Version classiqueVersion mobile

Premières communautés paysannes en Méditerranée occidentale

Jean Guilaine
Jean Courtin
Jean-Louis Roudil
et al.

Les débuts de l'élevage

The origins and relations of neolithic sheep and goats in the western mediterranean1

Hans-Peter Uerpmann


Le Néolithique en tant que période préhistorique est défini par la présence des animaux domestiques, particulièrement des ovins et/ou des caprinés. Les ancêtres sauvages de ces animaux vivent au Proche-Orient. Il n’a pas existé une domestication séparée en Méditerranée occidentale. Des éleveurs anciens de mouton sont arrivés dans le Midi de la France avant que le Néolithique cardial n’occupe la zone côtière.

Texte intégral

  • 1 Manuscript finished February 1984.
  • 1 For France see the overview aticles by Guilaine (1977) and Cour-tin (1978).

1Sheep and goat are characteristic animals used and herded by all rural societies around the Mediterranean. Zooarchaeological research has shown that the importance of these animals is not only a trait of the historic and protohistoric period, but that it started with those prehistoric cultures which are commonly called « Neolithic ». During this period, the remains of at least one of the two species are found in almost every site1. In fact, their presence is one of the most important common features of the Early Neolithic all around the Mediterranean and in the neighbouring regions of Europe and Asia. Whereas it remains questionable whether one or the other cultural complex without the remains of domestic sheep and/or goats may be called Neolithic or not, there can be no doubt that all those assemblages have to be designated as Neolithic which are asociated with these domesticates (and -of course-which do not belong to a later prehistoric period).

  • 2 In zoological nomenclature sheep and goats are classified together with some other animals as « cap (...)
  • 3 North Africa, although outside the natural distribution of wild sheep and goat, is another difficul (...)

2It is clear that the Neolothic as a particular phase of cultural evolution cannot be characterised by this single feature alone. However, when approaching things in a pragmatic way, the definition of the Neolithic by the presence of domestic sheep or goats is more easy to apply than the definitions based on the presence of pottery or ground stone tools. The only difficulties in the application of this definition are met with in areas where the wild ancestors of the domestic « ovicaprids2 » occur. These regions do not only contain the present ranges of wild sheep and goats, but also a vast area from where these animals have been expelled mainly by their domestic descendants and the people who exploit them. It is necessary also to consult the remnants from the time preceding the domestication of ovicaprids in order to define their original distribution3.

3The wild goat (Capra aegagrus) is found today in an area which extends from the Aegaean coast of Turkey to the mountains of Pakistan west of the Indus river. In the north, it reaches the center of the Caucasus. To the south, the slopes of the Taurus and Zagros mountains towards the Mediterranean, Mesopotamia and the Irano-Arabian Gulf form the présent limit. At the end of the Pleistocène, and during the Early Holocene, wild goats also existed in the mountains of the Levant. Today they are replaced in the southern part of this area by the Nubian ibex (Capra nubiana). There is no indication that true wild goats ever reached Europe. On the islands of the Eastern Mediterranean, separated from the mainland by waters deeper than the regressions of the Middle and Upper Pleistocene, the origin of some wild goat populations is most probably anthropogenic.

4For wild sheep, the situation is less clear. There still are different opinions on the taxonomy of the genus Ovis. At a time, this genus was divided into more than thirty species. Some authors later unified it under the single name of « Ovis ammon » (e.g. Haltenorth 1963). Today it is usually separated into two species : Ovis ammon in the Old World and Ovis canadensis in Eastern Siberia and North America (e.g. Pfeffer 1967, Cor-bet 1978),. By another approach six or seven forms are distinguished at the specifie level (Nadler et al. 1973). This last approach, which is based on cariotypes, is most suitable for our purposes. In this scheme the ancestor of the domestic sheep has its own specific name (Ovis orientalis). According to Nadler et al. (1973), Ovis orientalis comprises the wild sheep of Western Asia having a chromosome number (2n) of 54, which is identical to the chromosome number of domestic sheep. In the Elburz mountains and in Baluchistan these « Western Asiatic Mouflons » or « Red Sheep » slightly transgrade into another form, Ovis vignei or « Urial », which has 58 chromosomes. The cross-sections of the horns of the rams of urial and mouflon are quite different, and at least in northern Iran the border between the two groups cannot have shifted too much since the end of the Pleistocene (Uerpmann & Frey 1981). Apparently, the Urial was not important for the process of domestication. The Elburz, the deserts of Central Iran and Baluchistan thus form the eastern border of the natural range of the wild ancestor of domestic sheep. In the west there exists a remnant population of Ovis orientalis in southcentral Turkey. The main area, though, extends from eastern Turkey and the highlands of western Iran down to Baluchistan.

5In the Early and Middle Holocene the isolated population of southcentral Turkey must have been connec-ted to the main part of the range further east. In addition, the species extended further to the west reaching at least the western border of the central plains of Turkey. There is also evidence for a southwestern extension of the present area of distribution. Late Pleistocene and Early Holocene finds corne from the transitional zone between the Levantine mountains and the Syro-Arabian Plateau reaching as far down as southern Jordan and the Negev (Uerpmann in Gebel 1981 ; Davis et al. 1982). There is no serious indication that wild sheep ever existed in North Africa. The so called Barbary Sheep (Ammotragus lervia) is not a sheep but more closely rela-ted to the goat, and certainly it did not take part in the process of domestication.

  • 4 Especially postcranial remains of the chamoix (Rupicapra rupicapra) are easily mistaken for sheep r (...)

6The question whether wild sheep ever existed in Europe, is still difficult to answer. No convincing evidence has yet been published to prove that any wild Ovis species lived here at the end of the Pleistocene or during the Holocene. There is no room here to discuss the published evidence in any detail. However, it has to be mentio-ned that the findings, claimed to represent wild sheep, appear to be questionable either in their stratification or in their identification4. The most serious reason to believe in the former existence of wild sheep in Europe is the presence of the mouflon, Ovis « musimon », in Sar-dinia and Corsica. It has long been assumed, however, that this animal was not autochthonous on those islands (Clutton-Brock, pers. comm. in 1973; Vigne 1983). With regard to the depth of the sea between Corsica and Italy it is not likely at all that a modem form of Ovis could have reached the islands during the Late Pleistocene or in the Holocene without human aid. Most probably these animais represent ferai remains of those sheep brought to the Western Mediterranean by early herders at the beginning of the Neolithic period (see also Bunch, T. et al. 1978). Until convincing evidence is provided, one has to consider Europe as being outside the natural range of the genus Ovis.

7One could of course also use the existence of « pre-neolithic » sheep in Southern France as evidence for the presence of wild sheep in that area. However, the logic error inherent to this idea is obvious. The assumption that these sheep were wild because there was no domestication prior to the Neolithic is wrong in this case. The Neolithic there was not defined by its economy, but enti-rely by the existence or non-existence of pottery. The early sheep of Southern France therefore are rather « pre-pottery » than « pre-neolithic ». Except for the single incisor-tooth from the Sauveterrien of Gramari (Poulain 1971), which may well have moved down in the sediments, the contexts of all those sheepremains are dated to the time immediately before or after the turn from the seventh into the sixth millenium be. The domestication of sheep in the Middle East was finished about one thou-sand years earlier (Uerpmann 1979). It is also known that the expansion of early farmers and herders to the west started during the seventh millenium.

8Before 6000 be the islands of the eastern Mediterranean had been reached as well as eastern Greece - leaving no doubt that some form of marine transportation had been used to arrive at the new lands with families and livestock. Similar evidence from the central Mediterranean is increasing, and there is no reason why this movement - quite impoverished at the beginning - should not have reached the northern shores of the Western Mediterranean as well.

9It has already been stated by Ducos (1977) that the pre-pottery sheep from Châteauneuf-les-Martigues do not differ from domestic sheep in any respect. This should be corroborated through metrical analysis of their remains. However, statistical evaluation of the few measurable sheep bones from the « Castelnovien » of Châteauneuf is difficult, because no single element of the skeleton has provided enough measurements to calculate reliable means and standard deviations. One must take resort to methods like the calculation of size-indices (Uerpmann 1979 : app. 2) in order to obtain a body of data allowing for the comparison with other populations of early sheep. Fig. 1 contains a histogram for the distribution of the size-indices of aull sheep-remains from Châteauneuf. The pre-pottery specimens are marked separately at the bottom of the graph. It is clear from this figure that there was no size-difference between the sheep from the Castelnovien and the Cardial-Neolithic levels of the site.

10There is no room here to discuss the meaning of size-differences evolving during the process of domestication and exploitation of prehistoric sheep (cf. Uerpmann 1979). However, it is as well obvious from fig. 1 that primary domestication left quite a marked diminution of body-size in the sheep-populations exploited by Early Neolithic herders in the Middle-East. The upper two histograms of fig. 1 have been taken from Uerpmann (1979 : 98), where their contents is explained in detail. The large « Protoneolithic » sheep-remains of the first histogram come from the sites of Mureybit, Zawi Chemi Shanidar, Karim Shahir, and Asiab. All of them are Early Post-Pleistocene in date, and no climatic shift can be denounced to be responsible for the reduction in size, apparent in the sheep population of the Pre-Pottery Neolithic of Cayönü. Towards the Late Neolithic of the northern Levant there is no further size-reduction visible. The sheep remains represented by the third histogram of fig. 1, mainly come from the Halafian site of Shams ed-Din (Uerpmann 1982) with some few additions from the Amuq sites explored by Braidwood (Stampfli 1983). Whereas these domestic sheep were still as big as the earlier ones from the Pre-Pottery Neolithic, a further reduction in size can be observed with the sheep remains from Pottery Neolithic sites in Iran (histogram 4 of fig. 1). The distribution of size-indices in this histogram is not as regular as in the first three graphs. This may be due to the fact that the sheep-remains unified here are heterogenious, containing a large sample from the Neolithic layers of Beltcave in Northern Iran (Uerpmann & Frey 1981). This site is outside the natural range of Ovis orientalis, and it seems possible that the transgression of the original habitat was another factor favouring a diminution in size of the average animal.

Figure 1 : Distribution of Size-Indices in several Samples of Sheep Bones from Southwest Asia and the Western Mediterranean.

11In this respect it is most interesting that the sheep of Châteauneuf are of the same general size as the Iranian sample. The roughly contemporaneous domestic sheep of the northern Levant were slightly larger, and the same is true for the earlier sheep of the Pre-Pottery Neolithic, which must be regarded as ancestral to all mediterranean sheep. There is no doubt that the Châteauneuf sheep were far away from their wild progenitors. Another difference between the sample from Châteauneuf on one side and the oriental samples on the other is of some importance, too : the histogram for Châteauneuf does not show the long « tail » on the right side of the distribution. This « tail » is an indication for the influence of the wild population, which in Neolithic times still formed some sort of background to the domestic sheep of Western Asia. In Western Europe there was no wild population to provide this background. This is corroborated by the second example from Western Europe given in fig. 1 : the distribution of size-indices calculated for the sheep-remains from Cueva chica de los Murcielagos near Zuheros in Andalucia (Southern Spain) is similar in its width to the histogram given for Châteauneuf. A slight shift towards smaller values is visible, which can be explained chronologically with the slightly younger age and/or geographically with the lower latitude and the warmer climate of the Spanish site.

12In general, these results could be reproduced in a similar way using published measurements of Neolithic sheep-remains from the Western Mediterranean (Boess-neck & Von Den Driesch 1980 ; Perez Ripoll 1980). Unpublished measurements of the sheep-bones from Cueva de la Cariguela and Cueva de Nerja (comp. Uerpmann 1977) fall in the same range of variation. Unfortunately, metrical information is rarely published for Neolithic bone complexes from France. No measurements have yet been provided for the early sheep-remains from Gramari (Poulain 1971), Gazel and Dourgne (Geddes 1980). There is little doubt, however, that these finds will also prove the complete dependence of European Neolithic sheep on the centers of sheep domestication in Southwestern Asia.

  • 5 The situation at Cueva de Nerja is not clear. Both sheep and goal may be present in an « Epipalaeol (...)

13A similar dependance has always been assumed for the domestic goat of Western Europe. Remains of this animal used to accompany the Cardium pottery of the Early Neolithic around the Western Mediterranean, although they are always found in smaller numbers than sheep-bones. No goatbones have yet been described from pre-pottery contexts5. At Châteauneuf the first remains of Capra come from the lowest levels which also contain ceramics. The same observation is reported by Geddes (1980) for Grotte de Gazel.

  • 6 Calculated from the values basic to Diagr. 3 in Uerpmann 1979 : 104.

14The smaller number of goat-bones from most sites makes it even more difficult to analyse their metrical data than in the case of sheep. Even after transformation of all available measurements of goats from Châteauneuf into size-indices, frequencies are too low to use histograms as in fig. 1. Instead, one has to apply cumulative frequency graphs in order to demonstrate and compare the relative size of those goats. This has been done in fig. 2a, where 11 index-values from Châteauneuf are plotted against the smoothed cumulative curves for the relative frequency of size-indices in Early (1) and Late (2) Neolithic goat populations of the Middle East6. As in the case of the sheep, the values for the goats of Châteauneuf fit very well with the Pottery-Neolithic animals from Iran. The same is true for the second example given here, which is again from Cueva chica de los Murciélagos at Zuheros in Southern Spain (fig. 2b).

Figure 2a : Cumulative Representation of Size Indices of the Goat Bones from Châteauneuf-les-Martigues (o) compared to the respective Curves for Early (Aceramic) Neolithic (1) and Late (Pottery) Neolithic (2) Goat Bones from Southwest Asia.

Figure 2b : Cumulative Representation of Size Indices of the Goat Bones from Cuevahica, Zuheros (o) compared to the respective curves for Early Neolithic (1) and Late Neolithic (2) Goat Bones from Southwest Asia.

15Of course size alone is only a very rough indicator in determining relations between animal populations. In the case of the goat one further observation can be added, although this feature cannot be quantified yet : the horncores of female goats from Neolithic sites in the Western Mediterranean usually are « sabre-curved » (see also Boessneck & Von Den Driesch 1980 : 38). By the same time, twisted horns were normal in female goats of Levantine sites (comp. Uerpmann 1983). Apparently the sabre-horned goats of Neolithic Europe separated from the goat-population of the Middle East at an early stage when twisting of the horns was still less developed. With regard to the unsolved question of the origins of the Impressa-Pottery-Cultures in the Western Mediterranean it should be mentioned, therefore, that a direct derivation of the goats, kept by the users of Cardial Ceramic, from a contemporary Levantine stock, is unlikely.

16Unfortunately almost nothing is known about the history of sheep and goat in the Neolithic of Northern Africa. Italy, too, has not yet provided enough data on its Neolithic fauna in order to allow for a satisfactory evaluation of the pattern basic to the spread of domestic animals in this part of the world. At present, the available evidence seems only sufficient to state that sheep and goat were alien to the Western Mediterranean and that they have been brought there from the East as domestic animals by Neolithic man. How this happened and which routes were followed are questions which still require further research.



Boessneck (J.) & Von Den Driesch (A.) 1980 : Tierknochenfunde aus vier siidspanischen Höhlen. Studien uber friihe Tierknochenfunde von der Iberischen Halbinsel 7, 1-83.

Bunch (T.), N’Guyen (T.) & Lauvergne (J.) 1978 : Hemoglobins of the Corsico-Sardinian mouflon (Ovis musimon) and their implications for the origin of Hb A in domestic sheep (Ovis aries). Annales Genet. Sel. Anim. 10, 501-504.

Corbet (G.B.) 1978 : The mammals of the Palaearctic Region. London and Ithaca.

Davis (S.), Goring-Morris (N.), Gopher (A.) 1982 : Sheep bones from the Negev Epipalaeolithic. Paléorient 8, 87-93.

Ducos (P.) 1977 : Le Mouton de Châteauneuf-les-Martigues. In : L’Élevage en Méditerranée Occidentale. Paris, 77-85.

Gebel (H.-G.) 1981 : Aceramic Neolithic. Tübingen Atlas of the Middle East, Map B I 11.

Geddes (D.) 1980 : De la chasse au troupeau en Méditerranée occidentale. Archives d’Écologie Préhistorique 5, Toulouse.

Haltenorth (Th.) 1963 : Klassifikation der Säugeliere : Artiodactyla I. Handbuch der Zoologie 8, Belin, 1-167.

Nadler (C.-F.), Korobitsina (K.-V.), Hoffmann (R.-S.) & Vorontsow (N.-N.) 1973 : Cytogenetic differentiation, geographic distribution, and domestication of sheep (Ovis). Zeitschrifl fiir Saugetierkunde 38, 109-125.

Pellicer (M.) & Acosta (P.) 1982 : El neolitico Antiguo en Andalucia Occidental. In : Le Néolithique Ancien Méditerranéen. Sète, 49-60.

Pérez Ripoll (M.) 1980 : La fauna de vertebrados. In : Marti Oliver (B.) : Cova de l’Or (Beniarrés-Alicante) 2, Valencia, 193-255.

Pfeffer (P.) 1967 : Le Mouflon de Corse (Ovis musimon Schreber, 1782). Mammalia 31, Supplément, Paris.

Poplin (F.) 1979 : Origine du mouflon de Corse dans une nouvelle perspective paléontologique : par marronage. Annales Genet. Sel. Anim. 11, 133-143.

Poulain (Th.) 1971 : Le camp mésolithique de Gramari à Méthamis (Vaucluse) III Étude de la Faune. Gallia Préhistoire 14, 121-131.

Stampfli (H.-R.) 1983 : The Fauna of Jarmo with Notes on Animal Bones from Matarrah, the Amuq, and Karim Shahir. In : Braidwood, L.S., et al. (eds.) : Prehistoric Archeology Along the Zagros Flanks. O.I.P. vol. 105, Chicago, 431-483.

Uerpmann (H.-P.) 1977 : Élevage néolithique en Espagne. In : L’Élevage en Méditerranée Occidentale. Paris, 87-94.

Uerpmann (H.-P.) 1979 : Probleme der Neolithisierung des Mittelmeerraumes. Beihefte zum Tübinger Atlas des Vorderen Orients, Reihe B Nr. 28, Wiesbaden.

Uerpmann (H.-P.) 1982 : Faunal Remains from Shams ed-Din Tannira, a Haitian site in northern Syria. Berytus 30, 3-52.

Uerpmann (H.-P.) & Frey (W.) 1981 : Die Umgebung von Gar-e Kamar-band (Belt-Cave) und Gar-e Ali Tappe (Beh-Sar, Mazandaran, N-Iran) heute und im Spätpleistozan. In : Frey (W.) & Uerpmann (H.-P.) (eds.) : Contributions to the environmental History of the Middle East. Beihefte zum Tübinger Atlas des Vorderen Orients, Reihe A Nr. 8, Wiesbaden, 134-199.

Vigne (J.-D.) 1983 : Les mammifères terrestres non volants du Post-Glaciaire de Corse et leurs rapports avec l’homme. Thèse de 3e cycle, Paris.


1 For France see the overview aticles by Guilaine (1977) and Cour-tin (1978).

2 In zoological nomenclature sheep and goats are classified together with some other animals as « caprini ». The term « ovicaprid » does not exist in taxonomy. Nevertheless it will be used here as a name for the two domestic forms and their wild ancestors.

3 North Africa, although outside the natural distribution of wild sheep and goat, is another difficult area. This is due to the occurrence of other small ruminants with bones similar to those of the ovicaprids. With regard to the poor knowledge about the neolithic fauna of this area, it seems possible that the rarity of sheep and goat there is mainly the result of a certain reluctance to identify their remains. With increasing morphological knowledge this situation may change in the near future.

4 Especially postcranial remains of the chamoix (Rupicapra rupicapra) are easily mistaken for sheep remains.

5 The situation at Cueva de Nerja is not clear. Both sheep and goal may be present in an « Epipalaeolithic » layer, which in addition also provided remains of the domestic pig (Boessneck & Von Den Driesch 1980 : 20ff). The dating of this « Epipalaeolithic » is still unknown (Pelli-cier & Acosta 1982 : 54).

6 Calculated from the values basic to Diagr. 3 in Uerpmann 1979 : 104.

Notes de fin

1 Manuscript finished February 1984.

Table des illustrations

Légende Figure 1 : Distribution of Size-Indices in several Samples of Sheep Bones from Southwest Asia and the Western Mediterranean.
Fichier image/jpeg, 164k
Légende Figure 2a : Cumulative Representation of Size Indices of the Goat Bones from Châteauneuf-les-Martigues (o) compared to the respective Curves for Early (Aceramic) Neolithic (1) and Late (Pottery) Neolithic (2) Goat Bones from Southwest Asia.
Fichier image/jpeg, 80k
Légende Figure 2b : Cumulative Representation of Size Indices of the Goat Bones from Cuevahica, Zuheros (o) compared to the respective curves for Early Neolithic (1) and Late Neolithic (2) Goat Bones from Southwest Asia.
Fichier image/jpeg, 82k

© CNRS Éditions, 1987

Conditions d’utilisation :

Cette publication numérique est issue d’un traitement automatique par reconnaissance optique de caractères.
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search