Navigation – Plan du site
études et essais

A complex relationship:

Mozambique and the Comoro Islands in the 19th and 20th centuries
Edward A. Alpers
p. 73-96


Une relation complexe : le Mozambique et les Comores aux 19e et 20e siècles. -- Les Comores ont joué un rôle crucial dans l'histoire commerciale et religieuse du sud-ouest de l'océan Indien et en tant que trait d'union humain entre le continent africain, Madagascar et les Mascareignes. Dans cet article, je cherche à examiner ces trois aspects par rapport au Mozambique et dans le but de stimuler la recherche collective. Le premier aspect concerne les liens commerciaux étroits unissant les Comores au Mozambique. Alors que l'on a jusqu'ici mis l'accent sur le commerce des esclaves au cours du XIXe siècle, on ne doit pas sous-estimer la complexité et la prospérité du commerce de denrées alimentaires qui faisait également partie intégrante de ce même réseau régional. Ces relations sous-tendent les deux autres aspects abordés dans ce travail. Le second aspect concerne la diaspora africaine de l'océan Indien. J'analyse la façon dont des populations africaines déplacées ont été absorbées par les sociétés d'accueil, les moyens qu'elles ont utilisés pour à la fois maintenir et transformer leurs identités ainsi que les traditions qu'elles ont apportées avec elles. Le troisième aspect a trait à l'histoire de l'islam dans le nord du Mozambique au XIXe et au XXe siècle, histoire qui est intimement liée à celle des Comores.

Haut de page

Texte intégral

  • 1 This paper, now slightly revised, was originally presented at the "Colloque marquant le XXe anniv (...)

The Comoros have historically played a vital role in the commercial and religious history of the south-west Indian Ocean and as a human bridge between the African continent and Madagascar. In this paper I explore three aspects of this relationship that have hitherto been relatively ignored1.

The first topic examines the web of trading and political connections between the Comoros, Madagascar, and Mozambique. While much attention has been focused on the slave trade of this commercial circuit in the 19th century, we must not overlook the thriving, complex exchange of foodstuffs that was equally part of the same regional network. Both of these are well attested to in Portuguese sources and reveal how normal this traffic would have been regarded by the inhabitants of the coastal areas and islands of the northern half of the Mozambique Channel. These connections provided the foundation for the second and third issues that I will discuss in my paper.

The second aspect of this relationship focuses on the African diaspora in the Indian Ocean world. One of the features of the Indian Ocean diaspora that distinguishes it from the Atlantic world is the forced overseas migration of African labor to other parts of the African continent, including the offshore islands. There are or until recently were identifiable relocated populations from East Central Africa and the Tanzanian hinterland (mainly of "Nyasa", Yao, Makua, Ngindo, and Zigua background) as far afield as southern Somalia, the Kenya coast, Pemba, Zanzibar, the Comoros, and western Madagascar. I am interested in how these populations were absorbed into their host societies, the ways in which they maintained and transformed their own cultural identities, and the influences that they carried with them into these new historical situations. The Comoros are an integral part of this aspect of the diaspora, one that merits closer examination.

The third element concerns the history of Islam in northern Mozambique, which is intimately tied to the Comoros. This connection derives partly from the new commitment to Islamic education that marked the close of the nineteenth century in the face of aggressive European colonialism, but there is good evidence from earlier in the century for an especially tight Islamic network linking the dominant political lineage at Angoche to the Comoros. The critical moment, however, dates to the introduction of the Shadhiliyya Yashruti and the Qadiriyya turuq (sufi orders), both of which had strong Comorian connections, at Mozambique Island a century ago. Moreover, even during the later colonial period, Portuguese authorities looked to the Comoros for adjudication of problems within the Muslim community of Mozambique.

In what follows I attempt a preliminary reconstruction of each of these histories from the sources available to me and seek to identify future questions for research that arise from this initial effort. Above all, I hope to be able to stimulate new research involving collaboration with and among Comorian and other regional scholars on these important historical questions.

Trade and Politics: Mozambique and the Comoros in the Nineteenth Century

  • 2 For the earlier history of these exchanges, see NEWITT (1983).
  • 3 See, e.g., original letters in Arabic with Portuguese translations in Biblioteca Nacional de Lisb (...)
  • 4 For the era of the Malagasy raids, see J. MARTIN (1983, I: chap. 2), ALPERS (1977).
  • 5 With the notable exception of correspondence from the Sakalava interloper, Andriantsoli, to the G (...)
  • 6 Biblioteca Nacional de Lisboa, Reservados, Fundo Geral, Códice 8470, fl. 27, Sultane Chei Mondar (...)

There can be no meaningful appreciation of the history and culture of the Comoros in isolation from the human currents of the Mozambique Channel and the wider world of the western Indian Ocean. Towards the end of the eighteenth century these were severely disrupted and then transformed, first by the beginning of several decades of intermittent warfare by Malagasy maritime raiders against both the islands and the African littoral and, second, by the growth of slavery and slave trading throughout the sub-region2. One consequence of the threat of Malagasy raids was the regular exchange of intelligence through Muslim emissaries in the first decade of the new century between the Portuguese authorities at Mozambique Island, the Sultan of Nzwani, and the Queen of Bombetoka, the important trading entrepôt that was the predecessor of modern Mahajanga3. Although the era of Malagasy raids effectively ended in 1820, a sense of shared political vulnerability and convergence of commercial interests fostered the maintenance of official correspondence between the rulers of the Portuguese colony and the island sultanates for the next decade4. Indeed, virtually all of this correspondence relates either to matters of political intrigue and rivalries on the islands or to trading in provisions. Much of it involves letters between Portuguese officials at Mozambique Island and the Sultan of Nzwani or his representatives; only a few concern either Ngazidja or Mwali, while none are recorded from or to Mayotte5. Portuguese records of ships entering the port of Mozambique Island, on the other hand, record arrivals of ships from all four islands. For example, an early letter from the Sultan of Mwali to the Governor-General of Mozambique was precipitated by war between that island and Nzwani, its effective overlord, in 1828. A year later, the Sultan's assertion that the forces of Nzwani had suffered many more losses than had his own was confirmed by a letter from the Sultan of Nzwani to the same Portuguese official that claims some 400-500 deaths from this conflict. In addition, the Sultan of Nzwani sought to trade with the Portuguese, asking in particular for "some old slaves or small boys for my agriculture" 6. Many of these letters consist of professions of friendship, including offers of one party or the other to provide hospitality and services for the facilitation of trade between Mozambique and the Comoros.

  • 7 My thanks to Iain Walker (University of Sydney) for explaining this to me: email dates 17 August (...)
  • 8 See Public Record Office, London, Colonial Office 415/7, A. No. 172 (31-33), Captain Pilkinhome's (...)

Although the slave trade inevitably attracted most attention from European observers, it would be a mistake to consider that this was the only trade carried on between the Comoros and Mozambique. The islands supplied Mozambique with rice, sorghum, finger millet, some oats, beans, mung or green gram, coconuts, coconut oil, ghee, honey, goats, cattle, and sambo (from ntsambu in Shingazidja; Cyclas circinalis), a form of sago that was used only for feeding slaves7. We know from Portuguese accounts that these provisions were carried to Mozambique aboard vessels belonging to shipowners from the islands. Although most of these records refer only to ships making port at Mozambique Island, some are reported to have touched at the Kerimba Islands, as well8 (Prior 1819: 55; d'Avezac et al. 1848: 117, 120). A French visitor at Mwali in 1828 notes that ships from this island annually traded to both Mozambique Island and Quelimane (Leguével de Lacombe 1840: 331). This body of evidence also includes one instance of a Portuguese vessel bringing foodstuffs from the Comoros to Mozambique Island and one case of a local Indian trader seeking a license to trade for sambo at Ngazidja (Santana 1974: 981 #160, 15 #15).

  • 9 J. MARTIN (1983, I: 356-360, 587-588) argues that this incident was part of Mwinyi Mkuu's attempt (...)
  • 10 For the history of this period of famine, see NEWITT (1988).

Depending on local politics, however, relations were not always cordial; for example, traders from the Comoros were sometimes regarded as a threat to local agriculture by Portuguese officials because of the low price of their provisions. According to one source, Comorian vessels had not been admitted to the Portuguese roadstead before 1821-1822, when traders from Mwali first gained access, while the Governor-General in 1830 strongly opposed the role of Comorian trade at the Portuguese colonial capital (Santana 1967: 337-339 #5/2). In fact, it seems that European traders were no more welcome at Mwali than were Comorians at Mozambique. In 1828, when the French traveler Leguével de Lacombe was captured at Mwali, he discovered that Sultan Husayn, brother of the Sultan of Nzwani, hated all Europeans because one trader had sailed away without payment for the full cargo of goods he had taken on at Mwali, despite the fact that he had previously accommodated them with supplies of provisions and "slaves that my subjects went to buy for them on the coast of Africa" (Leguével de Lacombe 1840: 316-317; J. Martin 1983, I: citations in Index). Indeed, in 1831 tensions increased as a consequence of the capture of a Portuguese ship by Arab pirates. A letter from the Sultan of Nzwani to the Portuguese Governor-General in April of that year requests that the latter make peace with the Sultan of Quitangonha, an independent Swahili chiefdom immediately to the north of Mozambique Island, and regrets the Arab seizure of this Portuguese ship. Hoping to re-establish "the former friendship" that existed between Nzwani and the Portuguese, he specifically comments on the differences that existed between the Arabs and the people of Nzwani, "among whom there exists only a community of belief" (Santana 1974: 163-164; for further details, see Dubins 1972: 102-103, 105, 108-109). A month later, the Sultan of Ngazidja, Amad Bun Sahid Sualee (Ahmed b. Said Ali b. Soali, better known as Mwinyi Mkuu (1792-1875)), afforded protection to some of the Portuguese subjects whose vessel had been attacked at sea. According to his letter to the Portuguese Governor-General, the ship had been seized by one Abdallah b. Muhammad b. Suleiman Marzuqui, four of its crew killed, and the remainder carried to the port of Quitanda (Kitanda, Itsandra), which lay beyond his authority on Ngazidja as Sultan of Bambao. The survivers fled to the Sultan's port, which although it is not named would have been Moroni, from which he sought to repatriate them to Mozambique9 (Santana 1974: 159 #121). Whatever tensions and disruption to trade were produced by this incident, reality soon intervened. During a crisis precipitated in 1831 by the culmination of the extended period of devastating famine in south-eastern Africa, the Portuguese establishment at Mozambique Island depended for its very existence on the importation of foodstuffs. Accordingly, in June the Governor-General of Mozambique wrote a series of desperate letters to one of the rulers of Madagascar, the governor of Bombetoka, and to the rulers of Nzwani and Ngazidja, requesting them "to send their pangaios [small sailing vessels] with provisions and cattle to Mozambique". By September the Governor-General could report that the crisis had passed and that Mozambique enjoyed "very good understanding with the kings of Ngazidja and Nzwani" among other regional potentates10 (Santana 1967: 942-943 #9/1-3, 951-952 #15/1).

  • 11 See also Biblioteca Nacional de Lisboa, Reservados, Fundo Geral, Códice 8470, fl. 48.
  • 12 I have seen only a transcription of the letter in Portuguese from João Távares de Almeida to Said (...)
  • 13 There is a single reference to a pangaio embarking in January 1828 from Ibo, in the Kerimba Islan (...)

Nor did the commercial and political connections stop there. The Governor of Mwali in 1828 was a refugee from Sofala, an ancient center of Muslim commercial and religious activity on the southern coast of Mozambique, where his brother had defeated him in a local power struggle. So while he spoke neither French nor Malagasy, he is reported to have understood Portuguese (Leguével de Lacombe 1840: 322-323; J. Martin 1983, I: 126, 432-433 n. 49). Similarly, the Portuguese were entangled diplomatically in the political restructuring precipitated in north-western Madagascar and the Comoro Islands following the death of Radama I, the Merina defeat of the Sakalava kingdom of Boina, and the Malagasy invasions of Mwali and Mayotte (J. Martin 1983, I: 124-131)11. But by 1832 Portugal's fortunes were in steep decline in eastern Africa and the role they had played as potential counterbalance to competing forces in the Comoros was rapidly eclipsed by France, England, and Zanzibar under Sayyid Said ibn Sultan (r. 1804-1856). Although no one has searched the Mozambican and Portuguese archives for subsequent references to correspondence and traffic between the Portuguese at Mozambique and the Comoros, it seems unlikely that the regular exchanges of "official" letters during the first three decades of the nineteenth century were sustained after the early 1830s. A decade later, however, we know that at least two claimants to leadership at Nzwani, Said Hamza El Masela and Sultan Alawi M'titi, withdrew strategically to Mozambique when their political fortunes at home waned (J. Martin 1983, I: 163, 462; Oumar 1997: 75, 80). Occasional exchanges of official correspondence must certainly have occurred subsequently, although the only one that I have seen dates to mid-1861 and was addressed by the Governor- General of Mozambique to Said Omar El Masela, the ambitious Prince of Nzwani who spent at least some time at Mozambique where he learned to speak a little Portuguese12. Similarly, we have scant knowledge of the trade in provisions from the Comoros to Mozambique for these decades13.

By contrast, we know quite a bit about the slave trade and the exportation of so-called "free" labor or "engagés libres" from Mozambique to the Comoros, western Madagascar, and beyond to the Seychelles and Mascarene Islands that was stimulated by the rise of plantation economies in those places (Alpers 1975: 214; Shepherd 1980: 75-80; J. Martin 1983, I: 259- 260; Capela & Medeiros 1989: 268-270; Campbell 1981, 1989a, 1989b). Nancy Hafkin quotes some fascinating correspondence addressed to Molidi Vulai, the Captain-Major of Sancul, a mainland port to the south of Mozambique Island under nominal Portuguese suzerainty, that was captured in 1879 as part of the Portuguese anti-slave trade campaign. Among them, a letter from Sultan Abdallah III of Nzwani reveals the problems facing individuals who engaged in slave trading at this time, not least because of the French occupation of Mayotte in 1843. Others indicate difficulties in settling accounts without reliable agents in each place of business. Most illuminating is a letter from Sultan Said Bakr of Ngazija written from Moroni to Molidi Vulai on 5 January 1878:

"My agent Modohama just arrived here, having left three slaves there with Ambar. I ask you to keep an eye on them. We do not know each other except through letters, but I hope that you will watch over this collection. Your brother's young nephew is on Mohilla and Allah willing when he finishes his studies I will send him home to Mozambique. I repeat that I would like you to watch over my accounts. My father Sultan Ahmed passed away [in 1875]."

  • 14 Nancy Jane HAFKIN (1973: 56-57) square brackets in Hafkin's original citation of Sultan Said Bakr (...)

As Hafkin astutely observes: " Despite the fact that the two had never met, a sense of Swahili kinship and mutual reciprocity emerges from their correspondence." Moreover, she continues, although this group of letters is apparently unique, "the trade patterns it represents were common" as were the linkages among Muslim elites throughout the south-western Indian Ocean14.

  • 15 For the identification of Sito (Sitou in Martin), see Ministério das Colónias, Comissão de Cartog (...)

At the end of the century, before the Portuguese wars of conquest finally brought their East African territories under effective colonial rule, Mozambique continued to provide refuge for political refugees from the Comoros, as the French steadily imposed their own colonial domination over the islands. Thus, in early 1898, the deposed regent of Mwali, Mahmoud bin Said, and a handful of followers sailed away to the Mozambique coast, where they made land at Sito, just south of the Arimba peninsula on Mussongoma Bay, eventually sailing north to Ibo before finally reaching Zanzibar aboard a German steamship (J. Martin 1983, II: 223)15. But if by this time the commercial and political links between the Comoros and Mozambique had reached their limit, the religious ties had already begun to take on renewed vigor.

The African Diaspora in the Comoros

Before taking up the history of the Islamic network that spans the Mozambique Channel, let me turn to the fate and impact of those Africans who were sent as forced emmigrants, whether in bondage or as contract laborers, to the Comoros in the nineteenth century. It is not my intention to reconstruct the history of slavery in the Comoros during the nineteenth century. Gill Shepherd provides a useful overview of this subject that draws upon the most important published sources and touches upon the major issues involving slavery during this period. In general, she makes no distinctions among different ethnic groups within the slave populations of the islands, but in two instances she does include details that address the socio-cultural position of slaves from Mozambique. In her analysis of slavery and hierarchy, Shepherd (1980: 92) observes:

" Locally-born, fully-Islamised and culturally-integrated slaves, the wazalia, despised the wamakwa, the Makua tribesmen newly arrived from Mozambique. These were the only slaves for whom the blunt term `slave' was used--mrumwa (pl. warumwa)--with its literal meaning of `used person'."

  • 16 For evidence of modern racial discrimination by "Arabs" against "Africans", among whom must be in (...)
  • 17 SHEPHERD (1980) probably draws upon A. GEVREY (1980: 84), for filed teeth and tatooing. For an ex (...)

She notes further that urban-based, acculturated slaves had little social intercourse with rural, agricultural slaves, a pattern that is familiar from many other slave-owning societies16. In addition, African-born slaves were easily identifiable "with their filed teeth and tatoos" (Shepherd 1980: 81), a description that would fit many Makua and Makonde at this time and well into the twentieth century. Not surprisingly, new slaves did their best to assimilate by adopting local habits of dress, customs, and above all, Islam17. Furthermore, since at Nzwani the term makwa was clearly pejorative, it is not surprising that those to whom it was applied did their best to claim other identities (J. Martin 1983, I: 44; Robineau 1966a: 56-57, 1966b: 24-25). Nor should we wonder that Makua chafed under the burden of excessive tax collection by the Nzwani ruling class (J. Martin 1983, I: 53, 55). In 1856, rural Makua joined a larger body of Sakalava led by Bakari Koussou in a revolt against the French at Mayotte. In seeking to quell the rebellion, the French approached several notable community leaders, including one Namkopo, "a Mozambican chief greatly respected among this element of the population" (J. Martin 1983, I: 227, 493 n. 184, quoted at 230, 494 n. 197, which also mentions an African leader named Paria). Half a century later, in 1902, the chief spokesman for the disaffected workers who unsuccessfully challenged the authority of the autocratic Léon Humblot on Mwali, was a Makua overseer named Catchou (J. Martin 1983, II: 226- 227, 361 n. 270). Clearly, these scattered examples of rural rebellion reflect the objective class position of these workers, but class and ethnicity were closely intertwined in the Comoros. Nevertheless, not all Makua were despised agricultural slaves. For example, two of the wives of Andriantsoli, the Sakalava Sultan of Bweni on Mayotte in the 1830s, were Makua. In 1899, during the brief tenure of Henri Pobéguin as French Resident for Grande Comore (Ngazidja), a Makua named Cazambo briefly served as chief for the Société Humblot police force (ibid.: 176). So at least some Mozambican slaves were incorporated into the changing ruling elites of the Comoros or had some possibilities for upward mobility during this era (J. Martin 1983, I: 468 n. 150). One wonders whether any oral traditions exist that would help us reconstruct the histories of these individuals or others like them who have escaped notice in the European documentation.

  • 18 Guillain's figures are cited in C. ALLIBERT (1984: 116); for the original source, see GUILLAIN (1 (...)
  • 19 For my ideas about the M'Chambara, see MANDALA (1990: 167-180). For comparison, see the detailed (...)

One gauge of the demographic impact on the Comoros of the importation of unfree African labor during the nineteenth century comes from the sporadic population censuses on Mayotte after the French occupation in 1843. In that year, of an estimated population of 3,000 half were reckoned to be slaves, among whom many were undoubtedly mainland Africans. When Charles Guillain classified the population after emancipation in 1846, he counted 843 Makua (16.0%), 513 Makonde (9.7%), and 372 Mozambiques (7.1%) among the total of 5,268 individuals. Whatever the ethnographic accuracy of Guillain's Makua and Makonde enumeration, his third category probably reflected slaves shipped from the ports of Quelimane and, perhaps, Inhambane. The real value of his careful attention to origins, however, is to disaggregate the entirely artificial ethnonym of "Mozambiques". His figures are also valuable because they indicate that not all slaves, whom he totaled at 2,733 (51.8%), were from these new African immigrant groups, who represented 63.2% of the slave population and 32.8% of the total island population18. In 1852, Philibert Bonfils engineered a detailed census that also included information on ethnicity, although without Guillain's specific attention to African origins. In a total population of 6,191 organized into 131 villages, a total of 2,036 (32.9%) individuals were identified as Mozambiques (Makoas); in the two villages on the islet of Pamanzi, 157 of 707 souls (22.2%) were Mozambiques (J. Martin 1983, I: 206). The virtually unchanged total percentage of slaves from Mozambique in these two censuses suggests that Guillain's internal distinctions between Makua, Makonde, and Mozambiques accurately reflects the ports of origin from which slaves were shipped from the northern Mozambique coast to the Comoros: namely, the central littoral that included Angoche, Sanculo, and Quitangonha; the Cape Delgado/Kerimba Islands zone; and Quelimane. The next census for Mayotte dates to 1866 and includes a total of 3,716 Africans (31.7%) in a total population of 11,731. Of particular interest in this census is the demographic division by sex, which notes that men constituted 76.0% (2,824) of the African population in a society where men comprised 58.0% of the total population. This should not surprise us in what was by then a vigorous plantation economy where male agricultural labor was at a premium. Even more striking, however, are the 1867 figures for contract laborers working at Mayotte, of whom Mozambiques constituted the overwhelming majority, 2,245 (74.8%) out of 3,002, with Mozambican men outnumbering women by 1,951 to 294, or 86.9% of all Mozambican engagés. Gevrey (1980: 253, 277-278) also notes approvingly that "after having ended their contract, the majority renew, remain on the island, and augment the labor force". Despite the fact that he lumps all Africans together as Mozambiques in discussing Mayotte, Gevrey observes that although all slaves introduced from both the African coast and Madagascar are subsumed under the name of "Cafres" in the Comoros, the majority belong to three ethnic groups, "Makoua, Montchaoua and M'Chambara". The Makua identification is clear; that of Montchaoua can confidently be identified as the Yao, the important trading people of northwestern Mozambique and southern Malawi, and probably indicates shipment from Quelimane by this time; that of M'Chambara, however, is obscure, although my suspicion is that it may have been a name given to people coming down to Quelimane from the lower Tchiri valley (ibid.: 82)19.

  • 20 For a Makhuwa text from western Madagascar, see also GUEUNIER & KATUPHA (1991).

Although one cannot assume that what obtained at Mayotte also prevailed demographically on the other islands, at the end of the century Heudebert remarked upon the presence of "Makua, Makonde and the people of Nyasa", who certainly reflect both slave and "free labor" exports from Quelimane, among the African population of Ngazidja (Heudebert, cited in Allibert 1984: 116). And if the skewed gender balance of the African population at Mayotte roughly approximates that for the Comoros as a group, this would have had a significant impact on the transmission and retention of cultural traits by these diaspora populations, as also on the cultural impact of these groups on island societies. In addition, residential patterns that maintained social distance between urban and rural Comorians must be taken into account when seeking signs of African origins, as Shepherd suggests and others have equally noted (Tara & Woillet 1969: 355). According to Gevrey (1980: 82), "Their customs vary with the duration of their residence; for a while they retain their customs and the language of their country, but nearly all end by adopting the religion and habits of the Arabs". Not surprisingly, little remains of clear Mozambican origin in the languages of the Comoros. According to Sacleux's research, only a dozen words of Makhuwa have found a home in the Comorian lexicon, although his work suggests that there have been more retentions from Bantu languages of the Upper Zambezi around Sena (Sacleux, cited in Allibert 1984: 116, 152, 228). If this observation can be verified, it may reflect the fact that the last major group of African migrants to the Comoros embarked from Quelimane, despite the fact that the generic name for Mozambique slaves became Makua, for the people who dominated the earliest nineteenth century wave of forced labor migrants. Nevertheless, considering the evidence for the survival of Makhuwa-speaking communities and for the retention of Makhuwa folklore in western Madagascar and on Nosy Be well into the twentieth century, it is surprising that so little of this language has been recorded for the Comoros. According to Gueunier, in fact, at the end of the nineteenth century Pobéguin compiled a comparative lexicon of 250 words for a variety of languages spoken in the Comoros, among them "makoua" (Gueunier 1996: 313)20. So it may be that Makhuwa survivals were more significant a century ago than they would appear to be from Sacleux's oddly compiled dictionary, since Sacleux never visited the Comoros (Chamanga & Gueunier 1979: 23-24).

  • 21 See the paper of François Bensignor and Zainab Elyas, "Musique des Comores", on the web site http (...)
  • 22 Pierre VÉRIN (1994: 253) specifically mentions the important unpublished thesis of Damir ben Ali, (...)

Other examples of diaspora culture refer to music and dance. At Antananarivo in the early 1820s, Leguével de Lacombe noted that the "bobre africain", a widespread form of musical bow with a calabash resonator at its base and an instrument that he recognized from Mauritius and Réunion, "had been imported to Madagascar by the numerous Cafre and Mozambican slaves that the Arabs had brought there", a comment that would seem to be verified for the Comoros by Capmartin and Colin's observation in 1810 that Africans there played an instrument called the "violon Mozambique" (Leguével de Lacombe 1840: 111; Capmartin & Colin 1810: 149). Gevrey (1980: 84) provides further testimony to the African origins of this widely disseminated percussive stringed instrument. These last two sources also specifically identify the tam-tam as a drum of African origin. According to Bensignor and Elyas, during the nineteenth century new styles of music were introduced to the Comoros through the slave trade and contract labor from Africa. These new musical styles were often gender specific, women performing deba, lelemama, wadaha, and bora, men dancing mshogoro, shigoma, zifafa, and sambe. They note that the dance known as ikwadou is said to have been performed in the past by slaves, while today it is to the Comoros what maloya is to Réunion, "with its sense of reclaiming the past, its function of gathering ordinary people together and building solidarity"21. It will take further ethnomusicological research to determine the specific origins of these musical and dance genres, at least one of which--lelemama--clearly has wider regional provenance with connections to Manyema slave origins in eastern Congo through Zanzibar (Strobel 1979: 13-14)22.

  • 23 For trumba spirits in Mayotte the indispensable source is LAMBEK (1981: 152-159; 1993: 31-68). Al (...)
  • 24 There is clearly disagreement among different authorities on how the group of African spirits are (...)

Closely related to music and dance is spirit possession. Most of the modern literature for the Comoros on this topic focuses on Mayotte and on Malagasy-speakers. At the beginning of our century, however, one writer ended his discussion of Malagasy tromba spirit possession at Mayotte with the additional note that "The Makoa tromba is called `mouzouka' or `doungoumara'". This is confusing, at best, since both mzuka and dungumaro were widely dispersed Swahili possessing spirits in this century, but at least it points the way to further research among Makua descendant communities (Aujas 1911: 141)23. No less interesting for its reflection of wider Western Indian Ocean cultural exchange is the identification of African possessing spirits today on Ngazidja as sera or somali and their dichotomous opposition to Muslim spirits, called rauhan or rauhani, both as distinct from authoctonous possessing spirits bearing island-specific names and the Malagasy trumba (Chouzour 1994: 65-69; Vérin 1994: 131). As Blanchy et al. (1993, 1996: 124) note, "They represent the two polar values in the social hierarchy: Arab Muslim and African pagan, white and black". Moreover, unlike those uncontrollable forces represented by indigenous spirits, rauhan and sera or somali "are of external origin" (ibid.)24. Here the interesting research question is to determine when this category of spirit entered the cosmology of Ngazidja and whether it replaced an earlier category, such as, for the sake of argument, mozambique.

The Islamic Connection

  • 25 For a regional overview of this history, see ALPERS (2000).

The history of Islam in both the Comoros and Mozambique dates back many centuries, but both experienced a significant revival in the ardor of their belief and practice from the second half of the nineteenth century as part of the wider reform currents sweeping the world of Islam and the Muslim response to the forcible imposition of European colonial rule in Africa and the islands. In particular, the penetration and expansion of Islam throughout the coastal hinterland and far interior of northern Mozambique effectively dates primarily from this period. Although the first phase of this process of religious revitalization and conversion can be traced to the activities of individual walimu and mashaykh, the second phase and more extensive phase was primarily the work of two major turuq, the Shadhiliyya Yashruti and the Qadiriyya, several different branches of which took the lead in Mozambique. The groundwork for the introduction of the turuq was laid, however, in the commercial networks and personal relationships arising there from linking coastal Muslims in northern Mozambique to their coreligionists in the Comoros and Zanzibar25.

  • 26 Arquivo Histórico de Moçambique, Maputo, Secção Especial No 20, Cota S.E., 2 III p 6, Portugal, P (...)
  • 27 Melo Branquinho, "Relatório", op. cit., pp. 83-84.

Although literacy in Arabic and advanced Islamic learning were strictly limited by class in the Muslim communities of northern Mozambique, even before the era of the turuq they were not isolated from the wider world of Islam. By the early decades of the nineteenth century the Comoros had already established an important regional reputation as a center of learning, a reputation that was further enhanced by the close connections that evolved between the Comorian 'ulama and their counterparts in the rising Busaidi state of Zanzibar from the mid-1830s on (Pouwels 1987: chap. 6 & 8). For educated Muslims in northern Mozambique, the walimu of the Comoros acquired a special place of honor and important families sent their sons to study with these men when they could. In 1830, for example, a son of the Shaykh of Quitangonha was sent to study in the Comoros (Hafkin 1973: 48 n. 41). The most important community linking Mozambique and the Comoros through Islam was Angoche, which by the beginning of the twentieth century was a major center for Islamic activity. According to oral traditions regarding the life history of Musa Momadi Sabo, who later gained considerable economic and political notoriety as Musa Quanto, in his youth he traveled up-country along the northern bank of the Zambezi River to its confluence with the Lugenda River on a proselytizing mission with a relative who was a sharif and a haji. Upon returning to the coast, he circulated through the well-established Islamic commercial network of the south-west Indian Ocean to Mozambique Island, Zanzibar, the Comoros, and northwestern Madagascar before sailing home to Angoche in the mid-1850s to make his mature reputation (Machado 1920: 63; Lupi 1907: 183; Coutinho 1893: 10-11). Although the sources do not indicate the identity of Musa Quanto's kinsman, who was apparently the real force behind this religious peregrination, my guess is that he was probably Comorian. In fact, descendants of Comorian ancestors inhabit the entire coast from Angoche north to Mozambique Island. Furthermore, there is some additional indirect evidence to support this hypothesis in the oral traditions that identify a man from the Comoros named Hassan who married into one of the leading clan families at Angoche, the Murreiane, even though this story would seem to be entangled in more generalized regional stories of Shirazi origins26. Still another case of the close linkages between these coastal settlements and the Comoros dates to 1878, when the son of the Shaykh of Sancul--the nephew of Captain-Major Molidi Vulai--was sent to study at Mwila, as we have seen above (Hafkin 1973: 56). More generally, on the mainland opposite Mozambique Island members of the Murrapahine lineage of the Lucasse clan around Mossuril recognized similar Comorian links. According to Melo Branquinho, "They do not know how to give a good explanation for the family and clan ties to Iconi canton in the Comoros, of which many are natives who come to these lands". In one place he notes that they were descendants of Molidi Vulai, but later in his massive intelligence report he states that a chef named Murrapahine, who ruled the territory of Lunga during Molidi Vulai's time, was their progenitor, and that he in turn was descended on his father's side from "a native of Grand Comoro [Ngazidja], of the Morone [sic] clan, and on his mother's side from the Muanacha clan of Sangade"27. Each of these examples speaks to the intimate family and religious connections between coastal Mozambique and the Comoros, especially Ngazidja, during the nineteenth century, a subject that cries out for further research by scholars on both sides of the Mozambique Channel.

  • 28 See CARVALHO (1988: 61-63), where additional details of these fissions are included. There is no (...)
  • 29 Melo Branquinho, "Relatório", p. 370.

In contrast to this neglected theme, there is no need to reiterate the well known career of Shaykh Muhammad Ma'ruf b. Shaykh Ahmed b. Abi Bakr (1853-1905), a sharif who was born into a Hadrami family at Moroni. Shaykh Ma'ruf championed the Shadhiliyya Yashruti in the Comoros, where it had been introduced by 'Abdallah Darwish, and in north-western Madagascar. He preached its message to all who would listen, even though this often put him in conflict with leaders of the dominant Alawiyya tariqa. Fleeing to Zanzibar as a consequence of political persecution in about 1886, Shaykh Ma'ruf returned to Ngazidja following the exile of his political nemesis, Sultan Said Ali b. Said Omar, by the French in September 1893. In 1896 the Yashrutiyya was apparently brought independently to Mozambique Island by a trader from Moroni named Shaykh Amur ['Amir] b. Jimba, who was based at Zanzibar and also conducted business at Madagascar. The following year, perhaps ordered there by 'Abdallah Darwish, Shaykh Ma'ruf himself is said to have visited Mozambique Island for a month and granted ijazat (diplomas) to two men, Muhamade Amade Gulamo and Nemane b. Haji Ali Twalibo (also known as Nemane Haji Galibo) whom he left as co-leaders of the Shadhiliyya Yashruti. We do not know why Shaykh Ma'ruf passed over Shaykh Amur, but a contest for authority ensued in which the latter emerged as the dominant force in the tariqa at Mozambique Island. In 1898 Shaykh Ma'ruf entrusted his brother, Sayyid 'Ali b. Sheikh with the responsibility of resolving what must have been an unseemly struggle for power in this new outpost of his branch of the Yashrutiyya. Sayyid Ali issued three additional ijazat as khalifas to Ussufo Jamal Amur, Issufo Cassimo, and Sayyid Junhar b. Saide Amade. He also negotiated an agreement whereby Shaykh Amur would remain as head of the Yashrutiyya until his death, at which time leadership would pass back to Shaykh Anlauê b. Saide Abu Bakari, who was khalifa of the tariqa in the Comoros. Apparently this arrangement resolved the internal struggles of the Yashrutiyya at Mozambique Island for the next quarter-century, when Haji Muhamade Amade Gulamo, one of the two original khalifas appointed by Shaykh Ma'ruf in 1897, succeeded Shaykh Anlauê as khalifa in 1921. Internal competition soon resurfaced, however, and in 1924-1925 a new branch of the tariqa, the Shadhiliyya E'Madhania, was created by direct intervention from the center of the order at Medina. A decade later, in 1936, a third and final branch of the Yashrutiyya broke off to become the Shadhiliyya Itifaque28. Notwithstanding these divisions, in the late 1960s, the Yashrutiyya was still connected spiritually to the Comoros29.

  • 30 Ibid., pp. 353-354, 409; Shaykh Issa's Comorian roots are not mentioned in CARVALHO (1988: 63).
  • 31 Melo Branquinho, "Relatório", pp. 360-363, 366. This is the only reference I have seen to the Gaz (...)
  • 32 Ibid., p. 408.
  • 33 Ilha de Moçambique: Danze e canti della costa settentrionale, Folklore 7, Musiche dal Nuovo Mondo (...)

The Comoros were also a powerful external point of reference for those Mozambican Muslims who became adherents of the Qadiriyya tariqa. Thus, although the Qadiriyya was introduced at Mozambique Island from Zanzibar in 1904 or 1906, the shaykh whose mission it was to propagate the tariqa, Issa b. Ahmed, "was a native of Moroni canton [.. .] having completed his Qu'ran studies in the Comoros and, much later, in Zanzibar". Given the many connections between Ngazidja and the Mozambique coast, his decision to carry this tariqa to Mozambique Island cannot have been entirely coincidental. In any event, Shaykh Issa remained at Mozambique Island as leader of the Qadiriyya Sadate until he returned to Zanzibar, where he died, in August 1925, so that although his successor as shaykh, Khalifa Momade Arune, was a native of Cabaceira Pequena on the mainland opposite Mozambique Island, he would certainly have been well aware of the prominence of the Comoros in the life of his shaykh30. A quarter century later, the khalifa of the Qadiriyya Sadate was Sayyid Momade Sayyid Habibo, better known as Baguir or Bacre, the nephew of the recently deceased shaykh of the order. In the 1960s Baguir was regarded with suspicion by the Portuguese because of his well developed connections to the extensive network of Islamic leaders in eastern Africa. In 1950, Baguir left Mozambique Island to study in Zanzibar in the company of Shaykh Xará Abahassane b. Ahmed, a native of Zanzibar whose father was a Comorian, who had come to visit in the Portuguese territory; in Zanzibar he studied for three years with Shaykh Saide Omar b. Abdalia, also a native Zanzibar of Comorian descent. After studying for a further four years in the Hadramaut, with which the Arab families of the Comoros had multiple ties, Baguir returned to Mozambique Island, where he was regarded as heir apparent to leadership of the Qadiriyya Sadate. Indeed, until the revolution of 1964, Zanzibar was widely regarded as the center of Islamic learning in East Africa, overshadowing other regional places for study, such as the Gazila Islamic Center in the Comoros, which Portuguese authorities claimed was "considered in official documents as the third most important Islamic center in East Africa". Following the revolution, however, the spiritual line to the headquarters of the tariqa in Baghdad now ran through the Comoros, rather than Zanzibar31. Thus, with Zanzibar incorporated into a secular Tanzania and Islam fragmented in Madagascar, at the time of Melo Branquinho's report in 1969, he could state confidently that "The Comorians constitute the most important Islamic element in steady growth"32. No better mnemonic reminder of this living presence is the name given to one of the four single-headed hand drums used at Mozambique and Angoche in tufo, an Islamic musical genre performed by men: Ngazidja33.

  • 34 Arquivo Histórico de Moçambique, PP 1145, Associação Maometana Comoreana (Association mahométane (...)

Elsewhere in the Portuguese colony, by 1931 there were enough Comorians resident in the new colonial capital of Lourenço Marques (today Maputo), far from the historic commercial and religious circuit of the northern Mozambique Channel, to give rise to the formation of an Associação Maometana Comoreana (Association mahométane comorienne). According to the official statutes of this association, it was designed to promote the education and recreation of its members for their "moral, intellectual and physical development". Membership was open to "all adult Muslim individuals of 18 years of age, natives of the Comoro Islands, Madagascar, and Zanzibar, and those born in other places, but children of parents who are natives of those places and themselves Muslims"34. The activities of this organization and its members are still another topic that requires further research.

  • 35 For the full story, see ALPERS (1999). The following paragraph is based upon sources cited in thi (...)
  • 36 For a brief mention of this important individual, see CHOUZOUR (1994: 52).
  • 37 See, e.g., the fascinating preliminary study by J. KAGABO (1991).

The last aspect of the complex relationship between Comorian and Mozambican Muslims that I want to discuss dates to the period of the armed liberation struggle waged by Frelimo for the independence of Mozambique and the Portuguese response in seeking to win the hearts and minds of Mozambican Muslims against what they feared was a common, atheistic enemy35. By 1971 Portuguese intelligence services had initiated a multifaceted psychological warfare offensive that was designed to win the support of the Muslim leadership of Mozambique, in particular that of the turuq, who were regarded as a traditional, conservative force, against Frelimo. In August 1972 the Portuguese organized a carefully orchestrated gathering of twenty-one Islamic leaders, including those of all eight turuq (five Qadiriyya and three Shadhiliyya), at Mozambique Island to launch the officially published Portuguese translation of al-Bukhari's Hadith, one of the key texts of Islam. Part of the propaganda generated by this very public event was the publication of a letter to Sayyid Omar b. Ahmed b. Abu Bakr b. Sumait al-Alawi, Mufti of the Comoro Islands, from Shaykh Momade Said Mujabo, a key figure in the Portuguese psychological counter-offensive at Mozambique Island who was not affiliated with any particular tariqa, in which he declares his joy at this expression of Portuguese suppport for Islam. In addition to their potential role as a conservative force against the secular, radical nationalism of Frelimo, the Portuguese also considered the turuq to be a more reasonable form of Islam than that of the new generation of radical Islamic reformers who were beginning to make their presence felt in the southern urban centers of the colony, especially Lourenço Marques. At about this time, the "Wahhabis", as they were dubbed by Portuguese intelligence services, had launched an attack against the unorthodox character of the traditional Islam practiced by members of the eight turuq that dominated Mozambican Islam, focusing their criticism on the way in which funerals were conducted and, especially, on the peaceful resolution that had been effected by the mediation of Shaykh Momade Said Mujabo between those who prayed silently and those who expressed their piety by shouting. Fearing this threat to one of their staunchest new Muslim allies, the Portuguese called officially upon the Comorian Mufti, Sayyid Omar b. Ahmed, to resolve this dispute about bid'a (innovation) between the eight Mozambican turuq and their "Wahhabi" critics. His decision in favor of the traditionalists brought about their enthusiastic endorsement of the Portuguese translation of the Hadith, the statement of which is included as a preface to the final Government publication in 197236. What is missing in this reconstruction based on official Portuguese sources is any idea of how this Comorian intervention in the religious politics of Mozambique was regarded by the Mozambican Muslim leadership and how it figured into the rapidly evolving and conflicted politics of French decolonization and nationalism in the Comoros. More generally, although there has been some work done on the Islamic learned networks of northern East Africa, with extensions down to the Comoros, nothing of this sort has yet been attempted for the networks of the Mozambique Channel. It should be37.

Since achieving independence by quite different routes less than a month apart in June and July 1975, both Mozambique and the Comoros have experienced many difficulties and challenges. While I have no specific information on the survival or attrition of the historic connections between these two Muslim communities, it would be interesting to know how these may have affected the alleged logistics support provided by South Africa through the Comoros during Renamo's war against Frelimo during the 1980s (Finnegan 1992: 33-34; Vines 1991: 67-68; Vérin 1994: 214-216). Finally, considering the increasing acceptance and official recognition of Islam as a major national religion in democratic Mozambique, which has recently become an official member of the Organization of the Islamic Conference, it seems likely that relations with the Federal Islamic Republic of the Comoros will endure into the future at the levels of state diplomacy, turuq organizational links, and family alliances. Out of this web of relationships I hope that we shall see increasing attention paid to the study of this shared history.

University of California, Los Angeles.

Haut de page


Allibert, C.

1984 Mayotte: plaque tournante et microcosme de l'océan Indien occidental. Son histoire avant 1841 (Paris: Éditions Anthropos).

Alpers, E. A.

1975 Ivory and Slaves in East Central Africa (London: Heinemann; Berkeley: University of California Press).

1977 "Madagascar and Mozambique in the Nineteenth Century: the Era of the Sakalava Raids (1800-1820)", Omaly sy Anio, 5-6: 36-53.

1984 "The Western Indian Ocean as a Regional Food Network in the Nineteenth Century", ICIOS. II: International Conference on Indian Ocean Studies, Perth, Western Australia, 5-12 December, Proceedings Section A: Resources, Environment and Economic Development.

1999 "Islam in the Service of Colonialism? Portuguese Strategy during the Armed Liberation Struggle in Mozambique", Lusotopie 1999: Enjeux contemporains dans les espaces lusophones: 165-184.

2000 "East Central Africa", in N. Levtzion & R. L. Pouwels, eds, The History of Islam in Africa (Athens: Ohio University Press): 303-325.

Arnfred, S.

1999 "Female Identity Politics in a Period of Change? Muslim Women's Dance Associations in Northern Mozambique", Paper presented at the annual meeting of the African Studies Association, Philadelphia, November 1999.

Aujas, L.

1911 "Notes historiques et ethnographiques sur les Comores", Bulletin de l'Académie malgache, 9: 125-141.

Avezac, M. P. A. D' et al.

1848 Îles de l'Afrique (Paris: Firmin Didot frères).

Blanchy-Daurel, S.

1990 La vie quotidienne à Mayotte (Paris: L'Harmattan).

Blanchy, S., Cheikh, M., Said, M., Allaoui, M. & Issihaka, M.

1993 "Thérapies traditionnelles aux Comores", Cahiers des Sciences humaines, 29 (4): 763-790.

1996 "Rituels de protection dans l'archipel des Comores", Islam et Sociétés au Sud du Sahara, 10: 121-142.

Campbell, G.

1981 "Madagascar and the Slave Trade, 1810-1895", Journal of African History, 22: 203-227.

1989a "Madagascar and Mozambique in the Slave Trade of the Western Indian Ocean 1800-1861", in W. G. Clarence-Smith, ed., The Economics of the Indian Ocean Slave Trade in the Nineteenth Century (London-Totowa, NJ: F. Cass): 166-193.

1989b "The East African Slave Trade, 1861-1895: The `Southern' Complex", International Journal of African Historical Studies, 22: 1-26.

1997 "The Origins and Demography of Slaves in Nineteenth Century Madagascar: A Chapter in the History of the African Ancestry of the Malagasy", in I. Rakoto (textes réunis par), L'esclavage à Madagascar: aspects historiques et résurgences contemporaines. Actes du colloque international sur l'esclavage, Antananarivo, 24-28 septembre 1996 (Antananarivo: Institut de civilisations, musée d'Art et d'Archéologie): 5-37.

Capela, J. & Medeiros, E.

1989 "La traite au départ du Mozambique vers les îles françaises de l'océan Indien, 1720-1904", in U. Bissoondoyal & S. B. C. Servansing, eds, Slavery in South West Indian Ocean (Moka, Mauritius: Mahatma Gandhi Institute): 247-309.

Capmartin, J. & Colin, E.

1810 "Essai sur les îles Comores", in Annales des Voyages de la Géographie et de l'Histoire, 3.

Carvalho, Álvaro Pinto de

1988 "Notas para a História das Confrarias Islâmicas na Ilha de Moçambique", Arquivo (Maputo), 4: 59-66.

Chagnoux, H. & Haribou, A.

1980 Les Comores (Paris: Presses universitaires de France).

Chamanga, M. A. & Gueunier, N. J., eds

1979 Le dictionnaire comorien-français et français-comorien du R. P. Sacleux (Louvain: Peeters).

Chaudenson, R., dir.

1981 Encyclopédie de la Réunion. V. Les modes de vie (Saint-Denis: Livres- Réunion).

Chouzour, S.

1994 Le pouvoir de l'honneur. Tradition et contestation en Grande Comore (Paris: L'Harmattan).

Coutinho, J. de Azevedo

1893 Do Nyassa a Pemba (Lisboa: Typog. da Companhia Nacional Editora).

Craster, J. E. E.

1913 Pemba, the Spice Island of Zanzibar (London: Leipsic, T. F. Unwin).

Dubins, B.

1972 "A Political History of the Comoro Islands, 1795-1886", Ph.D. dissertation, Boston, Boston University.

Finnegan, W.

1992 A Complicated War: The Harrowing of Mozambique (Berkeley: University of California Press).

Froberville, E. de

1847 "Notes sur les m¤urs, coutumes et traditions des Amakoua, sur le commerce et la traite des esclaves dans l'Afrique orientale", Bulletin de la Société de Géographie, 8: 311-329.

Gevrey, A.

1980 [1870] Essai sur les Comores (Antananarivo: Association malgache d'archéologie).

Gray, R. F.

1969 "The Shetani Cult among the Segeju of Tanzania", in J. Beattie & J. Middleton, eds, Spirit Mediumship and Society in Africa (London: Routledge & Kegan Paul): 171-187.

Guébourg, J.-L.

1995 Espace et pouvoir en Grande Comore (Paris: L'Harmattan).

Gueunier, N. J.

1996 "Documents sur la langue makhuwa à Madagascar et aux Comores (fin XIXe-début XXe siècles)", in F. Rajaoson (coord.), Fanadevozana ou esclavage. Colloque international sur l'esclavage à Madagascar, 1996 (Antananarivo: Institut de civilisations, Musée d'art et d'archéologie): 309-321.

Gueunier, N. J. & Katupha, J. M.

1991 Contes de la côte ouest de Madagascar (Antananarivo: Ambozontany; Paris: Karthala).

Guillain, C.

1851 "Mayotte", Revue de l'Orient, 9.

Guy, P. & Cheik Amir, A. bin

1949 La vie et l'¤uvre du grand marabout des Comores, Said Mohamed ben Ahmad al Ma`arouf (Tananarive: Imprimerie officielle).

Hafkin, N. J.

1973 Trade, Society, and Politics in Northern Mozambique, c. 1753-1913, Ph.D. dissertation, Boston, Boston University.

Kagabo, J.

1991 "Réseaux d'ulama `swahili' et liens de parenté: une piste de recherche", in F. Le Guennec-Coppens & P. Caplan, dir., Les Swahili entre Afrique et Arabie (Paris: Karthala; Nairobi: CREDU): 59-72.

Lambek, M.

1981 Human Spirits: A Cultural Account of Trance in Mayotte (Cambridge-New York: Cambridge University Press).

1993 Knowledge and Practice in Mayotte: Local Discourses of Islam, Sorcery, and Spirit Possession (Toronto: University of Toronto Press).

Leguével de Lacombe, B. F.

1840 Voyage à Madagascar et aux îles Comores 1823 à 1830, t. II (Paris: L. Desessart).

Lupi, E. do Couto

1907 Angoche (Lisboa: Ministerio dos Negócios da Marinha e Ultramar).

Machado, C. R.

1920 "Mussa-Quanto o Namuali (O Napoleão de Angoche)", Boletim da Sociedade de Geografia de Lisboa, 38 (1-2): 54-70.

Mandala, E. C.

1990 Work and Control in a Peasant Economy. A History of the Lower Tchiri Valley in Malawi, 1859-1960 (Madison, WI: University of Wisconsin Press).

Martin, B. G.

1976 Muslim Brotherhoods in Nineteenth Century Africa (Cambridge-New York: Cambridge University Press).

Martin, J.

1983 Comores: quatre îles entre pirates et planteurs. I. Razzias malgaches et rivalités internationales (fin XVIIIe-1875). II. Genèse, vie et mort du protectorat (1875-1912) (Paris: L'Harmattan).

Newitt, M.

1983 "The Comoro Islands in the Indian Ocean Trade before the 19th Century", Cahiers d'Études africaines, XXIII (1-2), 89-90: 139-165.

1988 "Drought in Mozambique 1823-1831", Journal of Southern African Studies, 15 (1): 15-35.

Oumar, A. H.

1997 Histoire des Îles: Ha'Ngazidja, Hi'Ndzou'ani, Maïota et Mwali (St. Denis La Réunion: Djahazi éditions).

Pouwels, R. L.

1987 Horn and Crescent: Cultural Change and Traditional Islam on the East African Coast, 800-1900 (Cambridge-New York: Cambridge University Press).

Prior, J.

1819 Voyage along the Eastern Coast of Africa, to Mozambique, Johanna, and Quiloa.. . in the Nisus Frigate (London: Sir R. Phillips & co).

Robineau, C.

1966a Société et économie d'Anjouan (océan Indien) (Paris: Office de la recherche scientifique et technique outre-mer).

1966b "L'islam aux Comores: une étude d'histoire culturelle de l'île d'Anjouan", Revue de Madagascar, 35: 17-34.

Romero, J.

1860 Supplemento á Memoria Descriptiva e Estatistica do Districto de Cabo Delgado com uma Noticia á cerca do Estabelecimento da Colonia de Pemba (Lisboa: Typographia Universal).

Sacleux, C.

1939 Dictionnaire swahili-français (Paris: Institut d'ethnologie).

Santana, F., ed.

1964-1974 Documentação Avulsa Moçambicana do Arquivo Histórico Ultramarino. I, 1964; II, 1967; III, 1974 (Lisboa: Centro do Estudos Históricos Ultramarinos).

Shepherd, G.

1980 "The Comorians and the East African Slave Trade", in J. L. Watson, ed., Asian and African Systems of Slavery (Oxford: Basil Blackwell): 73-99.

Strobel, M.

1979 Muslim Women in Mombasa, 1890-1975 (New Haven: Yale University Press).

Tara, V. & Woillet, J.-C.

1969 Madagascar, Mascareignes et Comores (Paris: Société continentale d'éditions modernes illustrées).

Vérin, P.

1994 Les Comores (Paris: Karthala).

Vines, A.

1991 Renamo: Terrorism in Mozambique (London: Centre for Southern African Studies/University of York/James Currey; Bloomington-Indianapolis: Indiana University Press).

Haut de page


1 This paper, now slightly revised, was originally presented at the "Colloque marquant le XXe anniversaire de la création du CNDRS (Centre national de documentation et de recherche scientifique)", Moroni, République fédérale islamique des Comores, 27-28 January 1999. I am most grateful for the support and hospitality of the CNDRS during my visit, especially to its Director General at the time, Djaffar Mmadi, and to Hachimo Soendi, who guided me around the island of Ngazidja. Thanks also to Jean-Claude Penrad for his comments on the paper and for bringing it to the attention of the editors of this journal.
2 For the earlier history of these exchanges, see NEWITT (1983).
3 See, e.g., original letters in Arabic with Portuguese translations in Biblioteca Nacional de Lisboa, Reservados, Fundo Geral, Códice 8470, fls. 1, 3, 6 (a, b, e, f), 70, 95-96.
4 For the era of the Malagasy raids, see J. MARTIN (1983, I: chap. 2), ALPERS (1977).
5 With the notable exception of correspondence from the Sakalava interloper, Andriantsoli, to the Governor of Mozambique in 1829, for which see DUBINS (1972: 138).
6 Biblioteca Nacional de Lisboa, Reservados, Fundo Geral, Códice 8470, fl. 27, Sultane Chei Mondar Bon Sultan Chei Abo Bacar to Governor-General, 20 June 1828; fl. 37, Sultan of Anjouan, 11 December 1829. The odd wording of the Portuguese translation, " alguns escravos velhos ou moleques pequenos para a minha agricultura" , which I suspect is a mistake for something like " mature slaves or boys" in the Arabic original, makes a strong argument for modern translations of the documents in this important codex. For Shaykh Mukhtar bin Abubakr, see J. MARTIN (1983, I: citations in Index); the Sultan of Nzwani at this time was Abdallah II bin Alawi (1823-1836) (ibid.: 141 and citations in Index).
7 My thanks to Iain Walker (University of Sydney) for explaining this to me: email dates 17 August 1999. See also SACLEUX (1939: 690).
8 See Public Record Office, London, Colonial Office 415/7, A. No. 172 (31-33), Captain Pilkinhome's Journal. See also SANTANA 1964: 703 #82-83; 1967: 32 #47, 760 #251, 804 #38, 805 #41-42, 806 #45-47, 808 #52, 818 #80, 819 #83, 823 #101, 827 #117; and 1974: 152 #101, 678 #144, 752 #41 & #43, 978 #149, 979 #150 & #152, 980 #154, 982 #162 & #164, 985 #173, 991 #196, 998 #211, 999 #215, and 1000 #218 (The sign # indicates the number of the document; in Santana's collection of documents there are often more than one document on a page). For a broad discussion of the trade in foodstuffs, including information on western Madagascar and Comorian links to Zanzibar, see ALPERS (1984).
9 J. MARTIN (1983, I: 356-360, 587-588) argues that this incident was part of Mwinyi Mkuu's attempt to gain recognition from the Portuguese of his authority over the entire island, which as his letter clearly indicates, he did not exercise. DUBINS (1972: 207-208) confirms the identification of Quitanda.
10 For the history of this period of famine, see NEWITT (1988).
11 See also Biblioteca Nacional de Lisboa, Reservados, Fundo Geral, Códice 8470, fl. 48.
12 I have seen only a transcription of the letter in Portuguese from João Távares de Almeida to Said Omar, Moçambique, 24 July 1861. I am indebted to Professor Martin Ottenheimer, Kansas State University, for sharing his copy of this document, which he located in the Comoros. For Said Omar, see J. MARTIN (1983, I: 209-211 and subsequent citations).
13 There is a single reference to a pangaio embarking in January 1828 from Ibo, in the Kerimba Islands, for Nzwani in J. ROMERO (1860: 123).
14 Nancy Jane HAFKIN (1973: 56-57) square brackets in Hafkin's original citation of Sultan Said Bakr's letter. For more on this individual, the son of Sultan Mwinyi Mkuu of Bambao, see J. MARTIN (1983, I: Index).
15 For the identification of Sito (Sitou in Martin), see Ministério das Colónias, Comissão de Cartografia, Dicionário Corografico da Província de Moçambique, 1o Fascículo, "Territórios de Cabo Delgado (Companhia do Nyassa)", Lisboa, 1919, pp. 115 and 163.
16 For evidence of modern racial discrimination by "Arabs" against "Africans", among whom must be included the descendants of former slaves, at Ngazidja, see CHOUZOUR (1994: 68-69) and GUÉBOURG (1995: 197-198). For the very same phenomenom at Mayotte, where the terms of opposition are expressed as being between kabaila or noble and murumwa or slave, see BLANCHY-DAUREL (1990: 66).
17 SHEPHERD (1980) probably draws upon A. GEVREY (1980: 84), for filed teeth and tatooing. For an example of how easily recognizable were such slaves, see FROBERVILLE (1847).
18 Guillain's figures are cited in C. ALLIBERT (1984: 116); for the original source, see GUILLAIN (1851: 225).
19 For my ideas about the M'Chambara, see MANDALA (1990: 167-180). For comparison, see the detailed reconstruction of East African slaves sources for nineteenth century Madagascar by CAMPBELL (1997).
20 For a Makhuwa text from western Madagascar, see also GUEUNIER & KATUPHA (1991).
21 See the paper of François Bensignor and Zainab Elyas, "Musique des Comores", on the web site See also CHAGNOUX & HARIBOU (1980: 45-46). For maloya, see CHAUDENSON (1981: 101-103).
22 Pierre VÉRIN (1994: 253) specifically mentions the important unpublished thesis of Damir ben Ali, Musique et société aux Comores (Paris, École des hautes études en sciences sociales, 1977), which I have not yet seen.
23 For trumba spirits in Mayotte the indispensable source is LAMBEK (1981: 152-159; 1993: 31-68). Although LAMBEK addresses cultural diversity in Mayotte and includes an appendix on "Additional classes of possession spiritis in Mayotte" (1981: 186-190), he does not record any mainland African possession spirits. For references to mzuka and dungumaro in their wider Swahili context, see SACLEUX (1939: 658, 176). For rohani as a specifically Arab possessing spirit associated with the sea elsewhere in East Africa, see CRASTER (1913: 305) and GRAY (1969: 174).
24 There is clearly disagreement among different authorities on how the group of African spirits are named. Thus, while BLANCHY et al. (1993, 1996) call the African spirits somali, CHOUZOUR (1994: 65), who calls them sera, includes a specific spirit called Msomali among the rauhani group of spirits. For rohani as a specifically Arab possessing spirit associated with the sea elsewhere in East Africa, see CRASTER (1913: 305) and GRAY (1969: 174).
25 For a regional overview of this history, see ALPERS (2000).
26 Arquivo Histórico de Moçambique, Maputo, Secção Especial No 20, Cota S.E., 2 III p 6, Portugal, Província de Moçambique, SCCI, Prospecção das Forças Tradicionais -- Distrito de Moçambique, Secret, submitted by Director of the Services Fernando da Costa Freire, Lourenço Marques, 30 December 1969, enclosing José Alberto Gomes de Melo Branquinho, Nampula, 22 April 1969, "Relatório da Prospecção ao Distrito de Moçambique" (Um estudo das estruturas das hierarquias tradicionais e religiosas, e da situação político-social), hereafter Melo Branquinho, "Relatório", pp. 18, 343-346.
27 Melo Branquinho, "Relatório", op. cit., pp. 83-84.
28 See CARVALHO (1988: 61-63), where additional details of these fissions are included. There is no mention of Shaykh Ma'ruf's visit to Mozambique Island in the abridged translation of Sayid Ahmad b. `Abd al-Rahman's Manaqib al-Sayyid Muhammad b. Ahmad b. Abi Bakr al-Shadhili al-Yashruti (Cairo, 1934) by GUY & CHEIK AMIR (1949); indeed, its author claims (p. 11) that after Shaykh Ma'ruf's return from exile in Zanzibar, he made no further voyages; for references to his brother, see pp. 9-10. An alternative or perhaps simply mistaken version of these events, which cites a different oral source of information and dates the introduction of the Yashrutiyya at Mozambique Island to 1905, is included in Melo Branquinho, "Relatório", pp. 358-360. See also B. G. MARTIN (1976: 152-158). For the dating of Sultan Ali's exile, see J. MARTIN (1983, II: 152).
29 Melo Branquinho, "Relatório", p. 370.
30 Ibid., pp. 353-354, 409; Shaykh Issa's Comorian roots are not mentioned in CARVALHO (1988: 63).
31 Melo Branquinho, "Relatório", pp. 360-363, 366. This is the only reference I have seen to the Gazila center and welcome any suggestions regarding information on its history and activities.
32 Ibid., p. 408.
33 Ilha de Moçambique: Danze e canti della costa settentrionale, Folklore 7, Musiche dal Nuovo Mondo: Mozambico, compiled by Erasmo treglia (Instituto de Comunicação Social, Maputo; ARPAC, Arquivos do Património Cultural, Maputo; Centro Internazionale Cicrocevia & Sudnord Records, 1995), p. 8 of accompanying notes in Portuguese. Tracks 1, 5, and 11 on this CD are examples of tufo. For a fascinating preliminary analysis of tufo, see S. ARNFRED (1999). My thanks to the author for sharing her unpublished work with me.
34 Arquivo Histórico de Moçambique, PP 1145, Associação Maometana Comoreana (Association mahométane comorienne), Estatutos (Lourenço Marques, 5 May 1931), p. 5, Cap. I, Art. 1(a), and Cap. III, Art. 4.
35 For the full story, see ALPERS (1999). The following paragraph is based upon sources cited in this paper.
36 For a brief mention of this important individual, see CHOUZOUR (1994: 52).
37 See, e.g., the fascinating preliminary study by J. KAGABO (1991).
Haut de page

Pour citer cet article

Référence électronique

Edward A. Alpers, « A complex relationship: », Cahiers d’études africaines [En ligne], 161 | 2001, mis en ligne le 03 janvier 2001, consulté le 18 avril 2014. URL :

Haut de page

Droits d'auteur

© Cahiers d’Études africaines

Haut de page