^{page 345 note 1} I am most grateful to Professor Abdullahi Smith, Professor M. G. Smith, Muhammad al-Hajj, Dr. John Paden and others for their substantial criticisms of this paper in draft: I should add, however, that not all the interpretations given here are acceptable to them all. The paper is not intended to be definitive; far more detailed evidence for more areas is required. The research on which this paper is based was done first for the University of Ibadan during 1961–4 on grants from the Leverhulme Trust and the Nigerian Federal Ministry of Education; second, for the Northern History Research Scheme, Ahmadu Bello University, Zaria, during 1965–7. Throughout the six years I was greatly indebted to the many friends who gave me hospitality—and it is to them that I owe what knowledge I have of life and politics in the old Northern Region. The evidence for the nineteenth-century bureaucracy in Sokoto is given in my book The Sokoto Caliphate (London: Longmans, Green, 1967).Google Scholar

^{page 346 note 1} I am using ‘feudal’ and ‘bureaucratic’ not in any definitive sense, but to differentiate broadly between: (i) men of princely or aristocratic blood to whom large areas of people were allocated for government, tax-collection, and recruitment in time of war or politics, and (ii) men often of no genealogical distinction and linked to no particular area, serving as literate officials, clerks, advisers, mainly in the capital cities of the Caliph and Emirs but also in the larger towns of the countryside. As with most other crafts in Hausaland, sons of both groups, ‘feudal’ and ‘bureaucratic’, tended to follow their fathers' occupation if possible. The second group had undergone considerable training since childhood in order to be literate, and many became very learned in the Islamic sciences. But it was usual for sons of the first group also to be trained as scholars, the ideal type being the scholar-prince.

^{page 347 note 1} Though certain posts, such as vizier, judge, imam, or muḥtasib (inspector) were always held by freemen, not all offices were always held by scholars: sometimes they were held by servants of the Caliph or Emir, sometimes by his friends whose qualifications as scholars are not known. The distribution of jobs between slave and free was not necessarily fixed, though the number of state slaves probably increased during the nineteenth century, while the number of eunuchs may have declined. I am particularly interested here not in the slaves, but in the free scholar-officials for whom the gap between ideals and practice proved a problem, and for whom teaching and sufism offered an alternative career with its own status and authority.

^{page 349 note 1} The issues involved were not, of course, solely economic. One issue was probably Sokoto control; another was whether or not the somewhat forceful sons (there were 28 of them) of a previous Emir, ˓Abdullāhi, should take over the emirate. The Caliph had refused to appoint one of them, partly for personal reasons and partly, presumably, because he would have even less control over Kano unless he had as Emir a man who would depend on Sokoto for support against the ubiquitous sons of ˓Abdullāhi. The Caliph's vizier had strongly advised against risking civil war over the appointment, but was overruled. In the end, after the death of the two original claimants, the Caliph had to accept a son of ˓Abdullāhi as Emir, and a kind of rapprochement was effected.

^{page 351 note 1} Al-ḥājj ˓Umar had already played a politically ambiguous role in Bornu and had finally to be expelled from there. Later he was forced to leave Masina and Segu and, indirectly, even Futa Toro. His followers maintain that Muḥammad Bello became a Tijānī: this caused a bitter dispute, since the Tijāniyya would be ‘legitimate ’ in Sokoto, if this were so.

^{page 351 note 2} In these migrations, or in the memories of them, it is likely that the distinctions between Tijānī and Mahdist got blurred, not least since Tijānīs may also like Mahdists cross their arms in prayer. The dividing line, at a popular level, was probably fine enough for many to change allegiance either way. As persecution was mainly aimed at the Mahdists, it is likely that many used the label Tijānī to escape trouble. Perhaps this explains why a Tijānī scholar like Raji is sometimes called, inaccurately I think, a Mahdist. It is possible to believe in the imminent end of the world without being a Mahdist, so that the actions and anxieties of both Tijānīs and Mahdists seemed similar despite the difference in doctrine.

^{page 353 note 1} The genealogy, though written in the nineteenth century for the Shehu's family, was not treated very seriously in Sokoto itself. No one else bothered to claim it, though very many could have done so. The Sardauna's mother was from Numan, a pagan area in Muri emirate; the Sardauna's father, as Sarkin Raba, had been the Sokoto supervisor of Muri. As M. Sule Kumo has kindly pointed out to me, there is no slur attached to being the son of a concubine, either Islamically or locally—indeed in some Fulani areas it is thought to be a pre-requisite of a real emir. In traditional political appointments, it is probably important that a son of a concubine normally has no relatives on his mother's side.

^{page 354 note 1} Paden, J. N., ‘The Influence of Religious Elites on Political Culture and Community Integration in Kano, Nigeria ’ (Harvard, Ph.D. thesis, 1968)Google Scholar.

^{page 356 note 1} The implicit political reasons are too complex to discuss here. Suffice it to say that the Emir Sanusi, as head of the richest and most populous emirate (with the largest bloc of NPC parliamentary seats), and as an acknowledged leader of the Tijāniyya, was accustomed to considerable autonomy. Ironically, perhaps, the retiring British officials had declined to act against the Emir on the financial issue, asserting that the Emir had more power than they.