-scrapers, twisted bladelets) P 214e215þ >Age of tephra 0 214e215þ Sparse, non-diagnostic industry

Sources: Stratigraphic information and most dates are summarized from Farrand (2000), Perlès (1987, 1999, 2004), and Vitelli (1993, 1999). Volcanic ash (Campanian tephra) source is dated to 39.28 0.11 ka by 40Ar/39Ar (De Vivo et al., 2001). Cultural attributions and occupation phases are as defined by Perlès (n.d.) and in Perlès and Vanhaeren (2010); note that cultural terms derive from western Europe and that the Balkans industries are sufficiently distinctive from these and Near Eastern industries to require qualification (e.g., “Gravettoid” rather than Gravettian). Several hiatuses exist in the stratigraphic series, with a particularly long one between about 20e15 ka Cal. BP. More continuous occupations occurred thereafter, separated by several brief hiatuses. The occupation phases refer to a general framework for all trenches and are based on a composite of all artifactual, geological, and earlier zooarchaeological observations. Unit ranges for lithostrata and for occupation phases are not in perfect agreement in H1B, especially for the Mesolithic. a major, centralized residential hub during the Mesolithic occupa- (Alkanna, Lithospermum, and Anchusa)(Hansen, 1991; summarized tions with complete domestic groups in residence. Although the in Perlès, 1999), and greater land exposure. Under these conditions, Upper Paleolithic occupations clearly were small and probably erosion built-up the coastal land surfaces and contributed to the short-lived in comparison to those of the Mesolithic, it is not clear westward migration of the Franchthi embayment (Fig. 2; van Andel from the technological evidence whether the Upper Paleolithic and Shackleton, 1982; van Andel and Sutton, 1987). The Pleistocene occupations represent mainly brief hunting camps or small-scale Peloponnesian landscape was much more complex than the residential camps. Given that variations in site occupation intensity modern one, with large inland freshwater lakes and higher habitat can be the product of site function, numbers of people in residence, diversity in the surrounding sea (Fig. 1). Lower moisture during responses to territorial circumscription, or all of these factors, the glacial conditions reduced forests on the peninsula, and the interpretations drawn from detailed geological and technological enlarged coastal plains provided graze for European wild ass (Equus analyses must be checked against the subsistence data. Specifically, the diversity of subsistence activities represented, including the full scope of large and small game hunting and fishing, is important to interpretations about site function.

Previous environmental studies

The environmental history of the Franchthi Cave vicinity has been studied extensively. The cave is situated in a typical Medi- terranean-type environment at the junction of rolling limestone hills and coastal plain. Higher, craggy hills rise to the north. The modern shoreline lies only 11 m below the cave entrance, but the relationship between the site and the sea ﬂuctuated dramatically in the past, owing to a combination of sea eustasy, glacio-hydro isostasy, and minor local tectonic deformations (Jacobsen, 1969; van Andel and Lianos, 1983; van Andel and Sutton, 1987; Jackson, 1994; Lambeck, 1996). The LGM at approximately 20,000 years ago corresponds to the lowest shorelines (ca. 120 m) for the Argolid Peninsula and elsewhere (Fig. 2). The cave would have been about 7 km from the sea during the LGM (Jacobsen, 1969; van Andel and Sutton, 1987), but, in fact, this particular interval is poorly represented, if at all, in the Franchthi cultural sequence (Farrand, 2000). From ca. 70 ka to just before the LGM, sea levels varied between 50 m and 80 m (Fig. 2), and this was again the case Figure 2. Eastward shoreline migration of the embayment in front of Franchthi Cave from ca. 14 ka to 2.5 ka (Lambeck, 1996). from 18e12 ka and 10.5e8 ka, based on reconstructions by van Andel and Lianos The Upper Paleolithic occupations in Franchthi Cave coincided (1983) and Lambeck (1996). (F) Franchthi Cave; plant symbols indicate marshy with a generally cool dry climate regime, steppe vegetation areas. Author's personal copy

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 621 hydruntinus), red deer (Cervus elaphus), and aurochs (Bos primigenius). The post-LGM occupations at Franchthi Cave coincided with a milder, more humid climate and Mediterranean garrigue vegetation that included Pistacia, Prunus, Pyrus, wild cereals, and wild legumes (Hansen,1991; Perlès, 1999). The geography and ecology of the vicinity changed considerably with the post-LGM warming trend, slowly at first and accelerating after 14 ka. The Cycladian mega-islands broke into smaller islands by 12 ka and approached their modern configurations by about 10 ka. It is interesting in light of these developments that obsidian from the island of Melos occurs in Franchthi Cave only after about 11 ka BP (Perlès, 1987), implying that sea-worthy boats were required for its exploitation. A submerged Neolithic site in front of the cave indicates that the Aegean shoreline was still below the modern level between 7610 150 ka and 6220 130 ka BP (Jacobsen and Farrand, 1987; van Andel and Sutton, 1987; Gifford, 1990). Payne’s preliminary studies of the Franchthi faunas (Payne, 1975, 1982) documented a shift from predominantly open-land grazers during the Gravettoid occupations to woodland browser/ grazers among the ungulates and the notable addition of fish remains in the Late Paleolithic and Mesolithic, including a startling Figure 3. Plan of Franchthi Cave and locations of various excavation trenches inside and outside the cave. rise in large fish exploitation in the Upper Mesolithic (see also Rose, 1995). There are, however, no complete vertebrate faunal data sets to parallel the detailed technological (Perlès, 1987, 1990, 2004; Vitelli, 1993, 1999; Stroulia, 2010), geological (van Andel and widespread any one human occupation would have been within Sutton, 1987; Farrand, 2000), mollusk (Shackleton, 1988), and the cavern, and the horizontal representativeness of each trench is botanical studies (Hansen, 1991) that have been conducted at beyond our ability to judge. General stratigraphic connections Franchthi Cave. A study of the terrestrial mollusks is in press could be drawn among adjoining trenches (see Farrand, 2000), but (Whitney-Desautels, in press) and shows that the great majority of horizontal discontinuities at a finer scale abound and suggest that the land snail remains represent human food refuse. Unpublished humans concentrated their activities over smaller areas within the data on these remains are included in the presentation below. cave. Suffice it to say, the areas excavated within Franchthi Cave are A final note on the background to the Francthhi Cave excavation similar to the areas uncovered in other cave sites of similar age. project concerns the premise that each investigation of archaeo- A total of 17,975 identified vertebrate specimens (NISP) were logical or geological data and materials should develop its own recorded from H1B, 15,366 of which could be linked to human free-standing sequence of change (Jacobsen and Farrand, 1987). economic activities in the cave (SOM Appendix 1). The Final Later, a master chronology for each trench and a general synthesis Mesolithic and Neolithic materials that cap the H1B sequence are for the site was to be developed from these many independent stratigraphically mixed, and therefore could not be considered in threads (Perlès, n.d.). Our faunal study comes late in the history of this study. Most of the vertebrate remains from H1B are clearly the project and much has already been accomplished on other anthropogenic in origin, with few if any traces of damage from non- topics and materials from Franchthi Cave. We nonetheless post- human biological or geological agents (see below). There is pone the assignment of cultural labels to the faunal data until late a significant microfaunal component as well (SOM Appendix 2). in the presentation, so that the faunal results can play a clearer role Thorough recovery techniques were practiced during the exca- in fine-tuning the story of social and economic change at Franchthi vation of units 106e213 of H1B, including screening of all sedi- Cave. ments through 2.8 mm fine mesh (Diamant, 1979), which allowed small animal remains above this size threshold to be retrieved as Materials and methods consistently as those of large animals. The sediments were excavated in a series of shallow units within broader lithostratigraphic Trench H1B is the deepest trench of the HeH1 trench complex, layers (Jacobsen and Farrand, 1987; Farrand, 2000). The units vary with cultural deposits beginning about 3.3 m below the modern in thickness, averaging 4e5 cm deep, and cultural features were surface and reaching a depth of 9.71 m (Farrand, 2000). H1B was delineated within them. Faunal remains from a total of 107 vertical selected for this study because it contains the most complete, units were examined by this study, and more than 70% of these stratigraphically intact pre-Neolithic cultural sequence. In fact, four contained sizeable quantities of specimens. The total sediment excavation trenches in the cave yielded deep, intact deposits of volume excavated from the Mesolithic of the H1B trench is 5.07 m3, artifacts and faunal remainsdTrenches FAS, FAN, H1A, and H1B and 9.15 m3 from the Paleolithic (Farrand, 2000). (Fig. 3)dbut each is characterized by rich zones and gaps. The The distributions of burned specimens among fine vertical units Upper Paleolithic and most of the Mesolithic are best represented are used to evaluate the scale of potential disturbances to the in H1B and somewhat less so in H1A. Late Mesolithic to initial sediments. Experimental evidence shows that material buried Neolithic horizons are better represented in FAN, and the Final down to 5 cm below the ground surface in limestone cave sedi- Paleolithic through initial Neolithic in FAS. Trenches were also ments can be partly or wholly carbonized by a small surface fire placed in the “Paralia” area outside the cave, which contained whose center approaches 900C(Stiner et al., 1995). Importantly, ceramic Neolithic deposits (Trenches BeE, Fig. 3). this effect has a shallow reach into the sediments. Thus, while both The interior of the cave is roughly 1350 m2 (Farrand, 2000), and cultural and non-cultural materials may be burned in this way, each trench covers a very small percentage of this area; the area of hearth rebuilding activities often show considerable fidelity to Trench H1B, for example, is about 2.48 m2. It is not known how location inside caves (Meignen et al., 2001, 2007; Karkanas et al., Author's personal copy

622 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636

2004). Evidence for spatially discrete concentrations of burned broad diet. Because foragers can choose only from an extant array bone, indicative of hearth areas, therefore is taken as proof that the of foods in response to encounter and energy return rates, it is the deposits have maintained much of their spatial integrity. Post- relative evenness in foragers’ emphasis on high- and low-cost types depositional disturbances of the sediments instead would tend to that is of greatest interest in studies of diet breadth and economic blend or homogenize the distributions of burned materials both evolution. vertically and laterally. A single plane of prey categorization is required for the We then turn our attention to differentiating the prey of humans diversity analysis. In this study, only the specimens for which from other sources of bone in the cave. Burning damage (presence, genus (or family for birds and fish) could be identified are used in location on specimen, and intensity of damage; following Stiner the diversity analyses. This requirement of the data reduces the et al., 1995) was particularly important for this purpose. We available sample sizes for many of the assemblages. Although the assume that frequent accidental burning should indicate human- Inverse of Simpson’s Index is designed to control the influence of associated faunal remains, whereas burning of non-cultural mate- sample size through the application of a square-root function, rials should be less common on average. Also important to the a minimum assemblage size threshold must be set. The goal is to taphonomic analyses were observations of carnivore, rodent and identify a minimum assemblage size threshold without sacri- raptor damage (gnawing marks, digestive etching and rounding, ficing vertical resolution in the H1B trench unnecessarily. We punctures), tool marks ranging from fine incisions and scraping to experimented with lowest sample size (NISP) thresholds incre- hack marks and cone fractures, and general patterns of bone mentally to see how much prey diversity measurements were breakage that included the condition of break edges, contours, and affected. This procedure used regression statistics (Pearson’s r)to orientations (following Stiner, 1994, 2005; Fisher, 1995). examine the strength of the autocorrelation, slope, and intercept between total assemblage diversity and small game diversity, Terminology and quantification respectively (Table 2a). Though a minimum threshold of NISP > 99 would seem to be ideal from a statistical point of view, The vertebrate remains were identified with the help of refer- none of the outcomes that use thresholds down to NISP > 29 ence collections of the Wiener Laboratory of the American School of differ very much from one another. Another indication that Classical Studies in Athens (ASCSA) and electronic virtual image NISP > 29 is acceptable is the consistently independent relation libraries developed by the authors from personal collections and between the diversity patterns in the small game versus ungulate those of the Arizona State Museum in Tucson, Arizona. Large game fractions of the faunal series (Table 2b; and SOM Appendix 3). types in this study are ungulates. Small game types are defined as The major and minor shifts in the H1B series were observed to small-bodied animals of any sort, including fish, because size hold fast as the minimum sample size threshold was allowed to similarity is the basis from which differences in capture costs are vary between 30e100 NISP. Following a fine-grained analysis of ranked. The accounts of taxonomic abundance are based on NISP, or the faunas, we shift to a coarser view of the sequence in order to numbers of identified specimens (Grayson, 1984). In this study, consider the question of variation in occupation intensity. This NISP may include vertebrate specimens identified to species, genus, part of the study compares the accumulation rates of sediment, family, or order, as well as specimens identified only to a very bones, and land snails by lithostratigraphic layer, controlling for general taxon (e.g., carnivore, bony fish, chelonian, ungulate) variation in sediment volume and the overall duration of each provided that the body part could also be determined. Elements layer. Trends in prey diversity are also examined at this scale, found connected in anatomical position are recorded as separate using fewer but much larger faunal assemblages. Sediment elements, noting that they are connected. volume estimates (liters) and land snail counts (MNI) are The zooarchaeological presentation begins with simple abun- provided by Whitney-Desautels (in press), and the estimated dance comparisons (raw NISP and percentages of total NISP) for the intervals for lithostrata intervals are from Farrand (2000,pers. ungulates, carnivores, and small animals. Once the full vertebrate comm). The accumulation rate for faunal remains in a lithos- prey spectrum of the humans has been established from tapho- tratigraphic layer is calculated as follows: (specimen count / nomic results, the anthropogenic faunas are subjected to a series of sediment volume in liters) divided by the estimated interval over diversity analyses at fine and coarse scales in the stratigraphic which the layer formed. Layer interval or “duration” is an sequence. Published and some unpublished data on edible admittedly rough estimate of time elapsed, but it is useful mollusks from Franchthi Cave are also integrated into the presen- tations of prey abundance. Mollusk MNI (minimum number of individuals) is the ideal quantitative unit for comparisons to vertebrate NISP on account of the exceptionally high shell to soft Table 2 tissue weights of marine mollusks and great potential for species- Comparison of the effect of sample size (NISP) on assemblage diversity by unit using or genus-level identifications of small shell fragments (see Stiner, regression statistics.

1999, 2001). The land snail data from Whitney-Desautels (in 2 ’ fi Minimum N qualifying r Intercept Slope p press) are MNI counts, but Shackleton s (1988) shell sh data are sample size assemblages NISP values, which must suffice. a. Total assemblage evenness against small game evenness with minimum assemblage size raised progressively. Diversity analyses by fine vertical unit and lithostratum NISP > 29 78 0.614 0.624 0.369 0.001 NISP > 34 69 0.608 0.648 0.358 0.001 > The human-collected faunas are examined for trends in diet NISP 49 52 0.587 0.753 0.332 0.001 > ’ NISP 99 26 0.781 0.522 0.375 0.001 breadth (a.k.a.P prey choice), based on the Inverse of Simpson s r 2 ’ Index, or 1/ ( i) . The Inverse of Simpson s Index measures both b. Small game evenness against ungulate evenness with minimum assemblage size the richness (N-types) and evenness (proportionality) of prey types raised progressively. in the diet (Simpson, 1949; Levins, 1968). The index is especially NISP > 29 78 0.003 1.949 0.054 0.645 sensitive to changes in the evenness of prey choice, or the NISP > 34 69 0.006 1.906 0.08 0.542 NISP > 49 52 0.001 2.081 0.036 0.829 proportional (r) representation of different types (i). A low index NISP > 99 26 0.03 1.991 0.171 0.401 value indicates a narrow diet and a high value indicates a very Author's personal copy

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 623 because layer accumulation times vary by more than an order of correlation to preservation of relatively discrete hearth areas and magnitude through the stratigraphic series. limited post-depositional scattering of bone debris. Prey versus non-prey Caves as large and complex as Franchthi Results may attract a wide range of animal residents, which through natural mortality contribute non-cultural skeletal material to the Layer formation and vertebrate taphonomy deposits. Snakes, small raptors, owls, rock doves (pigeons), swifts, bats, small rodents, and snakes are common residents in Mediter- Some degree of post-depositional disturbance is typical in ranean caves, and all of these animals are represented in the Paleolithic cave sites (e.g., Goldberg and Bar-Yosef, 1998; Goldberg Franchthi faunas to some extent. Small items are especially prone to and Sherwood, 2006). Consideration of the scale of this phenom- accidental burning, because humans are less likely to remove them enon in Franchthi Cave is important on at least two counts: (1) from activity areas (Schiffer, 1983). Given the prevalence of burning assessing the intactness of the faunal series for detecting major diet damage in Franchthi Cave, particularly in the upper layers, shifts, and (2) establishing a basis for linking bones to human accidental “over-burning” of pre-existing debris may have been activities in the cave from burning patterns. Because we are con- a common occurrence. Even so, the remains of human prey cerned with general diet trends, macro-scale disturbances would generally should be burned more often than the basal rate of be problematic, but fine-scale disturbances (as documented by accidental burning in the faunas overall. The remains of resident Farrand, 2000) can be tolerated as long as the overall integrity of (i.e., non-prey) species should be burned at or below this the sequence is maintained. frequency baseline. To begin, the amount of bone (NISP) is fairly consistent among The burning damage results are presented by taxon in Table 3 units in the H1B stratigraphic sequence (Fig. 4), with the exception and as broader animal groups in Table 4, sub-divided between of higher quantities in units 161 through 155 (Final Paleolithic, in Strata U and V in recognition of the much higher burning rate in Stratum T). The frequencies of burning damage on bones show Stratum V and above. Table 4 shows that the distribution of burning a very different pattern in Fig. 4, increasing greatly above unit 150 damage across taxa is far from random. While some burning in Stratum V, and especially in the Mesolithic occupations of damage occurs on the bones of virtually all species, as expected, the Stratum W (units 147 and above; see Table 1). Overall, burning remains of some kinds of animals are burned substantially more damage is more than twice as common in the upper half of the often than others. Even taking into account the generally higher sequence (averaging 24% of vertebrate NISP) than in the Paleolithic rates of burning in the Mesolithic, it seems clear that the remains of units below (10%). The higher rates of bone burning in the Meso- ungulates, small carnivores, hares and hedgehogs (“small lithic units almost certainly stem from rapid hearth rebuilding mammal”), fish, and reptiles (chelonians) associate consistently (Farrand, 2000). The variation in burning rates among fine units is with human fire-tending activities in the cave. Tool marks are not correlated with variation in total vertebrate NISP (Pearson’s present on some of these remains, though never at high frequencies r ¼ 0.001, p ¼ 1, n assemblages ¼ 107). We attribute this lack of (Table 5). Gnawing damage from carnivores and rodents is extremely rare or absent. The carnivore remainsdall from small species with desirable furdare burned at exceptionally high rates, clearly linking them to human activities in the cave. Small- and medium-sized rodents, bats, small birds, snakes, and lizards instead appear to have been intrusive. Consistently low rates of burning (generally below 6% of NISP) are found for these microfaunal remains, both above and below the Final Paleo- lithiceMesolithic boundary (see Table 4). Snake bones are burned somewhat more often than expected (ca. 10%; Table 3), but in any case, snake remains are too few to have been of much economic significance at Franchthi Cave. The rates of burning on bird bones are more difficult to interpret, as they are consistently lower than for fish, turtles and tortoises, and hares and hedgehogs (Table 4). The largest ungulate bones are less often burned as well (Table 3), but the unwieldy size of these bones would encourage consumers to keep hearth-based activity surfaces free of such debris (Schiffer, 1972, 1983; Binford, 1978). Partridges, falconiform birds (diurnal raptors), and very large birds such as bustard are known to have been exploited extensively by humans at other Upper Paleolithic and Epipaleolithic sites in Greece and the Levant (i.e., at Klissoura 1 [Tomek and Bochenski, 2002; Bochenski and Tomek, in press]; partridges at Hayonim Cave [Munro, 2004]; and raptors at Meged Rockshelter [Kuhn et al., 2004]). At Franchthi, the bones of these kinds of birds are burned only at low to moderate rates (usually 7e10%), in contrast to fairly high rates of burning in Natufian faunas from Hayonim Cave, Israel (Munro, 2004). High fragmentation and transverse breaks on the Franchthi partridge bones nonetheless suggest a pattern of human interventiondonly 14% of the specimens are complete or nearly complete elements. Bird remains collected by owls or resulting from natural deaths in other contexts tend to be more complete Figure 4. Total vertebrate NISP and the frequency of burning damage (% NISP) by excavation unit in the H1B trench. Gray lines and terms indicate lithostratigraphic (Andrews, 1990; Kusmer, 1990; Hockett, 1996; Saavedra and layers following Farrand (2000). Simonetti, 1998; Bochenski, 2005). Some diurnal raptors are Author's personal copy

624 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636

Table 3 Total NISP and burning rates for selected taxa from the H1B faunal sequence, sub-divided at the depth where burning damage increases signiﬁcantly in the stratigraphic sequence.

Common name Latin name Units 213e152 (Strata QeU) Units 213e152 (Strata QeU) Units 151e106 (Strata VeX1) Units 151e106 (Strata VeX1)

NISP % Burned NISP % Burned UNGULATES Wild goat Capra sp. 97 11.3 7 0 Wild boar Sus scrofa 196 18.4 187 31 Aurochs (wild cattle) Bos primigenius 80 3.8 17 23.5 Red deer Cervus elaphus 861 6.2 630 28.3 European wild ass Equus hydruntinus 471 6.4 6 16.7 Very large ungulate Ungulata 23 8.7 14 14.3 Large ungulate Ungulata 968 7.5 608 39.8 Medium ungulate Ungulata 617 12 175 51.4

CARNIVORA Wild cat Felis silvestris 28 17.9 4 25 Lynx Felis [Lynx] lynx 52 36.5 24 16.7 Red fox Vulpes vulpes 462 12.1 639 33.8 Mustelids Mustelidae 13 38.5 35 20 Small carnivore Carnivora 149 10.7 107 31.8

SMALL MAMMALS European hare Lepus europaeus 1978 9 606 32.5 Hedgehog Erinaceus europaeus 25 18 104 19.2 Small mammala Mammalia 374 11.5 255 31

BIRDS Great bustard Otis tarda 16 0 2 0 Partridge Alectoris sp. 474 7.6 1 0 Bird of prey Falconiformes 16 6.3 3 0 Very large bird Aves 49 8.2 2 0 Large bird Aves 57 10.5 8 0 Medium bird Aves 756 5.4 55 3.6

FISH AND REPTILES Sea Breams Sparid 173 16.8 67 13.4 Blueﬁn tuna Thunnus thynnys 0 0 303 8.3 Barracuda Sphyraena sphyraena 0 0 103 2.9 Conger Eel Conger conger 0 0 6 16.7 Fish Indet. Pisces 1455 18.4 639 7.7 Fresh water turtle Emys/Maurymys 330 19.7 18 27.8 Land tortoise Testudo sp. 122 13.9 11 18.2

MICROFAUNA Large rodent Rodentia 801 5.5 27 37 Medium rodent Rodentia 104 2.9 10 30 Small rodent Rodentia 118 10.2 656 2.7 Bats Chiroptera 31 3.2 36 8.3 Snakes Colubridae 32 3.1 37 16.2 Pigeon Columba livia 166 4.8 24 4.2 Perching birds Passeriformes 199 6 50 4 Small bird Aves 200 5.5 49 0 Tiny bird Aves 33 9.1 3 0 11526 5528

a Notes: In this case, small mammal refers to remains in the size range of hares to hedgehogs only. The Mesolithic layers consistently associate with high burning rates. However, intact Mesolithic deposits begin with unit 147, while notable increases in burning frequencies begin at about unit 151. Hence, the latter division is used in this table.

Table 5 Frequency of cut marks and carnivore damage traces on various vertebrate Table 4 subsamples. Evidence of carnivore damage includes digestive etching, scoring, Average rates of burning by taxonomic group above and below the Final Paleo- crenellation, and puncture marks. lithiceMesolithic interface in the H1B trench. Taxon NISP % Tool marked % Carnivore damaged Taxon Units 213e152 Units 151e106 (Strata QeU) (Strata VeX1) Capra sp. 104 4.8 0.0 Sus scrofa 383 0.5 1.0 Average % burned Average % burned Medium-large ungulate 553 0.9 0.0 Ungulates 9 35 Equus hydruntinus 476 0.8 0.4 Carnivores 14 32 Cervus elaphus 1493 0.3 0.2 Hares and hedgehogs 10 31 Large ungulate 1577 2.3 0.3 Birds 6 3 Bos primigenius 97 0.0 0.0 Turtles and tortoises 18 24 Large-very large ungulate 24 8.3 0.0 Fish 18 8 Vulpes vulpes 1101 0.4 0.3 Microfauna 6 5 Small carnivore (indet.) 259 0.0 0.4 All taxa combined 10 24 Lepus europaeus 2605 0.2 0.5 Author's personal copy

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 625

Figure 5. Relative abundance of marine mollusks collected by humans in the H1B faunal series based on NISP data from Shackleton (1988: Appendix C.4). Monodonta, Patella, and possibly some of the bivalves (e.g., Cerastoderma) were used principally or exclusively as food; Columbella and Cyclope were used as ornaments; the uses of Cerithium and Murex are unclear but they may have served as fish bait. Very rare types are not illustrated. Gray lines and terms at left indicate lithostratigraphic layers. Shell images are not to scale. known to create fragmented avian assemblages, but the strong digestive acids of these carnivorous birds often leave a character- istic polish or corrosion on the bone surfaces (Brain,1981; Andrews, 1990; Hockett, 1996; Schmitt, 1997; Hockett and Bicho, 2000). No evidence of raptor digestion or punctures was found on the partridge or bustard bones from Franchthi Cave. Given the breakage patterns, absence of evidence for predation by carnivorous mammals or birds, and the importance of partridges in human diets at other contemporaneous sites, the partridges in Franchthi Cave are taken to represent human prey. Bustard remains are comparatively few in the H1B trench. These bones are highly fragmented (11.1% complete, n ¼ 18), and the incidence of burning is higher if the indeterminate very large bird bones (which can only be bustard) are taken into account (7.2%, n ¼ 69; Table 3). There are very few means by which bustards could have been introduced into the cave, given their exceptionally large bodysize (the largest flying bird at up to ca. 16 kg for males [Dunning, 1993]). Burning damage, and the fact that their bones are not easily broken, led us to conclude that the bustards were human prey as well. Rock doves (pigeons) are naturally attracted to caves and rock- shelters as nesting sites, and there are reasons to question whether or not they were a significant prey item in the Franchthi cultural sequence. Birds generally are much less important in the later faunas of the H1B series, yet the proportion of pigeons and closely related doves in the avifaunas increases significantly with time. Pigeon and dove bones are less frequently burned than those of most other birds in the Mesolithic deposits, and much less often burned than the bones of other vertebrates. Humans may not have eschewed pigeon meat entirely, but pigeons do not seem to have been an important economic resource at this site. To summarize thus far, the rates of burning on carnivore, hare and hedgehog, turtle and tortoise, fish, and ungulate remains are well above those interpreted to stem from secondary or accidental Figure 6. The relation between ungulate and small game abundance by unit in the burning. Supporting evidence includes the presence of cut marks H1B sequence as NISP. Arrowed bars indicate values in excess of the NISP scale. Gray and, on the ungulates, cone fractures and high frequencies of green lines and terms indicate lithostratigraphic layers. Author's personal copy

626 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 bone fractures (62e69% of long bone specimens). There is scant if species, particularly Cyclope and Columbella, instead were used as any evidence of carnivore gnawing or digestive damage (Table 5). ornaments (see Perlès and Vanhaeren, 2010). Partridges, bustards, and some of the raptors were also prey of Edible shellfish all but disappear above unit 153. Small Cyclope humans. Micro-vertebrates and certain other vertebrates appear to and Columbella ornamental shells dominate the marine mollusk have entered the cave voluntarily or as the prey of owls (see SOM assemblages thereafter, accompanied by small but persistent Appendix 2). quantities of ceriths. Humans modified the cerith shells mainly by Dense land snail middens occur in the Final Paleolithic and breaking the shell aperture and sometimes the spire (Perlès, pers. Mesolithic layers (Farrand, 2000; Whitney-Desautels, in press) and comm.). While it is possible that humans ate them, a compelling are composed predominantly or exclusively the edible species, alternative explanation is that the ceriths were used as line bait in Helix figulina. The middens resemble in density the well-known the manner that traditional fishermen in southern Greece have escargotières found elsewhere on the Mediterranean rim (Lubell, continued to use them into recent times (Katsanevakis et al., 2008). 2004). Marine shellfish were also exploited during the later part of the sequence (van Andel and Shackleton, 1982; Shackleton, Trends in prey abundance and diversity 1988), especially between 13.3e12.7 ka (Stratum T, Fig. 5). It is not clear from Shackleton’s (1988) study which marine mollusks The array of prey species in the Franchthi faunal series is truly were consumed as food and which as raw material for ornament- remarkable. A primary quantification of these remains is provided making. Considerable research on mollusk taphonomy has been in SOM Appendix 1 and the full set of diversity results in SOM undertaken since the publication of the Franchthi marine shells. Appendix 3. Two general prey groupsdungulates and small These newer studies document clear differences in the conditions gamedare common throughout the sequence (Fig. 6). Ungulates of shells put to ornamental versus culinary and other uses (compare dominate most units in terms of prey biomass, but the consistent Stiner, 1999; Henshilwood, 2004; Jerardino et al., 2009; Kuhn et al., presence of myriad small animals contrasts sharply with the 2009). It seems likely in retrospect that only the limpets (Patella hunting behaviors of Middle Paleolithic humans in southern Greece spp.) and turbans (Monodonta spp.) were significant food sources (Starkovich, 2009, 2011) and other Mediterranean areas where this among the marine mollusks in Franchthi. These shellfish would has been studied intensively (Stiner et al., 2000; Stiner, 2001). have been available from rocky shores at the northern end of Ungulates Six ungulate species occur in the H1B faunal Franchthi embayment (van Andel and Shackleton, 1982). Small seriesdred deer (Cervus elaphus), European wild ass (Equus hydrun- numbers of clams (Tapes), cockles (Cerastoderma), and possibly tinus), aurochs (Bos primigenius), wild pig (Sus scrofa), wild goat large ceriths (Cerithium spp.) and tops (Gibbula spp.) were also (Capra cf. ibex), and roe deer (Capreolus capreolus). Red deer is the consumed, and possibly Murex as well. The shells of very small most ubiquitous ungulate (Fig. 7). It is common in the earliest units

Figure 7. Ungulate species distributions* as the percent of all species-specific ungulate identifications and assemblage sizes (NISP) by unit in the H1B trench. (*) The percent data exclude many specimens that could be identified to ungulate body size class and skeletal element but not to species or genus; thus, the data presented are a fraction of the total ungulate NISP for each assemblage. Arrowed bars indicate values in excess of the NISP scale. Gray lines and terms at left indicate lithostratigraphic layers. Author's personal copy

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 627

(212e209, Stratum R), suggesting relatively mild, moist conditions. Table 6 Red deer gradually decline through units 193e186, where its repre- Diversity correlations (rs) for vertebrate prey groups to the temporal order of excavation units (¼time) in the H1B trench. sentation is sporadic and replaced in importance by wild ass. Above unit 185, Cervus rebounds and dominates the ungulate fraction for Prey groups (n assemblages) rs P the remainder of the sequence. Ungulate species diversity declines Ungulate taxa (n ¼ 102) 0.398 0.001 with time (Fig. 8, Table 6), though particularly above unit 151. Carnivore taxa (n ¼ 77) 0.116 w1 ¼ < < Carnivores Carnivore remains are comparatively rare in the Small game taxa (n 106) 0.284 0.01 p 0.001 Small game escape types (n ¼ 106) 0.088 w1 early part of the H1B sequence, but small fur-bearing species came Marine small game (n ¼ 54) 0.425 0.01 < p < 0.001 into heavy use above unit 167 (Fig. 9). The carnivore remains are Terrestrial small game (n ¼ 104) 0.341 0.001 overwhelmingly from red fox (Vulpes vulpes), but small quantities Note: Because the labeling system for units increases with excavation depth, of marten (Martes martes, rarely M. foina), wild cat (Felis silvestris), negative values reflect an increase in diversity values toward the recent, and posi- lynx (Lynx/Felis lynx), and badger (Meles meles) are also present. tive values reflect a decline in diversity values toward the recent. Two possible Canis sp. specimens and 1 leopard specimen (Panthera pardus) also were found in H1B. All of these carnivores may have been exploited foremost for raw materials but probably were also Marine resources came into importance in the second half of the eaten, as many core body parts are fractured and burned. Prelimi- sequence and included both shellfish and fish. The main types of nary taphonomic data indicate that the heads and some of the paws fish were sparids or sea breams (mostly gilthead, Sparus aurata), of foxes and wild cats remained in articulation after the rest of the along with barracuda (Sphyraena sphyraena), mullets (family body had been dismantled. We present the carnivores here in Mugilidae), sea bass (Dicentrarchus labrax), conger eel (Conger recognition of their general economic importance, but we do not conger), and bluefin tuna or tunny (Thunnus thynnys). Remains of include them in the list of staple meat sources. very small fish such as sardines are also present, but these are rare Carnivore diversity is difficult to evaluate in the H1B series even in flotation samples. because carnivores are important only above unit 164e167 or Stratum S2 (Fig. 8). Although the remains are dominated entirely by fox in most units, overall diversity (Table 6) generally increases above unit 165, perhaps simply because sample sizes increase. Small game The small game fraction of the H1B assemblages contains a wide variety of animals from terrestrial and marine sources. Terrestrial small game animals were hare (Lepus europaeus), hedgehog (Erinaceus europaeus), ground birds (partridges, mainly Alectoris sp., and great bustard, Otis tarda), Mediterranean spur-thighed tortoise (Testudo spp.), and pond turtles (Emys orbicularis or possibly Maurmys sp.). Of the small mammals, only hare was consistently important in the diet (Fig. 10). Birds were important during the Upper Paleolithic, but they decline dramatically above unit 152 (Stratum U), when partridges disappear entirely from the assemblages.

Figure 9. Genus-speciﬁc NISP* distributions in the H1B trench by excavation unit for Figure 8. Trends in diversity for major game groups based on the Inverse of Simpson’s the three most common carnivore groups: foxes, small cats, and mustelids. (*) These Index, by unit in the H1B trench (for assemblages with NISP > 29). Gray arrows data exclude many specimens that could be identiﬁed to carnivore body size class and highlight sudden increase in small game diversity in the faunal series. Gray horizontal skeletal element but not to species or genus. Arrowed bar indicates a value in excess of lines and terms at right indicate lithostratigraphic layers. the NISP scale. Gray lines and terms at left indicate lithostratigraphic layers. Author's personal copy

628 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636

Figure 10. The relative proportions (percent) by unit of small mammals, birds, chelonians, and fish within the small game assemblages, and counts by unit of vertebrate small game (NISP) in the H1B trench. Arrowed bars indicate values in excess of the NISP scale. Small mammals are mainly hare along with rare hedgehog remains. Game birds are partridge and/ or bustard, based on taphonomic indications. Edible shellfish (from Shackleton, 1988: Appendix C.4, cnt1) are mainly limpets (Patella) and turbans (Monodonta) but may include small numbers of clams (Tapes), cockles (Cerastoderma), Murex, and large forms of Gibbula. Land snails (Helix figulina) are given as MNI counts following Whitney-Desautels (in press). Gray lines and terms at left indicate lithostratigraphic layers.

Evidence of marine exploitation begins above unit 169, or in by notable increases in other previously rare prey on land and sea, Stratum S2. At first, marine fishing focused on inshore breams, but such as great bustard, land snails, shellfish, and pond turtles the importance of these fish was overtaken with time by barracuda (Fig. 10). Edible marine shellfish were important between units and eventually tunny (Fig. 11). The onset of fishing is accompanied 160e154. Use of tortoises and turtles at Franchthi Cave was minor,

Figure 11. Taxon-specific NISP* distributions in the H1B trench for the most common types of fish: sea breams (Sparidae), conger eel, barracuda, and tunny. (*) Counts are based on diagnostic head parts, axis, and terminal vertebrae only and therefore exclude many specimens that could be identified to skeletal element but not to species or genus. Arrowed bars indicate values in excess of the NISP scale. Gray lines and terms at left indicate lithostratigraphic layers. Author's personal copy

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 629

Figure 13. Plot of small game evenness against time (consecutive units) for “terrestrial” taxa (including small mammals, birds, and pond turtles) and “marine” taxa (fish) for assemblages with NISP > 29 in the H1B trench. Gray lines and terms indicate lithostratigraphic layers. Invertebrate prey (shellfish, land snails) are not included in this comparison. Figure 12. Genus-specific NISP* distributions in the H1B trench for tortoises and pond turtles. (*) Counts are based only on the most diagnostic bony parts and therefore represent a fraction of the total chelonian remains from each unit. Arrowed bars fl indicate values in excess of the NISP scale. Gray lines and terms at left indicate lith- hares predominate; (3) ying birds, mainly partridges and ostratigraphic layers. bustards; and (4) fast-swimming bony fish, here combining near- shore and offshore varieties. The four categories should correspond in some manner to differences in the amount of technological e except for the surge in abundance between units 172 158. Inter- investment needed to increase capture efficiency on land or sea. estingly, this brief burst in chelonian exploitation began with Some of the tunny in the Franchthi assemblages were moderately fi tortoises and nished with water-dwelling turtles, shifting in units large individuals, but they are considered with the other fish simply fi 163/162 (Fig. 12). Edible shell sh appear in the faunas very soon because tunny represent the final development in marine fishing. after pond turtles (161/160), as do the dense concentrations of land Though largely aquatic, pond turtles are grouped with slow snails (compare Figs. 9 and 11). terrestrial taxa in this analysis because they come from small water In addition to taxonomic distinctions, small game animals can bodies surrounded by land. fi be classi ed according to their habitat preferences and strategies There is no single trend in small game diversity, regardless of fl for eluding humans. It is the latter set of properties that in uence whether the data are examined from the viewpoint of taxonomic or captures costs (Table 7). Our escape-based typology for small cost-based categories (Table 6). This result contrasts markedly with the vertebrate prey uses four categories: (1) slow-moving prey such as long decline in ungulate diversity. There is, moreover, no correlation tortoises and turtles; (2) small mammals, in which fast-running between ungulate and small game taxonomic diversity through the

Table 7 Summary of small game predator avoidance and habitat preference traits.

Environment Predator avoidance traits Taxa Body size range (kg) Terrestrial Slow, cryptic Tortoises 0.3e2 Terrestrial/streams/ponds Slow crawling/swimming Pond turtles 0.3e1 Terrestrial Hiding, fast-running Hares 2e4 Terrestrial Hiding, fast ﬂying and running Partridges, bustards 0.4e8.5 Marine, littoral Sessile or slow, cryptic Marine shellﬁsh 0.05e0.1 Marine, near-shore Fast swimming Breams, eels, sea bass, etc. 0.3e3 Marine, off shore Fast swimming Barracuda, tunny 3e200

Notes: Data include tunny because they are aquatic, although they are not small. Bird weight estimates are from Snow and Perrins (1998); hare weights from Silva and Downing (1995); Mediterranean tortoise weights from Lambert (1982); pond turtle weights from Balázs and Györffy (2006) and Zuffi et al. (1999); tunny weights at Franchthi estimated from skeletal material by Rose (1995) and vary from small to very large; other fish and shellfish weights estimated by the authors. Author's personal copy

630 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 faunal series (r2 ¼ 0.153, p ¼ 1, n assemblages ¼ 101), meaning that the Site formation processes and economic intensification large and small game trends must be evaluated separately. The abrupt excursion of small game diversity into higher values The intensity of the human occupations in Franchthi Cave is an between units 172 and 163 in Fig. 8 (gray arrows) signals an abrupt important corollary to the results on increasing dietary breadth. rise in foraging costs. As diet breadth from terrestrial sources However, variable sedimentation rates are a potentially conflating peaked in this interval, people turned to the sea. Calculating small factor for interpretation (Jerardino, 1995). The fact that burning prey diversity for terrestrial and marine source separately (Fig. 13) damage increases greatly with time in the H1B sequence, for further reveals that terrestrial small game diversity was moderately example, may suggest a situation of reduced sedimentation in the high before unit 172, rose suddenly between units 172e163, then later periods, heavier use of the site by fire-building humans, or plummeted abruptly between units 163e147. Terrestrial small both of these. Changes in site function at the scale of residential game diversity rebounded somewhat above unit 147 and remained camps versus special use sites does not seem to be the explanation, so through the rest of the sequence. The short burst in terrestrial since small animals with diverse capture requirements and habitat small game diversity first identified in Fig. 8 clearly prefaces the preferences were exploited along with large game throughout the shift to marine foraging, which takes off above unit 168. H1B sequence. Most or all of the cultural horizons in Franchthi Stepping back to consider the faunal data by lithostratigraphic seem to represent residential camps on grounds that diverse layer removes most sample size issues (Table 8,andSOM economic activities are indicated throughout. Appendix 3a), while accumulation interval varies much more Occupation intensity here is examined in three ways: (1) (Table 9a). Ungulate diversity follows a single declining trend in burning patterns and distributions, (2) the relative accumulation this view of the data, as it did in the unit-based analysis. Rather rates of sediments, bones, and land snails as a function of time than one trend in small game diversity, there are two trends, one interval, and (3) trends in prey diversity as a function of the time following on the heels of the first. Terrestrial diversity builds interval (Table 9). Correcting for variation in sediment volumes and from Stratum R through T, collapses, and then climbs and stabi- lithostratum formation time ([specimen count / sediment volume lizes at a moderate level (Table 8). Marine diversity begins low in in liters] / duration), Fig. 14 shows that prey remains accumulated Strata T and U and increases through the rest of the sequence. very rapidly in Strata T and especially W, in parallel with the rapidly The onset of the diversity trend in marine fishing is accompanied accumulating sediment. According to Farrand (2000), there is by the significant inclusion of invertebratesdland snails and a strong anthropogenic component to the sediment composition of marine shellfishdinto the diet (Table 8b). (The somewhat these layers as well. The opposite is true for the early part of the anomalous results for Stratum W/X1 in Table 8 may be the result H1B sequence, where the rate of sediment to faunal accumulation of minor mixing with Neolithic deposits at the very top of the was high (Strata ReU) and the anthropogenic signal in the sedi- series.) ment matrix weak. Nearly identical patterns of variation are found

Table 8 Inverse of Simpson’s Index for various prey groupings by lithostratum, with shift from terrestrial to mixed marineeterrestrial diet indicated by horizontal line, and frequency data for land snails and edible shellﬁsh in the H1B trench.

a. General game groups:

Lithostratum Unit Taxon-speciﬁc All prey taxa Ungulate All Sg All Sg by range NISP for all prey taxa taxa escape traits Meso (W, X1e2) 106e116 700 4.70 1.80 2.38 1.64 Meso (W) 117e127 539 5.46 1.43 3.62 2.20 Meso (W) 128e140 538 4.84 1.66 2.43 1.67 Meso (W) 141e147 742 4.21 1.67 1.65 1.11 V 148e151 267 3.63 1.73 1.25 1.25 U 152e154 172 2.35 2.21 1.19 1.17 T 155e163 2457 6.87 2.45 3.20 2.88 S2 164e172 450 6.11 3.00 2.75 2.56 S1 173e180 370 2.42 3.08 1.53 1.52 R 181e212 2049 2.80 2.93 1.75 1.73 Q 213 5 2.27 e 1.60 1.60

b. By small game sub-groups and mollusk counts:

Lithostratum Unit Taxon-specific Taxon-specific Sg terrestrial Sg marine Land Edible shellfish range NISP for NISP for taxa taxa snail MNI NISP terrestrial Sg marine Sg Meso (W, X1-2) 106e116 119 329 1.20 1.49 ? 78 Meso (W) 117e127 127 94 1.73 2.14 3757 34 Meso (W) 128e140 175 38 1.72 2.03 5333 192 Meso (W) 141e147 218 8 1.54 1.28 5033 62 V 148e151 106 11 1.04 1.00 778 11 U 152e154 116 2 1.15 1.00 1834 256 T 155e163 813 162 4.01 1.06 26,924 251 S2 164e172 228 11 2.51 e 73 0 S1 173e180 287 2 1.51 e 45 0 R 181e212 1580 1 1.74 e 32 0 Q 213 3 0 1.00 ee 0

Notes: Data include only those specimens that could be identified to genus, or family in some instances. Land snail (Helix figulina) counts are from Whitney-Desautels (in press); low frequencies of snail shells in Strata ReS2 represent natural occurrences of these mollusks in the cave. Edible shellfish NISP counts (limpets, turbans, cockles, clams, Murex) are taken from Shackleton (1988: Append. C.4, cnt1). (Sg) small game animals; (?) no data. Author's personal copy

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 631

Table 9 Relative accumulation rate data (a) and calculations (b) for sediments, vertebrate remains (NISP, all and burned), and land snails (Helix ﬁgulina, MNI) by lithostratum in the H1B trench.

(a)

Lithostratum Unit range Sediment volume (liters) Duration in years Total vertebrate NISP % Vertebrate NISP burned Meso (W) 117e127 1158 140 1093 32 Meso (W) 128e140 1448 470 1160 36 Meso (W) 141e147 1473 80 1582 24 V 148e151 514 270 692 28 U 152e154 464 1410 508 5 T 155e163 1608 500 6492 14 S2 164e172 1795 1000 984 8 S1 173e180 1183 3000 629 7 R 181e212 3697 6000 3480 5 Q 213 529 ? 7 0

(b)

Unit Total vertebrate Total vertebrate Snail MNI/ Land snail MNI Burned vertebrate NISP/ Burned vertebrate range NISP/ NISP rate ( 1000) sediment rate ( 1000) sediment NISP rate ( 1000) sediment volume (liters) volume (liters) volume (liters) 117e127 0.94 5.835 3.24 23.174 0.31 2.190 128e140 0.8 1.599 3.68 7.836 0.29 0.623 141e147 1.07 11.278 3.42 42.710 0.25 3.182 148e151 1.35 3.314 1.51 5.606 0.37 1.391 152e154 1.09 0.619 3.95 2.803 0.05 0.037 155e163 4.04 7.427 16.74 33.487 0.56 1.123 164e172 0.55 0.415 0.04 0.041 0.05 0.046 173e180 0.53 0.156 0.04 0.013 0.04 0.012 181e212 0.94 0.140 0.01 0.001 0.05 0.009 213 0.01 ? 0.002 ? 0 ?

Notes: Sediment volumes and snail counts (MNI) are from Whitney-Desautels (in press: Table 7). Vertebrate remains include cultural and non-cultural types. Duration estimates are from Farrand (unpublished data). Rate per year ¼ (specimen count / liters of sediment)/duration of unit range. Rate per year is multiplied by 1000 to make comparisons manageable. among the strata for the rates of bone and snail accumulation and have low diversity, however, and the most diverse small game the incidence of burned bones (Fig. 14). assemblages in the younger layers all formed quickly (Strata T and Longer accumulation times raise the probability that many the W series). It is also clear that small game diversity rises and falls separate events and conditions will introduce diverse skeletal directly with the rate of cultural inputs to sediments. The first part remains to the sedimentary record. Increases in environmental of the Epigravettian in Stratum S2 is the sole exception in that the heterogeneity as a function of climate change can have similar assemblage is moderately more diverse than input rates would effects in raising diversity in faunal assemblages. If human diets predict. were expanding in response to reduced supplies of highly ranked The input analysis reveals that cultural material collected slowly foods, on the other hand, prey diversity could be exceptionally high relative to sedimentation in the earliest layers, with a first peak in even in strata that formed over very short periods. At Franchthi Stratum T (Fig. 14). The rate declined suddenly in Stratum U and Cave, ungulate diversity decreases toward the recent (see Table 8), then built-up again through Strata V and the W series (especially and it decreases consistently with lithostratum duration (Table 10 units 141e147). The bimodal pattern in cultural inputs through the and Fig. 15). The situation for small game in Fig. 15 is very stratigraphic sequence matches that for small game diversity different in that diversity shows no consistent relation to stratum (Fig. 14). This means that the intensity of the human occupations in duration. Most long interval (¼early) small game accumulations the cave increased in tandem with intensified use of costly small

Figure 14. Comparison of the relative accumulation rates of vertebrate remains, and snail shells by major lithostratigraphic layer, in the H1B trench. Sediment volume estimates (liters) and land snail counts (MNI) are from Whitney-Desautels (in press). Land snails are the edible species of Helix ﬁgulina. Author's personal copy

632 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636

Table 10 Argolid Peninsula and the powerful local influences of sea level and ’ Correlations (rs) of the Inverse of Simpson s Index of diversity to lithostratum terrestrial moisture on vegetation (Stiner et al., in press). Drier, duration for vertebrate prey groups in the H1B trench. open conditions and an expanded coastal plain would have favored Prey groups (n assemblages) rs p wild ass and aurochs. When conditions became wetter and the Ungulate taxa (n ¼ 10) 0.758 0.01 coastal plain contracted, red deer dominated the area, along with Carnivore taxa (n ¼ 10) 0.462 w1 wild pig. The environmental situation for ungulates on the ¼ w Small game taxa (n 10) 0.224 1 southern end of the peninsula differed significantly from the Small game escape types (n ¼ 10) 0.236 w1 Marine small game taxa (n ¼ 7) 0.189 w1 Peloponnesian interior. No fallow deer remains (Dama dama)were Terrestrial small game taxa (n ¼ 10) 0.212 w1 found in the H1B sequence of Franchthi, yet this was the dominant ungulate throughout the Upper Paleolithic and Mesolithic occupations at Klissoura Cave 1 (Koumouzelis et al., 2001; Starkovich, game resources. The expansion into marine resources mid-way 2009), where red deers were rare. Hare remains are abundant through the cultural sequence complicates the study of dietary throughout the Upper Paleolithic and Mesolithic occupations at breadth at Franchthi, mainly because the shift opened up an Klissoura Cave 1 (Starkovich, 2009), just as they are in the Franchthi entirely new range of foraging possibilities. Foragers seized upon series. these new possibilities in a step-wise fashion; the trends in small The character of the southern Argolid peninsula in combination animal exploitation follow the predictions of a prey choice model with marine transgression presented unique possibilities for die- for expanding dietary breadth in response to the decline of highest- tary expansion at Franchthi Cave. The coastal plain was reduced by ranked resources in the study area. Whether these local develop- half or more by the time of the Final Paleolithic and Mesolithic (see ments reflect a region-wide condition remains to be addressed. Fig. 2), forming the Franchthi embayment and an inlet with a marshy fringe below the cave (van Andel and Lianos, 1983; Discussion Jameson et al., 1994). There is no clear indication of a Younger Dryas cold oscillation in this area of southern Greece (compare “ ” The surf and turf variations in the faunas of Franchthi Cave Bottema, 1995; Karkanas, 2001; Lawson et al., 2004 on Thessaly). rekindle Jacobsen’s(1976)question of how much of the diet change The hills were covered by patchy scrub forest intermingled with can be explained by coastal evolution as opposed to cultural plants that included a variety of fruit- and nut-bearing trees, large- evolution. Any attempt at an answer must consider both the seeded grasses, and pulses. Foragers’ solutions to reduced avail- geological record of an evolving landscape and the theoretical ability of high-return terrestrial foods were to embrace the sea and distinction between foraging adjustments within an existing to work hard at processing high-cost foods such as plant seeds, adaptation and directional change as the result of natural selection. nuts, and land snails (Runnels, 2009). By the Mesolithic, people Several powerful forces have contributed to the cultural and were doing all of this. economic turnover at Franchthi Cave. This is not a simple story of The overarching trends in hunting and gathering are summa- continuous local evolution, as the most important cultural shifts rized in Fig. 16. As we have argued elsewhere (e.g., Stiner et al., associate with hiatuses in deposition (Payne, 1975; Perlès, 1999; 2000; Stiner, 2001; Munro, 2004), patterns of small game use are Farrand, 2000). Upper Paleolithic humans visited the cave until particularly sensitive indicators of changes in foraging efficiency in about 20 ka Cal. BP. After a long interval of limited or no human the Mediterranean basin. This is because the potential nutritional presence, occupations in the cave resumed around 15 ka Cal. BP. returns of small animals vary much more than those of large game, The dominant vegetation was steppe prior to the LGM, and after- owing to great differences in capture costs while body size varies wards a garrigue that supported a wide variety of edible plants. much less. The most striking shift in small game exploitation was Although the cave saw heavier use by humans in the later periods, the addition of marine prey to what was formerly a steadfastly most of the occupations appear to have been predominantly resi- terrestrial diet above unit 172, corresponding to the onset of the dential throughout the sequence. What has been less clear up to Epigravettian occupations (Strata S2 and T3-1). Foragers turned to now is whether or not the faunal trends at Franchthi Cave reflect aquatic resources in earnest by this time, gathering pond turtles changes in efficiency at the core of socioeconomic systems. Such from streams and marshes and shellfish from rocky marine shores changes would imply altered patterns of labor allocation, techno- (Shackleton, 1988), and angling for marine fish (Rose, 1995). At the logical investments, and other systemic features. The CervuseEquus dynamic in the Franchthi faunal sequence likely reflects climate-driven habitat changes on the southern

Figure 16. Summary of floral, faunal, and lithostratigraphic changes in the H1B stratigraphic sequence in Franchthi Cave. (j) main lithostratum shifts; (black fill) indicates presence and general frequency; for chelonians, (white fill) represents terrestrial tortoises, (black fill) represents mainly aquatic turtles. Information on Figure 15. Plot of diversity against duration of lithostratigraphic layer for ungulates marine shellfish summarized from Shackleton (1988), land snails from Farrand (2000), and for small game taxa. and macrobotanical remains from Hansen (1991). Author's personal copy

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 633 same time, land snails came into the diet, and they were exploited Table 6.2). There is somewhat less agreement among the faunal, in fantastic quantities during the Mesolithic (Farrand, 2000; geological, and artifactual trends in the Mesolithic occupations Whitney-Desautels, in press). Labor-intensive collection and pro- (unit 147 and above). cessing of small grass grains and legumes further distinguishes the Some of our faunal phases concur with a simpler scheme Mesolithic from the earlier periods (Hansen, 1991). developed by Payne (1975) on the basis of the H1A trench material The diet was particularly unstable during the Epigravettian (Gravettoid through Mesolithic), which recognized the shift from (units 170 through 162) and Final Paleolithic phases (units open-land faunas to those favoring moister conditions and the 159e148, Strata TeUeV). Marine fishing grew more important and rising intensity of fishing activities through the Mesolithic. The diversified thereafter. Fishing activities began with readily acces- differences we find to Payne’s scheme are explained partly by sible shoreline types and then expanded to include fish that inhabit lateral variation in the contents of the excavation trenches, such as open or deep water such as tunny weighing up to 200 kg (Rose, the absence of an Aurignacian component in H1A. However, 1995). It is not known whether the tunny were caught by fish- Payne’s preliminary study only considered selected elements for ermen from the shore or from boats on the open sea. Either way, large mammals and presence/absence (or very general expressions the cumulative outlay in equipment and work would have been of quantities) for the small vertebrates. Access to a much more significant and no doubt competed with other economic and social detailed dataset on game use and site inputs finally allows us to demands. address the environment-culture dynamic. Our analysis identifies eight (IeVIII) faunal phases in the H1B One indication that natural selection played a significant role in sequence, taking ecological and human behavioral indications into the dietary transformations, locally and possibly also at a regional account (Table 11). Faunal phases IV through VIII correspond well to scale, was the growing emphasis on low-return prey as staple meat changes both in lithostratigraphy and material culture, and to sources. This is expected to happen as the food supply shrinks in hiatuses in occupation of several centuries or more (Farrand, 2000: relation to the number of consumers, whether due to broad

Table 11 Major faunal phases within the H1B faunal sequence, including comparison to Payne’s (1975) preliminary results.

Proposed Lithostrata Key change(s) Cultural phase and faunal unit ranges other developments phases in H1B I 106e120 Focus on tunny fishing Upper Mesolithic; (W3, X1) tunny dominate the fish assemblages; continued open and deepwater fishing; cultural inputs very high relative to sediment accumulation; burning rates high; interval relates to Payne’s (1975) faunal zone ‘D20 in H1A trench. II 121e130 Open water/deep Mesolithic, sub-phase unclear or mixed; (W1eW2) water fishing barracuda and some tunny taken from open or deep marine waters, implying existence of sea-worthy water craft; interval relates to Payne’s faunal zone ‘D1’ in H1A trench. III 131e147 Increased burning rates; Lower Mesolithic; (W1eW2) accelerated cultural incidence of bone burning greatly increases; inputs seed exploitation suddenly becomes common; cultural inputs very high relative to sediment accumulation; interval relates to Payne’s faunal zone ‘D1’ in H1A trench. IV 148e151/153 Open-land taxa gone; Final Paleolithic; (U, V) red deer dominant ground birds, aurochs, and wild asses all but disappear from the faunal assemblages, implying a departure from open dry conditions; probably relates mainly to Payne’s faunal zone ‘D1’ in H1A trench. V 154e162 (T) Very high cultural Epigravettian, with double-backed bladelets and microburins; inputs first major burst in cultural inputs relative to sedimentation rate; largest faunal assemblages come from this section; ungulate diversity declines most in this interval; burning rates high; shift from terrestrial tortoise to mainly pond turtles; interval may relate to Payne’s faunal zone ‘C’ in H1A trench. VI 161e172 (S2) Small game diversity Epigravettian with microburins and La Mouillah points; excursion many changes in small game exploitation in this interval: harvesting of land snails and shellfish and coastal fishing begin in earnest; rodent activity and/or owl drops in cave increase significantly; interval generally relates to Payne’s faunal zone ‘B’ in H1A trench. VII 173e206 Wild ass and Gravettoid and some Aurignacian; (R, into S1) red deer wild ass as important or more important in the ungulate assemblages than red deer, implying open conditions; interval relates to Payne’s faunal zone ‘A’ in H1A trench. VIII 207e212 (R) Red deer dominant Aurignacian; red deer dominates a diverse ungulate assemblage, suggesting moister and possibly warmer conditions and a relatively heterogeneous environment; apparently pre-dates Payne’s faunal zones for H1A trench. Author's personal copy

634 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 environmental changes, demographic increase, or some combina- Cave 1 (Koumouzelis et al., 2001). Here, land snails became tion of these factors. The dietary trends at Franchthi Cave involved a significant food source only with layer III’ (Starkovich and Stiner, two sorts of responses. The first of these was increasing investments in press), which contains a Mediterranean backed blade industry in prey capture and processing technologies (boats, hooks, nets, (Koumouzelis et al., 2001; Kaczanowska et al., in press). Snails lines, snares, and other traps; see Perlès,1999). The second response became an exceptionally important food source in the Mesolithic. was a tendency to forage in an ever wider array of habitats or Hare exploitation also rose greatly in the Epigravettian and Meso- substrates. The order inwhich animal resources were integrated into lithic, overtaking (in NISP counts) fallow deer and all other large the diet follows an optimality prediction, which assumes that time is herbivores (Starkovich, 2009). Though the evidence takes a some- the main limiting currency (Stephens and Krebs, 1986). The marine what different form in this inland area, high-cost small animal trajectory is particularly revealing in this regard, since humans are resources were exploited more heavily with time at Klissoura Cave not naturally equipped to exploit free-swimming animals easily. The 1. The post-LGM and Mesolithic deposits in Caves 4 and 7 in the expansion into the fishing niche demanded increased expertise on Klissoura Gorge are said to contain snail remains as well marine waters, which the presence of Melos obsidian in Franchthi (Koumouzelis et al., 2004), but unfortunately these excavations confirms (Perlès, 1990), and new technological costs that had to be yielded scant faunal remains. subtracted from the net yield of fast-swimming prey. Nearlyall of the shifts in the meat diet suggest substantive declines in foraging Conclusion efficiency that could only be addressed through technological innovations. New patterns of cooperation must also have emerged We conclude from the above observations that the interaction of as the habitats exploited by foragers continued to diversify. By the profound landscape change and human foraging pressure accounts close of the Pleistocene, terrestrial sources may have been squeezed for the main trends in animal exploitation from the Upper Paleo- to the edge of sustainability through a combination of shrinking land lithic through Upper Mesolithic periods at Franchthi Cave. New area and reduced habitat complexity of inland areas. Human pop- economic systems evolved not only because sea level was rising, ulations may also have increased from the Upper Paleolithic to the but also because this and adjacent areas of Greece were experi- Mesolithic, though this is difficult to evaluate solely from cultural encing declines in preferred (highest-yield) meat supplies. The input rates in one site. Either way, forager groups had fewer options Mediterranean Sea is potentially a rich source of food, but this is to move when local resource supplies became strained. true only if the capture costs for fish can be overcome with tech- Does expanding diet breadth at Franchthi Cave reflect a localized nology, which also has costs. adaptation to changes in site function (specifically occupation The intensity of the occupations in Franchthi Cave increased in intensity), or can the tendency be linked to wider foraging conditions tandem with an intensified use of animal and plant resources. As in southern Greece? Hiatuses in the sequence also raise the question terrestrial food sources declined, the foragers of the southern Argolid of whether cultural changes relate to local dynamics or historical turned to the sea, first setting their sights on low-cost, highly shifts in cultural boundaries and distributions over large areas. accessible prey types and later pursuing littoral and large deepwater Answers to these questions come from several lines of infor- fish alike. Hunteregatherers were experiencing similar strains on mation. It is significant, first of all, that the later occupants of traditional food supplies in the interior lands. This fact implies that Franchthi Cave did not, or could not, simply shift their ranges foragers could not abandon the Argolid as local terrestrial resources deeper into the heart of the Peloponnese as the sea overtook low- became impoverished. At Franchthi, human diet breadth expanded lying coastal plains. It also is clear from our examination of small twice over the course of the Upper PaleolithiceMesolithic, first on game diversity that diet breadth increased before and after the land, and later via marine foraging, culminating in skilled harvesting addition of marine food sources to forager diets. Turning to the sea of some of the largest and most powerful fish in the Aegean Sea. brought only temporary relief. The trajectory we observe in aquatic exploitation at Franchthi follows a pattern broadly similar to those Acknowledgments of certain other world regions where increasing economic complexity accompanies expanding diet breadth in marine This work would not have been possible without the intellectual contexts (Llagostera, 1979; Kennett et al., 2007; see also Colonese and logistical assistance of Catherine Perlès (U Paris, Nanterre), K. D. et al.’s (in press) survey for the whole of the Mediterranean). On Vitelli (Indiana U, retired), and the colleagues of the Napflio the northern Channel Islands of Pacific North America, for example, Museum. We are very grateful to Catherine Perlès for her invitation expansion of diet breadth and intensified use of certain fish species to become part of the Franchthi project, and to Catherine, K.D., and paralleled increases in human population and investments in Bill Farrand (U Michigan, retired) for providing data and critical fishing technology (Kennett, 2005). Fishing at Franchthi does not comments to the manuscript. We also are very grateful to Nancy seem to have approached a situation of over-exploitation, as there Whitney-Desautels for access to her unpublished data on sediment is no evidence of “fishing down” the marine food web (i.e., volumes and land snail counts. We thank Mark Rose (now Editor of switching from long-lived, high trophic-level fish to short-lived, Archaeology Magazine), Richard Redding (U Michigan), Curtis very productive fish that exist at lower trophic levels; see Pauly Runnels (Boston U), Priscilla Murray (Boston U), Steve Kuhn (U et al., 1998). Nor is there any sign of fish body size diminution as Arizona), M. Koumouzelis and P. Karkanas (Ephoreia of Palae- documented by Broughton (1999) in San Francisco Bay. At oanthropology-Speleology, Athens), and Sherry Fox and colleagues Franchthi, foragers were utilizing marine habitats at full tilt by the at the American School for Classical Studies in Athens (ASCSA, Upper Mesolithic, but the quantities of tunny caught were not Greece) for sharing their insights with us and for valuable sup- enough to suppress the structure of tuna populations in the open porting information. This manuscript benefitted greatly from the sea (see Rose, 1995). It was the technological outlays and personal critical comments of the JHE editors, Curtis Runnels, Clive Gamble, risks on sea that were most prohibitive. Todd Whitelaw, and two anonymous reviewers. This first phase of Finally, we must look inland to evaluate the geographic scale of the faunal research was supported by a grant from the Institute for the dietary changesdat least some indications of resource inten- Aegean Prehistory (INSTAP) to N.D.M. and M.C.S., a grant to M.C.S. sification or diet breadth expansion should be obtained in adjacent on Pan-Mediterranean diet evolution from the National Science areas of the Peloponnese. There is for the moment only one detailed Foundation (BCS-0410654), and a University of Connecticut small record for comparison on the northern Argolid, the site of Klissoura grant to N.D.M. Author's personal copy

M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636 635

References Karkanas, P., Koumouzelis, M., Kozlowski, J.K., Sitlivy, V., Sobczyk, K., Berna, F., Weiner, S., 2004. The earliest evidence for clay hearths: Aurignacian features in Klisoura Cave 1, southern Greece. Antiquity 78, 513e525. Andrews, P., 1990. Owls, Caves and Fossils. University of Chicago Press, Chicago. Katsanevakis, S., Lefkaditou, E., Galinou-Mitsoudi, S., Koutsoubas, D., Zenetos, A., Balázs, E., Györffy, G., 2006. Investigation of the European pond turtle (Emys orbi- 2008. Molluscan species of minor commercial interest in Hellenic seas: distri- cularis Linnaeus, 1758) population living in a backwater near the river Tisza, bution, exploitation and conservation status. Mediterr. Marine Sci. 9, 77e118. Southern Hungary. Tiscia 35, 55e64. Kennett, D.J., 2005. The Island Chumash: The Behavioral Ecology of a Maritime Belfer-Cohen, A., 1991. The Natufian of the Levant. Ann. Rev. Anthropol. 20, 167e186. Society. University of California Press, Berkeley and Los Angeles. Binford, L.R., 1978. Nunamuit Ethnoarchaeology. Academic Press, New York. Kennett, D.J., Kennett, J.P., Erlandson, J.M., Cannariato, K.G., 2007. Human responses Bochenski, Z.M., 2005. Owls, diurnal raptors and humans: signatures on avian to Middle Holocene climate change on California’s Channel islands. Quatern. bones. In: O’Connor, T. (Ed.), Biosphere to Lithosphere. New Studies in Verte- Sci. Rev. 26, 351e367. brate Taphonomy. Oxbow Books, Oxford, pp. 31e45. Koumouzelis, M., Ginter, B., Kozlowski, J.K., Pawlikowski, M., Bar-Yosef, O., Bochenski, Z.M., Tomek, T. The birds of Klisoura Cave 1: a window into the Upper Albert, R.M., Litynska-Zajac, M., Stworzewicz, E., Wojtal, P., Lipecki, G., Tomek, T., Palaeolithic Greece. European Prehist, in press. Bochenski, Z., Pazdur, A., 2001. The early Upper Palaeolithic in Greece: the Bottema, S., 1995. The younger Dryas in the eastern Mediterranean. Quatern. Sci. excavations in Klisoura cave. J. Arch. Sci. 28, 515e539. Rev. 14, 883e891. Koumouzelis, M., Koz1owski, J., Koczanowska, M., 2004. End of the Paleolithic in the Brain, C.K., 1981. The Hunters or the Hunted? University of Chicago Press, Chicago. Argolid (Greece): excavations in cave 4 and cave 7 in the Klisoura Gorge. Broughton, J.M., 1999. Resource Depression and Intensification during the Late Eurasian Prehist. 2, 33e56. Holocene, San Francisco Bay: Evidence from the Emeryville Shellmound Koumouzelis, M., Kozlowski, J.K., Nowak, M., Sobczyk, K., Kaczanowski, M., Vertebrate Fauna. In: Anthropological Records, vol. 32. The University of Cal- Pawlikowski, M., Pazdur, A., 1996. Prehistoric settlement in the Klisoura Gorge, ifornia Press, Berkeley, CA. Argolid, Greece (excavations 1993, 1994). Préhistoire Européenne 8, 143e173. Colonese, A.C., Mannino, M.A., Bar-Yosef Mayer, D.E., Fa, D., Finlayson, J.C., Lubell, D., Kuhn, S.L., Belfer-Cohen, A., Barzilai, O., Stiner, M.C., Kerry, K.W., Munro, N.D., Bar- Stiner, M.C. Marine mollusc exploitation in Mediterranean prehistory: an Yosef Mayer, D.E., 2004. The Last Glacial Maximum at Meged Rockshelter, Upper overview. Quatern. Sci. Int., in press. Galilee, Israel. J. Isr. Prehist. Soc. 34, 5e47. Davis, S.J.M., 1982. Climatic change and the advent of domestication: the succession Kuhn, S.L., Stiner, M.C., Güleç, E., Özer, I., Yılmaz, H., Baykara, I., Açikkol, A., of ruminant artiodactyls in the late PleistoceneeHolocene in the Israel region. Goldberg, P., Martínez Molina, K., Ünay, E., Suata-Alpaslan, F., 2009. The Early Paléorient 8, 5e14. Upper Paleolithic occupations at Üçagızlı cave (Hatay, Turkey). J. Hum. Evol. 56, De Vivo, B., Rolandi, G., Gans, P.B., Calvert, A., Bohrson, W.A., Spera, F.J., Belkin, H.E., 87e113. 2001. New constraints on the pyroclastic eruptive history of the Campanian Kusmer, K.C., 1990. Taphonomy of owl pellet deposition. J. Paleontol 64, 629e637. volcanic Plain (Italy). Mineral. Petrol. 73, 47e65. Lambeck, K., 1996. Sea-level change and shore-line evolution in Aegean Greece Diamant, S., 1979. A short history of archaeological sieving at Franchthi Cave, since Upper Paleolithic time. Antiquity 70, 588e 611. Greece. J. Field Archaeol. 6, 203e217. Lambert, M.R.K., 1982. Studies on the growth, structure and abundance of the Dunning Jr., J.B., 1993. CRC Handbook of Avian Body Masses. CRC Press, Boca Raton, Mediterranean spur-thighed tortoise, Testudo graeca,infield populations. Florida. J. Zool. Lond 196, 165e189. Farrand, W.R., 2000. Depositional History of Franchthi Cave: Stratigraphy, Sedi- Lawson, I., Frogley, M., Bryant, C., Preece, R., Tzedakis, P., 2004. The Late-glacial and mentology, and Chronology. In: Excavations at Franchthi Cave, Greece, fasc. 12. Holocene environmental history of the Ioannina basin, north-west Greece. Indiana University Press, Indiana. Quatern. Sci. Rev. 23, 1599e1625. Fisher, J.W., 1995. Bone surface modifications in zooarchaeology. J. Archaeol. Meth. Levins, R., 1968. Evolution in Changing Environments: Some Theoretical Explora- Theor. 2, 7e68. tions. Princeton University Press, Princeton, NJ. Gifford, J.A., 1990. Analysis of submarine sediments off Franchthi Cave. In: Llagostera, A., 1979. 9700 years of maritime subsistence on the Pacific: an analysis Wilkinson, T.J., Duhon, S.T. (Eds.), Franchthi Paralia: The Sediments, Stratig- by means of bioindicators in the north of Chile. Am. Antiquity 44, 309e324. raphy, and Offshore Investigations. Excavations at Franchthi Cave, Greece, fasc. Lubell, D., 2004. Prehistoric edible land snails in the circum-Mediterranean: the 6. Indiana University Press, Indiana, pp. 85e116. archaeological evidence. In: Brugal, J.-P., Desse, J. (Eds.), Petits Animaux et Soci- Goldberg, P., Bar-Yosef, O., 1998. Site formation processes in Kebara and Hayonim étiés Humaines, du Complément Alimentaire aux Resources Utilitaires. XXIVe Caves and their significance in Levantine prehistoric caves. In: Akazawa, T., Rencontres Internationales d’Archéologie et d’Histoire d’Antibes, Antibes, pp. Aoki, K., Bar-Yosef, O. (Eds.), Neanderthals and Modern Humans in Western 77e98. Asia. Plenum Press, New York, pp. 107e125. Meignen, L., Bar-Yosef, O., Goldberg, P., Weiner, S., 2001. Le feu au Paléolithique Goldberg, P., Sherwood, S.C., 2006. Deciphering human prehistory through the moyen: recherches sur les structures de combustion et le statut des foyers. geoarchaeological study of cave sediments. Evol. Anthropol. 15, 20e36. L’exemple du Proche-Orient. Paléorient 26, 9e22. Grayson, D.K., 1984. Quantitative Zooarchaeology. Academic Press, Orlando, FL. Meignen, L., Goldberg, P., Bar-Yosef, O., 2007. The hearths at Kebara Cave and their Hansen, J.M., 1991. The Palaeoethnobotany of Franchthi Cave. In: Excavations at role in site formation processes. In: Bar-Yosef, O., Meignen, L. (Eds.), Kebara Franchthi Cave, Greece, fasc. 7. Indiana University Press, Indiana. Cave, Part 1. Peabody Museum of Archaeology and Ethnology Harvard Univer- Henshilwood, C., d’Errico, F., Vanhaeren, M., van Niekerk, K., Jacobs, Z., 2004. Middle sity, Cambridge, MA, pp. 91e122. Stone age shell beads from South Africa. Science 304, 404. Munro, N.D., 2004. Zooarchaeological measures of human hunting pressure and site Hockett, B.S., 1996. Corroded, thinned and polished bones created by golden Eagles occupation intensity in the Natufian of the southern Levant and the implica- (Aquila chrysaetos): taphonomic implications. J. Arch. Sci. 23, 587e591. tions for agricultural origins. Curr. Anth. 45, S5eS33. Hockett, B.S., Bicho, N.F., 2000. The rabbits of Picareiro cave: small mammal hunting Munro, N.D., 2009. Epipaleolithic subsistence intensification in the southern during the Late Upper Palaeolithic in the Portuguese Estremadura. J. Arch. Sci. 27, Levant: the faunal evidence. In: Richards, M.J., Hublin, J.-J. (Eds.), Integrating 715e723. Approaches to the Study of Paleolithic Subsistence. Springer, Netherlands, pp. Jackson, J., 1994. Active tectonics of the Aegean region. Ann. Rev. Earth Planet. Sci. 141e155. 22, 239e271. Panagopoulou, E., Karkanas, P., Tsartsidou, G., Kotjabopoulou, E., Harvati, K., Jacobsen, T.W., 1969. Excavations at Porto Cheli and vicinity, preliminary report, II: Ntinou, M., 2002e2004. Late Pleistocene archaeological and fossil human the Franchthi cave, 1967e1968. Hesperia 38, 343e381. evidence from Lakonis Cave, southern Greece. J. Field Arch. 29, 323e349. Jacobsen, T.W., 1976. 17,000 years of Greek prehistory. Sci. Am. 234, 76e87. Pauly, D., Christensen, V., Dalsgaard, J., Froese, R., Torres, F.J., 1998. Fishing down Jacobsen, T.W., 1981. Franchthi Cave and the beginning of settled village life in marine food webs. Science 279, 860e863. Greece. Hesperia 50, 303e319. Payne, S., 1975. Faunal change at Franchthi cave from 20,000 B.C.e3000 B.C. In: Jacobsen, T.W., 2000. Background of the Franchthi project. In: Farrand, W.R. (Ed.), Clason, A.T. (Ed.), Archaeozoological Studies. North Holland Publishing Co., Depositional History of Franchthi Cave: Stratigraphy, Sedimentology, and Amsterdam, pp. 120e131. Chronology. Excavations at Franchthi Cave, Greece, fasc. 12. Indiana University Payne, S., 1982. Faunal evidence for environmental/climatic change at Franchthi Press, Indiana, pp. 1e10. cave (Southern Argolid, Greece), 25,000 B.P.e5000 B.P. Preliminary results. In: Jacobsen, T.W., Farrand, W.R., 1987. Franchthi Cave and Paralia: Maps, Plans, and Bintliff, J.L., van Zeist, W. (Eds.), Palaeoclimates, Palaeoenvironments and Sections. In: Excavations at Franchthi Cave, Greece, fasc. 1. Indiana University Human Communities in the Eastern Mediterranean Region in Later Prehistory. Press, Indiana. British Archaeological Reports, Oxford, pp. 133e137. Jameson, M., van Andel, T.H., Runnels, C., 1994. A Greek Countryside: The Southern Perlès, C., 1987. Les Industries lithiques taillées de Franchthi (Argolide, Grèce). Tome Argolid from Prehistory to the Present Day. Stanford University Press, California. I: Présentation générale et industries Paléolithiques. In: Excavations at Jerardino, A.,1995. The problem with density values in archaeological analysis: a case Franchthi Cave, fasc. 3. Indiana University Press, Indiana. study from Tortoise Cave, Western Cape, South Africa. S. Afr. Arch. Bull. 50, 21e27. Perlès, C., 1990. Les Industries lithiques taillées de Franchthi (Argolide, Grèce). Tome Jerardino, A., Dewar, G., Navarro, R., 2009. Opportunistic subsistence strategies II: Les industries du Mésolithique et du Néolithique initial. In: Excavations at among late Holocene coastal hunter-gatherers, Elands Bay, South Africa. J. Isl. Franchthi Cave, Greece, fasc. 5. Indiana University Press, Indiana. Coastal Arch. 4, 37e60. Perlès, C., 1999. Long-term perspectives on the occupation of the Franchthi Cave: Kaczanowska, M., Koz1owski, J.K., Sobczyk, K., Wilczynski, J. Upper Palaeolithic human continuity and discontinuity. In: Bailey, G.N., Adam, E., Panagopoulou, E., occupations and material culture at Klissoura Cave 1. Eurasian Prehist, in press. Perlès, C., Zachos, K. (Eds.), The Palaeolithic Archaeology of Greece and Adjacent Karkanas, P., 2001. Site formation processes in Theopetra cave: a record of climatic Areas. Proceedings of the ICOPAG Conference, Ioannina, September 1994, no. 3. change during the Late Pleistocene and Early Holocene in Thessaly, Greece. British School at Athens Studies, pp. 311e318. Geoarch 16, 373e399. Author's personal copy

636 M.C. Stiner, N.D. Munro / Journal of Human Evolution 60 (2011) 618e636

Perlès, C., 2004. Les Industries lithiques taillées de Franchthi (Argolide, Grèce). Stiner, M.C., 1999. Trends in Paleolithic mollusk exploitation at Riparo Mochi (Balzi Tome III: Les industries du Néolithique ancient au Néolithique final. In: Exca- Rossi, Italy): food and ornaments from the Aurignacian through Epigravettian. vations at Franchthi Cave, Greece, fasc. 13. Indiana University Press, Indiana. Antiquity 73, 735e754. Perlès, C., From Hunting to Herding: a 35,000-year Chronology of Human Occu- Stiner, M.C., 2001. Thirty years on the “Broad spectrum Revolution” and Paleolithic pation at Franchthi Cave, n.d.(Argolid, Greece). demography. Proc. Natl. Acad. Sci. 98, 6993e6996. Perlès, C., Vanhaeren, M., 2010. Black Cyclope neritea marine shell ornaments in the Stiner, M.C., 2005. The Faunas of Hayonim Cave (Israel): a 200,000 Year Record of Upper Palaeolithic and Mesolithic of Franchthi (Argolid, Greece): arguments for Paleolithic Diet, Demography and Society. American School of Prehistoric Research, an intentional heat treatment. J. Field Arch. 35, 314e325. Bulletin 48, Peabody Museum Press, Harvard University, Cambridge, MA. Petit-Maire, N., Vrielinck, B., Bracco, J.-P., Brugal, J. Ph., Burollet, P.-F., Coudé- Stiner, M.C., Kozlowski, J., Kuhn, S.L., Karkanas, P., Koumouzelis, M. Klissoura Cave 1 Gaussen, G., Jalut, G., Lericolais, G., van Vliet-Lanoë, B., 2005. The Mediterranean and the Upper Paleolithic of southern Greece in cultural and ecological Basin: The Last Two Climatic Extremes. Explanatory Notes of the Maps. Maison contexts. Eurasian Prehist, in press. Méditerranéenne des Sciences de l’Homme and Agence Nationale pour la Stiner, M.C., Kuhn, S.L., Weiner, S., Bar-Yosef, O., 1995. Differential burning, recrys- Gestion des Déchets Radioactifs Parc de la Croix Blanche, Groupe Horizon, tallization, and fragmentation of archaeological bone. J. Arch. Sci. 22, 223e237. Gémenos, France. Stiner, M.C., Munro, N.D., Surovell, T.A., 2000. The tortoise and the hare: small game use, Rose, M., 1995. Fishing at Franchthi Cave, Greece: changing environments and the Broad Spectrum Revolution, and Paleolithic demography. Curr. Anth 41, 39e73. patterns of exploitation. Old World Archaeology Newsl. 18, 21e26. Stroulia, A., 2010. Flexible Stones: Ground Stone Tools from Franchthi Cave. In: Runnels, C., 2009. Mesolithic sites and surveys in Greece: a case study from the Excavations at Franchthi Cave, Greece, fasc. 14. Indiana University Press, Indiana. southern Argolid. J. Mediterr. Archaeol. 22, 57e73. Tchernov, E., 1992. Evolution of complexities, exploitation of the biosphere and Saavedra, B., Simonetti, S.A., 1998. Small mammal taphonomy: intraspecific bone zooarchaeology. Archaeozool 5, 9e42. assemblage comparisons between South and North American barn owl, Tyto Tchernov, E., 1993. The impact of sedentism on animal exploitation in the southern alba, populations. J. Arch. Sci. 25, 165e170. Levant. In: Buitenhuis, H., Clason, A.T. (Eds.), Archaeozoology of the Near East. Schiffer, M.B., 1972. Archaeological context and systemic context. Am. Antiquity 57, Universal Book Services, Leiden, Netherlands, pp. 10e26. 156e165. Tchernov, E., 1994. New comments on the biostratigraphy of the Middle and Upper Schiffer, M.B., 1983. Toward the identification of formation processes. Am. Antiquity Pleistocene of the southern Levant. In: Bar-Yosef, O., Kra, R. (Eds.), Late 48, 675e706. Quaternary Chronology and Paleoclimates of the Eastern Mediterranean. Schmitt, D.N., 1997. The taphonomy of golden eagle prey accumulations at Great RADIOCARBON. University of Arizona, Tucson, pp. 333e350. Basin Roosts. J. Ethnobiol. 15, 237e256. Tomek, T., Bochenski, Z.M., 2002. Bird scraps from a Greek table: the case of Klisoura Shackleton, J.C., 1988. Marine Molluscan Remains from Franchthi Cave. In: Exca- Cave. Acta Zoologica Cracoviensia 45, 133e138. vations at Franchthi Cave, Greece, fasc. 4. Indiana University Press, Indiana. van Andel, T.H., Lianos, H., 1983. Prehistoric and historic shorelines of the southern Silva, M., Downing, J.A., 1995. CRC Handbook of Mammalian Body Masses. CRC Argolid Peninsula: a subbottom profiler study. J. Naut. Arch. Underwater Press, Boca Raton, Florida. Exploration 12, 303e324. Simpson, E.H., 1949. Measurement of diversity. Nature 163, 688. van Andel, T.H., Shackleton, J.C., 1982. Late Paleolithic and Mesolithic coastlines of Snow, D.W., Perrins, C.M., 1998. The Birds of the Western Palearctic. In: Non- Greece and the Aegean. J. Field Arch. 9, 445e545. Passerines, vol. 1. Oxford University Press, Oxford. van Andel, T.H., Sutton, S.B., 1987. Landscape and People of the Franchthi Region. In: Starkovich, B., 2009. Dietary changes during the Upper Palaeolithic at Klissoura cave Excavations at Franchthi Cave, Greece, fasc. 2. Indiana University Press, Indiana. 1 (Prosymni), Peloponnese, Greece. Before Farming 2009/3 (article 4). Vitelli, K.D., 1993. Franchthi Neolithic Pottery, Volume 1, Classification and Ceramic Starkovich, B., 2011. Subsistence Change during the Middle to Upper Paleolithic at Phases 1 and 2. In: Excavations at Franchthi Cave, Greece, fasc. 8. Indiana Klissoura Cave 1 (Peloponnese, Greece). Ph.D. Dissertation, University of Arizona. University Press, Indiana. Starkovich, B.M., Stiner, M.C. Upper Paleolithic animal exploitation at Klissoura Cave Vitelli, K.D., 1999. Franchthi Neolithic Pottery, Volume 2, Classification and Ceramic 1 in southern Greece: dietary trends and mammal taphonomy. Eurasian Prehist, Phases 3e5. In: Excavations at Franchthi Cave, Greece, fasc. 10. Indiana in press. University Press, Indiana. Stephens, D.W., Krebs, J.R., 1986. Foraging Theory. Princeton University Press, Whitney-Desautels, N.A. The Freshwater and Landsnails of Franchthi Cave, Greece. Princeton, NJ. Indiana University Press, Indiana, in press. Stiner, M.C., 1994. Honor Among Thieves: a Zooarchaeological Study of Neandertal Zuffi, M.A., Odetti, F., Meozzi, P., 1999. Body size and clutch size in the European Ecology. Princeton University Press, Princeton, NJ. pond turtle (Emys orbicularis) from central Italy. J. Zool. 247, 139e143.