The History of African Village Chickens: an Archaeological and Molecular Perspective

The history of the introduction and dispersal of village chickens across the African continent is a subject of intense debate and speculation among scholars. Here, we synthesize and summarise the current scientific genetic and nongenetic knowledge in relation to the history of the species on the continent. Sociocultural, linguistic, archaeological and historic data all suggest a complex history for the species in Africa, characterized by multiple maritime and/or terrestrial introductions over time and several dispersal routes towards and within Africa. Molecular genetics information supports these observations and in addition suggests possible Asian centers of origin for African domestic chickens, including South Asia and Island Southeast Asia. However, both sets of data were until now too limited in their geographic scope, both within Africa and in comparison with chickens from Asia, to unravel the history of the species in detail. We anticipate that further continent-wide studies combining archaeological, ancient and/or modern genetic information may shed new insights on the history of the species. These will contribute to a deeper understanding of the history of trading networks and human interactions within Africa and between African and Asian societies, at the root of the development and expansion of African civilizations.

the history of the species. These will contribute to a deeper understanding of the history of trading networks and human interactions within Africa and between African and Asian societies, at the root of the development and expansion of African civilizations.
Keywords Africa . Chicken . Gallus gallus . Migration . Trading routes Introduction Domestic chickens are closely associated with humans, and they rely entirely upon humans for their dispersal and indirectly for their survival. The species are therefore important biological markers of agricultural, trade and cultural contacts between societies and civilizations. They are present across the African continent where free-range, scavenging village chickens are found in all agroecological zones, ranging from villages in the humid and subhumid tropical rain forests of West and Central Africa to the temperate highlands of East Africa and the arid and semi-arid regions of the Sahel and Kalahari deserts (http://dad.fao.org/; DAGRIS 2007). With an estimated total population of 1.6 billion at the end of 2010 (FAOSTAT 2012), they are the most abundant livestock species in Africa, contributing to a significant part of the continent's agricultural economy.
Although the African continent is rich in galliform species, the recognized main wild ancestor of domestic chickens, the red junglefowl Gallus gallus, is endemic to sub-Himalayan northern India, southern China and Southeast Asia (Delacour 1977), where the putative centers of domestication of the species are present (Tixier-Boichard et al. 2011). Therefore, domestic chickens, though abundant on the continent, are an introduced species from Asia. Whether or not the species was domesticated in a single geographic area in Southeast Asia (Fumihito et al. 1994(Fumihito et al. , 1996 or across the geographic range of the wild ancestor in both South and Southeast Asia (Liu et al. 2006, Kanginakudru et al. 2008) is still disputed today.
Opinions among scholars are divided concerning when the species arrived and the routes by which chickens entered and dispersed across the continent (Gifford-Gonzalez and Hanotte 2011). Archaeological data are patchy (Fuller et al. 2011) while molecular genetic information is just starting to emerge (Gifford-Gonzalez and Hanotte 2011).
In this paper, we review the information available today (sociocultural, linguistic, archaeological and molecular genetic) on the arrival and movement of chickens across the African continent, providing a proxy for our understanding of ancient and modern human interactions and migrations, including trading routes within Africa and between the African continent and civilizations across Asia.

Sociocultural and Linguistic Evidence
Although the primary reason for the introduction of chickens in Africa, whether ritual or alimentary, has remained a subject of active supposition (MacDonald and Edwards 1993), their arrival on the continent ushered in a major milestone in agricultural history and revolutionized the culture of most African societies. Not only were chickens an additional source of valuable animal protein with low husbandry requirements, they were also easily available as sacrificial animals for various sociocultural functions and economic transactions. More particularly, MacDonald (1992MacDonald ( , 1995a recognizes that in some modern-day African societies, the sociocultural significance of chickens is exceptional, indicating either a late arrival and quick integration of the species into African customs and traditions or a long-term presence on the continent (Table 1). Such high sociocultural significance raises important questions, such as when and why the species became so crucial in mythology, ritual and iconographic contexts (Guèye 1998;Sonaiya et al. 1999;Kondombo et al. 2003).
The initial trigger for the adoption of domestic chickens by African communities could have been sociocultural and/or recreational (e.g., cockfighting) rather than as a new source of food. Textual records indicate that chickens were already present in Egypt by the time of the Third Dynasty of Ur (c. 2,113-2,006 BC), and the oldest recognizable occurrence of the bird in art dates from the second half of the fourteenth century BC (Haller 1954;Smith 1965). Houlihan and Goodman (1986) suggest, however, that the first evidence for its consumption comes from the tomb of Petosiris, which dates to the fourth century BC, though this suggestion has been challenged by MacDonald and Edwards (1993). Houlihan and Goodman (1986) further suggest that chickens were probably brought to Egypt as creatures of curiosity and admiration because of their unusualness and that they did not become a regular feature of the Egyptian farmyard until the Ptolemaic period (304-330 BC). Domestic chickens could have been introduced to the African continent several times, fulfilling different societal needs at different time periods.
Linguistic evidence not only supports a deep embedding of chickens in African cultures and traditions, but also agrees with an early introduction and a complex history of arrival and dispersion of the species. More particularly, Williamson (2000) suggests that different geographic distributions, and spread of the three major roots of the word "fowl" across three language phyla, represent at least three separate introductions of domestic chickens to West Africa: two across Central Africa from the East Coast of Africa and one from North Africa across the Sahara. Blench (2008) noted that the origin of the common Malagasy name for chicken, akoho, cannot be assigned definitively to a Bantu or Austronesian language group. Rather, from his analysis, he concludes that all the terminologies relating to domestic animals in Madagascar appear to be derived from languages of the Swahili, with the Comoros Islands being a possible original source of chickens found in Madagascar.

Archaeological and Historic Evidence
Archaeological evidence for the occurrence of domestic chickens across Africa is summarised in Table 1 and Fig. 1. The data are relatively few, which may reveal that either the domestic species was uncommon in historic times or, perhaps more likely, that the species has remained largely unreported, although it may have been present in archaeological sites. Indeed, from an archaeological point of view, the situation is complicated by the osteological similarities between domestic chicken and African wild galliforms such as the francolin Francolinus sp. and guinea fowl Numididae sp. (MacDonald 1992). Guinea fowl is both a domestic and wild bird in most parts of Africa, and their skeletal remains, as well as those of the francolin, often occur concurrently with those of domestic chickens on many archaeological sites (MacDonald 1992). A protocol to discriminate bones of large galliforms does exist, but vital diagnostic parts are rare in most archaeological sites due to postdepositional processes that fragment fragile avian bones and prevent exact identification (MacDonald 1992).
The fourth to third millennia BC have been advanced as the period for the first arrival of domestic chickens in Africa (Chami 2001(Chami , 2007. However, these dates have been disputed as the osteological remains used in their calibration are nondiagnostic according to the recommended criteria (MacDonald 1992; Dueppen 2011). In North Africa, domestic chickens are depicted in Egypt from the second millennium BC (Houlihan and Goodman 1986). In sub-Saharan Africa, the earliest widely accepted evidence of domestic chickens dates to the mid-first millennium AD in Nubia and the late first millennium AD in East and South Africa (e.g., Horton and Mudida, 1993;Plug 1996a, b). In West Africa, several sites attest to the presence of chickens in the late first millennium AD (Table 1). More particularly, while in the Voltaic region evidence for chickens is scarce, researchers working at Daboya (Ghana) have suggested the presence of chickens throughout the first millennium AD (Shinnie and Kense 1989). Until recently, the oldest known, securely identified remains in the western part of the continent were dated to AD 500-850 from Jenne-Jeno in Mali (MacDonald 1995b). New evidence from Kirikongo (Burkina Faso) and Daboya indicates that chickens were already a significant component of the Iron Age economy in West Africa by the sixth century AD at the latest and probably several centuries earlier (Dueppen 2011) (Fig. 1).
Using geographic distribution and dating of the purportedly most ancient, undisputed archaeological evidence in the form of skeletal remains (c. 1,567-1,320 BC),  (1993) further suggest an independent diffusion of chickens into West Africa through the Sudano-Sahelian belt from the East African coast occurring much later, perhaps related to Indian Ocean trade networks. However, increasing evidence for greater antiquity than previously thought for the Indian Ocean trading network, based largely on the presence of Asian plant domesticates in Africa and of African domesticated plants in South and Southeast Asia, support an early possible entry point of domestic species along the East African coast (Fuller et al. 2011). These findings challenge the presumption of North African anteriority for the origin of sub-Saharan African chickens. A component of this early maritime trading network has been described in some detail in the Periplus of the Erythraean Sea from the early first century AD (Casson 1989;Cappers 2006). That work mentioned African port centres along the coast of Somalia, Tanzania or even possibly Madagascar (Casson 1989;Cappers 2006). Chami (1999) has also reported evidence confirming the great antiquity of trading networks in the Indian Ocean interaction sphere, lending support to the observations of the Periplus. As summarised in Blench (2010), it is likely that there were regular contacts between Island Southeast Asia and the East African coast prior to the first centuries AD, independent from the establishment of Austronesians in Madagascar around the fifth century AD  (Blench 2010). Lastly, the inclusion of East Africa into a maritime trading network extending eastward to Southwestern China is illustrated by the visit of Chinese fleets under the command of Cheng Ho off the coast of East Africa, although at a much more recent date, the beginning of the fifteenth century AD (Duyvendak 1939).
Undoubtedly, the complex and ancient interactions resulting from the Indian Ocean trading networks must have facilitated maritime and terrestrial intercontinental translocations of several domestic and nondomestic plant and animal species among Africa, the Middle East and South, Southeast and East Asia Fuller and Boivin 2009;Fuller et al. 2011). It would not be surprising that domestic chickens would have been an intrinsic part of these translocation packages in historic times and possibly much earlier than the introductions via the north of the continent.
The arrival of European settlers as part of the development of the terrestrial and maritime empires of several European nations across Africa and Asia (Portugal, Britain, etc.) from the fifteenth century AD onwards provided further opportunities for the arrival and movement of chickens within Africa. Also, since the second half of the twentieth century, the African continent, as other parts of the world, has witnessed the arrival of exotic chicken breeds developed for higher productivity (Kitalyi 1997), including for crossbreeding purposes with local flocks to improve meat and egg production. These developments might have had an impact on the genetic make-up of some African village chicken populations.

Molecular Genetic Perspectives
No molecular genetic study has so far attempted to illustrate and understand the genetic diversity and origin of African indigenous chickens at the continental level (Gifford-Gonzalez and Hanotte 2011). Also, there are no published data on ancient DNA sequence information for the same species in Africa. Currently, all molecular genetic studies have been limited to either individual countries or at best a few of them (Table 2 and 3). These genetic studies may be grouped in two separate sets according to the genetic markers of interest: microsatellite loci (Table 2) and mitochondrial DNA D-loop sequences ( Table 3).
The genetic diversity and relationships between chicken populations from several African countries have now been investigated using autosomal microsatellite markers ( Table 2). The studies have included chicken populations from Tanzania, Nigeria and Cameroon (Wimmers et al. 2000); Kenya and East Africa (Sudan, Ethiopia, Uganda) (Mwacharo et al. 2007); Zimbabwe, Malawi and Sudan (Muchadeyi et al. 2007); Ethiopia (Dessie 2007;Hassan et al. 2009;Dana 2011;Goraga et al. 2011); South Africa (Mtileni et al. 2010); Benin ; Ghana (Osei-Amponsah et al. 2010); Egypt (Eltanany et al. 2011); and Benin, Côte ďIvoire, Ghana, Morocco and Cameroon (Leroy et al. 2012). Although the markers used in these studies overlap to a large extent, with the recommended set of ISAG-FAO microsatellite markers (FAO 2011), no common samples or populations have been used across studies, and this constrains continent-wide analysis.
However, a few important general findings are emerging. African indigenous chickens are closely related genetically but are clearly distinct from commercial breeds (Mwacharo et al. 2007;Leroy et al. 2012). Their genetic diversities are roughly comparable across studies/populations and correspond to the values usually observed in other free-range scavenging chicken populations worldwide, but they are much higher than those observed in commercial breeds (e.g., Wimmers et al. 2000;Osei-Amponsah et al. 2010;Leroy et al. 2012). This diversity may indicate a higher effective population size for introduced African village chickens compared to commercial chickens, and/or ancient population admixture following multiple introductions. Genetic divergence between chicken populations is generally weak (e.g., Osei-Amponsah et al. 2010;Youssao et al. 2010), and it seems to reflect geographic distances between populations (see Mwacharo et al. 2007). However, Goraga et al. (2011) report genetic distinctions between groups of chicken populations from Ethiopia and Mwacharo et al. (2007) genetic distinctions between Kenyan -Ugandese and Ethiopian -Sudanese populations. Also, Leroy et al. (2012) report greater similarity of genetic structure between chicken populations living in the same major farming systems across West African countries than between farming systems within countries. These observations are compatible with a history of multiple introductions of different genetic pools of the species into the regions, chicken movement along specific trading routes, and/or long-term geographic isolation of these populations. Evidence of genetic introgression between commercial and local populations has recently been demonstrated using microsatellites for populations from Morocco and Cameroon, but not for populations from Benin, Ghana and Côte d'Ivoire (Leroy et al. 2012). Such introgression could be related to the long-term breeding improvement programmes that involve the supply of commercial breeds to local farmers for crossbreeding purposes. However, the impact of such introgressions is limited, possibly due to poor adaptation of exotic birds to village conditions and the preference of consumers for local chickens.
Only a few molecular genetics studies have now been published on African chickens using maternally inherited mitochondrial DNA (mtDNA) (Razafindraibe et al. 2008;Adebambo et al. 2010;Muchadeyi et al. 2008;Mtileni et al. 2011;  Mwacharo et al. 2011) (Table 3). Fortunately, these studies have all analysed a hypervariable region of the molecule, the D-loop, which allows comparison across studies. Unfortunately, there is no standardized nomenclature for the chicken mtDNA haplogroups identified across the African studies. So far, the largest study is the one by Mwacharo et al. (2011). It includes mtDNA sequences from 512 domestic chickens from four East African countries (Kenya, Ethiopia, Sudan and Uganda). For the sake of comparison, we have therefore adopted the nomenclature of Mwacharo et al. (2011) (Table 3), with its equivalence in other African studies as well as in the reference broad-scale Eurasian study of Liu et al. (2006) presented in supplementary Table S1. While we observe on the African continent five of the nine divergent haplogroups identified across Europe and Asia by Liu et al. (2006), two haplogroups, A and D, dominate the continent (Table 3). Haplogroup D clearly is found in all the African countries studied and is the most common in all countries with the exception of Madagascar, Zimbabwe and eastern Kenya. It is the only haplogroup so far identified in West Africa, represented there by Nigeria (Adebambo et al. 2010). The next commonest haplogroup is A, which is absent from Uganda, Sudan, Nigeria and Ethiopia but is the commonest in Madagascar, Malawi, Zimbabwe and eastern Kenya. The other haplogroups are observed at very low frequencies in all studies, being present in only one or two birds.
Through a detailed analysis of the geographic distribution of haplogroup A and D across East Africa, Mwacharo et al. (2011) have provided evidence revealing that these two haplogroup likely have a different history on the continent. It has been proposed that African haplogroup A originates from Southeast and/or East Asia (Muchadeyi et al. 2008;Mtileni et al. 2011;Mwacharo et al. 2011) and that it would have been the result of a maritime introduction along the East African coastline (Mwacharo et al. 2011) including Madagascar (Razafindraibe et al. 2008). Following the known history of the peopling of Madagascar, Razafindraibe et al. (2008) have suggested that its origin might be the Indonesian islands. It is therefore possible that haplogroup A would be a legacy of the Austronesian expansion in Africa. However, so far, no studies have attempted to address the relationship between African haplogroup A and the haplogroups observed in local chickens from the Indonesian islands.
Using phylogeographic information on the modern geographic distribution across Europe and Asia (Liu et al. 2006), the Indian subcontinent has been proposed as the initial center of origin for the haplogroup D found in Africa (Muchadeyi et al. 2008;Mtileni et al. 2011;Mwacharo et al. 2011). Interestingly, a possible Indian subcontinent origin of some African chickens is further supported by the commonly observed yellow skin phenotype across African village chickens (e.g., FAO 2009; Dana et al. 2010a;Youssao et al. 2010;Daikwo et al. 2011;Melesse and Negesse 2011;El-Safty 2012). This phenotype has recently been shown at the molecular level to be a legacy of successful introgression of the grey junglefowl Gallus sonneratii, a yet-tobe-domesticated Gallus species with restricted geographic range on the Indian subcontinent, into domestic chicken (Eriksson et al. 2008). However, whether or not these "crossbred" chickens of Indian origin reached the African continent through a maritime and/or terrestrial route remains unknown.
The possible center(s) of origin for chickens harbouring the mitochondrial DNA of haplogroups B, C and E, observed so far at low frequencies on the African continent (Table 3), is more speculative. Haplogroup B and C may have reached Africa following recent introductions of improved commercial chickens (Mwacharo et al. 2011). This suggestion is supported by the presence of identical or closely related haplotypes belonging to these haplogroups in European and commercial chickens (Muchadeyi et al. 2008;Dana et al. 2010b). Using microsatellite markers, Leroy et al. (2012) also demonstrated gene flow between commercial and local chickens in some sub-Saharan African countries. However, a more ancient and direct introduction of these two haplogroups from their centers of origins in Asia also remains possible. More data are required to further clarify the origin(s) of these two haplogroups on the African continent.
Haplogroup E is not observed in commercial chickens (Muchadeyi et al. 2008;Dana et al. 2010b). Both Muchadeyi et al. (2008) and Mwacharo et al. (2011), using the study of Liu et al. (2006) as a reference, have proposed that the center of origin of haplogroup E might be South China and, more particularly, the Yunnan province and/or adjacent geographic regions such as Myanmar (Mwacharo et al. 2011). The routes of introduction of this haplogroup into the African continent remain unknown, but its worldwide geographic distribution today (Liu et al. 2006) suggests that it was likely the consequence of a maritime introduction to the African continent. It raises the interesting hypothesis that its arrival might have followed the Chinese maritime trading expeditions across the Indian Ocean (Duyvendak 1939;Beaujard 2005;Mwacharo et al. 2011).

Conclusion and Future Directions
Sociocultural, linguistic, archaeological, historic and molecular genetic evidence are all in agreement in revealing a complex history of domestic chickens on the African continent, which may include multiple origins from the geographic range of the wild ancestor in Asia, multiple entry points and multiple routes of dispersion. Archaeological data support a relatively ancient presence of this domesticate in Africa, although several thousand years more recent than the earliest evidence of domesticated ruminants on the continent (Gifford-Gonzalez and Hanotte 2011). Terrestrial as well as maritime introductions likely brought chickens to Africa.
However, our knowledge of the history of African village chickens is still in its infancy with several important unknowns. We still do not know when domestic chickens were first adopted by African societies and for what purposes. We have, so far, little information on the likely number of waves that brought chickens into the continent and their time sequence. Also, the terrestrial and maritime dispersal routes of domestic chickens from their centers of origin in Asia to Africa remain largely speculative. Last but not least, our knowledge of the putative entry points and movements of chickens within the African continent is similarly incomplete.
We believe that the answers to these questions will rely on new archaeological and genetic information generated from inside and outside the African continent, with the most promising studies combining calibrated dating of archaeological bones, ancient DNA sequence information and joint analysis of archaeological and molecular genetic diversity data across Africa, Europe and Asia. New genetic marker analysis (autosomal markers, full genome and mitochondrial DNA sequences) will undoubtedly provide further information, as has now been illustrated for other African livestock species (e.g., cattle; see Gifford-Gonzalez and Hanotte 2011). In this context, it is important to take note of the fact that we do have at best incomplete molecular genetic information on the modern genetic diversity of indigenous chickens from the Arabian Peninsula, the Middle and the Near East, most of the Indian subcontinent, and large geographic areas in mainland and island Southeast Asia (e.g., Laos, Cambodia, the Philippines, Indonesia). Also, within the African continent, information on the genetic diversity of indigenous chickens is still too fragmentary to allow a comprehensive genetic analysis throughout the continent. Lastly, as reviewed in detail by others (Blench 2003;Boivin and Fuller 2009;Fuller and Boivin 2009;Fuller et al. 2011), chickens were only one of the agricultural species being voluntarily traded between Africa and Asia. Combining different lines of information from other agricultural and nonagricultural organisms that may have travelled as companion species will likely provide new insights and hypotheses about the history of African domestic village chickens.