¹ Contrary to practice in neighbouring Rwanda, no data on ethnic affiliation have been officially collected since independence. Belgian figures in 1956 listed 86·48 per cent Hutu, 12·39 per cent Tutsi, and 1·13 per cent Twa. While these computations have given rise to considerable–and heated–debate, the 85/15 proportion is given here as a mere indication of the relative weight of the two main ethnic groups.

² Le Monde (Paris), 20–21 September 1987.Google Scholar

³ See Chrétien, Jean-Pierre, Guichaoua, André, and Jeune, Gabriel Le, La Crise d'août 1988 au Burundi (Paris, 1989)Google Scholar and Amnesty International, Burundi: killings of children by government troops (London), 10 1988.Google Scholar

⁴ See Lemarchand, René, ‘The Report of the National Commission to Study the Question of National Unity in Burundi: a critical comment’, in The Journal of Modern African Studies (Cambridge), 27, 4, 12 1989, pp. 685–90,CrossRefGoogle Scholar and Reyntjens, Filip, Burundi 1972–1988: continuité et changement (Brussels, Centre d'étude et de documentation africaines, 1989), 5, pp. 65–71.Google Scholar

⁵ Amnesty International, Burundi: sectarian security forces violate human rights with impunity (London), 27 11 1992.Google Scholar

⁶ More information on these provisions and the problems of implementation can be found in Reyntjens, Filip, ‘L'Ingéniérie de l'unité nationale: quelques singularités de la constitution burundaise de 1992’, in Politique africaine (Paris), 47, 10 1992, pp. 141–6.Google Scholar

⁷ As the Constitution reaffirms the republican status of Burundi, the Parti royaliste parlementaire changed its name to Parti pour la reconciliation du peuple (P.R.P.) in order to be recognized.Google Scholar

⁸ Others considered opposed to the Charter of National Unity, and accused of waging armed attacks from neighbouring countries, were refused recognition, including the clandestine Parti pour la libération du peuple hutu (Palipehutu).Google Scholar

⁹ See, for example, Heilbrunn, John R., ‘Social Origins of National Conferences in Benin and Togo’, in The Journal of Modern African Studies, 31, 2, 06 1993, pp. 277–99.CrossRefGoogle Scholar

¹⁰ The fact that the presidential elections did not follow those for membership of the Assemblée nationale (as in a number of other countries) was due to Uprona's expectation that its vote would be increased by Buyoya's anticipated victory, since he was thought to be more popular than his party.Google Scholar

¹¹ The perception that Uprona was waging this campaign to discredit Frodebu out of sheer weakness was further enhanced by the refusal of the party's president to meet Frodebu's president in a radio-television debate on the truth of the allegations.

¹² While Frodebu did not mathematically need the votes of those supporting the R.P.B., the P.P., and the P.L., their significance should not be under-estimated. Faced with the attempts by Uprona to marginalise Frodebu, it allowed that party to escape isolation and to present Ndadaye as the candidate of a large opposition platform. The fact that Buyoya had desperately attempted to convince the leaders of these three parties to join his bid for the Presidency shows that he fully understood the symbolic importance of the alliance.

¹³ For instance, according to a despatch from the correspondent of Agence France Presse, dated 1 Junes 1992, ‘Major Pierre Buyoya should be elected in the first round with over 60% of the vote’. As it happened, the only accurate prediction was made by the outlawed Palipehutu party in ‘Prognostic sur les elections présidentielles au Burundi’, Bujumbura, 15 April 1993. Although this was only 1·5 per cent off the mark for Ndadaye and 0·7 per cent for Buyoya, it was grossly erroneous for several provinces.Google Scholar

¹⁴ While the foreign observers were sent by different governments and organisations, their monitoring operations were efficiently co-ordinated by U.N.D.P. representatives in Bujumbura. The national observers were commissioned by churches and various non-governmental bodies, including the human rights leagues.

¹⁵ National Democratic Institute for International Affairs, Communiqué de presse, Bujumbura, 2 06 1993.Google Scholar

¹⁶ All three Presidents have assumed power through military coups since the abolition of the monarchy in 1966. Only Jean-Baptiste Bagaza (1976–1987) sought the legitimacy conferred by the popular vote, but he was the sole candidate in the 1982 election.

¹⁷ De Standaard (Brussels), 1 June 1993.Google Scholar

¹⁸ The Minister of the Interior alleged that the list proposed by Frodebu in Cankuzo did not reflect ‘the diverse component parts of the Burundian population’, meaning that it was all-Hutu. However, in the same province Raddes was allowed to participate although its list was entirely composed of Tutsi.

¹⁹ Hence the article entitled ‘Elections présidentielles et législatives: danger de la montée du monopartisme’, in Le Citoyen (Bujumbura), 15–30 June 1993.Google Scholar

²⁰ Melchior Ndadaye was the president of the Rwanda-based Mouvement des éetudiants progressistes barundi (Bampere) before co-founding the Parti des travailleurs du Burundi (U.B.U.), of which he chaired the information department from 1979 to 1983, when he returned to Burundi.Google Scholar

²¹ This does not, of course, alter the fact that many Burundians voted for the candidate of their ethnic group, and for that reason alone.

²² Thibon, Christian, ‘L'Entrée en politique de la périphérie: des identités génératio-spatiales et une socialisation spécifiques’, in Cahiers d'histoire: revue de l'université de Pau-CREPAO (Pau, France), 7, 12 1993Google Scholar and ‘Évolution des systémes familiaux et tempéraments politiques au Burundi de 1900 á nos jours’, Symposium on ‘Histoire et développement’, Tananarivo, Madagascar, May 1992, published in Omaly sy Anio, the journal of the History Department, University of Tananarivo, 33–36, 1993.Google Scholar

²³ It should be noted that Kayanza, Makamba, and Urban Bujumbura are not classified as ‘centre’ or ‘periphery’ by Thibon.

²⁴ Cankuzo was already atypical during the referendum on the Charter of National Unity, and more research is needed to explain this deviance.

²⁵ For example: ‘Burundi Chooses a Hutu President’, in Gazet van Antwerpen (Antwerp), 4 June 1993; ‘Ethnic Changeover in Burundi’,Google Scholar in Le Monde, 4 June 1993; ‘Ethnic Changeover’, inGoogle ScholarJeune Afrique (Paris), 1692, 10–16 June 1993; ‘Hutu Success at the Parliamentary Elections’,Google Scholar in Le Soir (Brussels), 1 07 1993; ‘The Parliamentary Elections Confirm the Coming to Power of the Hutu Majority’,Google Scholar in Le Monde, 2 July 1993; and ‘Hutu Attain Power through the Ballot Box’,Google Scholar in De Volkskrant (Amsterdam), 5 07 1994.Google Scholar

²⁶ Letter of 9 July 1993 addressed to the Minister of the Interior, Bujumbura.

²⁷ The bimonthly Le Citoyen, 1–15 June 1993, claimed that ‘these people have elegantly worn the hat of UPRONA, but intelligently and subtly made propaganda for FRODEBU in absolute discretion’ (my translation).Google Scholar

²⁸ According to a diplomatic source in Bujumbura, July 1993, the military's suspicious about Frodebu were greatly reduced during the months leading up to the elections as a result of several discreet meeting between the leaders of the party and the army.

²⁹ Other unsuccessful armed interventions have been poorly documented. However, see La Semaine (Bujumbura), 3, 21 June 1993,Google Scholar and Isibo (Kigali), 103, 24 June–1 07 1993, for evidence of an attempted coup by a group of soldiers from Bururi during the night of 16–17 June 1993.Google Scholar

³⁰ This was the interpretation of a leading Belgian commentator: Cros, Marie-France, ‘Burundi: un putsch rassurant’, in La Libre Belgique (Brussels), 6 07 1993.Google Scholar

³¹ It should be noted that Article 9 of Law-Decree No. 1/010 of 15 April 1992 prohibits members of the armed forces, the police, and the judiciary from joining political parties.

³² The Provinces of Bubanza and Rural Bujumbura have three portfolios, three provinces hold two, six have one, while four (including the capital district) have no representative in the Cabinet.Google Scholar

³³ Paradoxically, given that the leaders of non-governmental organisations are now being enrolled in the activities of the state, Frodebu's victory is likely to result in a weakening of that same ‘civil society’ which has been a basis of its emergence as the major political force.

³⁴ Frodebu, , Melchior Ndadaye. Pour construire un Burundi nouveau: 46 propositions (Bujumbura, 05 1993).Google Scholar

³⁵ Although Pierre Buyoya realised that he would inevitably be replaced as President, as he told me in December 1991, he certainly did not anticipate that this would happen so quickly.Google Scholar According to Le Citoyen, 1–15 June 1993, Uprona envisaged a scenario that included Buyoya's replacement at the next elections in 1998.Google Scholar