Version classiqueVersion mobile
OpenEdition Books

Géoarchéologie des îles de la Méditerranée

Matthieu Ghilardi

Partie 1. Anthropisation et mutations paysagères à la transition Paléolithique/Néolithique / Anthropization and landscape changes during the Late Paleolithic/Neolithic transition

The insular ecology and palaeoenvironmental impacts of the domestic goat (Capra hircus) in Mediterranean Neolithization

Leppard Thomas P et Pilaar Birch Suzanne E


The Mediterranean islands presented substantial challenges to prehistoric colonists, not least their typically dry environments, low-elevation/high-relief geology, and small average size that discriminated against demographically secure communities. The introduction of the Neolithic “package” of domesticated plants and livestock along with ground stone tool technology and ceramics is often assumed to have been central in the adaptation of human communities to these island environments. The exact ecological consequences of the introduction of domesticates for semi-stable island biotas, and the impact pathways through which these consequences were affected, has only rarely been considered in detail. Here, we suggest that the introduction of livestock to the Mediterranean islands during the Early to Middle Holocene (11.7 ka bp – 4.2 ka bp), and especially the introduction of the domestic goat, would have had profound ecosystem effects. We consider the insular ecology of Capra hircus and suggest that its introduction into Quercus-dominated Mediterranean environments would have resulted in dramatically increased predation on plant genera that were not adapted for ovicaprid herbivory. This would have driven a series of corresponding changes, including the sheet erosion of topsoils and ecological release in competitor taxa. These changes – possibly including localized extinctions – may possibly have been more substantial than those effected via direct human agency.

Texte intégral


1Goats are, and have been for some time, quintessentially Mediterranean. In The Corrupting Sea, Horden and Purcell (2000) recall the fragmentary Goats of Eupolis, a fifth century comedy in which a chorus of the animals enumerates the various species of Mediterranean shrub that they are more than willing to eat. This suggests a dietary generalism acting as a foil to Mediterranean barrenness. It is appropriate, then, that the most barren fragments of terrestrial Mediterranean, the countless small, uninhabited islands, are known in Greek as oi tragonisi – the goat islands, a venerable association which first appears in Homer (Bevan and Conolly, 2013). Too circumscribed and dry for human habitation, deforested, and mineral poor, the only use these small islands have for humans is in providing temporary pasturage for domestic goats, naturally corralled and protected from predators by the sea itself (Figure 1). The goal of this chapter is to consider the intimate association of goats and hostile environments in the Mediterranean, and in particular of goats and islands; why should islands be suitable for goats but unsuitable for cattle, pigs, or polycropping? The answer is, of course, known in the abstract – the domestic goat does comparatively well in the type of dry environments exemplified by small, frequently karstic islands. We suggest, however, that a more detailed consideration of how and why Capra hircus is so ideally adapted to island environments might reveal interesting and as yet unexplored aspects of the process of Mediterranean island Neolithization. We begin by briefly reviewing the zooarchaeological data from some of the earliest sites associated with the permanent colonization of the Mediterranean islands. We identify some patterns that seem to indicate an overt reliance on ovicaprids (a) during relatively earlier island settlement rather than later, and (b) on smaller islands rather than larger. We consider the ecology of the domestic goat in particular, especially dietary preferences, in addition to recent studies of Capra hircus as an invasive species in other island contexts. We suggest that the extreme generalist capacity of goats (and to a lesser extent sheep), in conjunction with the ecological fragility of the smaller Mediterranean islands, meant both that maximizing ovicaprid pasturage on mediterraneoid maquis was an efficient means of maintaining a tenuous grasp on otherwise challenging niches, but also that such maximization would have had profoundly transformative effects on insular ecologies – such as deforestation and sediment transport – that might previously have been attributed solely to direct human intervention.

Figure 1: A typical Mediterranean scene: goats grazing and browsing Mediterranean scrub on the Aegean island of Gokceada (Imbros).
Photograph by S.E.P.B.

The zooarchaeology of Capra hircus in Mediterranean Neolithization

2The “Neolithic package”, and particularly the suite of Near Eastern domesticates, was crucial in facilitating the permanent colonization of the Mediterranean islands during the Holocene. These islands, excepting the very largest, tend to be dry, experiencing both low rainfall and poor water retention within soils. They also tend to have few alluvial formations and scrubby, xeric-tolerant native f loras. Several Near Eastern domesticates, notably goats (Capra hircus), barley varietals (Hordeum vulgare spp.), and multiple pulses, are often considered to be pre-adapted to dry source habitats (such as those found in the Taurus and Zagros), and thereby to the insular Mediterranean. This ability to exploit aridity-adapted species in otherwise inhospitable island environments, along- side risk-buffering strategies of storage and diversification employed by early Mediterranean farmers, substantially accounts for the long-term success of Neolithic colonization. This is in marked contrast to Epipalaeolithic and Mesolithic groups who, while demonstrably able to reach various Mediterranean islands – including Melos, Naxos, Crete, and Lesbos – nevertheless seem to have exploited them sporadically and seasonally, with none of the Mediterranean isolates large enough to support permanent itinerant occupation.

3While Mediterranean climatic conditions have fluctuated at a series of scales over the Holocene, and recognizing innate differences in farming practices and strategies, there is nonetheless a tendency for xeric-tolerant domesticated taxa to appear at higher frequencies in assemblages from early sites of island settlement. In particular, sheep and goat tend to comprise a larger portion of overall zooarchaeological assemblages on island as opposed to continental sites. In contrast, domestic ovicaprids may not have had a similar influence in circum-Mediterranean main- land environments due to the occupation of a similar niche by endemic chamois (Rupicapra rupicapra) and ibex (Capra ibex). Though traditionally thought of as high-altitude species (500-3000 m asl, Ferrari et al., 1988) favoring rugged terrain (Choisy, 1994), this may be due to modern habitat restriction caused by humans, and there is archaeological evidence for their presence at much lower elevations in the Late Pleistocene and early Holocene (Miracle and Sturdy, 1991; Phoca-Cosmetatou, 2002). Though both species are primarily grazers, they do include browse in their diet (Trutmann, 2009), and both taxa can be considered opportunistic generalists (Sturdy et al., 1997). Our review of the zooarchaeology in this short contribution is by no means exhaustive, but regularities in the data presented below are nonetheless evident. It is critical to note here that zooarchaeological distinction of ovicaprids from other taxa is well established, but the osteological determination of sheep versus goat within that category has been a topic of much discussion (Boessneck et al., 1964; Helmer and Rocheteau, 1994; Kratochvil, 1969; Prummel and Frisch, 1986). Though recent publications have evaluated in detail the efficacy of criteria for distinguishing between the two (Zeder and Lapham, 2010; Zeder and Pilaar, 2010), this should be borne in mind when interpreting published faunal data.

4On Crete, Jarman and Jarman (1968) noted the high proportion of ovicaprids relative to other taxa from the Neolithic levels of Knossos, the first (seventh millennium) settlement on the island (Figure 2 for sites mentioned in the text). Isaakidou agrees in a re-study (2008), charting a transition from ovicaprid dominance during the Neolithic (7000-3000 BC) to a system placing more emphasis on cattle husbandry in the Bronze Age (3000-1000 BC). Sheep seem to have been of greater significance than goats in Neolithic Knossos; in contrast, Final Neolithic (FN) Phaistos and other Cretan “marginal” FN and Early Bronze (EB) sites display both an overall reliance on ovicaprids, and high total frequencies of goats within this class (Halstead, 1996), with an apparent emphasis on butchering rather than secondary products. This preference for sheep and goats is paralleled elsewhere in the Aegean, despite the settlement of the Cyclades and Dodecanese in the fifth millennium substantially postdating that of Early Neolithic (EN) Knossos. In the Late-Final Neolithic Cyclades, Phoca-Cosmetatou (2011) records that the overwhelming majority of identifiable specimens (NISP) of recovered domesticates from island contexts are ovicaprid (LN Saliagos 84 per cent; LN Ftelia on Mykonos 85 per cent; FN Kefala on Keos 83 per cent), as opposed to about 55 per cent from equivalent Late-Final Neolithic mainland sites (the “flat extended” site at LN Makriyalos provides an illuminating contrast; Pappa et al., 2004). At the opposite end of the Mediterranean, Ramís (2014) tracks a similar transition in Balearic faunal assemblages from ovicaprid dominance immediately after colonization in the late third-early second millennium to a subsequent more balanced husbandry regime. Summing the Majorcan data from the sites of Ca na Cotxera, Coval Simó, Arenalet de Son Colom, Coval den Pap Rave, and Son Matge, ovicaprids constituted 95 per cent of the overall assemblage by NISP (n=2334) in the late third millennium, declining to 69 per cent (n=3603) in the late second (2014). Goats formed a majority of the ovicaprid assemblage and mortality profiles indicated selection for meat production; this, as Ramís notes, contrasts with comparable Corsican and Sardinian data in which reliance on ovicaprids is not as substantial (Vigne, 1988; Webster, 1996). The data from the Għar Dalam levels of Maltese Skorba also suggest an initial primary emphasis on sheep and goat husbandry (Trump, 1966). Cyprus, as so often, bucks any attempt to define a trend (even ignoring the controversial site of Aetokremnos), with scanty but telling evidence for the hunting (rather than herding) of wild boar (Sus scrofa) at the Cypro-PPNA (early-mid ninth millennium BC) site of Klimonas (Vigne et al., 2012). Though present in the Cypro-PPNB (8300-6900 BC), there is no clear emphasis on ovicaprid husbandry, with evidence for both continued reliance on wild red deer (Cervus elaphus) and the eventual disappearance of domestic cattle in the mid-Holocene (Knapp, 2013).

Figure 2: Sites and locations mentioned in the text.

5Tentative patterns emerge from this brief consideration of the data. Early island sites – ranging from ninth millennium in the Eastern Mediterranean to third millennium in the far West – tend to have faunal assemblages dominated by the remains of ovicaprids (Figure 3). These relative frequencies change over time, with cattle and pig playing more significant roles during later phases of occupation, although sheep and goat still constitute the majority of specimens. Two further points of discussion emerge from these observations. We speculate whether: (1) there is a correspondence between island size and the extent to which ovicaprids dominate faunal assemblages (on the early Cyclades, Balearics, and Maltese archipelago very dominant; on Crete less so, with admitted variability; on Cyprus and Sardinia less so again), and (2) whether goats predominate over sheep in areas “marginal” for Neolithic-Chalcolithic rainfed farming. These potential patterns may be related, with small-island reliance on ovicaprids corresponding to an actual strategic preference, reflecting other factors – rather than raw area itself – which relate to island size and promote overall marginality (e.g., reduced orographic rainfall or low pedological variability).

6So far, this is relatively unsurprising; sheep and goat are preferred in dry environments that are relatively hostile for rainfed farming. Indeed, it may be the case that, the smaller and more fragile the environment in question, the more complete the dominance of ovicaprids in the subsistence regime (Cherry, 1981). The spatial and social organization of these regimes, and the adoption of behaviors that boost resilience and mitigate risk, have been addressed comprehensively (Bogaard, 2005; Cherry, 1988; Halstead, 1990; 1996; 2006 and 2008; Isaakidou, 2008), and we do not intend here to develop these topics further. Less frequently considered, and consequently our theme for the remainder of this chapter, is the precise nature of the impacts introduced livestock (especially ovicaprids) had within island ecosystems, and the implications of these impacts for both endemic biotas and for incipient colonizing human populations. The rapaciousness of goats is proverbial (yet also, crucially, quantifiable), and the havoc they wreak in island environments elsewhere on the planet is consequential. This borne in mind, we ask: how would introduced ovicaprid livestock have affected previously isolated insular ecosystems in the Holocene Mediterranean; what would the likely outcomes have been in ecological and sedimentological terms; and how would this have transformed the landscape of subsistence choices? Reviewing the literature on the ecology of Capra hircus and studies of its record as an invasive species elsewhere, we suggest that the introduction of ovicaprids generally, and goats in particular, to the insular Mediterranean during the Neolithic would have generated sweeping environ- mental changes, driving not only localized extinctions (as well as promoting florescence in other taxa), but also landscape changes associated with overgrazing and browsing. It is argued that these unintentional landscape changes were more exaggerated, and with more significant long-term impacts, than has been appreciated.

Figure 3: Relative proportion of overall faunal assemblages of selected LN and FN Aegean sites comprised of sheep/goat, with island sites vs mainland sites distinguished with mainland sites in black, island sites in white and island vs mainland mean values indicated via small hashes and large hashes respectively.
Data from Halstead, 1996 and Phoca-Cosmetatou, 2011.

The ecology of Capra hircus and insular ecodynamics

7The domesticated goat is a small artiodactyl unusually catholic in its tastes, even by the standards of Ruminantia. Derived from one or more stocks of wild progenitors in Southwest Asia in the Early Holocene (Zeder and Hesse, 2000), with possible admixture from wild Capra spp. as the Neolithic lifestyle spread (Fernández et al., 2006), Capra hircus is comparatively better adapted to arid conditions than the other major Neolithic Southwest Asian ungulates, allowing it to tolerate greater variation in seasonality or elevation. It also exploits a wider variety of food species. Both sheep and goats will browse mature tree foliage as well as graze (unlike cattle), but this preference is pronounced in goats. Recent studies of the relative willingness of ovicaprids to consume shrub species that typically comprise Mediterranean maquis (Viburnum tinus, Quercus ilex, Pistacia lentiscus, Juniperus phoenicea, amongst others; Figure 4) recorded goats consuming approximately 25–75 per cent more per species of shrub (measured by grams consumed/kilograms of bodyweight) than sheep (Rogosic et al., 2006; Balasse and Ambrose, 2005 argue that this preference for C3 versus C4 pathway vegetation is distinguishable via stable carbon ratios from sheep and goat samples). This variable preference may be due to subtle differences in the biochemical organization of the digestive tracts of goats, which have a greater capacity to break down primary plant tissue such as fibers and proteins as well as tolerate secondary metabolites such as tannins (Hernández-Orduno et al., 2012). Some plant species typical of the Mediterranean biota experience relatively massive build-up of such metabolites, and an ability to tolerate them (possibly via saliva protein binding; Alonso-Díaz et al., 2012) allows goats to consume more plant material across a wider spectrum. This high tolerance of variability in environmental organization makes goats very competitive generalists. Capra hircus is capable of kidding between one and two young annually, with gestation lasting around 150 days (Peaker, 1978). This reproductive rate, in a situation in which other ecological constraints are relaxed, can lead to marked demographic growth. Island environments can and do provide such situations.

Figure 4: Variable preference for four species of Mediterranean shrub amongst sheep (open circles) and goat (closed circles) over a ten-day trial period. After Rogosic et al., 2006: Figure 1, reproduced with permission.

8As comparatively insulated ecologies, islands tend to be generally more hostile than mainland contexts to taxa that occupy higher tiers in large trophic structures. This hostility derives from the increasing fragility of trophic structures as area for primary production is reduced (Losos and Ricklefs, 2010). This has consequences at both ecological and evolutionary scales. Ecologically, islands tend to have fewer taxa occupying a given niche than in a comparable non- insular environment. Evolutionarily, insular basal foodweb taxa undergoing endemic radiation because of restricted genetic flow often develop traits – notably but not exclusively ecological naïveté – that would be maladaptive in other contexts (i.e., loss of anti-predator behaviors or morphologies). We might suppose, then, that introducing “…versatile generalist herbivores capable of surviving on grasses, forbs, browse, and even marine algae” (Chynoweth et al., 2013) would be to place the cat among the pigeons. This is supported by data from a series of island contexts. There are rich data on the ecodynamics of goats in insular environments. The impact of feral Capra hircus in the Hawaiian archipelago, for example, has been intensively studied (Chynoweth et al., 2013 and 2015). With a highly endemic flora and no native mammals to graze or browse, the more geologically recent Hawaiian islands (Maui and the Big Island) are especially exposed to goat herbivory, and this was recognized relatively early (Yocom, 1967). In the absence of competitors or herbivore-adapted plant taxa, goats have wreaked havoc on the native flora. The reduction of native Hawaiian plants has been achieved not only through direct herbivory, but also associated behaviors (such as trampling), and through non-native plants outcompeting endemic species following niche colonization in the wake of goat browsing (Wilcove et al., 1998). Similar effects of direct herbivory have been noted for the Galapagos and New Zealand. Suppression of populations of native flora is also evident from the California Channel Islands; this is of special interest, as the Channel Islands occupy a dry, mediterraneoid ecoregion. Santa Catalina retained total vegetation cover of 42 per cent in goat-free areas, in contrast to 27 per cent in areas with goat populations (Coblentz, 1978 with references). Moreover, environmental heterogeneity within naïve contexts fails to constrain the impacts of herbivory; despite evident dietary preferences, Chynoweth et al. (2015) found that variability in elevation, pedology, or flora on the Hawaiian Big Island did little to inhibit movements of feral flocks (although wet/dry seasonality imposed some structure to goat mobility). The long-term consequences of this constellation of feeding impacts can be terminal; feral Capra is implicated in the almost total eradication of the native flora of St. Helena in the South Atlantic (Cronk, 1989).

9A key recognition is that these deleterious impacts extend beyond those associated with direct goat herbivory on plants. Extensive grazing and browsing can drive both habitat fragmentation and habitat destruction, particularly but not exclusively in endemic scrub or woodland. Following the principles of community ecology, this sort of dynamism promotes change throughout the ecosystem. If less competitive, other taxa that also consume the species targeted by ovicaprids are likely to experience population decline and potentially localized extinction. This is also true of symbionts (especially arthropods), as well as taxa reliant on woodland for nesting and habitation. Pafilis et al. (2013) explore this causal chain in (modern) Mediterranean contexts, in which goat browsing and grazing reduces insect populations, thereby also reducing numbers of the lizard Podarcis gaigaea, dependent on insects for food.

10There are also pedological and sedimentological consequences of goat herbivory. Destruction of endemic woodland or scrub and its replacement with successor taxa (such as grasses) can reduce soil stability and integrity, as well as driving change in nutrient'n pathways. Yocom (1967) estimates a loss of 1.9 m of topsoil from Haleakalā Crater on Maui followed deforestation. Although other quantitative data on goat-derived soil loss on islands are hard to come by, recent work suggests that combination of over-grazing and over-browsing can drive sheet erosion, dependent on slope and precipitation patterns (Mwendera et al., 1997; Yong-Zhong et al., 2005). The effects of goats on soil composition as opposed to soil dynamics can be more nuanced but nonetheless substantial; goat-derived defoliation causes breakdown in established systems of nutrient cycling, systems which take time to recover even after the eradication of feral goats (Hata et al., 2014). We should bear in mind that a supposed direct causal pathway between “overgrazing” and “overbrowsing” and massive loss of sediment unrealistically fails to take into account variability in vegetation ecology, pedological regimes, and foraging behavior, and that a more subtle approach is necessary (Thornes, 2007). That said, comparative data clearly support a general model of a cluster of ovicaprid-forced impacts, one of which may be exacerbated soil-loss.

11The introduction of a hardy, xeric-adapted generalist ungulate that is capable of both browsing and grazing into environments where it faces an absence of competition and ecologically naïve food species creates sufficient conditions for ecological trauma (Chynoweth et al., 2015; Coblentz 1978). Introduced goat populations undergo various types of ecological release (from both predators and competitors) while simultaneously presented with varied prey species that have not developed structural defenses against predation by ungulates. Combined with the capacity of goats to tolerate very broad climatic parameters (in terms of rainfall, seasonality, and temperature), introduced populations can undergo very rapid r-type population growth; Watts and Conley (1984) report, for example, annual growth of 10–35 per cent. Depending on contextual factors, this has pronounced impacts on endemic insular floras, with the demography of species sometimes suppressed below thresholds of rarity that greatly increase the probability of localized extinction. These fluctuations have effects on symbiont and competitor taxa, driving down symbiont populations while prompting ecological release in competitors. This also stimulates change outside of the strictly biotic sphere, with soil composition and dynamics (i.e., integrity and A-horizon formation) deriving from intact nutrient pathways and from structural stability provided by woodland and scrub. This ecosystem-wide type of impact can be understood as a form of trophic or biophysical cascade process, the effects of which should be visible via a suite of palaeoenvironmental proxy data.

Discussion: Modeling the impacts of Capra hircus in the Holocene Mediterranean

12The palaeoenvironmental organization of the Mediterranean islands during the Holocene is known sufficiently (if unevenly) to the extent that we can predict likely types of impact deriving from the introduction of breeding populations of sheep and – especially – of goats. Despite long-term trends and short-term wobbles in Holocene Mediterranean climatic regimes deriving from different forcing mechanisms (Fletcher and Sánchez Goñi, 2008) the general post-Pleistocene tendency has been from less- to more-arid precipitation regimes. The corresponding general floral trend, though variable along environmental gradients, is from mixed deciduous and sclerophyllous woodlands to heath, semi-desert shrublands, and evergreen sclerophyllous woodland (Allen, 2003). Prior to initial Neolithic and Chalcolithic settlement, then, larger islands colonized earlier in the Holocene probably tended towards higher representation of deciduous taxa, whereas smaller islands colonized later tended towards evergreen sclerophyllous scrub and woodland. In terms of species composition, the Mediterranean islands during the Pleistocene veered towards extreme endemism (Palombo, 2008), with endemic faunal taxa such as Myotragus, Prolagus, Cygnus falconeri, Elephas creticus, E. creutzbergi, and Candiacervus as the most well-known – but not the only – examples. As common in island ecosystems, the insular Mediterranean in the initial Holocene was probably poor in terms of higher-tier taxa (Hunt and Schembri, 1999), with an endemic biota (flora as well as fauna) unused to the types of predation pressure exerted in mainland ecologies. Based, then, on knowing: (1) what ovicaprids (and especially domestic/feral Capra) do to previously isolated island ecologies; (2) that the Mediterranean islands were indeed not adapted to sheep and goat herbivory, and; (3) that these species were introduced to various islands during the Holocene and often preferred over other, less ecologically disruptive species, we are now in a position to suggest likely medium- to long-term effects of ovicaprid introduction. Chiefly, we would expect goat herbivory on a spectrum of plant species. Knowing that sheep prefer graze over browse, that both flocks and herds were maintained during island Neolithic colonization, and that small- scale hinterland “garden” agriculture likely provided substantial opportunities for sheep grazing (Halstead, 2008), it is possible that much goat herbivory occurred in marginal, otherwise-unexploited Quercus shrub or woodland habitat. In contexts in which predators were few or absent and browse relatively plentiful, we might expect high population growth rates and consequent over-browsing, driving habitat retreat and fragmentation and followed by invasion of weedy successor taxa. We suggest that this herbivory would have had three main corresponding effects. First, endemic fauna, reliant on woodland and shrubland for herbivory – such as deer or Myotragids – or as symbionts, would have been exposed to significant ecological pressure, driving down population numbers, potentially towards extinction thresholds. Second, the pressure of goat herbivory would radically alter the selective landscape for island flora, driving some species towards extinction and promoting anti-ungulate adaptations (such as thorniness) in others. Such adaptations include also retention of secondary metabolites in soft tissues which, despite the capacity of goats to digest even very high proportions of these, nonetheless clearly acts as a deterrent where more palatable soft plant tissue is available. The transition to Mediterranean xeric scrub on some smaller islands, then, may not be entirely climatically forced. Third, as (a) deciduous and sclerophyllous woodland fragmented and retreated and (b) as shrubby morphotypes were selected for over larger, woody morphotypes, we would expect soil stability on slopes to decrease. This would drive increased runoff via sheet erosion and cause increased alluviation in valley bottoms.


13While these observations are preliminary, we note some patterns in the palaeoenvironmental data which may support a model of landscape modification derived not through direct human impact, but rather through ovicaprid-induced landscape modification. In this context we highlight the astounding extinction rate of nearly 90 per cent for endemic fauna in the early- mid Holocene Mediterranean (Alcover et al., 1998). While this has been attributed to both climate change and to direct human predation (both no doubt causal factors to some extent), we speculate that massive niche-competition from goats in particular may in part explain the profound loss in Mediterranean island endemism, especially amongst Ruminantia, from the Pleistocene into the Holocene. In terms of grazing and browsing promoting soil instability, renewed interest in episodes of sedimentation suggests that any one-size-fits-all model for the Mediterranean, in the sense of Vita-Finzi, is probably inherently flawed. Nonetheless, recognizing that a finite number of types of human interaction with finite varieties of landscape may drive comparable biophysical processes seems to represent a conceptual compromise (Butzer, 2005).

14Evidence which might illuminate these processes, while comparatively rich basin-wide, remains sparse for the Mediterranean islands. A partial exception is Crete, where data hint at environmental changes in which herbivory played a role. On the Akrotiri peninsula, palynological data from the Limnes core indicates a general decline in arboreal taxa from the Early to Mid-Holocene, coincident with an increase in garrigue and domestic taxa (e.g., olive; Moody et al., 1996). The Delphinos core, taken near Rethymnon, suggests a possible transition from more a deciduous to a more evergreen (Quercus coccifera and Q. ilex) arboreal regime from 8300-7000 BP (Bottema and Sarpaki, 2003), a period in which Neolithic farmers were starting to colonize these more marginal Cretan environments (Halstead, 2008). There is clear variability between these cores, and between interpretations deriving from them (differences which, we wonder, might be explicable in terms of varying subsistence strategies as well as palynological taphonomy). Nonetheless, trying to correlate palynological and sedimentological events with types of human-mediated ecological process should be revealing. In this context, there may be potential to interpret the extensive Mid-Late Holocene terracing regimes on at least some Mediterranean islands both as an attempt to prevent soil-loss deriving from goat browsing and as an attempt to expand zones for poly- cropping and grazing sheep on fallow, pushing goats into ever-more marginal zones in which there was little sediment to be threatened by over-browsing. Though beyond the scope of this contribution, we note that while Bevan et al. (2013) see terracing activities as not necessarily driven by demographic pressure and vegetative recolonization of abandoned terracing so much as counteracting soil instability, we wonder whether terracing as one type of niche-construction may have been driven by different motivating factors at different points in the evolution of the Mediterranean lands- cape. In the study of Mediterranean Neolithization, scholarly attention has rightly been fixed on interactions between humans and domesticated species, and between humans and their environments, conceived broadly. We simply seek here to highlight the importance of domesticate-environmental interactions, the unintended ecological consequences of subsistence strategies, and the potential effects of such consequences in constraining choices available to Neolithic farmers. In doing so we endorse neither a fatalistic catastrophism (Horden and Purcell 2000), nor a teleological story of inevitable decay (Bevan and Conolly, 2013). Rather, we seek to refocus attention on the mutation of the Mediterranean islands into thoroughly domesticated landscapes as potentially driven, not just by direct human intervention, but by the introduction of invasive species which have had comparably trans- formative impacts in other insular contexts. Framing island colonization as a community Ecology problem, and considering Neolithic lifeways as modes of bypassing inherent trophic limitations in island ecosystems, is likely to be a productive avenue of future research.



ALCOVER J.A., SANS A., PALMER M., «The extent of extinctions of mammals on islands», Journal of Biogeography, 25, 1998, p.913-918.

ALLEN H.D., «Response of past and present Mediterranean ecosystems to environmental change», Progress in Physical Geography, 27, 2003, p.359-377.


C.M., «Amino acid profile of the protein from whole saliva of goats and sheep and its interaction with tannic acid and tannins extracted from the fodder of tropical plants», Small Ruminant Research, 103, 2012, p.69-74.

BALASSE M., AMBROSE S.H., «Distinguishing sheep and goats using dental morphology and stable carbon isotopes in C4 grassland environments», Journal of Archaeological Science, 32, 2005, p.691-702.

BEVAN A., CONOLLY J., Mediterranean Islands, Fragile Communities and Persistent Landscapes: Antikythera in Long-Term Perspective, Cambridge, Cambridge University Press, 2013.

BEVAN A., CONOLLY J., COLLEDGE S., FREDERICK C., PALMER C., SIDDALL R., STELLATOU A., «The long-term ecology of agricultural terraces and enclosed fields from Antikythera, Greece», Human Ecology, 41, 2013, p.255-272.

BOESSNECK J., MÜLLER H-H., TEICHERT M., «Osteologische Unterscheidungsmerkmal-zwischen Schaf (Ovis aries Linné) und Ziege (Capra hircus Linné)», Kuhn Archiv, 78, 1964, p.1-129.

BOGAARD A., «“Garden agriculture” and the nature of early farming in Europe and the Near East», World Archaeology, 37, 2005, p.177-196.

BOTTEMA S., SARPAKI A., «Environmental change in Crete: a 9000-year record of vegetation history and the effects of the Santorini eruption», The Holocene, 13, 2003, p.733-749.

BUTZER K.W., «Environmental history in the Mediterranean world: cross-disciplinary investigation of cause-and-effect for degradation and soil erosion», Journal of Archaeological Science, 32, 2005, p.1773-1800.

CHERRY J.F., «Pattern and process in the earliest colonization of the Mediterranean islands», Proceedings of the Prehistoric Society, 47, 1981, p.41-68.

CHERRY J.F., «Pastoralism and the role of animals in the pre- and proto- historic economies of the Aegean», in Whittaker C.R., Pastoral Economies in Classical Antiquity, Cambridge, The Cambridge Philological Society, 1988, p.6-34.

CHOISY J.P., «Reintroduction de bouquetins – Capra sp.: condition de reussite, choix des massifs, enseignements: l’exemple du Vercors», Ibex Journal of Mountain Ecology Speciale Gruppo Stambecco Europa, 1, 1994, p.15-33.

CHYNOWETH M.W., LITTON C.M., LEPCZYK C.A., HESS S.C., CORDELL S., «Biology and impacts of Pacific island invasive species. 9. Capra hircus, the feral goat (Mammalia: Bovidae)», Pacific Science, 67, 2013, p.141-156.

CHYNOWETH M.W., LEPCZYK C.A., LITTON C.M., HESS S.C., KELLNER J.R., CORDELL S., «Home range use and movement patterns of non-native feral goats in a tropical island montane dry landscape», PLOS One, 2015, DOI:10.1371/journal/pone.0119231

COBLENTZ B.E., «The effects of feral goats (Capra hircus) on island ecosystems», Biological Conservation, 13, 1978, p.279-286.

CRONK Q.C.B., «The past and present vegetation of St. Helena», Journal of Biogeography, 16, 1989, p.47-64.

FERNÁNDEZ H., HUGHES S., VIGNE J.-D., HELMER D., HODGINS G., MIQUEL C., HÄNNI C., LUIKART G., TABERLET P., «Divergent mtDNA lineages of goats in an Early Neolithic site, far from the initial domestication areas», Proceedings of the National Academy of Sciences, 103, 2006, p.15375-15379.

FERRARI C., ROSSI G., CAVANI C., «Summer food habits and quality of female, kid and subadult Apennine Chamois, Rupicapra pyrenaica ornata Neumann, 1899 (Artiodactyla, Bovidae). Z», Säugetierkunde, 53, 1988, p.170-177.

FLETCHER W.J., SÁNCHEZ GOÑI M.F., «Orbital-and sub-orbital climate impacts on vegetation of the western Mediterranean basin over the last 48,000 yr», Quaternary Research, 70, 2008, p.451-464.

GLASSER T. A., LANDAU S.Y., UNGAR E. D., PEREVOLOTSKY A., DVASH L., MUKLADA H., KABABYA D., WALKER J.W., «Foraging selectivity of three goat breeds in a Mediterranean shrubland», Small Ruminant Research, 102, 2012, p.7-12.

HATA K., KOHRI M., MORITA S., HIRADATE S., KACHI N., «Complex interrelationships among above ground biomass, soil chemical properties, and events caused by feral goats and their eradication in a grassland ecosystem of an island», Ecosystems, 17, 2014, p.1082-1094.

HERNÁNDEZ-ORDUÑO G., TORRES-ACOSTA J.F.J., SANDOVA L-CASTRO C. A., AGUILAR-CABALLERO A.J., CAPETILLO-LEAL C. M., ALONSO-DÍAZ M.A., «In cafeteria trials with tannin rich plants, tannins do not modify foliage preference of goats with browsing experience», Ethology, Ecology, and Evolution, 24, 2012, p.332-343.

HALSTEAD P., «Waste not, want not: traditional responses to crop failure in Greece», Rural History, 1, 1990, p.147-164.

HALSTEAD P., «Pastoralism or household herding? Problems of scale and specialisation in early Greek animal husbandry», World Archaeology, 28, 1996, p.20-42.

HALSTEAD P., «Sheep in the garden: the integration of crop and livestock husbandry in early farming regimes of Greece and southern Europe», in Serjeantson D., Field D., Animals in the Neolithic of Britain and Europe, Oxford, Oxbow, 2006, p.42-55.

HALSTEAD P., «Between a rock and a hard place: coping with marginal colonisation in the later Neolithic and Early Bronze Age of Crete and the Aegean», in Isaakidou V., Tomkins P., Escaping the Labyrinth: The Cretan Neolithic in Context, Oxford, Oxbow, 2008, p.229-257.

HELMER D., ROCHETEAU, M., Fiches d’Ostéologie Animale Pour l’Archéologie. Série B: Mammifères, Paris, CNRS, 1994.

HORDEN P., PURCELL N., The Corrupting Sea: A Study of Mediterranean History. Oxford, Blackwell, 2000.

HUNT C.O., SCHEMBRI P.J., «Quaternary environments and biogeography of the Maltese Islands», in Mifsud A., Schembri P.J., Facets of Maltese Prehistory, Malta, The Prehistoric Society of Malta, 1999.

ISAAKIDOU V., «“The Fauna and Economy of Neolithic Knossos” revisited», in Isaakidou V., Tomkins P., Escaping the Labyrinth: The Cretan Neolithic in Context, Oxford, Oxbow, 2008, p.90-114.

JARMAN M.R., JARMAN H.N., «The fauna and economy of Neolithic Knossos», Annual of the British School at Athens, 63, 1968, p.241-264.

KNAPP A.B., The Archaeology of Cyprus: From Earliest Prehistory through the Bronze Age, Cambridge, Cambridge University Press, 2013.

KRATOCHVIL Z., «Species criteria on the distal section of the tibia of Ovis ammonf. aries L. and Capra aegargus f. hircus L.», Acta Veterinaria, 38, 1969, p.483-490.

MIRACLE P., STURDY D., «Chamois and the karst of Herzegovina», Journal of Archaeological Science, 18, 1991, p.89-108.

MOODY J., RACKHAM O., RAPP G., «Environmental archaeology in prehistoric N.W. Crete», Journal of Field Archaeology, 23, 1996, p.273-297.

MWENDERA E.J., MOHAMMED SALEEM M.A., DIBABE A., «The effect of livestock grazing on surface runoff and soil erosion from sloping pasture lands in the Ethiopian highlands», Australian Journal of Experimental Agriculture, 37, 1997, p.421-430.

PAFILIS P., ANASTASIOU I., SAGONAS K., VALAKOS E.D., «Grazing by goats on islands affects the populations of an endemic Mediterranean lizard», Journal of Zoology, 290, 2013, p.255-264.

PALOMBO M., «Insularity and its effects», Quaternary International, 182, 2008, p.1-5.

PAPACHRISTOU T.G., PLATIS P.D., NASTIS A.S., «Foraging behaviour of cattle and goats in oak forest stands of varying coppicing age in Northern Greece», Small Ruminant Research, 59, 2005, p.181-189.

PAPPA M., HALSTEAD P., KOTSAKIS K., UREMKOTSOU D., «Evidence for large-scale feasting at Late Neolithic Makriyalos, northern Greece», in Halstead P., Barrett J., Food, Cuisine, and Society in Prehistoric Greece, Oxford, Oxbow, 2004, p.16-44.

PHOCA-COSMETATOU N., «A zooarchaeological reassessment of the habitat and ecology of the ibex (Capra ibex)», in R.C.G.M. Laurewerier, Plug I., The future from the past: Archaeozoology in conservation and heritage management. Proceedings of the 9th Conference of the International Council of Archaeozoology, Durham, August 2002, Oxford, Oxbow, 2002, p.64-78.

PHOCA-COSMETATOU N., «Initial occupation of the Cycladic islands in the Neolithic: strategies for survival», in Phoca-Cosmetatou, N., The First Mediterranean Islanders: Initial Occupation and Survival Strategies, Oxford, School of Archaeology, 2011, p.77-98.

Peaker M., «Gestation period and litter size in the goat», British Veterinary Journal, 134, 1978, p.379-383.

PRUMMEL W., FRISCH H-J., «A guide for the distinction of species, sex, and body side in bones of sheep and goat», Journal of Archaeological Science, 13, 1986, p.567-577.

RAMÍS D., «Early island exploitations: productive and subsistence strategies on the prehistoric Balearic islands», in Knapp, A.B., van Dommelen, P., The Cambridge Prehistory of the Bronze and Iron Age Mediterranean, Cambridge, Cambridge University Press, 2014, p.40-56.

ROGOSIC J., PFISTER J.A., PROVENZA F.D., GRBESA D., «Sheep and goat preferences for and nutritional value of Mediterranean maquis shrubs», Small Ruminant Research, 64, 2006, p.169-179.

STURDY D., WEBLEY D., BAILEY G., «The Palaeolithic geography of Epirus», in G. Bailey (ed.) Klithi: Palaeolithic settlement and Quaternary landscapes in Northwest Greece. Volume 2: Klithi in its local and regional setting. Cambridge: McDonald Institute for Archaeological Research, 1997, p.587-614.

THORNES J.B., «Modelling soil erosion by grazing: Recent developments and new approaches», Geographical Research, 45, 2007, p.13-26.

TRUMP D.H., Skorba: Excavations Carried Outon Behalf of the National Museum of Malta, 1961-1963, London, Society of Antiquaries of London, 1966.

TRUTMANN C., Diet composition of alpine chamois (Rupicapra rupicapra L.): is there evidence for forage competition to the alpine ibex (Capra ibex L.)? Unpublished Masters dissertation, epartment of Ecology, University of Zurich, 2009.

VIGNE J.-D., Les mammifères post-glaciaires de Corse. Étude archéozoologique, Paris, CNRS, 1988.

VIGNE J.-D., BRIOIS F., ZAZZO A., WILLCOX G., CUCCHI T., THIÉBAULT S., CARRÈRE I., FRANEL Y., TOUQUET R., MARTIN C., MOREAU C., COMBY C., GUILAINE J., «First wave of cultivators spread to Cyprus at least 10,600 y ago», Proceedings of the National Academy of Sciences, 109, 2012, p.8445-8449.

WATTS T., CONLEY W., «Reproductive potential and theoretical rates of increase for feral goat populations», Journal of Wildlife Management, 48, 1984, p.814-822.

WEBSTER G.S., A Prehistory of Sardinia 2300-500 BC, Sheffield, Sheffield Academic Press, 1996.

WILCOVE D.S., ROTHSTEIN D., DUBOW J., PHILLIPS A., LOSOS E., «Quantifying threats to imperiled species in the United States», BioScience, 48, 1998.p.607-615.

YOCOM C.F., «Ecology of feral goats in Haleakala National Park, Mauii, Hawaii», The American Midland Naturalist, 77, 1967, p.418-451.

YONG-ZHONG S., YU-LIN L., JIAN-YUAN C., WEN-ZHI Z., «Influences of continuous grazing and livestock exclusion on soil properties in a degraded sandy grassland, Inner Mongolia, northern China», Catena, 59, 2005, p.267-278.

ZEDER M.A., HESSE B., «The initial domestication of goats (Capra hircus) in the Zagros Mountains 10,000 years ago», Science, 287, 2000, p.2254-2257.

ZEDER M.A., LAPHAM H.A., «Assessing the reliability of criteria used to identify postcranial bones in sheep, Ovis, and goats, Capra», Journal of Archaeological Science, 37, 2010, p.2887-2905.

ZEDER M.A., PILAAR S.E., «Assessing the reliability of criteria used to identify mandibles and mandibular teeth in sheep, Ovis, and goats, Capra», Journal of Archaeological Science, 37, 2010, p.225-242.

Table des illustrations

Légende Figure 1: A typical Mediterranean scene: goats grazing and browsing Mediterranean scrub on the Aegean island of Gokceada (Imbros).Photograph by S.E.P.B.
Fichier image/jpeg, 704k
Légende Figure 2: Sites and locations mentioned in the text.
Fichier image/jpeg, 121k
Légende Figure 3: Relative proportion of overall faunal assemblages of selected LN and FN Aegean sites comprised of sheep/goat, with island sites vs mainland sites distinguished with mainland sites in black, island sites in white and island vs mainland mean values indicated via small hashes and large hashes respectively.Data from Halstead, 1996 and Phoca-Cosmetatou, 2011.
Fichier image/jpeg, 87k
Légende Figure 4: Variable preference for four species of Mediterranean shrub amongst sheep (open circles) and goat (closed circles) over a ten-day trial period. After Rogosic et al., 2006: Figure 1, reproduced with permission.
Fichier image/jpeg, 208k


Center for Cultural Analysis, Rutgers, The State University of New Jersey, Piscataway, NJ, USA

Department of Anthropology and Department of Geography, University of Georgia, Athens, GA, USA

© CNRS Éditions, 2016

Conditions d’utilisation :