Did disease constrain the spread of domestic dogs (Canis familiaris) into Sub-Saharan Africa?

Abstract

Dogs are not native to the African continent and appear to have been introduced there from the Middle East. While they spread relatively swiftly across North Africa — perhaps in association with a pastoralist economy — their expansion south of the Sahara appears to have been much slower and more staggered. Although their rarity in many Sub-Saharan faunal assemblages has sometimes been explained on taphonomic grounds, this paper argues that when they entered Sub-Saharan Africa dogs must have encountered a series of endemic diseases for which native mammalian species (including wild canids) acted as reservoirs. Having reviewed the archaeological evidence for their presence in Africa, it then discusses the challenges posed to dogs by five such diseases: canine tryopanosomiasis, canine babesiosis, African horse sickness, canine monocytic ehrlichiosis and visceral leishmaniasis. The paper concludes by considering possible ways in which the disease constraint model advanced here might be evaluated further and asks whether dogs may themselves have posed health risks to people or livestock and whether disease may also have constrained their expansion into other tropical parts of the world.

Le chien n’est pas originaire du continent africain et semble y avoir été introduit à partir du Moyen-Orient. Tandis que le chien se propagea assez rapidement à travers l'Afrique du Nord — peut-être en association avec une économie pastorale — son expansion au sud du Sahara semble avoir été beaucoup plus lente et progressive. Bien qu’on ait parfois expliqué par des raisons taphonomiques la rareté des restes canins dans de nombreux assemblages de faune subsahariens, cet article fait valoir qu’à leur arrivée en Afrique sub-saharienne les chiens ont certainement du rencontrer une série de maladies endémiques pour lesquelles les mammifères indigènes (y compris les canidés sauvages) agissaient comme réservoirs. Après avoir examiné les traces archéologiques de la présence de chiens en Afrique, l’article aborde alors les défis qui leur auront été posés par cinq de ces maladies: la tryopanosomiasis canine, la babésiose canine, la peste équine africaine, l’ehrlichiose monocytique canine et la leishmaniose viscérale. L’article conclut en avançant quelques propositions quant aux moyens par lesquels le modèle de contrainte des maladies avancé ici pourrait être évalué davantage. Nous nous demandons aussi si les chiens eux-mêmes auraient pu poser des risques pour la santé des personnes ou des animaux, et si des maladies ont également pu limiter l’expansion des chiens dans d'autres régions tropicales du monde.

Keywords:

• dogs
• expansion
• Sub-Saharan Africa
• disease challenges
• trypanosomiasis
• babesiosis
• African horse sickness
• monocytic ehrlichiosis
• visceral leishmaniasis
• wild canids

Acknowledgements

I should like to thank Tim Forssman, Kat Manning, Kevin MacDonald, Ina Plug, Sol Pomerantz, Kath Potgieter and Lyn Wadley for answering questions, providing me with access to otherwise difficult-to-obtain literature or commenting on a first draft of this paper. I am also grateful to Rachel King for the maps that accompany it and to Anne Haour for improving the accuracy of the French version of the abstract. The comments of three referees — Roger Blench, Diane Gifford-Gonzalez and Ann Horsburgh — further improved the text, but without incurring responsibility for any of the conclusions reached. Readers should note that, while I cite the results of experimental infections of dogs in this paper, I reject and condemn their use in such studies.

Notes on contributor

Peter Mitchell is Professor of African Archaeology at the University of Oxford, Tutor and Fellow in Archaeology at St Hugh's College, Oxford, and an Honorary Research Fellow of GAES, University of the Witwatersrand. As well as researching the archaeology of hunter-gatherers in southern Africa (where he has excavated extensively in Lesotho), he has a strong interest in the broader, comparative aspects of African prehistory. Recent books include The Oxford Handbook of African Archaeology (2013, co-edited with Paul Lane) and Horse Nations: The Worldwide Impact of the Horse on Indigenous Societies Post-1492 (2015), both published by Oxford University Press.

Notes

1. All radiocarbon dates cited in original sources have been calibrated for this paper using the OxCal programs, IntCal13 for the Northern Hemisphere and ShCal13 for the Southern.

2. A supposed dog canine tooth from an apparently first-millennium BC context at Ukunju Cave on the island of Juani in Tanzania's Mafia Archipelago (Chami 2006: 100) has not been reported in detail. It may thus prove to have no more weight than similar claims for pre-Christian era chickens from Zanzibar, now recognised as an exercise in the over-interpretation of poorly identified faunal remains (cf. Sinclair 2007).

3. Dogs are absent from the large and well-preserved faunal assemblage excavated at the early second-millennium AD town of Mahilaka in northwestern Madagascar (Solomon Pomerantz, pers. comm., 5 September 2014), though reported from Andranasoa at the opposite end of the island at broadly this time (Rasamuel 1984). Their absence from sites dating to the early/mid-second millennium cannot be due to any possible confusion with jackals, which do not exist on Madagascar, and may point to an introduction some time after the establishment of settled village communities from about AD 500 and long after the earliest, apparently forager colonisation ≥2500 years before that (Dewar et al. 2013). The terms for dog used in Malagasy appear to be of Bantu origin, suggesting that dogs were not introduced directly to the island from Southeast Asia (Blench 2008; pace Brown 2011).

4. Recent research indicates that the human parasite T. brucei rhodesiense of East and south-central Africa is not reproductively isolated from T. brucei brucei (Balmer et al. 2011) and that the same is also true of the West African human parasite, T. brucei gambiense (Goodhead et al. 2013). Additionally, there is genetic evidence that T. evansi is not a valid species, but should instead be considered a subspecies of T. brucei, T. brucei evansi, from which it evolved relatively recently (Lai et al. 2008).

5. T. brucei brucei is spread by G. fuscipes, G. morsitans, G. pallidipes, G. palpalis, G. swynnertoni and G. tachinoides, T. congolense by the last five of these, plus G. austeni, G. brevipalpis and G. longipalpis (Hoare 1970).

6. Speculation that an outbreak of Ehrlichia canis was responsible for a decline in African hunting dog numbers reported by Stevenson-Hamilton (1939) in South Africa's Kruger National Park in the 1920s and 1930s (van der Merwe 1959; Pienaar 1963) appears unfounded. No direct evidence of the parasite's presence in hunting dog populations was established, although the disease may have been responsible for the simultaneous deaths of domestic dogs living in the park (van Heerden 1979: 245). Subsequent surveys of African hunting dogs in Kruger have failed to find evidence of exposure to E. canis (Woodroffe and Ginsberg 1997: 71).

7. The dates cited by Faith (2014) for Equus algericus and E. mauritanicus from sites in Morocco are single, generalised radiocarbon dates with large standard deviations for the occupation levels within which the relevant specimens were recovered. They do not directly date the bones in question, something that is also true, to the best of my knowledge, of other specimens from early Holocene Capsian contexts in Algeria and Tunisia.

8. And the very fact that dogs are so widely evident in the faunal record of Eastern, south-central and southern Africa from the mid/late first millennium AD onwards, but not before, is itself suggestive that this pause — like the others I have discussed — is genuine. Certainly, it is extremely difficult to imagine what change in canine behaviour, faunal preservation or the relative identifiability of dog versus jackal bones might otherwise explain this chronological patterning. The broad temporal coincidence with the expansion across these areas of iron-using, relatively sedentary Farming Communities may well have provided longer-lived, larger settlements at which more dogs were kept/had a greater likelihood of entering the archaeological record, but still does not explain their total absence from earlier hunter-gatherer or pastoralist contexts in these parts of the continent.

9. The antiquity of this particular human/canine disease association has recently been revealed by recovery of a calcified cyst of the tapeworm Echinococcus granulosus from the 8000-year-old body of a woman from a hunter-gatherer-fisher context in the Lake Baikal region of Siberia (Waters-Rist et al. 2014).

10. In addition to a number of fox taxa, the relevant species include dholes (Cuon alpinus) in mainland Southeast Asia, Sumatra and Java, maned wolves (Chrysocyon brachyurus) in the grasslands and forests of southern Brazil and the Gran Chaco and bush dogs (Speothos venaticus) and short-eared dogs (Atelocynus melanotis) across Amazonia and much of the rest of Lowland South America north of the Gran Chaco savannas and the Pampas grasslands (Macdonald and Sillero-Zubiri 2007).