1Chickens (Gallus gallus) are the most common bird in the world. They are (and were) an attractive, adaptable, easily transported domesticate with much to offer in the ways of symbolism, entertainment and desirable products. In addition to their eggs and meat, they provide other useful materials: dung and feathers, which were likely to be of utility. Chicken dung is still sought after by gardeners, and a steady supply of it might have helped to support some forms of oasis agriculture. Although chicken management does require the use of specific husbandry skills, the birds can be raised in a range of environments and large quantities of water, space and other resources required by many mammalian species are not necessary for their upkeep. Despite these clear advantages and the praise heaped upon them by ancient writers, North African chickens remain largely uninvestigated from an archaeological perspective.

2Chickens are generally considered to have been introduced into North Africa in the early 1st millennium BC by Phoenician traders and colonists, having been present in the Eastern Mediterranean from the late 2nd millennium BC (Macdonald and Edwards 1993, 589; Nobis 1999, 584-585; Becker 2012-2013). However, their early presence has largely been identified through representations within tombs and as figurines (Fantar 1970; Camps 1992; Warden et al. 1990). Cockerels in particular have been considered in apotropaic roles; for instance, in protecting the people of Cyrenaica from the deadly basilisk or even as a form of embodiment for the dead (Pliny, N.H., 8.33; Camps 1992; Peyras 1993). However, in the later Roman and Byzantine periods, studies of animals in North Africa (where they have taken place) have concentrated upon larger, meatier domesticates in keeping with the greater economic focus of the research of these periods1. To some degree, the lack of scholarship on the roles of chickens and other poultry species in North Africa is the result of a combination of factors such as excavation and sampling strategies, a strong analytical focus within zooarchaeological study upon mammalian species, and a general perception of larger, meatier animals as more worthy of consideration.

3Across the continent as a whole, the chicken has been widely recognised as an important staple of African society, with studies considering genetic, linguistic, archaeological and historical data2. Chickens have also been used as one of many ways of examining long-distance relations and transmissions between Asia and Africa (Macdonald K.C. and R.H. 2000). It is then surprising that North Africa is not well-represented in these studies. This blind spot has fostered the view that there is no evidence or possibility that chickens may have made it into the Sahara (which they did) or were translocated to sub-Saharan Africa.

4Here, we use faunal and documentary evidence to examine past human-chicken relationships across North Africa and more fully illustrate the spectrum of these interactions. For both disciplinary and historical reasons, we make a distinction between Egypt and North Africa (by which we mean the lands within and immediately adjacent to modern Morocco, Algeria, Tunisia and Libya). Each strand of our investigation independently provides a slightly distorted vision of human-chicken relationships which, even when combined with other evidence, is frustratingly piecemeal. However, we demonstrate the significance of the chicken to the peoples of North Africa and show that they did not perceive chickens in consistent ways; instead, there was a diverse range of continuously developing relationships.

Sources of evidence

5In assembling this article, we have reviewed all available published faunal reports and some unpublished faunal reports from North Africa. These represent 43 archaeological sites and a total of 125 phased assemblages. Additionally, we gathered partial information from a further five sites, either from publications or personal communications, and we are aware of a further eight assemblages from the last two decades which remain unpublished. In total, there are 6,164 published chicken elements. This accounts for 7 % of the total number of identified specimens across all assemblages, and 87 % of all identified avian bones. In addition to the effects of the usual taphonomic issues, chickens were probably underrepresented in many of these assemblages in comparison to their mammalian counterparts because sieving wasn’t used and avian material was frequently not analysed3. In order to recover bird remains from archaeological sites, Serjeantson recommends that a mesh size no larger than 4 mm is used (Serjeantson 2009). Considering the lack of sieving, the number of chicken remains recovered is remarkable. Avian skeletal anatomy includes elements and morphology which are different and distinct from mammals. As a result, they can be challenging for an analyst trained primarily on mammalian taxa to identify. Another issue which bears mention is the substantial difficulty involved in distinguishing chickens (Gallus gallus) from guineafowls (family Numididae) and members of the francolin genera (Peliperdix, Dendroperdix, Scleroptila and Pternistis) from skeletal material in the absence of comparative collections and specialist manuals (e.g. Macdonald 1992; Tomek and Bochenski 2009). Although these species are collectively associated with other parts of Africa, trade relationships between these regions have existed for millennia and it is not beyond the realms of possibility that some individuals were traded into and across the Sahara.

6Although the sites we consider here range in date from the 10th century BC onward, the majority of assemblages (and the largest assemblages by NISP) date to the Late Roman or Byzantine period; far less excavation has been conducted on pre- or post-Roman sites. Another issue of note is the general tendency to focus upon coastal sites, especially large urban areas such as Carthage; these are not necessarily the locations where chicken husbandry is expected to be the most prominent (Redding 2015). Despite these issues, the data do show some evidence for the ongoing importance of the chicken to the peoples of North Africa after its introduction.

The introduction and spread of chickens

7On the basis of literary and visual representations, chickens first entered Africa through Egypt sometime in the mid-late 2nd millennium BC (certainly by 1120 BC) and are represented in a scatter of Nilotic funerary contexts thereafter4. The earliest chicken remains date no earlier than the Ptolemaic period (332-200 BC at Tell Maskhuta) (Boessneck 1986)5, and this has been taken to suggest that the bird was an exotic item or exclusive rarity (Macdonald and Edwards 1993)6. By the early 1st millennium BC, the species appears to have been firmly established across the Levant as well as in parts of Turkey and Greece7. It is at this point that chickens may have been transported around the western Mediterranean and translocated to new environments, probably as part of the wave of colonial activity that characterised this period. The earliest reported chickens from the central Mediterranean are from a cremation burial near Castel Gandolfo, south of Rome; these date to the 9th century (De Grossi Mazzorin 2005). In the western Mediterranean, chickens have been positively identified in 9th/8th century deposits from Castillo de Dona Blanca, north of Gibraltar (Hernandez Carrasquilla 1992). In both the Iberian and Italian peninsulas, there are multiple sites with chicken remains present which date to the 1st millennium BC.

8The earliest reported chicken skeletal material from a North African site is from Mogador, a Phoenician colony on Morocco’s Atlantic coast (Becker 2012-2013) (fig. 1, shaded symbols). Forty-eight elements have been identified as certain or likely to be chicken; twenty-four of these were excavated from stratified contexts which date to the 7th-5th century BC. Although a 20th century prison near part of the site may have disturbed some contexts8, this date is consistent with trade along the Mediterranean coastline by seafaring communities. Indeed, even an earlier 9th-8th century date range is plausible, but there are very few excavated contexts from North Africa of this date. The first depictions of chickens in North Africa are slightly later, and date to the 6th century. These are a group of twenty-five small bronze figurines from the sanctuary of Demeter and Persephone at the Greek colony of Cyrene (Warden et al., 1990). Five of these depictions represent cockerels, and two portray hens.

Fig. 1. Archaeological excavations with identified chicken faunal remains.

9In the 1st millennium BC, chickens have only been identified at coastal sites. At Carthage, which is by far the most extensively excavated city in North Africa, twenty-one bones have been identified from the 5th-4th century BC layers from the Hamburg excavations below the Byrsa hill. Chicken elements are also present in excavations on the Rue Ibn Chabaat and they have been identified in a 6th century BC context from the Bir Massouda site nearby (Nobis 1999; V. Linseele, pers. comm.). Other sites with early chicken include Sidi Khrebish (Cyrenaica) and Lixus (Morocco) (Barker 1979; Grau Almero 2001). The coastal distribution may be a reflection of the overwhelming bias in excavations targeting Phoenician and Punic colonies and a lack of attention to possible indigenous settlements, but it is noteworthy that chicken elements were not identified in the pre-Roman levels of the Tunisian site of Althiburos (S. Valenzuela-Lamas, pers. comm., and see in this volume).

10The early 1st millennium AD appears to represent a marked change in the presence of chickens (fig. 1, white symbols): the majority of assemblages contain chicken elements, regardless of location. Further evidence for the spread of the species is present in the Sahara. In the 1st millennium BC, sites such as Zinkekra and Fewet have no evidence of chickens, although the analysed assemblages are very small (Van Der Veen and Westley 2010, 521; Fothergill 2012; Alhaique 2013). However, perhaps in keeping with burgeoning trans-Saharan trade involving the Garamantes, chicken elements have been identified in contexts dating from the 2nd century AD and onward at the urban site of Jarma, the Garamantian capital (Holmes and Grant 2013; Holmes 2013).

11Blench has used linguistic studies to assert that chickens likely spread south of the Sahara from at least the start of the 1st millennium AD; however, this argument is not yet supported by the excavated animal bone evidence (Macdonald and Blench 2000). The site of Kirikongo in Burkina Faso has a number of probable chicken bones as well as eggshell dating to the 1st to 5th century AD, but the first definitively-identified chicken elements date somewhat later to the 6th to 7th century AD (Dueppen 2011). In Senegal, chickens have been identified at Arondo in the 5th-7th century layers as well as in Nigeria at Ngala, phase I (pre-AD 700) (Watson 1999; Linseele, 2007, 24, 336). By the 7th to 8th century, chickens as a species had been embraced along the entire southern edge of the Sahara.

Translocating Chickens

12Chickens are a portable and adaptable bird that can be rapidly transported by ship or over land. The earliest diffusion of chickens in the Mediterranean was almost certainly by seafaring groups (whether they were Phoenicians, Greeks, or other peoples). Considering the likely degree of maritime travel and trade during this period, the issue of whether the chicken entered North Africa from the Levant or Egypt may be a moot point. It is probable that chickens were transported on ships that visited multiple Mediterranean ports; in addition to featuring as part of an initial colonisation package, they were doubtless traded as commodities from time to time.

13Despite the urban, coastal focus of North African excavation, it is difficult to know how long the chicken remained an urban bird in these coastal colonies due to a lack of comparative data. At this time, Althiburos is the only inland site in the Maghreb with a faunal assemblage from the 1st millennium BC. The absence of chicken could imply that they were not traded externally, that there was no interest in adopting them outside of these coastal colonies, or that their presence was not reflected in the excavated materials. The first two options may potentially hint at an indigenous perspective of the chicken as directly associated with and perhaps as a symbol of the colonisers (and bring to mind the symbolic role of the bird in the Greek colony at Cyrene as well as the depiction of a cockerel on the coinage of Himera, Sicily). However, it would be fascinating to know if chickens were husbanded in rural settlements in proximity to the colonies, or if they were adopted by indigenous communities living in closer proximity to the coast. A detailed comparison with other parts of the Mediterranean could provide valuable insight into this possibility. Redding has argued that we should expect to find chickens primarily in villages and farms (Redding 2015). Certainly, the very few excavated rural sites in North Africa do have chicken present in most cases, but none of these date to earlier than the 1st century AD (e.g. Van Der Veen et al. 1996).

14The translocation of the chicken to the Sahara and West Africa presents an entirely different challenge. MacDonald and Edwards initially ruled out the Saharan route, and scholars have generally favoured an Indian Ocean introduction of chickens which then diffused into West Africa (e.g. Macdonald and Edwards 1993; Redding 2015; Mwacharo et al. 2013). However, early evidence for chickens in East Africa is disputed. The presence of chickens at Jarma by the 2nd century AD and evidence of trans-Saharan goods originating from Kissi, Burkina Faso as well as the movement and presence of other domesticates (the horse, donkey, and camel) gives a strong case for reassessment of the possibility of a trans-Saharan translocation of chickens. Transportation of chickens has recently been discussed by Redding with regard to the Middle East and Egypt and by Fothergill and colleagues with regard to trans-Saharan trade (Redding 2015, Fothergill et al. 2020). The comparatively low water requirements and ease with which nomadic peoples are able to transport chickens is a foundational component of Redding’s argument for the adoption of the species in Egypt. Those same features make the chicken an ideal candidate for transport via trans-Saharan routes. Certainly, the rapid spread of the species in some regions reflects the portability and adaptability of this avian domesticate and suggests that a variety of transport mechanisms (e.g. by donkey or camel as described by Redding and Fothergill et al.) could have moved them with success.

Representations of chickens

15There are a number of recognised representations of chickens from North Africa in the 1st millennium BC (fig. 2). The cock in particular had taken on specific symbolic significance with regard to religious and funerary contexts amongst the Liby-Phoenician populations of cap Bon and the Northern Tunisian Tell, as reflected by depictions of chickens from haounet (chambered tombs) (Stone 2007). In Tomb 8 at Jebel Mlezza, two cockerels with exaggerated claws, beak and feathers were painted next to a depiction of a walled town and a tower mausoleum (Cintas and Gobert 1939, 190-197). At Jebel el Mangoub, two cockerels with prominent beaks, conspicuous combs and formidable spurs stand above the chambers of Tomb 26 (Ghaki 1999). Yet another portrayal of a male chicken is present in Chamber 49 of the necropolis of Menzel Temime.

Fig. 2. Archaeological sites with representations of chickens (with assemblages containing chicken-circles).

16An unusual form of chicken depiction is found at Jebel Zabouj and Latrech: line drawings of tower mausolea which are topped by stylised cockerels (Longerstay 1993; Camps and Longerstay 2000). A similar, heavily rendered cockerel atop what appears to be a Garamantian-style pyramid tomb was described in the Tassili n’Ajjer in the Algerian Sahara (Camps 1992, 46-48). Although none of the tombs surmounted by cockerels survives, a poem inscribed on the mausoleum of the Flavii near Kasserine tells of “the trembling wings of the rooster on the top, which, I think, flies above the highest cloud” (Pillinger 2013). It is clear from the structure of the rest of the inscription that this phrase was added after the poem was initially inscribed; it seems as though the Latin poet was recalled by Flavius Secundus and directed to incorporate this important North African motif into the final version of the poem.

17Cockerels also featured on material culture found within tombs, especially the Punic necropoleis of Carthage. These include terracotta tablets and bronze rings from the Colline de Saint-Monique necropolis, on ceramics from Serapeum and Dermech, some as early as the 7th-6th century BC and inscribed glass amulets from Bordj Djedid (Delattre 1894, 431-432; 1898, 625; Vassel 1921, 52).

18Depictions of cockerels in relation to funerary contexts are known to be present in Egypt and Etruria, but they do not seem to be so frequently portrayed amongst other Phoenician/Punic groups in areas such as Sardinia, Spain, and Sicily. The representations of chickens in Northern Tunisia are distinct in appearance, which raises the possibility of a definitively Carthaginian or North African conceptualisation of a chicken during this period (which may have persisted into Roman times). Interment of chickens with the dead has been reported at Setif and Kerkouane, though the accounts are somewhat patchy and the bones have seemingly not been subsequently analysed (Stirling 2004)9.

19Chickens in non-funerary contexts are also well-documented; we have already mentioned the link between Demeter and chickens at Cyrene (Warden et al. 1990). Several votive stele of late Punic and early Roman date are known that feature depictions of cockerels, including at Tipasa, in the vicinity of and in Carthage (including the Ghorfa group) (Vassel 1921, 48-52; Camps 1992, 35-37, 42, 49). These stele are distinctive to the region and mix a range of iconography in which chickens feature prominently. Furthermore, zoomorphic terracotta figurines and clay tablets have been recovered from Carthage and Sousse (Vassel 1921, 48-52). In the Roman period, cockerels continue to be important with regard to some religious activities and beliefs. On the great altar in the Severan Forum of Cuicul, cockerels are represented as one of the animals to be sacrificed (as opposed to pigs) (Ferrié E.B. and J.-N. 1994). Additionally, a statue of Asclepius with a cockerel perched upon his head has been described from the Carthage region, which Benseddik interprets as a specifically Punic representation (Benseddik 1997, 149; Heuzey 1890, 119-121).

20Studies of faunal remains from sanctuaries in North Africa remain rare, and archaeozoological data cannot yet be used to supplement other findings. A lead tablet deposited in a tomb at Bir el-Djebbana mentions the sacrifice of a cockerel in relation to the good fortune of a number of chariot horses (Delattre 1888, 295-296)10. Yet, despite this direct account, there is very little in the way of chicken bone evidence from these contexts, whereas they are common elsewhere in the Mediterranean and can dominate assemblages from sanctuary contexts. For the western Mediterranean during the 1st millennium BC, the Sicilian sanctuary to Zeus Meilichios yielded a few chicken remains and the Punic sanctuary of Demeter at Cuccuru s’Arriu, Sardinia, had many bird bones present. Similar contexts from North Africa might therefore be good candidates for the recovery of chicken elements (Gabrici 1927, 161 cited in Urquhart 2010, 258; Van Dommelen and López-Bertran, 2013, 289). A single chicken bone from the sanctuary of Demeter and Persephone at Cyrene, Libya, is noted by Crabtree (1990, 119), but not discussed further.

21Human-chicken relationships in the Maghreb appear to be intimately tied to the sights and sounds of the chicken and the cockerel in particular; for example, Aelian (3.31) advised that the crow of a cockerel would protect you from the ‘Libyan monsters’ the lion and the basilisk. Aggressive and resplendent, cocks are interpreted as guardians or souls of the dead and as divinely-connected creatures. Although it is tempting to envision colourful cocks strutting around the necropoleis and sanctuaries of North Africa, it is more likely that they were kept as part of flocks and occupied more than a symbolic role.

Chicken husbandry

22Chickens are covered in detail in the works of the agronomists Varro, Columella and Palladius, with the former listing African fowl (which may be guinea fowl) as one of the three main kinds11. Varro further references the Carthaginian writer Mago whilst discussing the feeding of poultry, doves, bees, etc. which suggests that chicken husbandry, as a form of agricultural practice, was established (at least in northern Tunisia) by the mid-2nd century BC (Varro, De re rustica, 3.2.13). From this, we might surmise that chickens increasingly became thought of as a farm animal, kept for eggs and meat, in addition to its ongoing, sacred role from the late 1st millennium BC onward.

23As productive as the villa estates so beloved by the agronomists undoubtedly were, they are ill-suited parallels to many of the arid landscapes of North Africa and the Sahara. However, the chicken was undoubtedly established as an important, reliable staple of communities throughout the Maghreb and various oasis belts to the south, and is common amongst Berber communities12. A hitherto overlooked source of evidence as to the importance of chickens in past societies comes from Egypt’s western desert and the Upper Nile. At Kellis in the Dakleh oasis, one of the most significant finds was the so-called Kellis Agricultural Account Book (KAB) which provides a detailed record of the produce collected from tenants in that settlement during the mid-360s or 370s AD.

24It is perhaps surprising that there are 67 different mentions of chicken and further mentions of eggs in this short book, consisting of eight double-sided palm boards (Bagnall 1997). Along with wheat and barley, chickens were one of the key items of rent collected from tenant farmers. In these accounts, a chicken is worth 3-4 matia of wheat (c. 6-8l.) making them an affordable (and easily taxed) animal, available to all classes of society. In terms of the transaction accounts, chickens are not only the item in which some rents or dues are routinely paid, but are used as a basic unit of exchange which is sometimes broken down into fractions and substitutions are often made, e.g. “Pisecthis… to wit: through Sarapas 1 chicken, in money” and “Kele 1 ¼ chickens to wit: for must [new wine], ¼ chicken” (Bagnall 1997, 111). It is apparent from the dealings recorded in the KAB that the majority of men and women involved in account transactions were from peasant families who were often in arrears. Also repeatedly mentioned in the KAB is a group called ὀρνɩθίοιϛ or ornithioi, ‘poultrymen’, who were paid at regular intervals in barley or cumin (ibid., 129, 153). Although it is not clear precisely which type of cumin is referred to, it is tempting to contemplate the possibility that it was incorporated into the feed of poultry at Kellis, particularly considering the positive effects upon egg laying productivity and immune response which have been documented in modern chicken populations whose diets were supplemented with black cumin (Aydin et al. 2008; Abbas and Ahmed 2010; Al-Mufarrej 2014). If this were the case, supplementation of chicken feed would suggest that poultry husbandry practices at Kellis were complex and that a certain level of expertise existed amongst the ornithioi. The exact duties of these ‘poultrymen’ are unclear, but the term is clear evidence of poultry specialists and that at least some men were involved in the occupation. This runs counter to assumptions that chicken husbandry in the ancient world was generally “women’s work”, as a statement by Palladius implies: “there is not a woman who does not know how to raise chickens” (Op. ag. 1.27.1).

25The comparative presentation of other animals in the KAB is also instructive. The only other likely food animals referred to are piglets (twice) (Bagnall 1997). Chickens make up 71 % of all animals which are mentioned in the accounts. A cache of ostraca (inscribed pot sherds) recovered from Kellis presents a similar proportion of references to chickens. Out of nineteen mentions of animals, twelve refer to chickens (ibid.). Furthermore, one of these ostraca contains an order for twenty eggs, thus lending some support to the otherwise vague assertion that the chickens in the area were of an egg-laying variety due to a paucity of male chickens and because the examined skeletal remains were of a relatively small size (ibid., 41).

26Another cache of 58 ostraca comes from the site of ancient Itos in Upper Egypt (Boyaval 1965; www.trismegistos.org/archive/349). In this case, we have remarkable evidence for a poultry production centre dealing in chickens, guinea fowl and eggs. The ostraca represent 33 orders, 8 receipts and 9 registrations for outgoing products; these are mainly addressed to Haremephis, the poultry farmer and later, the estate manager. In addition to providing further evidence for day-to-day interactions and exchanges involving chickens, it is clear that Haremephis was communicating and coordinating with a network of other poultry farmers, some of whom were from local areas and others who were located in the metropolis.

27A broad spectrum of society is encountered in these oasis accounts. In addition to illustrating the suitability of chickens for the arid and semi-arid environments found throughout North Africa, these accounts show how deeply ingrained chickens were in the everyday lives of oasis inhabitants. Chickens and their eggs may have provided a major source of protein, but they were also inexpensive resources that people could grow and transport, trade, eat and use for sacrifice.

Chicken populations and distribution

28The textual material suggests that chickens were one of the key domesticates in antiquity, but they do not show differences between regions or time periods; indeed they can imply a single timeless form of husbandry. The faunal remains and other archaeological evidence are therefore critical for examining the scale and distribution of the species, despite the often severe excavation biases and provisos on the quality of the archaeological record which were discussed above. Chickens are found at 53 % of all sites and 80 % of sites with more than 1000 bones, but they are by no means evenly distributed. To characterise this, we have created the following index: chicken index = NISP of chickens / (NISP of cattle + NISP of sheep/goat + NISP of pigs) and plotted this against the mid-point of each assemblage13. The resulting pattern is striking (fig. 3). The late antique period (c. AD 400-650) is very clearly marked by a dramatic increase in the proportion of chicken. At the ‘Ecclesiastical Complex’ of Carthage (Phases 1-4: 7th century AD), chicken elements are 30.3% of faunal NISP. Given the substantial taphonomic and sampling bias against the preservation, recovery, and identification of chicken elements, this implies a remarkably high level of chicken presence which is rarely encountered in archaeological assemblages. A correspondence analysis of different taxa by phased assemblage demonstrates the strong skew towards Late Roman and Northern Tunisian assemblages (fig. 4). Although chickens are found everywhere in North Africa by the 1st millennium AD, it is in the coastal towns of the central Mediterranean such as Carthage and Leptiminus that they are most numerous. Our correspondence analysis further highlights that this trend is also associated with higher percentages of pig bones.

Fig. 3. Graph showing average chicken index through time.

Fig. 4. Correspondence analysis of North African faunal assemblages grouped by a) region and b) period.

29In her study of Byzantine animal husbandry, Kroll noted that in the early Byzantine Empire Carthage had the largest percentage of chickens at any site apart from two desert forts on the Dead Sea (Kroll 2012). She also noted that southern Italy also had a high percentage of chickens, with the species accounting for c. 15 % of the NISP of domestic species in both regions. A similar pattern has been noted for north Italy (Rottoli 2014). Kroll attributed this to a form of urban animal husbandry for which chickens, which can be kept in backyards, small living spaces, or even cages, are well-suited. The corresponding increase in pigs (and especially piglets) also implies a distinctive urban dimension to husbandry practices. Yet, it is not clear why this period should be different to earlier periods in which the cities were larger in size. A decrease in the rural supply of chickens could be one contributing factor, but a change in tastes and preferences may also come into play (regardless of whether there was less meat available, people were choosing to consume more chicken). It might be productive to consider whether changes in species representation reflect alterations to the ways in which people bought and consumed meat (and who was able to do so).

Chicken Size

30This high proportion of urban bones provides an opportunity to consider the metrics of chickens from these contexts. A total of 116 elements have been recorded from the sites of Carthage, Jarma, Mogador and Setif (Nobis 1999; Becker 2012-2013; Holmes 2013; Slopsma et al. 2009; King 1991). Additionally, Günter Nobis, in his 1999 faunal report on Carthaginian material provided mean, minimum and maximum values for the GL of the major elements, a further 121 bones (Nobis 1999). These elements are far from representative of North Africa as a whole, with 68 from late antique Carthage alone, but they nevertheless supply a useful way to contemplate changes in chicken husbandry practices.

31Measurements from 1st millennium BC chicken elements are sparse (table 1), but the metrics from Carthage are comparable in size to those from Pantanello (and elsewhere in Italy), although their wing elements are possibly slightly longer (Gal 2010; De Grossi Mazzorin 2005) (fig. 5). Of the four measured elements from Mogador, only the partial humeri have comparable data, with their SD and Bd measurements larger than those reported for Italy (these individuals may be males). From this very small sample, the North African chickens appear to be the largest currently known from the Mediterranean in this period.

32For the 1st to 3rd century AD, there are measurements for six bones from Carthage, six from Mogador and three from Jarma. Two of the bones from Jarma are very small and certainly female (Femur GL: 66.34 and Carpometacarpus GL: 28.97). The third element, a tarsometatarsus, is either a small male or a large female. This is reminiscent of the description of chicken remains in the Dakhleh oasis, which were described as small in size and interpreted primarily as egg layers (Bagnall 1997, 41). Of the Mogador bones, one tibiotarsus is very small (Dd: 10) and the other elements (and those from Carthage) are of similar size to chickens from Italy, but they show a marked increase in size from the earlier chickens; the chickens from late antiquity are of comparable size to the earlier chickens.

Table 1. Metrics from North African chicken elements dating to the 1st millennium BC (all measurements are converted to the protocols described by Cohen and Serjeantson 1996, 109-112).

Fig. 5. Comparison of log-scaled length measurement for a) wing elements and b) leg elements.

33From the limited data, it is difficult to assess characteristics of the North African chicken population, but in Late Roman Carthage there appears to be a sex ratio of approximately two females for every male. This is quite a high proportion of males, and may be indicative that young males were reared for meat, probably along with egg-layers (the high numbers of bones in this period certainly suggests that chickens were eaten regularly). This may be related to the presence of capons (neutered male chickens bred for their fatty meat) for which there is indirect evidence in the form of a curse tablet from Carthage14. Widespread caponisation could account for a higher proportion of elements sexed as male in North African assemblages.

Discussion

34Whereas the Punic iconography of chickens is heavily stylised and linked to the funerary and religious worlds, the late Roman and Byzantine depictions, especially those in mosaics, are far more naturalised and agrarian; the later chicken is much more recognisable with regard to modern conceptualisations of domestic poultry. Mosaics from Tunisia depict plump hens and cockerels in the company of other farmed birds: pigeons, ducks and geese. The trend for increasing numbers and larger sizes of chickens seems to take place within an increasingly domestic context of depiction.

35The apparent growth in chicken production may hint at changes in animal and meat economies. The Late Roman period in North Africa brought important changes, including the adoption of a new religion in the 4th century. Christianity not only introduced new ideas about the role of animals, but banned activities such as animal sacrifice. However, we must keep in mind that the majority of our evidence comes from urban sites. In the small assemblages from the Libyan Valleys, which are mostly late Roman in date, re-analysis by the first author has identified previously unreported chicken bones, but these appear quite small and are relatively few in number. Perhaps urban areas developed different tastes (e.g. for suckling pig as found at Carthage) (Nobis 1999) or became increasingly dependent on urban animal husbandry – to which both chickens and pigs are much better suited than sheep, goats and cattle. Late antique North Africa as a whole became increasingly unstable and independent. Cities shrank in size and large fortifications were constructed. Rural centres developed greater independence and defences were built to protect these areas as well. The patterns present in the chicken data may therefore indicate a fracturing of urban and rural economies and animal-human relationships.

36Yet, it is also necessary to consider these sweeping changes within a broader perspective. Other Mediterranean cities, such as Caesarea, have similarly high numbers of chicken and chicken is abundant at both rural and urban sites in Southern Italy (Kroll 2012). It is not clear at this stage whether the trends in Late Roman North African faunal assemblages result more from regional factors or wider Mediterranean trends. The relationship between urban and rural animal husbandry strategies (as in so much of North African archaeology) remains obscured by a lack of excavated rural and inland sites. Without correcting this bias, it is not possible to elaborate further on this crucial issue. There does, however, seem to be scope for greater exploration of related changes affecting the wider Mediterranean world.

Conclusions

37In this paper, we have used archaeozoological, written and iconographic evidence to examine North African human-chicken relationships and have highlighted a diverse range of interactions. We can now add an important North African dimension to discussion of the translocation of the chicken. Within our data, there appears to have been three loose phases. Firstly, in the early 1st millennium BC, from Egypt and the Levant along the Mediterranean littoral and down the Nile, incorporating the Maghreb and the Italian and Iberian peninsulas. Secondly, in the late 1st millennium BC, inland away from colonial foundations into the Maghreb proper and into the rural economies. Thirdly, it was then carried along trading routes into the Saharan oases, as far as the Garamantian kingdom in Fazzān by the 2nd century AD and, assuming a Saharan instead of East African translocation, into West Africa from Lake Chad to the Upper Senegal Valley by the 6th/7th century AD.

38Chickens are an excellent source of protein, but the other roles which the chicken played in North Africa suggest that this was probably not the sole reason for the movement and adoption of the species. These simultaneous sets of human-chicken relationships shifted in nature over time, but were unlikely to have been monolithic in any sense. Redding has suggested that the chicken was adopted due to the efficiency with which it produces protein (in comparison with mammalian domesticates) (Redding 2015). Whereas Sykes suggests that when chickens were first moved into new areas they would have been exotics and rarely consumed; instead she has argued that cockfighting was a major motivating factor in driving the dispersal of the species (Sykes 2012). With respect to North Africa specifically, we agree that chickens may have been perceived as an efficient and convenient source of protein, but have encountered no evidence for cockfighting15. In common with the rest of the Mediterranean, the majority of our earliest evidence is in the form of either depictions in funerary and/or religious contexts or burials with chickens. We propose that the chicken was viewed in multiple ways (as symbol, sacred, sacrifice, and producer of dung and feathers, in addition to protein and perhaps entertainment) and that this multi-faceted quality ensured its popularity across diverse groups throughout North Africa. In addition, the accessibility and affordability of the chicken probably made the species a very attractive addition to households at different economic levels.

39In the Phoenician and Punic periods, there were small numbers of maritime-travelling chickens that may not have contributed much in a culinary sense, but possessed a crucial role in mediating with gods and the journey to the afterlife. The chicken was apparently perceived as a fierce, noisy and exotic bird with magnificent plumage, sharp spurs and razor talons. Other evidence illustrates the role of the chicken in sacred rites. In Egypt’s western desert and Fazzān, the chicken was an important part of sedentary oasis life, providing a useful source of eggs and meat (and probably dung and feathers) as well as underpinning the local economy. Specialists in poultry ensured the productivity of an affordable and reliable bird, but one that was small in stature. With regard to Carthage and other Late Roman cities, a third series of relationships between urbanite and urban bird is evident. Plump and colourful, these chickens graced the mosaics of the elite and featured in parables from the new Christianity, but they also provided a way of feeding a city with limited space and a substantial maritime population which was becoming increasingly detached from inland communities16.