Tuesday, 31 March 2009

EATING BATS ON PEMBA ISLAND: A Local Innovation or Cultural Borrowing?

by Martin Walsh

[This is the text of an article orginally published in October 1995 in Mvita: Newsletter of the Regional Centre for the Study of Archaeology in Eastern and Southern Africa (Mombasa), 6: 15-18. I wrote this at short notice during a visit home to Mombasa in September 1995, and still have unpublished notes on the ethnotaxonomy of bats and bat-hunting on Pemba. The historical questions raised in the article are discussed at greater in length in a 2007 paper 'Island Subsistence: Hunting, Trapping and the Translocation of Wildlife in the Western Indian Ocean', Azania, 42 (Special issue: Stephanie Wynne-Jones (ed.) The Indian Ocean as a Cultural Community): 83-113 (with an online appendix: 'Island Mammal Lists and Local Names').]


The Swahili are virtually unique among the peoples of East Africa in their possession of a maritime culture. Their early adaptation to the exploitation of marine resources, coupled with a growing involvement in the trade networks of the western Indian Ocean, both enabled and prompted them to establish settlements along more than a thousand miles of coastline and offshore islands. The only other peoples to share in this spectacular maritime expansion were the Mwani and Comorians, speakers of closely related Sabaki Bantu languages. At the same time the Swahili and their close relatives did not live off the sea alone. They also carried with them an Eastern Bantu inheritance of mixed farming, cultivating grain and root crops and keeping cattle and other livestock.

Though it is sometimes easy to forget, hunting and gathering were also a component in this inheritance and therefore the eclectic economy and varied diet of the Swahili. Not surprisingly, hunting and gathering plays much the same role in rural Swahili communities as it does among other Eastern Bantu, and the main differences (including differences among the Swahili themselves) are ones of degree. Swahili methods of hunting and trapping are little different from those of their mainland neighbours, and the range of permissible animal foods is very similar, except for the Islamic prohibition upon pig-meat (which does not stop pigs being hunted as vermin and fed to dogs). Birds, rodents and herbivores are generally fair game: carnivores, primates, reptiles and amphibians are not. The Swahili, like most East African Bantu, have fewer dietary restrictions than Cushitic pastoralists and mixed farmers and the speakers of Nilotic lang-uages who have been influenced by them (and who, like some northern Swahili clans of Eastern Cushitic origin, do not eat fish). However, they are nowhere near as omnivorous as formerly “pure” hunter-gatherers like the Hadza.

Bats are usually considered inedible by Bantu-speakers, a prohibition which is often rationalised with reference to their nocturnal habits and associations with witchcraft and/or death. The Swahili inhabitants of Pemba island, which lies off the northern coast of Tanzania, provide a striking exception to this rule. Pembans appear to be unique among the Swahili and other people of the East African coast in that they hunt and eat bats, and do so with great relish. Their anomalous behaviour in this respect demands an explanation. It seems extremely unlikely that this is an inherited practice which has survived on Pemba but has been discarded by other Swahili-speakers. Two main possibilities remain: either that Pembans have (for reasons which also need to be clarified) innovated the practice, or that they have, at some point in their past, borrowed it from a people who did hunt and eat bats. As we shall see, it is not easy to decide between these two competing hypotheses or different variants of the borrowing hypothesis. Before discussing these at greater length, however, I will first describe the Pemba practice in more detail.

Hunting Bats for Food on Pemba

At least 11 species of bat (Order Chiroptera), probably more, are found on Pemba island. The Swahili-speaking inhabitants of Pemba refer to them collectively as p’opo. Like most other East African Bantu, Pembans usually classify bats as birds (ndege), though close familiarity with these flying mammals leads some people to question this. Not all bats are hunted for food. The largest, tastiest, and most sought-after species is the endemic Pemba Flying Fox, Pteropus voeltzkowi. The second-largest species on the island, the Straw-coloured Fruit Bat, Eidolon helvum (subspecies helvum), which is commonly thought to be an immature flying fox, is also widely hunted, though said not to be quite so tasty. Other edible bats include Decken’s Horseshoe Bat, Rhinolophus deckenii. The importance of bats in Pemban culture is reflected in the fact that there are more than 20 different names for particular kinds in use on the island. The majority of these are alternative names for the principal edible species, underlining their role in the local diet.

Many of the smaller species, however, are not hunted for food. In some places on the island colonies of bats, visible or invisible, are regarded as spirits or their guardians: the Egyptian or Long-haired Rousette, Rousettus aegyptiacus (sub-species leachii), has been definitely identified as one subject of such beliefs. An extraordinary plague of malevolent spirits which swept throughout Pemba during and after Ramadhan in 1995, and subsequently migrated south to Unguja island, also had a nominal association with bats. These spirits were identified as manifestations of a single evil entity, called p’opo bawa, “the bat’s wing”. This designation is said to be purely metaphorical and based upon descriptions of the frightening shadow cast by the nocturnal attacker and his spiritual legions. This is not the first time, however, that p’opo bawa or a plague of the same name has appeared on Pemba, and it may well be that a more substantive association with bats was once known.

A variety of methods are employed to hunt and trap bats, either during the daytime when they are roosting or at night when they are feeding. Until recently shotguns were widely used by hunters, though the rising cost of ammunition and the increasing difficulty of obtaining guns and licenses for them (especially since the introduction of multi-party politics) has much curbed this practice. The most common bat-hunting method now is to shoot them with a catapult, a favourite occupation of groups of boys and youths. Throwing-sticks are sometimes used, but recognised as much less accurate and effective. A more widespread practice is to trap bats by baiting a tangle of thorn branches (from misoo trees) with ripe fruit and lodging this trap high in a tree on which they are feeding. On the eastern coast of Pemba small groups of men sometimes chop down Borassus Palms containing roosting colonies of the Straw-coloured Fruit-bat, enabling them to kill up to 100 bats in one go. The destructiveness of this practice gives some idea of the lengths to which Pembans will go to obtain bats, which form an occasional substitute for fish in the local diet. After the bats have been skinned and their heads and wings removed large catches are usually shared out among the hunters who then take them home for the family pot (they are boiled like chickens). Surplus bats are given to relatives and friends and only occasionally sold.

Until the 19th century the greater part of Pemba was covered by forests: hence its Arabic epithet, “The Green Island”. Settlement appears to have been concentrated along the coast, where shifting cultivation was practised on the coral rag, and rice and other crops were grown on a more permanent basis in the well-watered valleys which penetrate inland. Given the abundance of natural habitat, bat-hunting was presumably once a sustainable activity. Unfortunately this is no longer the case. The large-scale clearing of the forests for clove plantations and the continuing expansion of population and agriculture have destroyed much of the habitat of tree-roosting species. Uncontrolled hunting with shotguns in the 20th century has taken a further heavy toll upon bat populations, especially that of the endemic Pemba Flying Fox, which is now listed as an endangered species. It remains to be seen whether current efforts by conservationists and the Zanzibar government to ensure its survival will meet with any success. Persuading Pembans not to hunt bats is no easy task.

A Pemban Innovation?

Relatively little archaeological work has been undertaken on Pemba. Although we know that Pemba has been settled by Swahili-speakers since at least the second half of the first millennium, we have no evidence to tell us how long bats have been a component in the Pemban diet. However, we can speculate on the origins of this apparently anomalous practice.

One possibility is that the hunting of bats for food is a Pemban innovation. We then have to ask what prompted the inhabitants of Pemba to modify their dietary practice to include a creature which other Swahili-speakers, not to mention other East African Bantu, classify as inedible. The abundance of large bats on Pemba relative to other natural sources of meat could provide an explanation. Pemba supports a limited mammalian fauna, apparently as a consequence of its distance and long period of geological separation from the mainland (though human impacts upon a more extensive original fauna cannot be entirely ruled out as a contributory factor). There is, for example, only one medium-sized herbivore, the Pemba Blue Duiker (Cephalophus monticola, subspecies pembae). Bats, however, comprise a much larger proportion of the mammalian fauna than they do on the mainland: roughly half of the total number of species of land mammal found on Pemba are bats. This is presumably because of the relative ease with which they have been able to colonise the island and thrive in the absence of significant predators. The Pemba Flying Fox is one such colonist, the westernmost representative of a genus with Austronesian affinities and endemic forms on a number of other Indian Ocean islands.

The Swahili settlers of Pemba encountered an abundance of bats relative to other mammals, and one bat in particular which was larger and had more meat than any other which they had encountered before. Under these circumstances the inclusion of bats in the Pemban diet could be conceived as a simple response to locally available opportunities, initially either as a result of pressure during periods of food shortage or perhaps purely as a means of varying the diet. The inherited classification of bats as birds may have assisted in this transition, enabling Pembans to overcome any scruples they may have originally had as to their edibility. A similar explanation can be posited for another Pemban dietary innovation: the consumption of freshwater terrapins, kobe (specifically the Eastern Hinged Terrapin, Pelusios castanoides). In this case the obvious similarity of these reptiles to edible sea turtles (an early addition to the Swahili diet) may have eased their acceptance as an item of food (other Swahili-speakers and their mainland relatives regard the thought of eating kobe, a term which includes land tortoises, with disgust).

There are a number of problems, however, with this argument. The lack of meat protein sources on Pemba has not seen them switch to eating other mammals normally avoided by the Swahili. Vervet monkeys (Cercopithecus aethiops, subspecies nesiotes), for example, are abundant on the island and readily eaten by Makonde immigrants but not by their Swahili hosts (with the exception of isolated and much-criticised individuals prepared to break this dietary taboo). Meanwhile, other island-dwelling Swahili with similar opportunities for eating fruit-bats have not taken up the Pemba practice. The Straw-coloured Fruit-bats of Unguja island are generally only eaten by people of Pemban origin. Likewise, although Mafia island and the adjoining islets support a species of flying fox related to that found on Pemba, the local inhabitants declare that they are only eaten by visitors from Pemba. While these examples do not rule out the possibility that Pembans innovated in this regard, they do make it rather more difficult to explain.

A Borrowed Practice?

The alternative hypothesis is that the Swahili-speakers of Pemba adopted bat-eating from a people who already hunted and ate them. If this were the case, then the fact that they were able to overcome their inherited scruples suggests that this was the result of prolonged interaction, and possibly the incorporation of a group of bat-eaters into the Swahili population. Who might these people have been? There are at least three main candidates, and the case for each is discussed below.

In one of the published references to bat-eating on Pemba it is remarked that one species was considered a delicacy by the Arabs on the island. This raises the possibility that Pembans have turned to hunting bats under the influence of recent (especially 19th century) Arab immigrants, whose descendants are still to be found scattered throughout the clove plantation areas of Pemba. However, there are obvious objections to this argument. As far as I am aware, the consumption of bats is not common practice in Oman or other parts of the south Arabian peninsula from where Pemba’s Arabs originally came. Moreover, Arab immigrants else-where on the Swahili coast have not introduced this practice, nor are they reported to eat bats at all. Under these circumstances it is difficult to explain how they might have fostered a habit which appears to be deeply ingrained in Pemban culture, as widespread in the traditional fishing villages as it is in the mixed settlements of the plant-ation areas. It seems more likely that the opposite process has taken place, and that Arab settlers have adopted existing Pemban practice.

The second possibility is that Pembans adopted bat-hunting under the influence of an earlier population of hunter-gatherers. The problem with this argument is that at present we have no evidence for the existence of such a population on Pemba, though it is a tantalising thought. We do, however, know that Pembans, in common with other Swahili-speakers, had close contact with hunter-gatherers on the northern coast before they migrated and sailed to the south. A number of Swahili words, including some zoological terms, appeared to have been borrowed from the ancestors of the Southern Cushitic-speaking Dahalo, a group of (former) hunter-gatherers in the north of the Tana River delta. Other words, including the generic name for bats (proto-Sabaki *mpopo), derive from the Eastern Cushitic-speaking Aweera (or Boni), a group of hunter-gatherers who live further to the north, on both sides of the Kenya-Somalia border. If bat-eating had likewise derived from one or other of these groups, then we would expect it to be more widely distributed than it is or to find evidence that it was so in the past. Unfortunately this evidence is lacking, as is information on whether or not the Dahalo and/or Aweera actually eat bats, or have ever done so in the past.

Alternatively we could assume that the early Pembans had separate contact with these or other bat-eating hunter-gatherers on the mainland before they migrated to Pemba. Again, we have no other evidence for this. Most of the Pemba Swahili names for different kinds of bats, including the Pemba Flying Fox, are composite and descriptive terms (like p’opo-maembe, “the mango bat”) whose transparent etymologies do not provide any useful clues in this respect.

A third possibility is that the Pembans derived their bat-eating habit from early settlers from Borneo, in other words the same people who settled the island of Madagascar. Suggestions that these people might have had any direct impact upon the East African coast are usually treated with considerable scepticism. There are, however, an increasing number of indications that this might not be as far-fetched a thesis as previously thought. Swahili has at least a sprinkling of words which appear to have originated on the other side of the Indian Ocean: these include the generic term for fish-poison (utupa) and the plants bearing it (mtupa), and the name for a kind of raft (sap’a in Pemba and other dialects). There is also some botanical and zoological evidence for this connection. The presence of the Wild Banana, Musa acuminata, on Pemba (and nowhere else in the western Indian Ocean) is perhaps the best known example. The problem with most of this evidence, however, is that given our current state of knowledge it is extremely diff-icult to distinguish between direct introductions and later, secondary, introductions from Madagascar, where the Swahili also founded coastal settlements.

Even so it is interesting to note that the Ma’anyan, the Bornean people whose language is most closely related to those of Madagascar, hunt and eat fruit-bats, as do many of the peoples of Borneo and elsewhere in the Indonesian archipelago. Not surprisingly, therefore, bats are widely eaten in Madagascar, while Malagasy-speakers are thought to have taken this practice with them to the Comoro islands, where the largest species, Livingstone’s Giant Fruit-bat (Pteropus living-stonii) is threatened with extinction. Whether or not close relatives of the Ma’anyan and/or Malagasy ever settled on Pemba (bringing with them a propensity for eating bats) remains to be determined: it seems, however, to be a possibility which is worth investigating further.


The origin of the Pemban practice of hunting bats for food is a puzzle which has yet to be solved, though the answer probably lies in one or other of the solutions proposed above. Further comparative ethnographic and linguistic work may provide additional clues, but the most telling evidence probably lies buried in the ground, awaiting careful excavation and analysis by archaeologists (though the current practice of discarding the heads and wings of bats before the torso is taken home to be cooked suggests that it may be difficult to identify bats themselves in an archaeological context).

Inconsequential though it may seem, the Pemba case raises a whole series of unresolved issues about the historical emergence of the Swahili and their distinctive economy. As we have seen, the role of hunter-gatherers as well as possible Bornean or early Malagasy contacts in Swahili history is poorly understood. The chronology of early Swahili settlement on Pemba and other Indian Ocean islands, not to mention the coast itself, is still very much the subject of guesswork, though recent archaeological and linguistic research has given us a number of pointers. The precise relation between the Swahili and their fellow long-distance migrants, the Mwani and the Comorians, is also something of a mystery, and in this case recent linguistic work appears only to have made it a more complicated one. The Pemba case provides an analogy for what is perhaps the key problem in understanding the emergence and development of these related peoples: how did they learn to exploit the marine resources of the coast? To what extent did they simply innovate and adapt their existing practices and technologies? To what extent did they adopt or absorb the practices and technologies of others? And who were these others?


In addition to the many Pembans who have shared their knowledge of bats and bat-hunting with me (and which I hope to describe in greater detail in future papers), I would like to give special thanks to Tony Archer, Tuula Kurikka, Abigail Entwistle and Nadia Corp for the information and many ideas on this subject which they too have provided.

[The Pemba Flying Fox and some of the other bats mentioned in this article are illustrated on a poster produced by The Field Museum (Chicago) / Critical Ecosystem Partnership Fund: Popo wa Pemba / Bats of Pemba]


by Martin Walsh and Helle Goldman

[text (with updated references) of an article originally published in 2003 in Nature East Africa, 33 (1/2): 14-16. An abridged version, 'The Zanzibar Leopard – Dead or Alive?', was published in 2004 in Tanzanian Affairs, 77: 20-23]

The Zanzibar Leopard, Panthera pardus adersi, is an elusive and possibly extinct subspecies endemic to Unguja (Zanzibar) Island. It has presumably been evolving in isolation from other leopards since at least the end of the last Ice Age, when Unguja was separated from the Tanzanian mainland by rising sea levels. The “founder effect” (genetic characteristics of the marooned population) and/or adaptation to local island conditions produced a smaller leopard than its continental relatives and one which “changed its spots”, or rather saw its more numerous rosettes partially disintegrate into spots (Pakenham, 1984; Kingdon, 1989).

Not much is known about the biology of the Zanzibar Leopard. Visitors to the natural history section of the Zanzibar Museum will be familiar with the stuffed and rather faded specimen kept in a display case there together with an old black and white photograph of a leopard trap. Apart from scraps of pelt furtively kept by hunters, to date we have only located five other skins: three in the Natural History Museum in London and two in the Harvard Museum of Comparative Zoology in Cambridge, Massachusetts. Of these only the type specimen in London and the two Harvard skins are accompanied by their skulls.

The Zanzibar Leopard’s behaviour is also poorly understood. It has never been studied in the wild and the last time a researcher claimed in print to have seen one was in the early 1980s. Rural Zanzibaris’ descriptions of the leopard and its habits are coloured by the widespread belief that an alarming number of these carnivores are kept by witches (wachawi) and sent by them to harm or otherwise harass their fellow villagers. This belief comes together with an elaborate package of ideas about how leopards are bred, trained, exchanged and sent to do the evil bidding of their owners. For local farmers this supplies a neat explanation for predation by leopards on livestock and humans, and more generally for their appearance “out of place” in the vicinity of farms and villages (Goldman and Walsh, 1997).

The growth of human population and agriculture in the 20th century was largely responsible for this state of affairs, as people encroached on the habitat of leopards and the animals they preyed upon. Increasing conflict with leopards and the fear that this generated led to a series of campaigns to exterminate them. These were localised at first, but became island-wide after the Zanzibar Revolution, when a combined anti-witchcraft and leopard-killing campaign was launched under the leadership of Unguja’s most famous witch-finder, Kitanzi. The long-term result of this campaign and the subsequent classification of leopards as “vermin” was to bring them to the brink of extinction (Walsh and Goldman, 2003).

The available evidence suggests that when we began our joint research on the Zanzibar Leopard in the mid-1990s there were still a few of these elusive animals remaining (Goldman and Walsh, 2002). Now we can’t be so sure. Most zoologists think that this island leopard is extinct: indeed some of them already thought so when we began our joint study in 1996. This pessimistic conclusion scotched subsequent proposals for a conservation initiative targeting the Zanzibar Leopard: if they were gone or going then there wasn’t much point in trying to do anything about it - apart from supporting the habitat conservation initiatives that were already underway on the island.

By contrast, the majority of people who live and work on the “coral rag” lands of southern and eastern of Unguja, including government staff and conservationists, believe that the Zanzibar Leopard has not been completely exterminated. Claims of sightings abound, as do reports of other evidence for leopards’ continued presence on the island and their nefarious use by witches. Many of these reports are difficult to evaluate and impossible to verify independently. So far none of the cases that we have investigated over the past two years (2002-03) has produced confirmation of a sighting or other leopard signs.

The recent scientific “discovery” of the Zanzibar Servaline Genet, Genetta servalina archeri, previously known only to islanders themselves, suggests that perhaps Unguja has yet to give up all of its zoological secrets. This small carnivore, another island endemic, was first described from an old skin and skull obtained in 1995. Its status was uncertain until a number of individuals were photo-trapped in January 2003 (Goldman and Winther-Hansen, 2003a; 2003b). If the Zanzibar Leopard survives, then similar standards of proof will have to be applied for any record to be acceptable to the scientific community. Otherwise most of us will get no closer to it than that faded museum specimen and those colourful cryptozoological narratives.

In addition to all of the people and institutions acknowledged in our original (1997) report, we would like to thank Daphne Hills in the Zoology Department (Mammal Group) of the Natural History Museum, London, and Judith Chupasko and Mark Omura in the Mammal Department, Harvard Museum of Comparative Zoology, for facilitating the examination and recording of specimens.


Goldman, H. V. and Walsh, M. T. 1997. A Leopard in Jeopardy: An Anthropological Survey of Practices and Beliefs Which Threaten the Survival of the Zanzibar Leopard (Panthera pardus adersi), Zanzibar Forestry Technical Paper No. 63, Jozani Chwaka Bay Conservation Project, Commission for Natural Resources, Zanzibar.

Goldman, H. V. and Walsh, M. T. 2002. ‘Is the Zanzibar Leopard (Panthera pardus adersi) Extinct?’, Journal of East African Natural History 91 (1/2): 15-25. [with separate map]

Goldman, H. V. and Winther-Hansen, J. 2003a. The Small Carnivores of Unguja: Results of a Photo-trapping Survey in Jozani Forest Reserve, Zanzibar, Tanzania. Tromsø: privately printed.

Goldman, H. V. and Winther-Hansen, J. 2003b. ‘First Photographs of the Zanzibar Servaline Genet, Genetta servalina archeri, and Other Endemic Subspecies on the Island of Unguja, Tanzania’, Small Carnivore Conservation 29: 1-4.

Kingdon, J. 1989. Island Africa: The Evolution of Africa's Rare Animals and Plants. Princeton: Princeton University Press.

Pakenham, R. H. W. 1984. The Mammals of Zanzibar and Pemba Islands. Harpenden: privately printed.

Walsh, M. T. and Goldman, H. V. 2003. ‘Killing the King: Political Imperatives and the Extermination of the Zanzibar Leopard’, paper presented to the International Symposium on Le Symbolisme des animaux: l’animal “clef de voûte” dans la tradition orale et les interactions homme-nature, Paris (Villejuif), France, 12-14 November 2003.