A Leopard In Jeopardy: An Anthropological Survey Of Practices And Beliefs Which Threaten The Survival Of The Zanzibar Leopard (panthera Pardus Adersi)

Conservation section, Commission for Natural Resources, Maruhubi

A Leopard in Jeopardy: An anthropological survey of practices and beliefs which threaten the survival of the Zanzibar leopard (Panthera pardus adersi) Dr. Helle V. Goldman Dr. Martin T. Walsh 1997 Forestry Technical Paper No. 63

A study funded and supported by: Jozani-Chwaka Bay Conservation Project, Commission for Natural Resources, Zanzibar, Tanzania

Preface The following report presents the findings of a study on the Zanzibar leopard undertaken in the first three weeks of July 1996. This work was contracted by the Jozani-Chwaka Bay Conservation Project (JCBCP), a partnership between the Government of Zanzibar and CARE Tanzania, and funded by the Government of Austria. Provisional results were proffered in a verbal presentation given in the Forestry Sector (FS) of the Commission for Natural Resources (CNR) on Friday 19 July 1996, and in an end-of-fieldwork summary report (Walsh 1996) which was produced shortly afterward. The research team comprised four persons: Dr. Martin T. Walsh (consultant anthropologist and team leader), Dr. Helle V. Goldman (anthropologist, technical advisor JCBCP), Ali Ali Mwinyi (Wildlife Officer, FS), and Suleiman Iddi Hamadi (former Secretary and current Assistant Secretary of the Wasasi wa Kitaifa, or National Hunters). The research undertaken by the team included the following principal components; a) review of pertinent literature and documentation, including official files, both current and in the Zanzibar National Archives; b) formal and informal meetings with resource persons in Zanzibar town, including representatives of relevant government institutions; and c) interviews and discussions with individual Zanzibaris, including past and present hunters, and a cross-section of villagers an townspeople, both men and women. A large part of the study was devoted to semi-structured interviews with hunters in villages throughout Zanzibar. Fieldwork during the second week was facilitated by the division of the team into two pairs (MTW/SIH and HVG/AAM) who worked independently in different locations. Fieldwork has continued, on a sporadic basis, by Goldman up to the completion of this paper (March 1997). The research team consisted of four people; this report was jointly written by Walsh and Goldman. The tireless assistance of Mwinyi and Hamadi are gratefully acknowledged, and we also thank other staff at the CNR, JCBCP, and other government organs who helped in various ways.

Executive summary A Leopard in Jeopardy 1. The status of the Zanzibar leopard The Zanzibar leopard is a little known and possibly endemic subspecies found on Unguja Island, Zanzibar. 2. Current distribution The current status of the Zanzibar leopard is controversial. Some authorities claim that it is extinct. However, on the basis of recent interviews with a wide array of informants and the examination of hunting records, this study suggests that though on the brink of extinction, a small population of leopards remains extant. 3. Legalities Laws imposed during the British colonial era offered leopards and other wildlife some measure of protection. Though such legislation has never been explicitly contravened during post-colonial decades, the government of Zanzibar has encouraged the extermination of the Zanzibar leopard, because it is defined as "vermin" and because of its association with witchcraft. 4. Leopard-keeping and witchcraft It is a pervasive Zanzibari belief that some leopards are kept by witches, often organised into "clubs," who magically control the leopards and use them to carry out evil errands. Motivations for leopard keeping by witches include: • to terrorise others, chiefly in order to gain "respect" • to get chickens, goats, and other foods • to guard wealth • to profit from stud fees and the sale of cubs The beliefs that kept leopards are magically protected and that leopard-keepers retaliate against leopard hunters seems to have prevented some hunters from killing leopards; at the very least it has resulted in an under-reporting of leopard killings. Leopards are associated with some "sacred sites," another aspect of the role of leopards in the local belief system. The authors argue that there is no real evidence for leopard keeping. We maintain that leopards cannot be controlled through magical means. 5. Other reasons for killing leopards Though there is some fear of retaliation by leopard keepers, Zanzibari hunters do kill leopards, mainly to profit from the sale of skins and other leopard parts.

iii

6. Leopard hunting Leopards were hunted throughout the colonial period, but it was with the Kitanzi campaign of the mid- to late 1960's that leopard numbers were seriously diminished as a result of hunting. Leopards are currently hunted by these groups, among which there is some conflict: • National Hunters, who are supported by the government of Zanzibar in an effort to wipe out "vermin" and who also hunt for pleasure • Village Hunters, who are based in rural communities and who also seek to rid the countryside of dangerous animals and agricultural pests • individual village-based hunters, who supplement their incomes through the sale of meat and animal parts • private town-based hunters, who hunt for profit and sport 7. Leopard conservation? The authors argue that more research and a conservation program are urgently required in order to stave off the imminent extermination of the Zanzibar leopard. Though there are numerous obstacles to leopard conservation, we discuss potential avenues through which it can be demonstrated that leopard conservation is in the best interests of the people of Zanzibar: • international conservation organisations promote the protection of leopards worldwide • the Zanzibar leopard may prove to be a distinct subspecies and is therefore important in terms of global biodiversity • leopards are associated with "sacred sites" in Zanzibar • the leopard is known locally as "king" • Zanzibaris are increasingly aware that conservation attracts tourists Appendix A: Recent evidence for the presence of leopards Reports of sightings, killings, and other evidence of leopards are presented in table form. These data, albeit sometimes of dubious validity, suggest that leopards remain in Zanzibar and that they are concentrated in the east-central and southern regions of the island. Appendix B: Statistics of leopards killed Data culled from various documents are presented in table form, offering further evidence for a continuing leopard presence and suggesting that National Hunters have made a significant dent in the leopard population. Appendix C: Leopard classification and linguistics Leopards are known by a variety of terms in Zanzibar. The meaning and etymology of leopard terminology are discussed in detail.

iv

Contents Preface

page

ii

Executive summary

iii

A leopard in jeopardy 1. The status of the Zanzibar leopard 2. Current distribution 3. Legalities 4. Leopard keeping and witchcraft 4.1 Leopard clubs 4.2 Why witches keep leopards 4.2.1 To terrorise others 4.2.2 Other reasons witches keep leopards 4.3 Fear and consequences of killing leopards 4.4 Leopards and "sacred sites" 4.5 Are there really kept leopards and leopard keepers? 5. Other reasons for killing leopards 6. Leopard hunting 6.1 Colonial period 6.2 Kitanzi campaign 6.3 National Hunters 6.4 Village Hunters 6.5 Individual village-based hunters 6.6 Private town-based hunters 7. Leopard conservation?

1 1 1 4 5 6 7 7 9 10 12 13 15 16 16 17 17 20 21 21 21

Appendix A: Recent evidence for the presence of leopards

24

Appendix B: Statistics of leopards killed

31

Appendix C: Leopard classification and linguistics 1. Introduction 2. Glossary of named varieties 3. Glossary of euphemisms and other expressions

37 37 39 50

Figures

55

Literature cited

56

v

A Leopard in Jeopardy 1. The status of the Zanzibar leopard Woefully little is known about the Zanzibar leopard; indeed, its very existence is a question subject to controversy. Evidence for the animal's continued existence is limited to (primarily) second-hand reports of sightings, the occasional pugmark or scat, and the periodic appearance of leopard skins offered for sale to tourists and others. Though many Zanzibaris claim that numerous leopards are kept in captivity by evil-doers, this practice has not been confirmed and no such leopard has been examined by qualified experts or disinterested persons. Though some are skeptical of the current persistence of the felid, there is no doubt that leopards were found on Unguja island during the first half of this century. Four skins documented to have come from Zanzibar are housed in museums, including two in the British Museum (one, taken by Aders, constitutes the "type" specimen), one in Harvard's Museum of Comparative Zoology (a photograph of which appears in Pakenham [1984]), and a mounted specimen in the Zanzibar Museum of Natural History (figures 4-6). Unguja is a continental island, separated from the mainland when sea levels rose at the end of the last ice age some 10-15,000 years ago. A number of the island's mammalian fauna are endemic subspecies as a result this isolation. Apparent morphological differences (its small size and distinctive coat pattern; figures 3-9) have led some authorities to treat the Zanzibar leopard as a separate subspecies, Panthera pardus adersi: Pocock (1932), Pakenham (1947), Swinnerton and Hayman (1951), and, more latterly, Kingdon (1977, 1989) and Alden et al (1995). A recent study of molecular genetic variation among leopards (Miththapala et al 1996) did not include an evaluation of the Zanzibar leopard due a dearth of appropriate samples. It is of interest to note that though Miththapala et al argue for the reevaluation of leopard subspecies from some 27 currently used trinomial designations to a mere six, both island populations (Sri Lankan and Javan) examined in the study were found to constitute phylogenetically distinct groups. Future research on the Zanzibar leopard may well add a seventh subspecies to the list. 2. Current distribution Several sources list the Zanzibar leopard as extinct or likely to have been extirpated (Hes 1991; Miththapala et al 1996; Nowell and Jackson 1996). On the basis of our fieldwork, however, we believe that this is not yet the case. Research into the question of the persistence of the Zanzibar leopard is severely hampered by the felid's association with witchcraft, rendering many of those who may have encountered leopards in recent times reluctant to reveal such information. For many Zanzibaris, merely to mention a leopard sighting, much less details regarding the animal's appearance, precise whereabouts, and so on, bodes ill. Compounding this

1

difficulty, there is the opposite tendency among some informants toward boastful exaggeration in order to secure the researcher's continuing attention and perhaps to enhance prestige among other Zanzibaris present. Finally, lack of knowledge about the island's wildlife, which has become increasingly scarce over the last decades, may lead other informants to identify inaccurately the animals they fleetingly observe. Knowledgeable informants warned us to take some reported sightings with a grain of salt since a small or immature Zanzibar leopard can be confused with Unguja's second largest wild carnivore, the African civet. Even experienced hunters admitted to having mistaken leopards for antelopes—by the light of a headlamp their reflective eyes are similar heights above the ground—so presumably antelopes have been mistaken for leopards on occasion. Of the research team, only one (SIH) has seen (and killed) a leopard for himself. Another (HVG) visited several caves in heavily thicketed areas called Migombani and Kiwandani (east of Kitogani and near Mtule) that were believed by local informants to have been regularly inhabited by leopards. The "evidence" is thus preponderously indirect and anectodal. Having registered these caveats, we nonetheless argue that there have been enough reliable sightings by knowledgeable Zanzibaris to demonstrate that the leopard continues to exist on Unguja. Map 1, covering the period 1990-96, summarises the results of our research, as well as studies by others (Archer 1994; Marshall 1994; Selkow 1995). (For details, see Appendices A and B.) The reader can assess the quality of the evidence for himself. However, if one conservatively discounts sightings of leopards, sightings of pugmarks and scat (by both Zanzibaris and foreigners), and reports of attacks on domestic animals, it is clear that even on the basis of documented leopard kills (Appendix B) alone, leopards inhabited Unguja up to 1995. If leopards are indeed extinct, they were wiped out only a year or two ago. It seems reasonable, however, not to dismiss all reports of recent sightings and attacks, and therefore to conclude that leopards have not yet been extirpated in Zanzibar. Though reports of leopard sightings span the years 1990 to 1996 along the north-eastern coast of the island (Nungwi south to Chwaka), leopards appear to be concentrated in the central-western, and southern regions of Unguja. These areas are also the most heavily forested with natural vegetation (i.e., not plantations of imported tree species such as coconut and clove) and harbour the densest populations of antelope and monkeys (Williams et al 1996; Othman and Goldman 1996). The highest concentration of leopards appears to be in the vicinity of the Jozani Forest Reserve, including Charawe, Unguja Ukuu, Pete, Jozani, Kitogani, Muungoni, Mtule, and Ukongoroni. In Pete, for example, separate sightings were reported six times, by different informants over the course of several months in 1996. There is a compelling internal consistency to these reports from the Pete area: in June, a female leopard was seen at Kichanga; in September, an adult leopard was seen near Kichanga (southern Pete, near the mangrove swamp); in October a female leopard and two cubs were seen near Kilimani Latusi (by the main road and adjacent to Jozani Forest); during November an adult was spotted on four separate occasions in various places around Pete; and in late November a foreigner carrying out a study of red colobus monkeys found two pugmarks near Kichanga, apparently an

2

3

adult leopard and an immature one. It seems fairly certain then, that a female leopard inhabits the Pete area, which is almost completely surrounded by forest and "coral rag" thicket. If the reported sightings are accurate, this female produced a litter of young in mid-1996, but only one survived to the end of the year. Whether this leopard is the same one that has been seen in the forest and thickets near Jozani village, or even Kitogani, cannot be known. Because we know virtually nothing of the ranging patterns of Zanzibar leopards, it is imprudent to attempt population estimates by simply tallying up all sightings in different areas: a single leopard may range across a very large area in search of prey and suitable denning sites; two or more leopards may occupy overlapping home ranges. Bailey (1993) has shown for leopards in Kruger National Park that home range sizes vary markedly across habitats of differing quality and also that home ranges overlap to a considerable degree. Furthermore, without evidence of leopard feeding habits, prey densities (a key factor in assessing habitat quality) cannot be used as the basis of estimating home range sizes. Until more is known about these and related aspects of leopard behaviour in Zanzibar, a firm estimate of the population cannot be hazarded. However, it is safe to say that the leopard population has declined precipitously over the course of the last two or three decades as a result of habitat destruction and hunting. Numerous informants related how leopards have been reduced in number, how some twenty or thirty years ago it was not safe to travel at night in a certain area, or that leopard calls were once heard regularly in the forest, but are no longer. Given Unguja's small size, the island may never have supported a very large number of leopards even before it was inhabited by people. In fact, something just above the minimum viable population for the species may have been the most Unguja ever supported. After humans immigrated to the island, small reductions, as a result of forest-clearing for cultivation, firewood, and settlement, were apparently tolerated by the leopard population. But much heavier pressure, as appears to have occurred since the 1960's as a result of a population explosion and concerted efforts to eliminate the leopard, cannot be withstood indefinitely. Though we can offer no hard numbers, it is our contention that the leopard population in Unguja has probably already dropped well below its ability to sustain itself in the long term without active efforts on our part to conserve it. 3. Legalities British colonial authorities recognised at an early stage that the Zanzibar leopard was vulnerable. In 1919 it was placed on the Wild Animals Protection Decree (CAP. 128 of 1919), a schedule of animals—mainly endemic species and subspecies—whose killing and utilisation was prohibited without explicit permission. Possession of leopard skins and other parts was also proscribed. With this measure of protection, leopards appear to have thrived, especially in the coral rag forests and thickets on the north, east, and south of the island. As the human population grew and the farming frontier expanded, attacks upon people and domestic stock increased in frequency. In 1950 the government bowed to popular pressure and removed the Zanzibar leopard from the protected list by issuing the Zanzibar Leopard Exception

4

Order (G.N. 30 of 1950). The Exception Order included provisions that leopards could not be killed with the aid of head lamps. However, according to our information, many leopards continue to be hunted in this way. In the aftermath of the Zanzibar Revolution of 1964 this and other provisions of the Protection Decree and Exception Order were generally either forgotten or ignored. The culling of leopards began in earnest shortly after the revolution when the Revolutionary Government sponsored the leopard-killing and witch-finding campaign of Kitanzi (see section 6.2, below). The government of Zanzibar has sanctioned leopard hunting in another way, by subsidising the National Hunters who are, in some ways, the heirs of Kitanzi (see section 6.3, below). Zanzibar, a semi-autonomous state within the nation of Tanzania, is not party to the Convention of International Trade in Endangered Species of Wild Fauna and Flora (CITES), in which the leopard is listed in Appendix I,1 although Tanzania is party to CITES-endorsed hunting quotas for leopards and other endangered or vulnerable species. Since the hasty cobbling together of the newly independent nations of Tanganyika and Zanzibar in 1964, Zanzibar's status vis-à-vis the Union Government has remained ambiguous and Zanzibar continues to test the as yet undefined limits of its autonomy. Whether Zanzibar is technically a party to Tanzania's CITES or other international wildlife conservation agreements will not be clarified until the question is raised, at the risk of stirring up once again the controversial issue of Zanzibari autonomy. The legal situation in Zanzibar is thus rather muddled in two respects. First, British colonial laws remain "on the books" and have never been contravened by post-Independence laws; yet the government continues to sanction leopard killing by National Hunters and imposes no restrictions whatever on the hunting of leopards by village and individual hunters. Second, though part of Tanzania, Zanzibar has never participated in any international agreements regarding wildlife protection. Indeed, knowledge of such agreements is scarce among Zanzibari officials. 4. Leopard keeping and witchcraft Early in 1921, a leopard made its appearance in my district, and in a few nights accounted for eighteen goats. I went out to try and get it, and sat up all night for in a tree over a kill, but it was of no use, and the [Zanzibaris]... said I might have saved myself the trouble, as an Mchawi had fuga-ed (tamed) the leopard and it would go where he told it. [Ingrams 1931:471]

1 CITES is a treaty which regulates and restricts trade in wildlife. Its signatory nations now number

122. Appendix I includes: "all species threatened with extinction which are or may be affected by trade. Trade in specimens of these species must be subject to particularly strict regulation in order not to endanger further their survival and must only be authorised in exceptional circumstances" (Nowell and Jackson 1996:221).

5

It is generally believed that though some leopards are completely wild animals which diligently avoid people, other leopards are "kept" by certain nefarious individuals. Informants suggest various proportions of wild to kept leopards, but agree that wild leopards can become kept leopards—when leopard keepers gain control over them through the use of magic—and kept leopards can become feral— when a leopard keeper dies without transferring his leopard to another keeper and the animal simply returns to the forest. A "leopard keeper" (mfugaji, mfuga chui, mwenye chui), who may be female but is likelier a man, either inherits his leopard from an elder relative, is given or sold the leopard by a friend, kinsman, or more distant associate, or habituates a wild leopard. Wild leopards (i.e., those without owners) are habituated by leaving out food to which has been added magic/medicine (dawa). According to one informant, a novice acquires co-ownership of a leopard in this way: enlisting the aid of an experienced leopard keeper, the would-be leopard owner goes with his friend out into the forest, where the keeper summons his leopard. The owner's assurances overcome the leopard's reluctance to draw near a stranger and over the course of several days the animal is habituated to the new person. Finally, the owner commands the leopard to obey his new associate and the leopard's tail waves to show its consent. The novice is now a co-owner of a leopard. 4.1 Leopard clubs It is believed that a single leopard is commonly shared by several keepers, who may or may not be related and who may or may not reside in the same community. For example, though it is said that there are no leopards on Tumbatu Island, lying off the northwest shore of Unguja, it is also believed that some Tumbatu people may have "shares" in leopards on the main island. The notion of leopard-keeping "clubs" (vyama) is consistent with the more general Zanzibari belief (held even in Pemba Island, where there are no leopards) that witches are highly social, if amoral, beings organised into cohesive guilds with particular initiation processes, membership rules, headquarters (invisible to the uninitiated), hierarchies, group activities, and so forth (Goldman 1996). Apropos of this, it is interesting to note that some informants likened leopards themselves to witches in that, for example, they can become invisible, and that leopard keepers are said to keep leopards as other witches keep spirits (majini, pepo, masheitani). Leopards are more commonly kept by groups of people, rather than individuals, as a way of protecting the owners from retribution. If a person believes that he has been the victim of witchcraft, including harassment by a kept leopard, he may announce that he will read a curse (halbadhiri) that will doom the perpetrator, e.g., leopard possessor, to death.2 In such cases, a single leopard owner will receive the full measure of the curse. However, a group of leopard keepers can shift their shared animal from owner to owner, and village to village, thereby outmaneuvering or outwitting the curse so that it fails to find its target. It can readily be seen how this serves to explain why, in some instances, after an incident of harassment by a

2 See also Ingrams (1931:472).

6

leopard and a subsequent recitation of the curse no likely candidate in a circle of neighbouring community falls fatally ill immediately afterward. Leopard keepers exercise control over their animals through the regular supply of magically doctored food to them. Several informants claim to have spied leopard keepers provisioning leopards at some isolated spot in the bush. Kept leopards may reside in the bush, in which case they are summoned by their owners who knock sticks against stumps or other objects. This activity is another which informants claim to have seen and, along with food provisioning, is commonly cited as evidence for the practice of leopard keeping. Other leopards are thought to reside in their keepers' homes, usually in a special room and often under a bed. According to some informants, the sight of a leopard entering a home or disappearing in the immediate vicinity of a house is not uncommon and constitutes clear evidence of leopard keeping. The translocation of kept leopards, as when an owner inhabiting one village "lends" the leopard to a co-owner residing elsewhere, is effected through the transference of small charm (pingu). Third parties are sometimes unwittingly involved.3 Aside from the use of doctored food and such charms, the precise mechanisms through which kept leopards are supposed to be controlled—e.g., sent by their owners to harm particular people at particular places—remain unelaborated. However, kept leopards are often adorned with anklets, ear-rings, collars, and even women's cloth wraps, and this may play some role in their manipulation. 4.2 Why witches keep leopards Kept leopards are deployed by their owners to achieve a number of objectives: to terrorise people; to obtain food for the owner; to guard wealth; to guard livestock; and to breed so that the offspring may be sold for profit. 4.2.1 To terrorise others Leopard keepers, like the witches (wachawi) among whom they are commonly classed, are inherently evil persons who derive pleasure from causing others anguish and pain. This sense of pleasure is apparently magnified when the leopard keeper, through the shrewd guesses of his co-villagers (perhaps aided by divination) and through the thinly veiled hints dropped by the keeper himself, is widely suspected of owning a leopard. Hence, informants almost unanimously stated that the prime motivation for keeping leopards is to scare (kuwatisha watu, kuwaogopesha watu) or

3 An informant in Paje related the following story: In Makunduchi, a child was asked by her grandfather to fetch a small package from a man in Jambiani. Unbeknownst to her, the package contained an object that attracted the leopard and compelled it to walk parallel to her, albeit concealed in the thicket. When the girl shifted her tiny parcel from the right side of her cloth wrap to the left, the leopard crossed the road in front of her, continuing to follow alongside her (but in the bush) until she reached her elder. When she arrived, she told her grandfather that she had been frightened by a leopard on her journey. The child fell ill shortly afterward, on account of having mentioned what she had seen.

7

harass people (kuwasumbua watu, kuwakera watu) and, significantly, to enjoy some measure of implicit recognition for this. A contradiction manifests itself here. Leopard keepers are putatively proud of the damage they do, like to be feared by family and neighbours, and therefore enjoy reputations as "great/powerful people" (wakubwa) or even "kings" (wafalme), as leopards themselves are known. At the same time, however, leopard keepers are said to be extremely secretive, wishing to conceal their activities so that they will not be punished via a curse or other sort of retribution (such as, during the Kitanzi campaign, arrest and imprisonment). It is interesting to note here that informants who claimed to have no knowledge whatever of leopards (such as their appearance and behaviour) were as likely to be identified as leopard keepers by others as those informants who related some information about leopards or leopard keeping. In the first instance, it was said to be a matter of the keeper/informant wishing to conceal his evil activities, and in the second a matter of the keeper/informant aiming subtly to imply that he was himself a keeper. In short, both reticence and garrulity are construed as evidence of leopard keeping. It should be stressed that no informant actually came out and admitted that he was a leopard keeper to the researchers. Rather, such information was gathered from others in private, out of the hearing of the reputed keeper. Out of fear, only in the rarest circumstances do people directly accuse others of leopard keeping; likewise, it appears that open admissions are quite rare. Harassment by kept leopards can take several forms, ranging from the animal simply allowing itself to be seen at a distance, to entering a village, mauling livestock, and savaging human beings. The object of harassment may be a single individual, a family, or an entire community. One informant explained that because an individual is embedded within a community, a leopard keeper cannot harass one person without frightening great numbers of others. Indeed, he continued, there have been cases of entire villages being abandoned as a result of particularly harrowing bouts of leopard harassment. Informants related numerous instances of leopard sightings, attacks on livestock, and attacks on people (see Appendix A for most recent instances). People were reported to have been killed by leopards in Jambiani, Uroa, and Ndudu. Indeed, one among our informants was a man who, as a boy of perhaps five or six was dragged by a leopard from the flimsy shack in which he was sleeping. Scars on his leg and the back of his head serve as constant reminders of how dangerous leopards are.4 Leopard attacks on cattle were reported in Nungwi, Chwaka, and Unguja Ukuu. Chickens and guinea fowl were reported killed by leopards in 4 This man lives in Uroa. His story was confirmed by others in the village and many details added by his mother. The child's wounds were treated at hospital, but after returning home he began to vomit up leopard hairs. The attack was diagnosed by a diviner/traditional healer (mganga) at Bwejuu as case of bewitchment involving the boy's natural father who, never married to the child's mother, was bitter about having no legal rights to his biological son. The motive for the man's sending his leopard to harm the boy was essentially: "since I can't have him, no one can." The boy received a course of treatment from the healer, including small cuts in his skin into which medicine/magic (dawa) was rubbed (this would protect him from further leopard attacks), and he soon recuperated fully. A curse was recited. The boy's natural father died soon after.

8

Chwaka and Paje. Most of these events appear to have occurred during colonial times, more than thirty years before the present, but the cow mauling in Nungwi occurred in 1989 or 1990 and three goats were killed in Chwaka5 just a few months before our field visit (1996). It is widely believed that the mere sighting of a leopard causes the observer to become instantly ill and a particularly close encounter will result in leopard hairs appearing in the victim's vomit or faeces. In one case we heard of, when a man tried to drive a leopard off his cow by hitting him, his arm became paralysed and has remained so ever since. The only remedy for afflictions of this sort are traditional means of curing which typically involve divination in order to determine the source of the evil. Divination is not always necessary however: it is believed that after someone has seen a leopard, its keeper usually visits the victim the following day, greeting him effusively and enquiring solicitously about his health and welfare. Such conduct—what an outsider might consider part and parcel of normal Zanzibari social behaviour—is considered a flimsily disguised menace that serves to inform the victim that his outwardly amiable interlocutor was the person responsible for the leopard encounter and that it would behoove the victim not to cross the keeper again in any way. Although currently the practice of deploying leopards to terrify people has entirely ill effects on the community at large, it formerly had more positive social consequences according to some informants. Leopard keeping is said to have once remained strictly in the hands elders who used their leopards to enforce proper behaviour by punishing the disrespectful, especially misbehaving youths. Thus an elder exercised his responsibilities as moral guardian of his community by punishing (kumtia adhabu) the miscreant and enforcing respectful conduct from him (kumheshimika). At the same time, the elder enhanced his own status as a powerful person. His wish to instill fear seemed to have both selfish and selfless motives, but these were compatible with one another. It should be noted that in Zanzibar the notion of "respect" (heshima) is intimately bound up with fear since a highly respected person is a powerful person and therefore also somewhat dangerous: elders, who once controlled access to cultivable land and housing sites and who arranged marriages, were respected because they could punish their juniors by withholding key resources. In this light, it makes sense that older people are more commonly believed to be witches. Their age, experience, and numerous social connections empower them, and this is sometimes manifested through witchcraft. As elsewhere, power is morally ambiguous in Zanzibar and people are profoundly ambivalent about those who dominate them. It is described as a sad change that elders inexorably lost their monopoly on leopard keeping. Younger, more essentially self-interested people became involved as traditional patterns of authority and systems of social control eroded. 4.2.2 Other reasons witches keep leopards Though leopard keeping may have once played a role in sanctioning immoral conduct, it is believed that current motivations for keeping leopards are limited to 5 The informant claimed that the goats were discovered with their hearts and livers torn out.

9

terrorizing others (for self-interest only), obtaining food, guarding wealth, and earning money through the sale of cubs. Having already discussed at some length what may be called the "terror motivation," we move on to the rest. Kept leopards are said to go out and kill fowl or livestock and bring these prized food items back to their owners. This is fairly straightforward and little need be added here except to point out how this belief serves to explain why leopards sometimes attack domestic animals. It is also said that kept leopards guard their owners' wealth, typically by residing in a room in which valuables are stored. One among our informants6 claimed that he himself, as a boy, had met a leopard in the home of his guardian's parents. It was quite tame and allowed the child to stroke it. According to the man, this leopard was jointly owned by several people in the village, who sent the animal back and forth among them. Some say that leopards are kept even in the main town, where wealthy urbanites use them to protect their money and jewels. At night, the leopards are released to search for food or to harass people. In the countryside, kept leopards are believed to guard their owners' livestock. In the past, we were told, leopards even guarded cultivated fields, protecting crops from marauding bush pigs and antelopes. Finally, some entrepreneurial keepers breed their leopards in order to sell the cubs to novice leopard keepers. Breeding itself is sometimes a risky endeavor; we were told one story of a male leopard owned by a guild in Uzi. A club in Muyuni wanted to breed their female with it, but before the pairing could take place, the Uzi male was killed by hunters. The Muyuni people extracted a fine of 3,000/- (this was quite a sum in about 1972) from the Uzi islanders for failing to protect their stud. 4.3 Fear and consequences of killing leopards Due to their association with witchcraft and the threats they pose to life, limb, and livestock, leopards are hated and feared by Zanzibaris. It is somewhat ironic that this same association with witchcraft may have offered leopards some measure of protection. More certainly, the fear of killing a kept leopard and consequent retaliation by the animal's keeper has resulted in the under-reporting of leopard killings. In one case, a National Hunter reputedly shot a leopard7 and, perhaps because word got round of what he had done, he has been "bleeding like a woman" ever since that time. In another case,8 a leopard approached the shack in which a married couple were staying while they cultivated their fields. The man grabbed his gun, shot the animal, and took it to the police. The leopard's keeper soon learned of what had happened and bewitched the man. He became quite ill and had to visit a number of specialists before he finally recovered.9 6 In Unguja Ukuu, Kae Pwani. 7 In the forest near the coast at Mchangani, reported by an informant in Paje. 8 Reported by a woman in Kiwengwa. 9 Two other cases are pertinent. According to an informant from Dimani, during the Kitanzi

campaign a hunter tried to shoot a leopard, failed at the first two attempts, but succeeded on the third try and the animal was struck dead. However, the hunter's gun barrel split during the event and the

10

Several hunters with whom we spoke related stories of having killed leopards "by accident" and then, out of fear of discovery by leopard keepers, burying or burning the carcases and hastening from the scene. In one case, our informant revealed that whilst hunting alone one night about 30 years ago he killed a leopard near Uzi. Afraid, he deposited it in a hole and burned it. Despite his precautions, he became ill, his hands and feet partially paralysed. He only recovered after being treated by a traditional healer.10 One of our informants, an elderly former Village Hunter in Paje, admitted to having killed two leopards in his lifetime; he destroyed both carcases and was glad that there had been no witness to the first killing and only a small child to the second.11 This is one of the reasons why some hunters say they prefer to hunt alone rather than in national or local hunting groups. It is said that "there are secrets in the forest" (mwituni kuna siri), that is, what one does alone in the forest is unlikely to become general knowledge. Some of the hunters we spoke with claimed to have avoided killing leopards they encountered due to the momentous perils involved. Though some stated that a leopard owner need only touch the body of his killed leopard and the hunter will rumble like a leopard or even die, other informants emphasised instead that certain body parts must be removed. The larynx (makoromo, koromeo) or the throat (koo), for example, must be excised lest the owner discover the carcase and use the parts as the basis for a very powerful hex on the leopard killer. In addition, a leopard killer sometimes removes such other parts as the claws which can be employed to concoct a magical recipe that will protect him from future leopard encounters. Such prophylactic measures are also applied to livestock. Commonly used in hunting, dogs, too can be protected, so that leopards attempting to defend themselves against the dogs find that their claws will not come unsheathed. Apropos of this, Pakenham mentioned magic aimed at protecting hunters and livestock in the Chwaka area (1947). Finally, it should be noted here that kept leopards are themselves believed to be protected by magic. A number of informants claimed that this magic causes hunters' guns to jam, though others said this was not so. It is worth mentioning that a great many guns currently employed in Zanzibar are quite old; some are made locally, using hollow window bars as barrels. Malfunctions are probably fairly frequent. hunter fell unconscious. He recovered three days later. Another time, a pig hunter killed a leopard at night in Unguja Ukuu, thinking it was an antelope. The owner bewitched him so that whenever he closed his eyes he saw a leopard. He visited many specialists, and was finally cured by one in Tanga, on the mainland. 10 This man, from Unguja Ukuu, did not attempt to divine the identity of the leopard keeper and did not perform the curse; it was enough that he recovered and he did not want to stir things up again. 11 The first killing took place shortly after Independence, perhaps 1966. The then sheha of Paje shot a leopard and though it was badly wounded, it ran off. Some two or three months later, this leopard, much diminished in strength, approached the informant's dwelling. His dogs set upon it and the man dispatched it with his spear. Afraid that it might have been a kept leopard, he burned the carcase and did not mention the matter. A few years afterward, he was hunting antelope and guinea fowl the forest near Paje, accompanied by a small boy. Returning late in the evening, the light of his head torch caught what he thought was an antelope. After shooting it, he saw that he had in fact killed a leopard. He dug a hole and buried it.

11

4.4 Leopards and "sacred sites" Compounding the complexity of the beliefs surrounding leopards, the association of leopards with mizimu—commonly translated as "sacred sites"—demands attention, for in this case the association is somewhat positive. Mizimu (sing. mzimu) are locations where people petition God for assistance in cases of illness, barrenness, poverty, or other forms of misfortune. These locations are typically sited in the bush and may consist of a grave, a cave, a large or unusually shaped tree, or some other distinctive feature of the landscape. They are sometimes marked with stone cairns and small, rough hearths where incense (ubani) has been burned. Shards of clay vessels, the remains of bowls in which traditional rice dishes (wali) have been brought to the sites, are also sometimes found at these places. Particular kinship groups (koo) are said to "own" (wanamiliki) particular mizimu; they are described as the "owners" (wamiliki) of the place. Each site is believed to be inhabited by spirits (masheitani) which have for generations been associated with a particular kinship group. These spirits are of the type known as "family" or "ancestral" spirits (masheitani wa jadi). They are "inherited" by the members of subsequent generations, though they do not equally affect all family members and they sometimes skip generations (see Goldman 1996). When a person becomes ill, cannot find a spouse, watches his livestock die one by one, or suffers some other kind of adversity, he (or she) consults a diviner who may diagnose the misfortune as having resulted from the individual's neglect of his family spirits and their mzimu. Many years may have elapsed since they were last propitiated, and, angered, the spirits are vexing the individual to remind him of his duties toward them. The individual then prepares a dish of rice boiled in coconut milk and brings it to his mzimu, where, sometimes joined by some near relatives, he consumes the rice, burns incense, and asks God to help him. At the same time, of course, he is propitiating the spirits, although this aspect of the ritual generally remains implicit only: to claim explicitly that one is asking spirits for forgiveness and assistance makes many Muslims uneasy, though this is certainly the core of mizimu rites. After he has finished, the individual leaves the bowl and some food in it, and returns homeward. Mizimu are protected areas; no one may cut the bush surrounding these sites for approximately 100 metres in all directions. (Thus, one way of recognizing mizimu is to look for high stands of old trees, perhaps surrounded by patches of cut thicket.) It is believed that if one cuts trees near a mzimu, the illicit harvester will suffer, perhaps by an attack of bees, an accident, or illness.12 Or, in some cases, he will encounter a leopard. Many mizimu are caves that have been formed through the leeching of limestone in pockets of the fossil coral deposits that underlie the island's thin layer of soil. Some of these caves are quite sizable. A number of mizimu caves in Kiwandani and Magombani, heavily thicketed areas east of the village of Kitogani in south12 However, honey may be taken from these areas by anyone who wishes to do so and no special

permission is required. One of the caves mentioned here contained a ladder which had been propped up against the cave wall to reach a beehive.

12

central Unguja, are believed by local informants to be inhabited, off and on, by leopards. It can readily be apprehended how mizimu propitiation rites, which include leaving food in caves occupied by leopards, is connected to beliefs concerning leopard keeping.13 But it should be noted that our Kitogani informants believed that the leopards in these caves are often wild ones. Apparently, not all mizimu leopards are kept leopards, yet they are indirectly associated with particular family groups through their association with those families' sacred sites. Moreover, these leopards assist in the protection of these sites, by helping to scare off intruders. 4.5 Are there really kept leopards and leopard keepers? Having found no evidence to the contrary (and at the risk of being accused of unfashionable positivism), we take the view that the notion that witches employ leopards to do their evil bidding must be placed alongside other widely believed but scientifically unsubstantiated beliefs—such as the ideas that witches can turn themselves into animals, can become invisible, and can fly. If scientifically unfounded, these beliefs are profoundly interesting from a sociological perspective. As with witchcraft and its associated ideas more generally, once a basic set of assumptions about leopard keeping is accepted, then ideas concerning leopard keeping are sustained by an unbreachable internal logic: all contingencies can be explained and the basic assertions remain unchallenged. For example, if one accepts (a) that leopard keepers retaliate against those who harm their leopards and (b) that retaliation may not occur if the killer successfully conceals his act (or removes certain body parts), then whether or not a leopard killer falls ill, the belief system is sustained. If he remains healthy, then it must be because he took proper precautions. If he sickens, it means that the leopard keeper found out somehow (perhaps through supernatural means). Thus, the leopard killer can remain well or fall ill; either outcome buttresses the belief system. In some cases the "logic" is clearly circular. For example, Zanzibaris explain that the difference between wild and kept leopards is that the former are exceedingly farouche, avoiding human contact at all costs and having no taste for livestock, while the latter harbour no fear of people and prefer to dine on human foods. Thus, any leopard that is seen in the vicinity of a settlement is typically deemed a kept leopard, and any leopard activity near a village is a form of diabolical harassment. One of the researchers suggested that, under dire circumstances (e.g., gunshot injury, old age, water shortage, absence of wild prey), a wild leopard might overcome its natural fear of humans and their villages in order to obtain access to water or food. Villagers claimed that this was quite impossible since only kept leopards come near people. In other instances contrary evidence is simply ignored. For example, a number of informants stated that the mere sight of a leopard, even at a distance, causes the observer to become instantly ill. Several detailed stories were offered as proof. On the other hand, plenty of our informants described having seen leopards but failed to then mention that they fell ominously sick afterward.

13 Ingrams saw a dish containing the remains of food inside a cave at Kufile, in the southern part of

the island, and was told that the dish had been placed there by the leopard's owner (1931:471).

13

Having said that we do not believe that witches are magically manipulating leopards, we must also add that it may well be that some Zanzibaris have wholly or partly habituated wild leopards by feeding them in much the same way that red colobus monkeys have become quite accustomed to human visitors in the vicinity of Jozani Forest. There is nothing magical about this. Likewise, it is within the realm of possibility that some people are actually breeding leopards, just as suni antelopes are bred at Zala Animal Park, in Kitogani.14 However, it must be stressed that habituating or possibly even breeding leopards is not what is meant, in Zanzibar, by "leopard keeping" (kufuga chui). Rather, the phrase implies a supernatural relationship between animal and malevolent owner, who is able to dispatch his leopard to carry out various activities at his whim. Furthermore, given the tendency of some people to relish a reputation for being powerful—for being "respected"—it is conceivable that some individual Zanzibaris have allowed the proliferation of rumours about themselves as dangerous leopard keepers. Indeed, such a man or woman may even start this sort of gossip himself, opportunistically taking advantage of a neighbour's chance sighting of a leopard to raise himself up a notch in the local hierarchy of prestigious, or feared, persons. Recall that a leopard keeper, like the leopard itself, is sometimes refered to as a "king" (mfalme). Finally, that some informants have deliberately misled researchers, both past and present, cannot be ruled out entirely. The idea seems to have spread across the island that foreign researchers are willing to pay quite a bit of money (the figure cited is typically a neat 100,000 Tsh15) to catch sight of a (kept) leopard. Though several people have pursued these offers, to our knowledge no such sighting has ever been made by a foreigner. As observed much earlier, Zanzibari ideas about leopard keeping cannot be fully appreciated except within the greater context of a complex cosmology concerning personhood, good and evil, and causes of misfortune. It is characteristic of Zanzibar, as in other kinship-based societies, that misfortune (e.g., illness, injury, loss of livestock) is generally interpreted within a socially relevant framework. That is, misfortune is often attributed—directly or indirectly—to other people rather than to chance and, in this case, to the natural and predictable behaviour patterns of animals which have no objectively demonstrable connection to human society at all. Thus, an illness that occurs after a leopard encounter is ascribed to a leopard keeper, who is probably a neighbour and perhaps even a relative of the victim.16 The illness (or attack, or loss of stock) then becomes comprehensible in light of animosities that may have their roots in long-standing property conflicts, marital and paternity disputes, and such basic human factors as jealousy, envy, and greed. In this way, the unfortunate event acquires social meaning and can be acted upon through culturally prescribed means such as divination, curses, and traditional methods of curing.

14 Leopards do breed fairly well in captivity under favourable conditions. 15 About 165 US$. 16 See the case of the boy who was attacked by a leopard supposedly sent by his own father, footnote

4.

14

These modes of interpretation and resolution are changing in the face of increasing urbanisation, tourism, commercialisation, and other linked trends which result in smaller family units and independence from elders (and therefore less fear of and respect for them). The interpretive framework that was so effective in the rural regions of Unguja is gradually replaced by a more "modern" one that relies more heavily on western scientific explanations. Though the change is far from complete, it is significant that informants commonly explained that leopard keeping is waning because the younger generation is no longer interested in learning traditional practices from their elders. The reduction in leopard keeping and other forms of witchcraft—antisocial practices—is linked to an ebbing interest in socially approved traditional curing methods and other forms of indigenous knowledge. And though the majority of informants stated quite clearly that they wish leopards were entirely exterminated in Unguja, a few younger people considered that leopard conservation might serve as a useful draw for the tourists who throng the island. 5. Other reasons for killing leopards We have seen that the two main reasons for killing leopards—their association with witches and their occasional attacks on people and livestock—are really one. The leopard is at the same time "vermin" and an evil being. These reasons alone would amply account for why the animals are killed, yet two other reasons merit attention: the use of leopard parts in traditional curing practices and the sale of leopard skins. As noted earlier, leopard parts serve as vital ingredients in anti-leopard recipes. Besides this, leopard parts are also employed in the treatment of other complaints. Claws, tufts of fur, and strips of skin all have a variety of medicinal uses and are often kept by the hunter, given away to friends, or sold. Leopard parts are, it is said, available in one of the town's specialty herbalists' shop, though they are not offered for sale openly. However, it would appear that few if any leopards are killed specifically to obtain such magico-medicinal ingredients. Rather, these parts are probably taken opportunistically, while the leopards are killed for other reasons. But hunting leopards in order to sell their skins is a motivation that surely stands on its own given the large sums of money that are reputedly involved. One quote we obtained was 30,000 Tsh (50 US$), which is nearly double the monthly salary of a mid-level civil servant. During and for some years after the Kitanzi campaign, leopard skins were delivered to the government, to be sold on what was then an open market. At some point this system broke down, leaving no official guidelines at all regarding the disposal of skins. The black market has filled this gap. Though skins have been offered for sale to tourists in Zanzibar,17 after passing through one or two middlemen (and rising in price accordingly), most skins are said to wend their way to the Tanzanian mainland, en route, presumably to somewhere else (the Gulf/Arabian Peninsula was suggested as one likely destination). The 17 For example, a skin was offered for sale to two South Africans visiting Paje in 1995. They declined

the offer. Another skin was brought to the Principal Curator of the Zanzibar National Museum in December 1996. This too was refused.

15

Zanzibar end of this trade is evidently not very well organised—hunters may sometimes wait for a month or longer before they can find a buyer—and this probably reflects the fact that the local supply is very limited and many of the skins of inferior quality and size to those which can be found on the mainland. Nonetheless, this trade in skins gives hunters an important incentive to kill leopards: some hunters earn most of their living from the sale of wildlife products, and leopard skins are among the most highly prized of these. 6. Leopard hunting Leopards are almost always killed by male hunters, usually hunting with dogs and shotguns or, in some areas (for example, the far south) with spears. A number are killed at night by hunters equipped, illegally, with head torches. Leopards are actively sought out or killed when the opportunity presents itself on account of their status as agricultural "vermin" (wanyama wa haribifu)—because they occasionally take livestock—as well as their association with witchcraft. Sometimes, mistaken for antelope or other animals, they are killed "accidentally." 6.1 Colonial period Pakenham noted that the Zanzibar leopard "occasionally becomes a nuisance by killing live-stock, although generally speaking he amply atones for such misdeeds by subsisting on monkeys, dik-dik, and pig" (1947:31). This statement reflected the general policy of the British government with respect to leopards. They appear not have been ranked as "vermin"—as were vervet and Sykes' monkeys, suni antelope ("dik-dik"), bushpigs, crows, and rodents—and were therefore not subject to the extermination campaigns that were aimed at eliminating these agricultural pests from the islands (Pakenham 1947; see also Wilson 1940; Borsa 1987). Leopard hunting was delimited by the various wildlife protection laws that the British enacted, as described in an earlier section. However, leopards were certainly killed during this period, both legally and illegally. Pakenham (1984) includes a list, given to him by an Arab District Commissioner, of 23 leopards known to have been killed between 1939 and 1943.18 These were, presumably, licit kills as were the leopard hunts that informants today recall as having involved British and other European hunters. Informants remember that they themselves were not permitted to hunt leopards, nor to carry fire-arms, but were employed to assist in leopard hunts: the British took the leopards and the Zanzibaris were allowed to take any other animal that was killed during the hunt (excluding legally protected species). Of course, we have no way of knowing how many leopards were taken illegally. Whether the British enacted laws to "protect" the leopard in order to preserve for themselves the sport of game hunting is a question that merits further investigation, but cannot be answered here. As elsewhere in East Africa, definitions of "poacher" and "hunter" typically broke down neatly along colour lines. 18 Two in the area of Mkokotoni; 2 in Dimani; 1 in Kikungwi; 1 in Kufile; 10 in Uroa; 1 in Mavuu

Makunduchi; 1 in Koani; 1 in Umbuji; 2 in Mtende; 1 in Fumba; and 1 in Muyuni.

16

6.2 Kitanzi campaign Kitanzi, whose nickname comes from the word for "snare" and, by extension, "dangerous person," was from Makunduchi, in the southernmost part of Unguja. In about 1964, he emerged as a leader of a hunting group which specialised in antelopes and, then, of a special leopard-hunting group. His reputation for being a powerful hunter and magician grew. His precise status vis-à-vis the government is unclear, but at some point in the mid-60's he was empowered to carry out a very concentrated anti-leopard and anti-witch campaign. Traditional healers (waganga) were also singled out for arrest. This campaign lasted into the early 1970's and was focused in the south-central portion of the island. During the campaign, Kitanzi apparently never hunted himself, but dispensed magic (dawa) to others on the team so that they could "see" leopards and catch their equally elusive owners; the consumption of leopards by some of Kitanzi's hunters may have been part of this magical treatment. The success of the crusade was perceived to lie largely in Kitanzi's own formidable powers as a witch who was able to vanquish other witches. Leopards were sometimes trapped in baited wooden cages19 and then stoned, clubbed, or speared to death. Some of our informants recall seeing leopards for the first time under these circumstances. One informant described how arrests of reputed leopard keepers took place: Kitanzi's team would knock at the door of a supposed keeper and demand to see the man's leopard. When the man vehemently denied that he had a leopard, he would be arrested for noncompliance.20 The homes of suspected witches, leopard keepers, and traditional health practitioners were typically torn apart in the search for "evidence." Kitanzi's campaign played a key role in hardening popular and official attitudes towards leopards and in burying any vestigial recognition of the laws still nominally in existence that protected leopards from indiscriminate killing. Not surprisingly, there are no records from Kitanzi's days so we do not know how many leopards were killed nor how many people were arrested. However, the Zanzibaris with whom we spoke recall that leopards were markedly reduced after this campaign, as was witchcraft and sorcery.21 6.3 The National Hunters The National Hunters (Wasasi wa Kitaifa) is a nationally organised body of hunters which has its origins in the colonial period. Though formed in the late 1950's, this group became more active following independence when the Revolutionary Government agreed to provide transport, fuel, and shotgun cartridges to the hunting group.22 Concerned as they were with the extermination of agricultural vermin, the 19 Probably similar in design to that shown in the photograph in Pakenham (1947). 20 We were told that in one case, Kitanzi arranged the capture of a witch from Bumbwini, but when

the vehicle reached gaol the witch was no longer inside it. 21 A similar campaign, but one aimed not at leopards themselves but at witches and evil spirits, was

more recently carried out by Tokelo, from about 1990-92. 22 Note that the first chairman of the National Hunters was also on the Revolutionary Council.

17

National Hunters were then under the Ministry of Agriculture, Livestock, and Natural Resources. In 1986, it was switched from MALNR to the new Department of Regional Administration. In the past, official hunts were also organised at lower levels of the administrative hierarchy—there were Regional and District hunts—but the former system has largely broken down, and only National Hunts are held with any regularity. Membership in the group is open to all Zanzibari hunters—no fees are required—and many members also belong to Village Hunting groups (see below) or hunt on an individual basis. Not all members participate in all hunts; "membership" may also be very short term, perhaps confined to only a few hunts. Thus, a man may identify himself as a Village Hunter on one occasion or as a National Hunter another time, depending on the circumstances. The National Hunters' official task is to carry out vermin control in the countryside, sometimes following a written request for assistance from local leaders. The hunts usually take place on Sundays (sometimes beginning Saturday evenings and continuing through the night) at prearranged locations where the National Hunters may also join up with Village Hunters. The explicit prime target of most of these hunts are bushpigs, though they are sometimes organised against Sykes' monkeys. Other animals, such as antelopes, vervet monkeys, civets, pythons, mongooses, and leopards are also taken.23 While the explicit raison d'être of the National Hunters is to assist rural Zanzibaris by reducing agricultural vermin,24 sport and financial gain from the sale of wildlife products are among the primary motives of the varying number of individuals who take part in its activities. It is not clear precisely what happens to the skins of leopards taken by National Hunters to town, or who profits from the sale of these hides. Steered by a committee that includes a chairman and secretary as well as untitled hunters, the National Hunters are not tightly regulated by the government. The government does not, for example, set the National Hunters hunting quotas or attempt to verify the number of kills they report, nor the species that they report having killed. The National Hunters are, however, required to keep records of their activities in the form of individual hunt reports and annual summaries. Record keeping among the National Hunters is spotty and a comparison of leopard kill statistics from different National Hunt sources reveals inconsistencies and lacunae (see Appendix B for references and details). For example, individual hunt reports show that a total of 10 leopards were killed between 1983 and 1995, while the annual summaries—purportedly compiled on the basis of the hunt reports—show a total of 57 for this same period. The National Hunters annual summaries also include some data on Village Hunts, as reported to the National Hunters by leaders of the Village Hunters, though this information too appears rather perfunctory (locations and 23 The sole animal which is not hunted is the endemic Zanzibar red colobus monkey, which is

actively protected in Zanzibar. There was an incident in 1996 during which a red colobus monkey was killed during a National Hunt. This seems to have been accidental. 24 National Hunters documents, such as annual summaries and requests for fuel, typically include such lines as Shime wasasi kufika kwa wingi usasi hizo ili kufanyikisha mtakula ("Come, let the hunters number many in this campaign to achieve food self-sufficiency") and Shime tuwapige vita wanyama waharibifu ("Come, let us wage war on vermin").

18

dates of kills are typically absent). One set of National Hunters documents that are well-furnished is formal requests for cash advances or reimbursements to cover fuel and cartridge expenses. According to what statistics are available, National Hunters are responsible for having killed more leopards than any other category of hunters (Table 1), perhaps as many as 115 in the last decade (or, minimally, 92). This is almost twice as many as Village Hunters reported having taken. Examining the records temporally is also illuminating: the greatest numbers of leopards seem to have been hunted from 1988 to 1994, when between 11 and 18 leopards were reported killed by National or Village Hunters per annum. In the last two years, the number of leopards reported killed by these organizations has plummeted. Though it is tempting to attribute this to declining number of leopards, other factors must be considered as well. For example, budget restraints have always been a problem for the National Hunters but may be more pronounced now than in the past. Moreover, it must be reiterated that record keeping is rather irregular: the low numbers for 1995 and 1996 are partly the result of inadequate documentation. Nonetheless, the hypothesis that National Hunters are killing fewer leopards these days due to actual reductions in the leopard population is backed up by National Hunters themselves (SIH in particular), as well as by village informants, both hunters and others. Table 1. Summary of leopards killed, compiled from National Hunters documents (see Appendix B for references) and field interviews (see also Appendix A). number of kills year 1996

MAXIMUM Nat'l Hunters

to midJuly

4

1995 1994 1993 1992 1991 1990 1989 1988 1987 1986 1985

1 13 13 12 18 12 15 11 5 4 7 115

total

19

1 12 8 6 2 5 12 11 5 4 7 73

combined Village NH and Hunters Village other

1 4 6 16 7 3

37

4

3

4

1 12 11 9 17 10 11 2 5 4 7 92

1

1

MINIMUM

Nat'l Village Hunters Hunters

combined NH and Village other

3 1 11 7 5 2 3 8 2 5 4 7 56

1 3 4 15 7 3

32

1

1

3

6.4 Village Hunters Another important body of hunters is the Village Hunters (Wasasi wa Kona25). These consist of groups of hunters affiliated with particular villages or clusters of settlements. As most of them see it, the chief task of these hunters—most of whom are also farmers—is to reduce agricultural vermin in their particular areas. Nonetheless, the sale of wildlife products (especially the meat of mini-antelope) provides some of these hunters with their principal source of income, though the regularity and intensity with which they hunt varies. Village Hunts can take place on any day, but are typically held on Thursdays. In a case of a Village Hunt and a National Hunt being slated for the same day in the same area, the National Hunt always takes precedence and the Village Hunt is postponed. Several neighbouring groups of Village Hunters often band together and hunt as large combined parties, usually with one village acting as host to two or three others. Combined hunting parties of, say, five or six communities are known as Wasasi wa Kona Kubwa. The role such combined hunts play in cementing cooperative relations between villages, and the hitorical origins of these hunts, merit further study. Village Hunters have also hunted with National Hunters but this is quite rare. More commonly, some men are affiliated with both groups. But though ostensibly united by the common goal of exterminating agricultural vermin and thereby furthering the development of the farmer, National and Village Hunters (as well as individual hunters; below) are also beset by conflicts of interests, at least according to some Village Hunters. We were told that while Village Hunters (and individual village-based hunters)—as farmers themselves—have vested interests in exterminating vermin, National Hunters are a boisterous and raucous group of primarily townsmen who are more interested in hunting for its entertainment value. One elderly Village Hunter complained that National Hunters leave animal carcases to rot in the fields where they have been shot, pull up cassava tubers and pick oranges, papaya, and coconuts for on-the-spot snacks without asking cultivators' permission, and generally disrupt village life.26 It must also be considered that as Zanzibar's wildlife ("vermin") decreases it becomes, ironically, an increasingly scarce natural resource over which competition emerges. For example, though antelope are considered vermin, their meat (priced below beef) is in high demand especially during the feasts following Ramadhan. Some village-based hunters have recognised that it is in their own best economic 25 Kona is a word of obscure etymological origin. One exegesis was suggested: that kona derives from ukoo, the largest unit of traditional Zanzibari kinship groupings. If this is indeed correct, then these hunters may have originally been organised on the basis of common kinship rather than residence per se, though kinship and residence overlapped significantly since residence choices were largely governed by, or rationalised in terms of, kinship. 26 This man, from Paje, also accused National Hunters of falsifying their records in order to account for all spent shotgun cartridges while concealing red colobus kills and, during the antelope closed hunting season, antelope kills. In February 1997, one of the authors (HVG) discovered National Hunters hunting in a Forest Reserve (Unguja Ukuu), where hunting is prohibited. They were also carrying shotguns, a hunting tool that was banned during that period—the newly implemented closed antelope hunting season.

20

interests to ensure sustainable populations of antelope in their areas: their value as game is beginning to outweigh their liability as crop-eaters. A similar trend may influence leopard hunting. Some informants expressed irritation when telling of private-town based hunters and National Hunters who hunt leopards in their villages' forests and thickets without first asking the permission of local leaders. Certain that these town-based hunters are profiting hugely from the sale of leopard skins, these village residents resent their intrusion. These conflicts will surely be exacerbated as more villagers adjudge the possibility that leopards and other wildlife may represent attractions for tourists in years to come. Several informants, mainly young men, expressed interest in capturing leopards in order to display them to paying visitors. Though we do not consider this an appropriate conservation measure, the significance of this idea as an indicator of profound changes in attitudes toward wildlife cannot be overlooked. 6.5 Individual village-based hunters Much of what has been said for Village Hunters applies to individual rural hunters, who hunt by themselves, though a few additional remarks are warranted. Some rural hunters prefer to hunt alone because they believe that single hunters are usually more successful, on a per capita basis, than groups of hunters who cannot avoid making noise, and therefore frightening off animals, as they move through the bush. A few hunters claimed that they hunt alone because, as noted earlier, "there are secrets in the forest," i.e., what one does kills in the bush can remain his private business. This is, of course, particularly relevant to killing leopards. It is also important to observe that individual rural hunters (like private townbased hunters; see below) are under no requirement to report their kills. It is therefore impossible to know what impact these hunters have, in comparison to National and Village Hunters, on wildlife populations. 6.6 Private town-based hunters There is one well-known group of private hunters based in Mlandege, Zanzibar Town, who typically set up camp in favoured locations and hunt there for a number of days. (Currently they use two camps regularly; they have been excluded from a third site by local villagers.)27 This small group of hunters sometimes employs local village hunters, and may also be joined from time to time (as in National Hunts) by visiting sport hunters from Muscat or elsewhere in the Arabian Peninsula. As with individual village-based hunters, we do not know how many leopards or other animals private town-based hunters take. 7. Leopard conservation? Though a census of Zanzibar leopards has never been undertaken, it is clear that leopard numbers have decreased significantly in the last decades. The two main 27 It seems likely that hunters caught illegally hunting in the Jozani Forest Reserve in April 1997

included members of this group.

21

reasons for this decline are: (1) habitat destruction as the growing human population clears bush and forest for house-building sites, cultivation, and firewood; and (2) hunting by various groups of hunters because the felid is associated with witchcraft, occasionally mauls people and livestock, and its skin and other parts are valuable. An educated guess is that the Zanzibar leopard population has probably reached a critical level: without human intervention it will be completely eliminated well within the next ten years. Two key questions remain. First, should the leopard be conserved? Second, if so, how? Though these queries were beyond the scope of our research—an investigation of local attitudes and practices concerning leopards— we believe that some recommendations should be attempted here. Whether the Zanzibar leopard merits conservation efforts is a complex issue. One might first consider the status of leopards worldwide. First listed in CITES' Appendix I (where it remains) in 1973, the status of leopards has been controversial as some have argued that leopards are not endangered while others have vehemently disagreed (Nowell and Jackson 1996). The IUCN Cat Specialist Group currently lists the leopard in Global Category 5a, which is a mid-level priority ranking, though the leopard's Regional ranking for Sub-Saharan Africa is higher, in Category 4 (ibid). In terms of preserving the world's biodiversity, the importance of the Zanzibar leopard can arguably depend on the precise degree of its genetic divergence from leopards elsewhere. As observed earlier, there is a good chance that the Zanzibar leopard, like the leopards of Java and Sri Lanka, does indeed constitute a distinct subspecies. To this end, further research is clearly required and we recommend that genetic samples be collected and analysed by experts in the very near future. Even if the Zanzibar leopard is determined to be a distinct subspecies, and therefore to be conserved, it may be discovered that there are simply too few leopards remaining to make it worth the expenditure that a conservation programme would entail. Leopards are known to be among the most flexible and adaptable of felids, showing surprising resistance to habitat encroachment and hunting, yet pressures on the leopard in Zanzibar may prove insurmountable at this late hour. Again, more research is required and we strongly recommend a study of leopard numbers and distribution that is based not merely on second-hand reports of sightings, but rather on tracking, photo-trapping, faecal analysis, and other such methods. Our anthropological survey will, it is hoped, provide useful background material for future demographic and ecological research. Finally, but by no means least important, Zanzibari attitudes must be reckoned with: any conservation programme will ultimately fail without the consent and active participation of Zanzibaris themselves. This includes village hunters and farmers, local-level leaders, National Hunters, government bodies such as the Ministry of Agriculture, Livestock and Natural Resources, and the police. The picture is, as we have portrayed it here, rather grim, yet we have also uncovered several bright points that may, if carefully developed, serve leopard conservation. One of these points is the association of leopards with "sacred sites." With support from government agencies, these traditionally protected groves might be expanded in order to provide good-sized pockets of land in which hunting and tree-cutting are

22

prohibited. Though these areas are too small to provide leopards with adequate prey, at least some suitable den sites would be protected. Another point is the reputation of the Zanzibar leopard as "king" (mfalme). Even as our informants vilified leopards, many showed a grudging admiration for the animals' size, strength, and cunning. The Zanzibar leopard could and should be a source of national pride. Linked to this is the growing awareness (especially among the younger generation) that the increasing numbers of tourists flocking to Zanzibar are keenly interested in wildlife and conservation more generally. These ideas are still in their infancy and are limited to inappropriate suggestions for trapping and exhibiting leopards in cages. But, if carefully nurtured, such ideas may develop into more sophisticated understandings of conservation as other, more appropriate, outlets for eco-tourists are presented as alternatives. We further recommend that leopard conservation be tied to the conservation of other species, such as Ader's duiker and Zanzibar red colobus monkeys, as the island's rare, endangered, and endemic wildlife are less and less regarded as impediments to rural production, but rather as valuable natural resources which must be protected in their wild state.

23

Appendix A: Recent evidence for the presence of leopards place

Tumbatu

last report of leopards' presence none

Nungwi

c. 1991

Archer (1994) was told of a leopard sighting near Nungwi about three years earlier. According to our own informant, leopards have disappeared from Nungwi. He recalled that cattle were mauled just outside the village by a leopard 6-7 years ago (c. 1989-90).

Matemwe

c. 1991

A leopard was last seen in the area 5 years ago, as reported by a villager to R. Wild (personal communication).

Mchangani

1993?

A leopard was killed in 1993 by Village Hunters from Mchangani and Pangeni (see Appendix B). This incident may, however, have taken place in the latter area, or even elsewhere.

Upenja

1995

In February 1995 the District Commissioner of Unguja North B forwarded a letter to the National Hunters asking them to take action against a leopard which was harassing people and livestock in Upenja. The National Hunters were not, however, able to organise a hunt and did not respond to the request. Selkow (1995) was told by a hunter from Upenja that a cow had been injured on its flanks by a leopard 5 months previously (around the end of 1994); the same hunter also reported that a donkey had been killed two years before (c. 1993).

Paangeni

1995

In May, or thereabouts, the National Hunters received a letter from the authorities in Pangeni asking them to take action against a leopard which was killing cattle in a government-owned herd. They did not respond due to difficulties in organising a hunt.

Pongwe

1995

Selkow (1995) was told by an 18-year old woman that, in the company of two other Pongwe women, she had seen a leopard on 18 April 1995: "During the morning at the quarry near Jasini, they saw it from approximately 50 metres away, coughing and coming at them, and it had the shoulder height of a goat. She was unable to give any further details because she was scared and ran away."

Ndudu

1996

A 10-year old boy reported seeing a leopard on the road which leads south to Uroa: this was about two months earlier (around May). It was in the morning when he was going to school with his elder sister, who took to her heels in fright. Another informant, a woman, told us that she saw a female leopard and cub at Matuni about a year ago (mid-1995). National Hunt records indicate that a leopard was killed at Ndudu in 1994.

24

additional information

Tumbatu Island has no leopards, 'kept' or otherwise, according to our informant from Nungwi. There are, to our knowledge, no reports of leopards ever having lived on the island.

Uroa

1996?

Informants in Ndudu reported that a leopard had been seen a few months earlier (early 1996) to the south of Ndudu, near Uroa. However, our informants in Uroa itself did not of this incident nor did they cite any other recent sightings. Selkow (1995) was told by a man in Uroa that he had seen a leopard take three waterfowl three months earlier (January or February 1995). National Hunt records show that two leopards were killed in Uroa in 1994.

Chwaka

1996

Three goats were killed by a leopard in early July. This happened after dark, at around 7:30 pm. The leopard was said to have torn out only the hearts and livers of the goats, and was alleged to belong to a local keeper. National Hunt records show that a leopard was killed in Chwaka in 1994 and another one in the Ufufuma forest in 1991.

Umbuji

?

SIH believes one or more leopards are present in this area.

Jendele

1994?

National Hunt records show that a leopard was killed in Jendele in 1994. However, this may have been the same as the Hanyengwamchana record of the same year. An earlier kill occurred in 1989.

Hanyegwamchana

1994

A leopard was killed during a National Hunt on 2 October. Four dogs were injured by the leopard.

Bumbwini

?

According to SIH, leopards were one present in Bumbwini, but in small numbers.

Kidoti

?

According to SIH, leopards were present in this area in the past.

Cheju

?

SIH believes there to be a kept leopard in this area. According to his National Hunt records, a leopard was killed there in 1992.

Ubago

1994

A leopard was killed during a National Hunt on December 4.

Zanzibar Town

1996

According to SIH, in February or so, a leopard was sighted at Makadara, near Uwanja wa Lumumba. In about November, 1995, a dog was attacked at night at the home of a cartwright. The dog was clawed but managed to escape. The leopard fled when people were alerted, and its alleged owner was later accused and threatened. We were also told of an incident in Saateni 5 years ago (c. 1991), when a leopard broke into a chicken house at night and killed around 20 birds.

Fuoni

?

SIH believes there is one kept leopard in Fuoni and another in the FuoniKidutani area.

Nyamanzi

?

SIH believes there is one kept leopard here.

25

Dimani

1996

On c. 21 April, three leopard cubs were reported to have been killed and fed to their dogs by some young hunters. According to an active local hunter, leopards are common in the area, though there are far fewer than there were in the past. They are seen most often on the coral rag, for example at a place called Kwa Haji Hassan, where there is mixture of high and low trees. Sometimes they are seen on the main road at night. He had not seen one himself for 3 months (since around March), since he been busy cultivating. Otherwise, he expects to see a leopard on about every fifth day of hunting (he hunts on alternate days, which gives around three sightings a month). By contrast, one 39-year old man we spoke to in the village had never seen a leopard.

Fumba

?

SIH believes there is a kept leopard in Fumba.

Uzi

c. 1988

Our informants agreed there are no leopards left on Uzi Island. One hunter reported that leopards disappeared 8 years ago (c. 1988), and that he had not seen or heard one himself for 20 years. Another informant estimated that before they were hunted out, the island supported 10-15 "wild" leopards.

Unguja Ukuu

1996

National Hunters saw a leopard in Kaebona in June, during the day. SIH believes this to be a kept leopard. One of our informants had seen a set of leopard pugmarks about 3 months before the interview (April or so). He also reported seeing a leopard about 3 years ago (c. 1993). This was when he was hunting alone, at night, in the bush. Marshall (1994) recorded a single sighting in November 1994. Another informant said that there are now very few leopards, if any, around Unguja Ukuu, and suggested that they may moved to other parts of the island. There is, he said, an alleged leopard keeper in Tindini, and possibly another in Uzina Kwero. A third villager told us that there are no leopards remaining in the area: he used to hear their calls, but not longer does. National Hunt records show three leopard kills in 1987.

26

Pete

1996

According to SIH, hunters reported sighting a leopard on two occasions in June, both times in broad daylight. A leopard is reported to have been sighted by a villager at Kichanga, Pete, in September 1996, after the conclusion of our survey. In October, a (presumably female) leopard and two cubs were seen by villagers at Kilimani Latusi, near the main road and this was assumed to be the same leopard as that seen at Kichanga earlier. In November, a leopard was reportedly seen at Mungwi on three separate occasions by a villager. On November 28, a foreigner doing research on red colobus monkeys saw pugmarks of an adult leopard and a smaller one, at Kichanga. One of our informants, a 32-year old hunter, reported seeing leopards about five times every year. His last sighting, around June of 1996, was in the bush at Kichanga: he was not hunting, and the leopard, a female, lay down when she saw him. About 3 years ago (c. 1993), he saw leopard cubs during the dry season, in the daytime. He estimated that there are now possibly three kept leopards in the area. He added that he had seen leopards on the way to both Ukongoroni (twice) and Paje. Another informant said that he had not seen a leopard for 4-5 years (since c. 1991-92), when there were a number of raids on chicken coops. Likewise he had not heard leopards calling for many years. Nonetheless, he thought that there were still leopards in the Pete area, possibly as many as ten. Marshall (1994) only gathered one record from Pete for the 1990-94 period: a call heard in 1992. A leopard was in fact killed by National Hunters in Pete in November 1992.

Jozani

1996

Leopard pugmarks have been observed in Jozani Forest on a number of occasions: most recently in June 1996, on the path to Wangwani, by R. Wild, the Conservation Advisor of the Jozani-Chwaka Bay Conservation Project, and A. Archer, an experienced hunter and naturalist (figures 1012). Our informant believes that there are possibly two leopards in the area, ranging between Jozani, Charawe, an Ukongoroni. Marshall (1994) was told of four leopard sightings in 1994, together with one observation of pugmarks. Seven sightings and one observation of pugmarks were reported from 1993. Our own informant in Jozani village remarked that he had not heard leopards in the area for some years. The last time he saw a leopard was in Muyuni, about 7 years ago (c. 1989); otherwise he saw leopards regularly in about 1985, at Mapandani, where the road forks to Charawe and Ukongoroni. National Hunt records report a leopard killed in Jozani in 1993.

Bungikerenge

1994

A leopard was killed by National Hunters in January 1994 and another one in 1993. A leopard was also killed by National Hunters at Kiwandani (not the Kiwandani near Kitogani) in December 1992.

27

Kitogani

1996

One informant noted that he still hears leopards calling and estimated that there are four of them in Kitogani and Muuongoni combined (three "wild" and one "kept"). The last leopard killed in the area, he said, was about 3 years ago. Although he remembered this as the work of an individual hunter, it may be the same as the National Hunters kill recorded on 2 January 1994. He also recalled the case of a youth mauled by a leopard about 5 years ago (c. 1991) in a valley north of the road. This youth, who was cultivating at the time, had run to investigate the cries of his dog, which was in the process of being killed by the leopard. Marshall (1994) received reports of two sightings in 1994, with pugmarks and scat also being observed. The only other evidence he collected for the period from 1990 onward was a sighting in 1991.

Muungoni

1996

A 33-year old hunter reported hearing calls, albeit infrequently: the last occasion was 2 months earlier (in May), when he was cultivating. He has also seen leopard scat through to the present and estimated that there are four leopards in the Muungoni area. He has only definitely seen a leopard once, four years ago (1992). While out hunting just before dawn, he heard a grumbling sound and saw a leopard with a partly eaten Zanzibar red colobus monkey. SIH, who believes there are two kept leopards in Muungoni, related that while during a hunt at night (with a full moon) in 1995, four dogs chased a leopard into a house. The hunters abandoned the chase, believing that this must be a kept leopard. Another informant reported that some local children had found leopard cubs in a hollow about 3 years earlier (c. 1993) and people subsequently avoided the area. He thought that leopards had been very few in number from the 1970's onwards, but over the past year have begun to increase in the area. According to National Hunt Records, individual leopards were killed at Muungoni during hunts in March 1994 (when three dogs were mauled by a leopard after it had been shot) and August 1993. Four leopard kills were attributed to Village Hunters in Muungoni in 1991.

Charawe

1994

SIH has not heard of any leopards currently in this area. Marshall (1994) recorded a single sighting in 1994, as well as a report of pugmarks. A leopard call was heard the previous year and two sightings reported for 1992. Archer (1994) records the sighting of a leopard near Charawe in 1994. Two leopards were reported killed in Charawe in 1993. However, the hunter we interviewed in Charawe claimed not to have heard a leopard since he last hunted one, 10 years ago.

Ukongoroni

1996

We were told by one villager that leopards can often be seen at night along the path to Bwejuu and he also offered to show us a site with some leopard bones. At least one sighting in Ukongoroni was reported from 1995, when a leopard was seen during a National Hunt. National Hunters were said to have seen leopards three times in recent years. Marshall's (1994) informants reported seeing pugmarks (once), scat (twice), and hearing calls (once) in 1994, while there had also been a sighting in 1993. Our older local informants, however, claimed that there are no longer any leopards in the area. By contrast, a number of hunters in villages to the south-west remarked that Ukongoroni is the area in which they would most expect to see leopards. National Hunt records indicate that a leopard was killed in 1993. SIH believes that there are two kept leopards in Ukongoroni.

28

Dongwe

1994

Marshall (1994) noted a single sighting in October 1994, with observations of pugmarks and scat the following month. Single sightings were also reported for each of the preceding 4 years (1990-93). Archer reported that a hunter had seen a leopard near Dongwe at night in 1994.

Bwejuu

1994

Marshall's (1994) informants reported pugmarks and scat in 1994. They also recalled a sighting in 1991. Archer (1994) recorded that a hunter had encountered a leopard early in the morning at Ziwani, on the path between Ukongoroni and Bwejuu, in 1994 (compare the Ukongoroni observations). Two leopards were reported killed in Bwejuu in 1989.

Mtule

1996

In is rumoured that a party of "Omani" (wamanga) hunters from Mlandege in Zanzibar Town killed a leopard in Mtule (between Kitogani and Paje) in March. A male leopard was seen by National Hunters in south Mtule in January 1996. The same leopard had also been seen in the same area during a National Hunt the previous month, i.e., December 1995. This was at night, when it attacked and injured a hunter's dog. In about September of the same year, a group of hunters following Ader's duiker also saw a leopard. This particular individual was reputed to be owned by a man in Paje who cuts wood in the area. Before this, a leopard was killed in Mtule during a National Hunt in 1994 and three leopards were reported killed in 1990.

Paje

1994

Marshall (1994) was told of a leopard sighting in the Paje area in 1994; earlier the same year pugmarks and scat had been observed. Individual sightings were also reported for 1993, 1992, and 1990. Selkow was told of an attack upon a cow in Paje, but does not indicate how recently this may have taken place. Our own Paje informants did not volunteer any recent sightings, though a hunter from Jozani had heard of a young man who had been mauled by a leopard on the path to Paje about 6 years ago (c. 1990). A South African couple were offered a leopard skin in Paje in 1994, but refused to consider purchasing it. This skin could, of course, have come from another area. The most recent leopard kill identified as taking place in Paje as such was in 1989.

Jambiani

1995

A leopard was killed during a National Hunt in April: one dog was injured. An earlier National Hunt, in 1994, claimed one leopard, while two were killed in 1988. Our informants did nor volunteer any recent sightings.

Muyuni B

1995

Our informants in this village agreed that leopards are no longer present in the area. One active hunter reported that he had not seen or heard evidence of leopards for some years: although some people claim to have seen them, he is sure they are only African civets. He last saw a leopard 8 years ago (c. 1988), when he took part in a kill together with other local hunters in the Muyuni valley area. Our other local informant, no longer active as a hunter, assterted that a leopard had not been killed in the area of 15-20 years.

Muyuni C

1993

Archer (1994) reported that leopard sightings 'go back from one to three years' here.

29

Kizimkazi

1996

Our informant in Dimbini, Kizimkazi, reported that a leopard had been seen in the daytime about 3 months ago (c. April). He though that the largest numbers of leopards were at Kibuteni (mentioned by a number of leopards as a centre for leopard keepers) and in the forest north of Muyuni A, where Ader's duiker are lso found. National Hunt records indicate that a leopard was killed in Kizimkazi during a hunt in 1993.

Makunduchi

1994

A hunter from this area told Selkow (1995) that 6 months earlier (late 1994) a leopard had killed a cow. One of our informants in Kajengwa claimed that leopards have been absent from the immediate Makunduchi area for about 3 years (since c. 1993), as evidenced by the fact that chickens and goats straying into the bush remain unharmed. He though it possible that one or two could be found in the south. He had killed a leopard himself, to the west of Makunduchi, in 1994. He was one of a party of five to six hunters with 18 dogs, hunting at night. The same hunter had also shot a leopard on the way to Jambiani in 1990, in the daytime. Our other local informant had not seen or heard a leopard for about 6 years (since c. 1990), when he saw one in a field.

30

Appendix B: Statistics of leopards killed Table B.1. Comparison of leopard kill statistics from different sources (includes National Hunts and Village Hunts). Year

Hunt reports*

Annual summaries**

1985-93 summary (typed)***

1995 1994 1993 1991 1990 1989 1988 1987 1986 1985 1984 1983

1 4 1 1 0 1 0 0 0 0 0 0

-5 6 17 9 11 0 0 0 0 0 0

--6 0 0 1 9 0 0 3 ---

totals

10

57

19

1985-94 summary (ms.)/SIH records**** -5 6 1 3 3 2 5 4 7 --40

* Hunt reports are tallies of individual National Hunts. These are kept in the Department of Regional Administration. ** Annual summaries, stored with individual hunt reports, are tallies of all National Hunts carried out in a year and include some information for Village Hunts as well. ***The 1985-93 summary (typed) is contained in the Ripoti ya utekelezaji wa kazi za kitengo cha uhifadhi mwezi wa Januari 1995, which is in the Monthly Report File of the Zanzibar National Archives. ****The 1985-94 summary (ms.) is a handwritten document, by the former secretary of the National Hunters (SIH). It is in the Wildlife Facts File at the Zanzibar National Archives. SIH has the notes on which this manuscript is based.

Tables B.2. to B.13. Details of known leopards killings, compiled from sources cited above and informants' recollections, from 1985 to mid-July 1996. Table B.2. Leopards killed in 1985. date

place

undated undated undated total in 1985

Makunduchi Mtende Kibuteni

31

leopards killed 3 2 2 7

(source of information), hunters involved, additional observations (SIH records). National Hunters. (SIH records). National Hunters. (SIH records). National Hunters.

Table B.3. Leopards killed in 1986. date

place

undated undated total in 1986

Muyuni Kizimkazi

leopards killed 2 2 4

(source of information), hunters involved, additional observations (SIH records). National Hunters. (SIH records). National Hunters.

leopards killed 2 1

(source of information), hunters involved, additional observations (SIH records). National Hunters. (SIH records). National Hunters.

2

(SIH records). National Hunters. SIH recalls the case of a leopard killed in Pete by National Hunters c. 1987. He was later accosted by the leopard's alleged keeper who told him that the National Hunters would have no further success in hunting bushpigs and other game in this area and this indeed turned out to be the case. This was therefore presumably the second of the two leopards killed in Pete this year. It may well have been the last leopard killed here, an event which one of our Pete informants also said had occured about 10 years ago.

Table B.4. Leopards killed in 1987. date

place

undated undated

Kaebona, Unguja Ukuu Shangani, Kae Pwani, Unguja Ukuu Pete

undated

total in 1987

5

Table B.5. Leopards killed in 1988. date

place

undated undated

Jambiani unrecorded

total in 1988

32

leopards killed 2 9

11

(source of information), hunters involved, additional observations (SIH records). National Hunters. (1985-93 summary [typed]). National Hunters. One or two of these may be the Jambiani kills, above. The 1985-94 summary (ms.) shows only two leopards. at least 9 if the typed version is correct; as few as 2 if the ms. is correct.

Table B.6. Leopards killed in 1989. date

place

16/11 10/9-23/12

Jendele unrecorded

undated

somewhere in Uzi, Unguja Ukuu, or Pete somewhere in Bwejuu, Paje, or Jambiani somewhere in Muungoni, Muyuni, Kizimkazi, or Makunduchi

1

undated

Paje

1

undated

Bwejuu

2

undated undated

total in 1989

leopards killed 1 8

1 1

15

(source of information), hunters involved, additional observations (hunt report). National Hunters. (annual summary). National Hunters. One of these could be the Jendele kill of 16/11. A Kitogani informant recalled a National Hunt kill in 1989 at Manyuni. (annual summary). Village Hunters of these villages. (annual summary). Village Hunters. (annual summary). Village Hunters. This could possibly be the kill which one of our informants reported has having taken place in the Muyuni valley 8 years ago. (SIH records). National Hunters. This could be one of the kills recorded in annual summary. (SIH records). National Hunters. One or both could be kills recorded in annual summary. At least 11 if records are consolidated.

Table B.7. Leopards killed in 1990. date

place

7/10-31/12 undated

unrecorded somewhere in Muyuni, Kizimkazi, or Makunduchi

undated

somewhere in Ndijani, Dunga, Umbuji, or Bambi somewhere in Charawe, Chwaka, or Uroa somewhere in Kinyasini, Upenja, Pangeni, or Kiwengwa Mtule (between Kitogani and Paje)

undated undated

undated

total in 1990

33

leopards killed 2 3

1

(source of information), hunters involved, additional observations (annual summary). National Hunters (annual summary). Village Hunters. Note that the summary is dated 7/10-31/12, so some of these and other kills recorded in it could date from the last week of 1989. One of the kills involving Village Hunters from Makunduchi could be the same as that reported by our Kajengwa informant to have been shot by him to the north of the village, on the way to Jambiani. This took place in 1990. (annual summary). Village Hunters.

2

(annual summary). Village Hunters.

1

(annual summary). Village Hunters.

3

(SIH records). National Hunters. One or two of these kills cold be the same as those made by the National Hunters unrecorded locations between 7/10 and 31/12. At least 10 if records are consolidated.

12

Table B.8. Leopards killed in 1991. date

place

24-25/7

Ufufuma forest (east of Hanyegwamchana), Chwaka unrecorded Muungoni

15/7-31/12 undated undated

undated

undated

undated total in 1991

leopards killed 1

(source of information), hunters involved, additional observations (hunt report). Village Hunters (konakubwa).

2 4

(annual summary). National Hunters. (annual summary and SIH records). Village Hunters. (annual summary). Village Hunters.

somewhere in Makunduchi, Jambiani, Paje, or Bwejuu somewhere in Charawe, Chwaka, or Uroa

7

somewhere in Mkwajuni or Chaani (prob. on coral rag to the east) Pangeni or Kiwengwa

1

(annual summary). Village Hunters. One of these may be the kill made in Ufufuma on 2425/7. (annual summary). Village Hunters.

1 18

(annual summary). Village Hunters. At least 17 if records are consolidated.

leopards killed 1 1 1

(source of information), hunters involved, additional observations (annual summary). National Hunters. (annual summary). National Hunters. (annual summary). National Hunters. This, or one of the earlier unlocated National Hunts may have been that observed by an informant from Kiwengwa. He saw a leopard which had been shot at Pango Sinashara, in the bush near Kiwengwa. (annual summary). National Hunters. (hunt report). National Hunters. This is possibly the last one recorded in the annual summary. (annual summary). Village Hunters of Uzi, Unguja Ukuu, Pete, Muungoni, Bwejuu, Paje, Jambiani, Makunduchi, Kizimkazi, and Muyuni. One or two of these leopards may be the ones recorded in individual hunt reports. (annual summary). Village Hunters.

2

Table B.9. Leopards killed in 1992. date

place

1/1-14/6 15/6-11/10 12/10-31/12

unrecorded unrecorded unrecorded

15/11 ?/12

Pete Kiwandani (near Bungi Kerenge)

1 1

undated

somewhere in southern Unguja

3

undated

somewhere in Charawe, Ukongoroni, Chwaka, or Uroa somewhere in central Unguja

2

Cheju

1

undated

undated total in 1992

34

1

12

(annual summary). Village Hunters from bambi, Umbuji, Ndijani, Dunga, and Mkorogo (east of Fuoni and northeast of Jumbi). (SIH records). National Hunters. Possibly one of the above. At least 9 if records are consolidated.

Table B.10. Leopards killed in 1993. date

place

6-7/2

Kizimkazi

17/4 1/8 undated

Ukongoroni Muungoni somewhere in Muyuni, Kibuteni, Kizimkazi, or Makunduchi

1 1 2

undated

somewhere in Chwaka, Marumbi, or Uroa somewhere in Mchangani or Pangeni (south of Upenja) Charawe Ukongoroni

1

(source of information), hunters involved, additional observations (hunt report). Combined party of National and Village Hunters. (annual summary). National Hunters. (annual summary). National Hunters. (annual summary). Village Hunters. One of these is possible the same as that recorded killed by National and Village Hunters in Kizimkazi, 6-7/2. (annual summary). Village Hunters.

1

(annual summary). Village Hunters.

2 2

Jozani Bungi Kerenge

1 1 13

(SIH records). National Hunters. (SIH records). National Hunters. One of these is probably the same as that recorded killed on 17/4 in the annual summary. (SIH records). National Hunters. (SIH records). National Hunters. At least 11 if records are consolidated.

undated

undated undated

undated undated total in 1993

leopards killed 1

Table B.11. Leopards killed in 1994. date

place

2/1 22-32/1 6/2 13/3 2/10 6/10 4/12

Kitogani Bungi Kerenge Mtule Muungoni Hanyegwamchana Jambiani Ubago (c. 10 km west of Town) somewhere in Muyuni, Kizimkazi, and Makunduchi

undated

undated undated undated undated

total in 1994

35

Ndudu Uroa Chwaka Jendele

leopards killed 1 1 1 1 1 1 1 1

1 2 1 1

13

(source of information), hunters involved, additional observations (annual summary). National Hunters. (hunt report). National Hunters. (annual summary). National Hunters. (hunt report). National Hunters. (hunt report). National Hunters. (annual summary). National Hunters. (hunt report and annual summary). National Hunters. (annual summary). Village Hunters. This is possibly the leopard reported by one of our Makunduchi informants. The leopard was killed west of the settled area. (SIH records). National Hunters. (SIH records). National Hunters. (SIH records). National Hunters. (SIH records). National Hunters. Possibly the same as the leopard killed in Hanyegwamchana on 2/10. 12 if two of the records can be consolidated.

Table B.12. Leopards killed in 1995. date

place

17-18/4 total in 1995

Jambiani

leopards killed 1 1

(source of information), hunters involved, additional observations (hunt report). National Hunters.

(source of information), hunters involved, additional observations One of our informants, an individual hunter in Dimani, reported having killed three cubs. According to an unverified report, "Omani" (Wamanga) hunters from Mlandege in Town killed a leopard in Mtule in March. At least 3 if Mtule report is discounted.

Table B.13. Leopards killed in 1996. date

place

c. 21/4

Dimani

leopards killed 3

?/3

Mtule

1

total in 1996 to mid-July

36

4

Appendix C: Leopard classification and linguistics (Martin T. Walsh) 1. Introduction The classification and naming of the Zanzibar Leopard and its perceived varieties is a complex subject which previous investigators have barely done justice. In this appendix we will examine this subject in some detail. Although it provides some clues to the range of morphological variation which exists in Unguja’s leopard population, readers are warned that it does not produce information that can be easily translated into zoological terms, though we have tried to indicate some of the ways in which it may. It does, however, provide a number of insights into local ethnozoology and its linguistic history, fascinating topics in their own right. Given the threatened status of the Zanzibar Leopard and diminishing knowledge of it on the island, we have felt it important to record what we can of indigenous taxonomy, adding our own deductions about its historical background. We do so not only to help preserve this knowledge before it is lost, but also in view of the possibility that some of this information may be used in environmental education, especially in efforts designed to preserve the Zanzibar Leopard from extinction. During our interviews with hunters and other Swahili speakers we recorded 25 different Swahili names or expressions for leopards, including the generic chui. One additional term was found in a published vocabulary of the Makunduchi dialect, bringing the total to 26. This is a surprisingly large number of names for a single species, and we have no doubt that a more extensive survey, including more work in central and northern Unguja, would add further to the list. Thirteen of the terms we recorded were given to us as names for different types or varieties of leopard which our informants recognised or had heard of: the remainder are euphemisms and other miscellaneous terms used by hunters to refer to leopards in specific contexts. No single informant produced the whole list, or anything approaching it: hunters named two or three varieties of leopard at most, and only one informant could recall three euphemisms. Very few hunters produced the same classification or list of names, even when they hailed from the same village. Even when they did agree on names for one or more types of leopard, their descriptions of what these names represented often differed considerably. The overall picture which emerges is one of considerable taxonomic and linguistic heterogeneity, to some extent fragmentation. There seem to be three main reasons for this, as follows: (1) Lack of close familiarity with leopards. A relatively small proportion of Unguja’s population have seen the Zanzibar Leopard at close quarters, whether alive or dead in the form of carcasses or skins. Rather fewer people have an active interest in differentiating one kind of leopard from another: the most obvious category of those who do being hunters. The current generation of hunters has had even less exposure to leopards than those who were hunting twenty and more years ago, and it was apparent that our younger informants were often much less well informed than their 37

elders. In many areas there are also fewer committed hunters than there were in the past, and cooperation between hunters in different communities and at different levels is clearly not what it was. As leopards and leopard kills and sightings have become rarer, and opportunities for sharing information about them have diminished, the pool of indigenous knowledge about them has become increasingly impoverished - and this is reflected in the fragmentation of taxonomies and terminologies. (2) Fear of leopard keepers and association with leopard keeping. Hunters are often wary of disseminating the knowledge which they do possess about leopards. They have two good reasons to be secretive: fear of the wrath of leopard keepers who they believe may target them as a threat to their nefarious activities; and the fear that they may be accused of being leopard keepers themselves (how else could they know so much about leopards?). This last consideration appears to have discouraged some of our own informants from revealing their knowledge, especially when interviewed in groups (which we therefore tried to avoid). When talking about leopards in front of others, hunters frequently employ euphemisms to conceal the subject of conversation. Careless talk kills, or at least may lead to affliction of one kind or another (who knows when a leopard keeper might be listening?). The net result, or rather one of the results, of this culture of sensitivity surrounding discourse about leopards is to discourage the communication and dissemination of knowledge about them, including the relevant terminologies, at local village level as well as more widely. (3) Complex dialect geography and history. Although the vast majority of Unguja’s inhabitants are first language speakers of Swahili, and Zanzibar town is the home of the standardised version of this language, the dialect geography of the island is notably complex. This is no doubt a result of the fact that Unguja has been subject to numerous external influences, including different epsiodes of immigration, as well as internal political shifts and local movements of population, in the course of its history. Linguists have traditionally divided the Swahili dialects of the island into three, reflecting a parallel ethnographic classification. These dialects / clusters are (a) Unguja, the speech of Zanzibar town and the plantation areas in the west and centre of the island, which is expanding at the expense of the other dialects; (b) Tumbatu, spoken on Tumbatu Island and, to a diminishing degree, in other parts of the north; and (c) the dialect cluster of southern and eastern Unguja, including Makunduchi (Kae) and related local varieties, formerly subsumed under the (now derogatory) ethnic group label Hadimu (Nurse and Hinnebusch 1993). The last of these clearly requires more careful description on a village by village basis, a task yet to be undertaken comprehensively by linguists. Although many features of local speech have been buried under the advance of the main Unguja dialect, a considerable amount of lexical variation has survived, notably in specialised terminologies such as that being considered here. Other aspects of ethnobiological classification in this region show a similar pattern of local fragmentation: these include the indigenous taxonomies of the smaller carnivores, snakes, and plants in

38

general. The impact of local linguistic history and differentiation upon leopard taxonomy is equally evident. Given the simultaneous operation of these three factors, it is perhaps not surprising that the classification and naming of leopards lacks any overall coherence, even at local level. Indeed, some of our informants offered taxonomies and terminologies which appeared to be the idiosyncratic products of their own idiolects, leaving us to wonder whether they had been assembled for the first time under pressure from our questioning. Only a more detailed survey could confirm or deny this. 2. Glossary of Named Varieties The following is a glossary of all the names we recorded, with notes on informants’ descriptions of the named types and the meanings and inferred etymologies of the names. bete (5/6?)28 [Chwaka]29 kibete (7/8) [Charawe] The term bete was elicited from a relatively young informant in Chwaka, who gave it as the name for a short but robust leopard, in contrast to mwanzi (see below). An older hunter in Charawe, who initially classified leopards into two main types (bungala and kariuki), subsequently used the cognate term kibete to describe small leopards in general. He did so on the prompting of one of his interviewers (SIH), offering it as a synonym for the term kichigi, which was introduced by the latter. The Standard Swahili dictionaries define kibete, and its alternative kibeti, as a person or animal which is smaller or shorter than usual; a pygmy, or dwarf (Johnson 1939; Taasisi ya Uchunguzi wa Kiswahili 1981). The application of this term to small leopards is therefore quite straightforward. Its etymology, however, is less transparent. There is no obvious link with the other meaning of kibeti (and beti (5/6)) as a small, usually leather, pouch, especially if it is true, as the Standard Swahili-English Dictionary suggests, that this is a loanword from English ‘belt’ (Johnson 1939). bungala (5/6) [Charawe] This term was given as the name of a type of leopard by a single informant, the hunter from Charawe referred to above. He described bungala as the largest and longest of two named types, contrasting it with the smaller kariuki, and adding that its coat is characteristically shiny, and that this variety is preferred by leopard keepers. According to SIH, the name bungala, like mkonge, is more normally applied to the largest male and female Bush Pigs, Potamochoerus porcus. Given that only one 28 Noun class. 29 Area of current usage.

39

informant volunteered bungala as a term for leopards, the suggestion that it has been adopted from porcine terminology seems quite plausible. bungala is also used to describe varieties of rice, banana, and sugar-cane, and according to the Standard Swahili-English Dictionary is a (presumably relatively recent) loanword meaning ‘Bengal’, as in the name of the Indian region (Johnson 1939). It is not immediately clear why the same term should be applied to large pigs (and hence leopards): the most likely explanation would be that these mammals share something in common (large size? colour?) with one or more of the bungala cultivars. chui (9/10) [all Swahili speakers] This is the general name and cover term for all leopards. Cognate forms of the same name are found all along the Swahili-speaking coast, and it is also the Standard Swahili term for leopards. The proto-Sabaki form of this name has been reconstructed as *ncuWi, derived from a Common Bantu root with the same meaning (Nurse and Hinnebusch 1993). Evidently, and as might be expected, the Swahili and their Bantu-speaking ancestors have remained familiar with leopards throughout their history, hence the retention of this common name. Not surprisingly, chui occurs as an element in place names in Unguja, including Chuini (‘the place of the leopard(s)’), Machui (meaning ‘many leopards’, possibly ‘large’ or ‘bad leopards’), and Kijibwe Chui (‘little rock of the leopard(s)’). (chui) asili (9/10) [Kizimkazi] One informant, from Dimbani, Kizimkazi, used the name chui asili for leopards which are large, have predominantly yellowish coats, and are fiercer than the other type he identified (chui uwanda, small and ‘red’ in colour). He also volunteered that chui asili is the type which is most often kept by leopard keepers. asili is a loanword from Arabic (Johnson 1939), and in this context, used adjectivally, means ‘original’ or ‘typical’. chui asili is evidently a substitute for one of the more widely used terms, either kisutu or konge. (chui) mwanzi (9/10) [Chwaka, Mangapwani, Ndudu, Pete, Zanzibar town] The name chui mwanzi was first recorded by Pakenham at Mangapwani, who placed it on his list of ‘Animals reported but undiscovered’ and remarked ‘Possibly a local name for Panthera pardus but unlikely’ (1984). Marshall (1994), who only worked to the south and east of Jozani Forest, did not record mwanzi as a leopard name, but Selkow (1995) did, probably because he also worked in villages to the north of Jozani. From informants’ descriptions he surmised that the mwanzi was the longest of three types, ‘having slender hind quarters and a bulkier chest and fore limbs. It has a predominantly yellow coat with black spots. However, two of his interviewees described it as mostly “black”’. Our own research suggests that the name mwanzi has a predominantly northern distribution, and earlier records tend to support this. A woman informant in Ndudu contrasted mwanzi with ngawa, describing it as long and coloured like the

40

kisutu variety of kanga cloth (pointing to an example which was yellow and green in colour). A male informant in Chwaka contrasted it with the short and robust bete, noting that it was long and slender, its coat coloured yellow with large black spots. A hunter in Pete differentiated it from both chui unyasi and kisutu, remarking that it had smaller spots than the latter. SIH, meanwhile, possesses a small piece of a skin which he describes as mwanzi, different in colour (a richer yellow) from both mkonge and kichigi. This piece of skin was taken from a leopard kill in Muyuni in 1986. However, none of the informants he showed this piece to identified it as being from a mwanzi leopard, though two described it as belonging to the (m)konge variety (Kitogani and Muungoni). These observations support the thesis that mwanzi is a term from northern and / or central Unguja, referring to the larger type or types of leopard which are usually called konge or kisutu in the south. The primary meaning of mwanzi (3/4) is ‘bamboo’, and on Unguja in particular the Golden Bamboo, Bambusa vulgaris, originally of Asian origin (Williams 1949; Purseglove 1972). The Swahili name for this and related African plants can be traced back to proto-Sabaki *mulanzi and even further back into the Bantu past (Nurse and Hinnebusch 1993). The use of the term mwanzi, ‘bamboo’, to describe a particular kind of leopard is probably analogous to its use in chelonian terminologies. The name kasa mwanzi is used by fishermen in both northern and southern Unguja (as well as on Pemba) to describe one or more species / variety of sea turtle (the primary referent of kasa being the Green Turtle, Chelonia mydas). The perceived similarity with bamboo probably relates to the colour and pattern of the sea turtles’ scales, and presumably much the same logic applies in the case of leopards. As it happens, some turtles are also compared directly with leopards, ng’amba chui being the name for a particular type of Hawksbill Turtle, Eretmochelys imbricata, while the name kasa chui (species unidentified) has also been recorded (Clark and Khatib 1993). (chui) unyasi (9/10) [Muyuni B, Pete, Uzi] The name chui unyasi, usually abbreviated to unyasi, has a relatively restricted distribution to the west and south-west of Jozani Forest. It was elicited from all three of our informants in Pete, two of whom contrasted it with kisutu, and one with kisutu and mwanzi. In Muyuni B one informant contrasted unyasi with konga, while in Uzi another informant differentiated it from both mkonge and kisutu. There was general agreement that unyasi was the smallest of the types, although one informant described it as long in the body. There was also some measure of agreement about the description of its coat: yellowish; more yellow than the larger type and with smaller spots; red, with small (i.e. smaller than otherwise) white patches; reddish, the colour of (dry) grass; darker in colour than the larger variety, with no white showing. Given the comparatively broad and overlapping reference of Swahili colour terms, these descriptions seem to reflect a pattern in which the leopard’s rosettes have disintegrated to an even greater degree than in the larger type(s), apparently a function of smaller body size. The primary meaning of unyasi (11/10), usually heard in the plural (nyasi), is ‘coarse grass’, including the dry grass sometimes used in thatching. It derives from a

41

proto-Sabaki root, *lunyasi, with a long Bantu history (Nurse and Hinnebush 1993). Its application to this type of leopard would appear to be based upon similarity of colour, perhaps comparing the tawny patches on the leopard’s coat with the colour of dry grass or straw, a connection made explicit by one of our informants. The motivation behind the name chui unyasi is therefore probably parallel to that behind the name chui mwanzi, albeit describing a different type. (chui) uwanda (9/10) [Kizimkazi, Muyuni A and B, Uzi] This term overlaps with chui unyasi in its use and distribution, extending further to the south-west of the island. Selkow (1995), who recorded uwanda as a leopard name only in Muyuni A and B, assumed it to be synonymous with konge, which he classed as intermediate in size between the other two types he recognised. However, he was unable to elicit a coherent description of this type, and our own evidence indicates that he was mistaken in collapsing uwanda and konge into one and ascribing an intermediate position to either of them. One informant, from Uzi, contrasted chui uwanda with kisutu, describing it as smaller and shorter, with small red patches on its coat. He emphasised that leopards with this description were fully mature, and that the difference between the chui uwanda and kisutu was not one of sex. A second informant, from Dimbani, Kizimkazi, contrasted chui uwanda with chui asili, describing the former as small and red in colour. These are broadly similar to the descriptions of chui unyasi, and we can assume that informants were referring to much the same combination of features. In its primary sense uwanda is a geographical term referring to Unguja’s bushed grassland, ‘small areas of thicket in grassland’, some of which may be natural and others which are the result of human agricultural activities in the past (Williams et al. 1996). Uwanda is apparently a loanword into the southern Swahili dialects of Unguja (and Pemba) from northern Swahili, ultimately deriving from proto-Sabaki *luWanja (with the reconstructed meaning of ‘open area’) and an earlier Bantu source (Derek Nurse, personal communication; Nurse and Hinnebusch 1993). It is not immediately clear why the smaller leopards should be named after the uwanda: presumably either because they are thought to frequent these areas, or in recognition of a similarity in colour between the leopards and the dominant vegetation of this habitat (much as chui unyasi is named for the grass it resembles). futizi (5/6?) [Jambiani] This term was recorded from a single informant in Jambiani, who described futizi as a leopard more gracile than the robust and large kisutu, but less so than the tall and slender konge. The etymology of futizi is obscure. kariuki (5/6) [Charawe] This term was elicited from a hunter in Charawe, who contrasted kariuki with bungala, describing it as a leopard which was small or medium in size, and patterned with patches of red, rather like the skin of the African Rock Python, Python sebae.

42

Although he noted that leopard keepers prefer the larger, longer and sleeker bungala, he claimed also to have seen a kariuki with its keeper. Although SIH knew of the term bungala (as a name for large Bush Pigs), he had never heard of kariuki, and neither had anyone else we asked. The unusual shape of kariuki, including the initial element ka-, which is the class 13 prefix (with diminutive sense) in many East African Bantu languages (excluding Swahili), suggests that this is a recent loanword from a mainland language (compare the well-known Kikuyu personal name, Kariuki). kichigi (7/8) [Makunduchi, Unguja Ukuu, Zanzibar town] Selkow (1995) recorded kichigi as a Makunduchi name for the type of leopard otherwise known as kisutu, describing it as the shortest variety, with a predominantly yellowish coat. Although he was right in identifying the provenance of this term, he was wrong in identifying it with the kisutu. Both of our informants from Kajengwa, Makunduchi, contrasted kichigi with konge (in one case with two varieties of konge). One of these informants described the kichigi as short and red in colour, with no patches of white, while the other described it simply as small and red. The latter recalled killing four vichigi on different occasions: one of these was male, one female, and the other two unsexed. SIH likewise contrasted kichigi (small) and mkonge (large), adding mwanzi as a third type differentiated by its colour. A fourth informant, a relatively young man from Tindini, Unguja Ukuu, contrasted kichigi with kisutu and a third type which he could not name. He described the kichigi as small, short and very fierce, coloured black and white with no yellow. These various descriptions, especially those from Makunduchi, put kichigi in the same general category as chui unyasi and chui uwanda. According to SIH, kichigi and mkonge are also applied to Bush Pigs, Potamochoerus porcus, the former to small pigs and the latter (like bungala) to large males and females. The greater abundance of pigs, which provides hunters with more frequent opportunities to distinguish between them, gives the impression that these terms, including kichigi, were applied to pigs before they were applied to leopards. This need not necessarily be the case, however, as an examination of the probable origin of the name kichigi indicates. The Standard Swahili-English Dictionary defines chigi as ‘the name of a small yellow bird’ (Johnson 1939). On Pemba island chigi (and its cognate shigi) is the cover term for a number of different species of small passerine, including the Olive Sunbird, Nectarinia olivacea (chigi-asali), the Pemba White-eye, Zosterops senegalensis ssp. vaughani (chigi-manjano), and the Black and White Mannikin, Lonchura bicolor, and Bronze Mannikin, L. cucullata (both called chigi-tongo). It also appears as an element in a name for the Zitting Cisticola, Cisticola juncidis (kichonga-chigi) (Koenders 1992; scientific nomenclature modified with reference to Britton [ed.] 1980). Unfortunately we have less information on the referents of chigi on Unguja. The Kamusi ya Kiswahili Sanifu (Taasisi ya Uchunguzi wa Kiswahili 1981) repeats the unhelpful definition given by Johnson (1939), but also gives chozi as a synonym for chigi. Chozi is the common cover term for sunbirds: the Olive Sunbird (see above) and the Collared Sunbird, Anthrepetes collari, are the only ones present on Unguja which might be described as ‘yellow’. Maimu (1982), however, gives chigi as the

43

Swahili name for the Yellow-fronted Canary, Serinus mozambicus. This bird is relatively uncommon in the wild on Unguja, but is a popular cage-bird in Zanzibar town (Pakenham 1979). Whichever of these species is called chigi on Unguja, the bird name is probably the source of kichigi, meaning ‘chigi-like’, as applied to both small pigs and leopards. The obvious point of comparison is small size. Coloration provides an additional point of resemblance between at least some of these birds, with their yellow (breast) feathers, and more deeply coloured variants of the Zanzibar Leopard. This increases the probability that kichigi was originally coined to describe small leopards rather than small pigs, though the contrary thesis cannot be entirely ruled out. kisutu (7/8) [Dimani, Jambiani, Jozani, Kitogani, Muungoni, Pete, Unguja Ukuu,Uzi] sutu (5/6?) [Kitogani] kisutu (and sutu, used with emphatic sense by one informant) was elicited more frequently than any other name for a variety of leopard. It appears to be widely used by hunters in the south of the island, though not in the Makunduchi area. Informants offered the following descriptions of the kisutu: - different in colour - with black and yellow spots - from other (unnamed) types (Dimani) - large and robust in contrast to the tall and slender konge, futizi being intermediate between the two (Jambiani) - coloured like ngawa, the African Civet (Viverra civetta), with black predominating (Jozani) - coloured black and yellow, in contrast to the black and white konge, which is also longer and taller: although smaller than the konge its skin has a better market, because of its better colour. Many visutu, according to this informant, are females, and he has seen them most often between Kitogani and Paje (Kitogani) - showing more white than the konge, whose spots reduce the areas of white background colour (Muungoni) - large and yellow in contrast to the smaller and darker unyasi, which has no white patches (Pete) - with large patches of colour and black stripes “like a zebra”, contrasting with both the yellowish unyasi and mwanzi, which has smaller marks (Pete)

44

- distinct (though he could not specify how) from the unyasi, which is red with small white patches (Pete) - tall, long and yellow, with largish black spots, in contrast to kichigi, which is small, short, very fierce and all black and white with no yellow colouring (Unguja Ukuu) - the fiercest type, with large spots of every colour, smaller than the sisal leaf-coloured mkonge and larger than the reddish coloured unyasi (Uzi) - larger, whiter and with bigger patches of colour than chui uwanda, which is shorter and redder in colour. visutu, which may be either male or female, are stronger and more often kept by leopard keepers (Uzi) In so far as these different descriptions overlap, they suggest a good-sized leopard whose coat conforms more closely than that of other types to the typical mainland pattern, possessing relatively well-spaced rosettes or spot-like condensations of the sane on a yellowish ground (which is neither as pale as on some of the larger leopards nor has the tawny colour of smaller leopards in which the rosettes have further disintegrated). Marshall (1994) was led to describe the kisutu in similar terms, though he only contrasted it with the larger konge. Selkow (1995) also referred to the predominantly yellowish background colour of the kisutu, though, as already noted above, he confused it with the smaller kichigi. The name kisutu has an interesting origin. Its most likely source is the Southern Cushitic language Dahalo, or rather an earlier form thereof. In contemporary Dahalo the name shúti refers to the Spotted Hyena, Crocuta crocuta (Ehret et al. 1989), and can be traced back to a proto-Southern Cushitic root which may have originally referred to the Hunting Dog, Lycaon pictus, or another spotted or splotched carnivore (Ehret 1980). Dahalo is spoken by a diminishing number of former hunter-gatherers in the north of the Tana River delta in Kenya. They are close neighbours of the Swahili-speaking peoples of the Lamu archipelago, and live at the southern end of the stretch of coast on which the original Swahili homeland was most probably located, around one and half thousand years ago. On present evidence it is not clear at what stage the Dahalo term was borrowed into Swahili: whether before or after the breakup of the proto-Swahili community into northern and southern dialect groups. It may be that kisutu (root -sutu) was a secondary borrowing into one or more of the southern dialects on Unguja from one of the northern dialects of Swahili. It is also difficult to say at what stage in its history the meaning of the term shifted, and in which direction. One possibility is that shúti originally meant ‘leopard’ in Dahalo, and was borrowed by Swahili before its meaning changed. The simplest hypothesis, however, is that the meaning of the Dahalo term has remained constant, and only changed after it had been adopted into Swahili. The shift in reference from ‘spotted hyena’ to ‘spotted leopard’ would have had an obvious motivation in the Unguja context, where Spotted Hyenas do not exist. This might also explain the otherwise unusual assignment of the Swahili name to the noun class pair 7/8, suggesting that the

45

original sense of kisutu was ‘(the leopard which is) like a spottted hyena’. The Zanzibar Leopard differs from its mainland cousins in the spot-like condensation of its rosettes, and its coat can therefore be compared more readily with that of the Spotted Hyena. It is not difficult to imagine that this was precisely the connection drawn by the early Swahili settlers on Unguja island when they first encountered its distinctive leopard. Kisutu is also the common name of a particular design of kanga, the rectangular pieces of cloth worn by women. Most Swahili speakers, on Pemba as well as Unguja, know this name, but have no idea that it also refers to leopards. A woman informant in Ndudu even described the leopard she called mwanzi as being coloured like the kisutu kanga, pointing to an example which was green and yellow. Closer examination, however, indicates that the kanga was originally named after the leopard, and not vice versa. The defining feature of the kisutu type of kanga is a central pattern of two alternating motifs: flower-heads and crosses or plus signs, variously elaborated, but unbroken by other motifs. The colours of the kanga are incidental to this design. While the Standard Swahili dictionary suggests a rather dubious Arabic etymology for kisutu (Johnson 1939), it seems more likely that this kind of kanga was named after its schematic resemblance to the pattern on a leopard’s coat, much as the name kanga itself is derived from the similarity of early prints to the spotted plumage of guineafowls (on Unguja the Helmeted Guineafowl, Numida meleagris, and the Kenya Crested Guineafowl, Guttera pucherani, both called kanga). According to one informant, a pair of kanga with the kisutu design is traditionally given to a bride by her parents, to be worn during her wedding. The origin of this practice, said to be confined to Unguja, is obscure, and though there is no explicit reference to leopards or leopard-skins in this context, it is tempting to suggest that some kind of connection, historical or otherwise, is implied. It might be added here that kanga with a more obvious resemblance to leopard-skins - for example with swirling and overlapping red and black feathery motifs on a yellow ground - are actually called chui, ‘leopard’, but have no ceremonial significance. The term kisutu also occurs in the saying ‘nimeona kisutu, mwenye kisutu sijamwona’, which roughly translates as ‘I have seen a mark, but not the marker’. This expression might be used, for example, when seeing betel which someone has spat out, but not the person who did the spitting. This usage probably represents a further shift in the meaning of kisutu, from spotted leopard to a spot, sign or mark in general, though the extended sense appears only to have been retained in this particular saying. Given the elusiveness of the Zanzibar leopard, and even more so of their alleged keepers, this expression seems doubly appropriate (‘I have seen a spotted leopard, but not its owner’!). koko (5/6) [Paje] A single informant in Paje used the name koko to describe leopards which are small and short in the back, in contrast to the long mkonge. He noted, however, that leopards cannot be readily distinguished on the basis of their colour or build: colour may vary that of the earth a leopard has been rolling in, and either koko or mkonge

46

might be stout or thin, depending on how it had been feeding. The spot pattern, he added, is also very variable. koko is more usually used to describe feral dogs, mbwa koko in the usage of SIH and other informants. Chum (1994) defines koko as a ‘small and thin’ kind of dog, ‘a dog of inferior quality’, in the Makunduchi dialect of Kae, adding a second definition of koko as ‘dirt’. The Standard Swahili-English Dictionary records the primary meaning of koko in this context as ‘bush, undergrowth, jungle, such as in a mangrove swamp’, linking it with mkoko, the name of a common mangrove species, Rhizophora mucronata (Johnson 1939). The same stem is evident in Sangu ilixoxo, ‘wild animal (in general)’ (Walsh 1985). Sangu is a member of the Southern Highlands group of East African Bantu languages, and the occurrence of the same root in cognate form in that area suggests that koko is an inherited Swahili term, originally meaning either ‘bush, wild place’ or ‘wild animal’. However, given the limited use of this as a leopard name on Unguja, it is probable that it has been adopted directly from koko in its locally restricted sense of ‘wild dog’. Feral dogs are typically underfed and scrawny, hence the ‘small and thin’ of Chum’s definition and the ready application of this term (in at least this single case) to leopards of unimpressive size. konge (5/6) [Jambiani, Kitogani, Makunduchi, Muungoni] konga (5/6) [Muyuni B] mkonge (3/4) [Kitogani, Paje, Uzi, Zanzibar town] konge and mkonge are cognate terms, and konga presumably an idiosyncratically skewed version of the same. As the name(s) for a recognisable kind of leopard, we recorded them almost as frequently as kisutu, over a large area of southern Unguja, which included Makunduchi and Paje, but excluded Jozani, Pete and Unguja Ukuu. Informants provided the following definitions: konge - tall and slender, contrasting with the large and robust kisutu and intermediate futizi (Jambiani) - tall and long, predominantly black and white and colour, in contrast to the short, black and yellow kisutu. According to this informant, who had seen a male konge killed at Manyuni (on the coral rag on the way to Paje) in 1989, they are most often seen between Kitogani and Charawe / Ukongoroni (whereas kisutu are more common due east of Kitogani). He noted, however, that there is nothing fixed about these types, which can interbreed (Kitogani) - long, colour a mixture of red, black and a little white, in contrast to the short and red kichigi, which has no white on its coat (Makunduchi) - this informant described two kinds of konge: (a) black with large spots; and (b) white with small black spots. He contrasted both of these with the small and red kichigi (Makunduchi)

47

- long, with spots which reduce the amount of white background, whereas kisutu shows more white (Muungoni) konga - big, with more and larger spots than the small and yellower unyasi. konga are more common than unyasi, and a lot of those killed have been males (Muyuni B) mkonge - long, coloured yellow with some white, in contrast to the (unnamed) short type, which is bluish in colour (Kitogani) - long in contrast to koko, which is smaller and short in the back (Paje) - the largest of three types, coloured like a mkonge (sisal) leaf, whereas kisutu has large spots of every colour and unyasi is reddish, coloured like (dry) grass (Uzi) - large, in contrast to the smaller kichigi and differently coloured mwanzi (SIH, Zanzibar town) The common feature in these descriptions is that konge / mkonge is always taken to be larger (taller and / or longer) than the types it is contrasted with. There also seems to be some agreement that its coat has a paler background colour than other leopards, though this ‘white’ ground may in some cases be largely obscured by densely packed black spots. Marshall’s (1994) informants led him to believe that konge are ‘almost all black with only very slight indications of spots present’. This view was not shared by most of our own informants, though they did agree with Marshall’s sources that the konge is longer and taller than the kisutu (in cases where kisutu was recognised as a contrasting type). There is clearly no basis for Selkow’s (1995) hypothesis that konge describes a melanistic leopard, and it remains doubtful whether such ‘black panthers’ are to be found on Unguja at all. According to SIH, mkonge is also used as a name for large male and female Bush Pigs, Potamochoerus porcus. As in the case of kichigi (applied to small pigs), it is not immediately clear which mammal, leopard or pig, the term originally referred to, though mkonge is evidently more widely understood as a name for pigs. One of our informants (in Uzi) suggested that some leopards are called mkonge because of the resemblance between the background colour of their coats and the leaves of the plant(s) with the same name. mkonge (3/4) is the common Swahili name for a number of different plants, especially members of the Agavaceae family. On Unguja its referents include Canthium bibracteatum (Siex 1995), Sansevieria kirkii (Williams 1949), and cultivated sisal, Agave sisalana, the leaves of which provide a light-coloured fibre. At first sight this seems to be a plausible derivation, if only

48

because it parallels the plant-derived etymologies of two other leopard names, chui mwanzi and chui unyasi. A more likely source, however, is the name for another species of carnivore, which appears in the Swahili dictionaries as konje, though no animal with this name has been recorded on Unguja. The Standard Swahili-English Dictionary gives konje, ‘a small animal like a fox’, as well as the compound form kalakonje, ‘a kind of wild cat’, a definition it shares with kala (Johnson 1939). The Kamusi ya Kiswahili Sanifu defines konje as ‘a small animal like a fox / jackal (mbweha)’, and also refers the reader to kala (‘a small animal with a body like a large cat or cane-rat, which is very fond of eating chicks’), kalakonje (not defined), and njuzi (for which there is no separate entry) (Taasisi ya Uchunguzi wa Kiswahili 1981). Maimu (1982) does not record konje, but has kala as a name for the Small-spotted Genet, Genetta genetta, and suzi describing the Serval, Felis serval (1982). Variants of njuzi / suzi, with the same or similar meaning, are widespread in East African Bantu, and need not concern us further here (cf. terms recorded in Swynnerton 1946). Variants of kala are also widespread on the mainland, where they mainly refer to one or other species of mongoose, and appear to be linked to a reconstructed proto-Rift Southern Cushitic root (*kilambay-, ‘small long-tailed carnivore sp.’: Ehret 1980). In Giriama, one of the Mijikenda dialects on the Kenyan coast, the name kala has been identified with the White-tailed Mongoose, Ichneumia albicauda (Costich 1977). In Rabai, another Mijikenda dialect, kalakonje has been described as the ‘same species as the gala [sic], but somewhat larger and different in colour’, with a white body resembling Sansevieria fibres (Krapf and Rebmann 1887). This brings us back to the etymology which was presented earlier. If the folk etymology recorded by Krapf and Rebmann were correct, it is extremely unlikely that the qualifier konje would be detached from the compound and applied to the description of a quite different animal. The dictionaries suggest the identification of Swahili konje with a kind of fox or jackal: Swynnerton speculated that this may be the Bat-eared Fox, Otocyon megalotis (1946). Given that this is the only fox-like mammal present in East Africa, and the only member of the family Canidae which does not have a well-known Swahili name, this speculation may well be justified. There are, of course, no Bat-eared Foxes, or indeed truly wild canids of any description, on Unguja. The motivation for applying the inherited name for the Bat-eared Fox to a kind of leopard would have been much the same as that in the case of kisutu, already discussed above. The Bat-eared Fox has a relatively uniform dark grey-yellowish coat, and by lending its name to leopards without the characteristic pattern of rosettes, the Swahili settlers of Unguja may have been highlighting the contrast between this type and the more obviously spotted varieties of the Zanzibar Leopard. ngawa (9/10) [Ndudu] Only one informant, a woman in Ndudu, gave ngawa as the name for a type of leopard, contrasting it with the longer and yellower mwanzi. She described the ngawa leopard as patterned and coloured like the African Civet, Viverra civetta, the common name of which on Unguja is ngawa. The further etymology of this name,

49

which is sometimes heard in the compound form fungo ngawa, is obscure. fungo is an inherited term (proto-Sabaki *mfungo, from an earlier Bantu root: Nurse and Hinnebusch 1993), variants of which are usually applied to the African Civet. On Unguja (and Pemba), however, its primary reference is to the introduced and somewhat smaller Javan Civet, Viverricula indica. Hunters admit that it is possible to mistake the African Civet for a leopard, especially if sighted fleetingly, in poor light, and / or by an inexperienced observer. The use of its name for smaller leopards provides another illustration of the ease with which animal names can be transferred from one to another, though in this case the novel usage does not appear to have a wide currency. 3. Glossary of Euphemisms and Other Expressions bange (5/6?) [Jambiani] This term was given as a euphemism for ‘leopard’ by an elderly informant in Jambiani. Its linguistic origin is obscure: bange is also recorded as the name of an unidentified kind of fish (Johnson 1939; Taasisi 1981), but it is not clear how this might connect, if at all, with its use as a euphemism for leopards. bwana mkubwa (1/2) [Muyuni B] According to the single informant in Muyuni B who reported this expression, it was once used by leopard keepers when calling or referring to their kept male leopards. The literal meaning of bwana mkubwa is ‘big man’ or ‘boss’, a phrase which can be used either respectfully or with a touch of irony, depending upon the context. bwana on its own means ‘sir’, and by extension ‘husband’: mkubwa, from the root -kubwa, is a qualifying adjective meaning ‘big’ (or in this context ‘important’). The linguistic history of both of these terms is uncertain. chuma (7/8) [Makunduchi] A single informant in Kajengwa, Makunduchi, gave this term as a euphemism for leopards. The primary meaning of chuma in Swahili is ‘iron’, which can be traced back to proto-Sabaki *kyuma and an earlier Common Bantu form with the same or similar reference (Nurse and Hinnebusch 1993). chuma is also used figuratively of people in expressions such as ‘chuma cha mtu’, ‘a strong, tough-minded person’ (cf. Johnson 1939). This last usage is obviously the immediate source of the euphemism, which refers to the strength of leopards, Unguja’s primary carnivores and occasional predators of humans as well as their livestock. dume (uyo) (5/6) [Muyuni B, Pete] The use of this expression as a euphemism for leopards is perfectly transparent. Swahili dume refers primarily to male animals or plants, and in the case of animals

50

may also be used to suggest their superior size and strength (cf. Johnson 1939). It derives ultimately from Common Bantu via proto-Sabaki *ilume (Nurse and Hinnebusch 1993). According to the informant from Pete who gave dume as a euphemism for leopards, it usually occurs in the phrase ‘dume uyo!’, ‘that male (i.e. large, dangerous leopard)!’. Our informant in Muyuni B reported the same phrase, but said that it was once used by leopard keepers to refer to their kept male leopards, in the same way that bwana mkubwa was. keke (5/6) [Kitogani, Muungoni, Makunduchi] This term was elicited in two different areas, Makunduchi and Kitogani / Muungoni, and has possibly spread to the latter from the former. Both of our informants in Kajengwa, on the northern side of Makunduchi, gave keke as a euphemism for leopards. One informant said specifically that it is used to conceal the subject of discussion from women and children. The other Kajengwa informant suggested that this euphemism has its origin in the word makeke, which means ‘nervousness’ in the local dialect (the equivalent in Standard Swahili is wasiwasi), a state which would afflict anyone frightened by the sight of a leopard. keke (‘leopard’) and makeke (‘uncertainty’) are listed under a single entry in Chum’s (1994) vocabulary of Kae, the Makunduchi dialect, and the derivation is certainly plausible. The etymology of the root -keke is, however, obscure, and on the basis of existing evidence we cannot rule out the possibility that semantic innovation took place in the opposite direction (deriving makeke from keke, the description of a state of fear from the name of a fearful animal). While our Kitogani informant gave keke simply as a euphemism for leopards in general, the much younger hunter from Muungoni described keke as the largest and longest of the three varieties of leopard which he had heard about. He was, however, unable to name the two smaller types, and had some difficulty in recalling keke itself. It seems most likely that he had only remembered the euphemism, and incorrectly elevated this to the status of a named variety. Other hunters in this general area did not volunteer keke as a euphemism (or named variety), and this reinforces the thesis that this term has its origin in Makunduchi and the far south-east of the island. masharubu (-/6) [Uzi] Only one informant, a retired hunter in Uzi, gave masharubu as a euphemism for leopards, noting that in context its referent was unmistakeable. sharubu, plural masharubu, is the Standard Swahili term for a moustache, and with reference to leopards might be better translated as ‘whiskers’. As well as having a restricted distribution, this euphemism is probably of relatively recent historical origin, sharubu being a loanword from Arabic (Johnson 1939).

51

mfalme (1/2) [Kitogani, Muungoni, Pete] A number of informants used this term to signal the preeminence of leopards among Unguja’s wild fauna, rather than as a euphemism to avoid direct reference to leopards. The primary meaning of mfalme is ‘king, ruler’: it is presumably derived from an earlier Swahili form *mufalume (hence the personal name Mfaume), the provenance of which is unknown. The expression ‘chui ni mfalme’, ‘the leopard is the king’, was used by hunters on a number of occasions. This usage seems to centre in the area to the south of Jozani Forest, in other words close to the heart of the Kitanzi campaign. One informant in Kitogani stated that “chui alikuwa ni mfalme kwa Zanzibar”, “the leopard was Zanzibar’s ruler”, while another said that in the past leopards were “ufalme wa kiswahili”, “(traditional) Swahili royalty”. An informant in Pete, referring to the historical taboo upon eating leopard meat, commented that “humli mfalme”, “you don’t eat the king”. It is tempting to read these statements as verbal vestiges of a forgotten royal symbolism: unfortunately we have no other evidence to suggest that leopards played such a symbolic role in Unguja’s pre-Omani polities, though it is a possibility which cannot be ruled out. Otherwise it might be noted that at least one former Sultan of Zanzibar is said to have participated actively in leopard hunts. mirime (-/4?) [Jambiani] This euphemism was elicited from the same elderly informant in Jambiani who gave bange. It is possibly related to the Standard Swahili-English Dictionary’s mirimo, ‘the secrets of the medicine men, wizards, &c.’ (Johnson 1939), referring to leopards as the instruments of wachawi, witches. The etymology of mirimo is obscure. nyolopa (9/10) [Makunduchi] The only record of this term is in Chum’s (1994) vocabulary of Kae, the Makunduchi dialect, where it is glossed simply as ‘leopard’ and exemplified in the phrase ‘nyolopa kanafugwa’, ‘the leopard is being kept’. None of our own informants recognised this term, including two hunters from Kajengwa in north Makunduchi. Nonetheless, we take Chum’s record as reliable, though it is not clear from his entry whether the term nyolopa should be understood as the name of a variety of leopard, a name for kept leopards, a euphemism (like his keke), or a combination of these (perhaps a name which became a euphemism and subsequently lost its currency?). Linguistically nyolopa is possibly the most unexpected term in our list of leopard names and euphemisms. Its nearest linguistic relatives would appear to be terms of similar shape and related meaning in southern Tanzania. *inyalupala is the reconstructed term for ‘lion’ in proto-Southern Highlands: Southern Highlands being the name for a group of Bantu languages, including Hehe, Bena, Sangu, Wanji, Kinga, Pangwa and Manda, spoken to the north and east of Lake Malawi (Nurse 1988). Likewise the regular reflex of the inherited form in contemporary languages in this group is inyalupala (cf. Swynnerton 1946). Derek Nurse (personal communication) has also drawn attention to the existence of another parallel form,

52

nyalubwe, also meaning ‘lion’, in Mwera, a little-known language spoken by about 3,000 people in the vicinity of Mbamba Bay on the eastern shore of Lake Malawi. This term has also been reported from the other side of Lake Malawi: according to the big game hunter Peter Hathaway Capstick nyalubwe was his nickname in the Nyanja language of Zambia, in this case actually meaning ‘leopard’ (Capstick 1983). As Nurse points out, it is difficult to explain the relationship between these terms: they are obviously similar but not cognate, so have presumably spread as loans from one language to another and / or from a lost fourth source. The apparent antiquity of the Southern Highlands form, and the physical proximity of Mwera, suggest an origin in this general area. However, while it is easy to explain a semantic shift from ‘lion’ to ‘leopard’ following the arrival of the term on Unguja island (where leopards are the largest carnivores), it is not very easy to see how it might have reached there, and Makunduchi in particular. One possibility is that it was brought back along early trade routes from the coast into the interior, from Kilwa to the Southern Highlands and Lake Malawi region, and was later carried over to Makunduchi from Kilwa or elsewhere on the coast. It seems unlikely to have been brought to Zanzibar in the nineteenth century: most loans from the interior in this period can be found in the Unguja dialect and its derivative, Standard Swahili. In the absence of corroborating evidence, however, this is no more than speculation. shambi-shambi (5/6?) [Pete] This term was given by a single informant at Pete, who suggested that both shambi-shambi and shwambu were the names of varieties of leopard rather than euphemisms as such. He described shambi-shambi as referring to a leopard similar in appearance to ngawa, the African Civet, Viverra civetta. The origin of this term is obscure. shwambu (5/6) [Jambiani, Kitogani, Pete] This was the most frequently reported euphemism (followed by keke). It was elicited from all three of our informants in Pete, as well as one from Kitogani and one from Jambiani, suggesting that its widest currency is in the former village. One Pete hunter specified that the name shwambu is used when a leopard is heard calling and is presumed to be dangerous. The youngest and least knowledgeable of the Pete hunters, who knew the name kisutu but could not provide a description of it, gave shwambu (and also shambi-shambi, see above) as the name of leopards which are very long and have clearly defined black patches and larger areas of white than others. The balance of evidence suggests that he was confused. Nonetheless, the etymology of shwambu is unknown, and it may be that it originally referred to a particular variety of leopard. twiga (9/10) [Kitogani] According to the single Kitogani informant who gave this term, twiga was introduced as a euphemism for leopard meat when hunters began eating it in the

53

course of the Kitanzi campaign in the 1960s. twiga is Standard Swahili for ‘giraffe’, deriving from the reconstructed proto-Sabaki *ntwiga and an earlier East African Bantu form with the same meaning (Nurse and Hinnebusch 1993). There are, of course, no giraffes on Unguja. It is not clear what led to the adoption of ‘giraffe’ as a euphemism for ‘leopard meat’: one possible motivation, in addition to the humorous image which it conjures up, may have been a perceived similarity in the colour and marking of the two animals. Given that the consumption of leopard meat was previously unheard of and remained tabooed to many hunters, it is not surprising that those at the physical centre of the Kitanzi campaign should have felt the need to conceal their participation in this novel practice from their fellow villagers, including suspected leopard keepers.

54

Figures 1. Photograph of a leopard trap in Chwaka, by Pakenham (1947).

2. Photograph of a leopard trap, in Zanzibar Natural History Museum.

3. Representation of Zanzibar leopard (top) and typical East African leopard (bottom), by Kingdon (1989:27). Note small size of spots and rosettes on Zanzibar leopard. Figure not available for this PDF.

55

4-6. Mounted Zanzibar leopard in Zanzibar Natural History Museum. This leopard was shot at Kisakasaka by the Hon. W. Grazebrook, M.C. No year is given.

56

7-9. Fragment of a leopard pelt, purportedly from a Zanzibar leopard. Pelt in the possession of a Zanzibari hunter. Original report included two additional photographs.

10-12. Pugmarks that Kenyan naturalist and hunter, A. Archer, identified as belonging to leopard. Found on Wangwani path, northern Jozani Forest, in June 1996. The middle photograph shows two overlapping prints. Figures not available for this PDF.

The scarred knee of a Zanzibari who as a boy was attacked by a leopard, described on p. 8. This photograph not in the original report.

57

Literature cited Alden, P. C., R. D. Estes, D. Schlitter, and B. McBride. 1995. National Audubon Society Field Guide to African Wildlife. New York: Alfred A. Knopf. Archer, A. L. 1994. A survey of hunting techniques and the results thereof on two species of duiker and the suni antelopes in Zanzibar. Zanzibar: FS, CNR/ FINNIDA. Bailey, T. N. 1993. The African Leopard: Ecology and Behavior of a Solitary Felid. New York: Columbia University Press. Borsa, F. 1987. Coral Rag Agriculture: A Compilation of Secondary Data up to 1984. Zanzibar National Archives file no. BA 108/17. Britton, P. L., ed. 1980. Birds of East Africa: Their Habitat, Status and Distribution. Nairobi: East Africa Natural History Society. Capstick, P. H. 1983. Death in the Dark Continent. New York: St. Martin's Press. Chum, H. 1994. Msamiati wa Pekee wa Kikae: Kae Specific Vocabulary. Uppsala: Nordic Association of African Studies. Clark, F. and A. A. Khatib. 1993. Sea Turtles in Zanzibar: Status, Distribution, Management Options and Local Perspectives. Zanzibar Environmental Study Series, No. 15b. Zanzibar: Commission for Lands and Environment. Costich, D. E. 1977. A checklist of mammals in Gedi National Park with KiGiriama names. East Africa Natural History Society Bulletin. Jan-Feb: 12-13. Ehret, C. 1980. The Historical Reconstruction of Southern Cushitic Phonology and Vocabulary. Berlin: Dietrich Reimer. Ehret, C., E. D. Elderkin, and D. Nurse. Dahalo lexis and its sources. Afrikanistische Arbeitspapiere. 18:5-49. Goldman, H. V. 1996. A Comparative Study of Swahili in Two Rural Communities in Pemba, Zanzibar, Tanzania. Ph.D. dissertation, New York University. Hes, L. 1991. The Leopards of Londolozi. Cape Town: Struik Winchester. Ingrams, W. H. 1931 Zanzibar: Its History and Its People. London: Frank Cass and Co. Johnson, F. 1939. A Standard Swahili-English Dictionary. Oxford: Oxford University Press. Khamis, K. A. 1995. Report on the status of Zanzibar leopards from 15th Dec. 1994 to June 1995 in different times at Zanzibar. Moshi: Mweka College of African Wildlife Management. Kingdon, J. 1977. East African Mammals: An Atlas of Evolution in Africa. Vol. IIIA, Carnivores. Chicago: University of Chicago Press. 1989. Island Africa: The Evolution of Africa's Rare Animals and Plants. Princeton: Princeton University Press. Koenders, L. 1992. Fauna of Pemba Island: A Checklist of Animals Occurring on Pemba Island, Tanzania. Publication no. 1, Wildlife Conservation Society of Tanzania. Krapf, l. and J. Rebmann. 1887. Nika-English Dictionary, T. H. Sparshott, ed. London: Society for Promoting Christian Knowledge. Marshall, S. 1994. The status of the Zanzibar leopard. Zanzibar: FS, CNR.

58

Miththapala, S., J. Seidensticker, and S. J. O'Brien. 1996. Phylogeographic subspecies recognition in leopards (Panthera pardus): molecular genetic variation. Conservation Biology. 10(4):1115-1132. National Hunters, various years. Hunt reports and annual summaries kept in the Department of Regional Administration, Zanzibar. Nowell, K. and P. Jackson, eds. 1996. Wild Cats. Gland, Switzerland: IUCN. Nurse, D. 1988. The diachronic background to the language communities of southwestern Tanzania. Sprache und Geschichte in Afrika. 9:15-115. 1996. Personal communication to Martin Walsh. Nurse, D. and T. J. Hinnebusch. 1993. Swahili and Sabaki: A Linguistic History. Berkeley: University of California Press. Othman, W. J. and H. V. Goldman. 1996. Red colobus census of Unguja Island, Zanzibar, 1996: partial results and preliminary analysis of questionnaires distributed to all shehias. Zanzibar: FS, CNR/JCBCP. Pakenham, R. H. W. 1947. Land tenure among the Wahadimu at Chwaka, Zanzibar Island. Zanzibar: Government Printer. 1979. The Birds of Zanzibar and Pemba: An Annotated Check-list. BOU Checklist no. 2. London: British Ornithologist's Union. 1984. The Mammals of Zanzibar and Pemba Islands. Harpendon: printed privately. Pockock, R. I. 1932. The leopards of Africa. Proceedings of the Zoological Society of London, 542-591. Purseglove, J. W. 1972. Tropical Crops: Monocotyledons. Harlow: Longman Scientific and Technical. Selkow, B. 1995. A survey of villager perceptions of the Zanzibar leopard. Zanzibar: FS, CNR. Siex, K. S. 1995. The Zanzibar Red Colobus Monkey (Procolobus kirkii): Ecology, Demography and Use of Cocos nucifera. MSc dissertation, University of Florida. Swinnerton, G. H. 1946. Vernacular names for some of the better-known mammals in the Central Province, Tanganyika Territory. Tanganyika Notes and Records. 21: 2138. Swynnerton, G.H. and Hayman, R. W. 1951. A checklist of the land mammals of the Tanganyika Territory and the Zanzibar Protectorate. Journal of the East Africa Natural History Society. 20: 274-392. Taasisi ya Uchunguzi wa Kiswahili. 1981. Kamusi ya Kiswahili Sanifu. Dar es Salaam: Oxford University Press. Walsh, M. T. 1985. Shisango Dictionary. Unpublished ms. 1996. The Zanzibar leopard: an anthropological survey. Zanzibar: FS/JCBCP. Williams, A., A. A. Mwinyi, and S. J. Ali. 1996. A Population survey of the miniantelope—Ader's Duiker (Cephalophus adersi), Blue Duiker (Cephalophus monticola sundevalli), Suni (Neotragus moschatus moschatus)—of Unguja, Zanzibar. Zanzibar: FS, CNR/JCBCP. Williams, R. O. 1949. The Useful and Ornamental Plants in Zanzibar and Pemba. Zanzibar: Zanzibar Protectorate. Wilson, F. B. 1940 (possibly 1939). Notes on peasant agriculture and industries in Zanzibar Island. (Zanzibar National Archives file no. AU2/54.)

59

Zanzibar Goverment. 1995. Ripoti ya utekelezaji wa kazi za kitengo cha uhifadhi mwezi wa Januari 1995. (Kept in the Wildlife Facts File at the Zanzibar National Archives.)

60