¹ I would like to express my gratitude for the invaluable teaching of Antonio Mujira, Joaquim Kateleku, Simao Kakiniki, Manuel Tutetuluta, Kadjende, Kaluwembe and Kalungula, which provided a foundation for this enquiry.

² A week after birth a Kwandu receives a second name, the otchoto or “altar name”. The infant's crying serves as a pretext to consult a soothsayer who must divine which ancestor made him cry, and it is always a deceased patrilinear relative. After this, the child's father together with a “sister” (daughter of his father or of his father's brother) must find a piece of lundau root (this tree grows beside the cemetery) and thread it on a strip of leather to make a necklace. Near the otchoto, which is the place within the dwelling complex sacred to the cult of ancestors the necklace is place round the infant's neck. The newborn receives the name of his ancestor in this ceremony and from then on re-incarnates him symbolically. The “altar name” may only be pronounced by those who actually knew the dead ancestor, and on certain occasions, the celebration of puberty, or during the funeral of one such users of the “altar name”, the young person is referred to by his father and aunt, who conferred the name, in terms of the relationship which connected them to his ancestor.

³ The bual was at the top of the Kwandu hierarchy. This status was reached “when age whitened the hair and teeth began to fall out”. Nevertheless, wisdom and power would anticipate time and a man with exceptional knowledge of the past or the owner of a large flock could become bual.

⁴ Although there were also matriclans—eanda, membership of which was through the female line, but the father's clan was always invoked when a Kwandu made a pledge or wished to ensure the fulfilment of a particular desire.

⁵ When other sons married they became “neolocals”. At times they went to live with a parallel cousin, son of their father's brother, who was himself a second son.

⁶ The role of head of the matriclan was inherited by the brother nearest to him in age. Following that it passed to the son of his eldest sister, following the invariable logic of matrilinearity.

⁷ The number of individuals linked by the matriclan was much greater than in the veto. This resulted from the fact that the matriclan was frequently common to various ethnic sub-groups such as Kwandu, Dimba, Himba, Cuvale, Hacavona, Gendelengu, and Tchilenge-Muso.

⁸ This “office” was limited specifically to participation in the “collective tribunal”, presiding over the ceremony of change of clan, and to directing ceremonies of female initiation involving the young girls of his own clan.

⁹ On the rare occasions when this system of proof was used, the process hardly ever ran its full course. Normally the guilty person would confess during the night for fear that the animal would be mistaken and implicate an innocent Kwandu who might then have recourse to a witch-doctor or eanga in revenge, thus bringing a worse punishment upon the culprit.

¹⁰ It was here that in 1784 Sergeant-Major Gregorio Jose Mendes contacted the Kwandu “during an exploration in the Sertoes of Benguela". I lived among them in the 1970s in the region of the Serra da Neve.

¹¹ Some Kwandu could not adapt to this more sedentary way of life, preferring to continue their wanderings and surviving as gatherers. They were forced at times to eat certain forbidden “magical” animals such as lizards which caused their expulsion from Kwandu membership so that they became Twa or Kwiss, men without a people.

¹² The Munda derived from the Muhila was belonged to the ethnic group Nhaneca-Humbe. Their farming technology was quite advanced, they used manure as a fertilizer, practiced irrigation and tilling as well as sowing. They cultivated small maize and had a few cattle. The Munda tended towards a centralisation of power and kept a dependent link, although distant, with the chief of the Muhile. They had a matrilineal system and transference of property was through the female line, the preferred heir being the first-born son of the eldest sister.

¹³ One alteration in the power concept concerned land. While the Kwandu still considered that the land had no owner, cultivated areas belonged to the sower until the crop had been harvested.

¹⁴ It is interesting to note that the Kwiss, although no longer Kwandu could still belong to the veto. They were not connected by any matrilineal ties since the himo only appears after the division of the group.

¹⁵ Surplus products could be exchanged for cloth or cattle. These animals would belong to the husband but the first calf would be completely owned by the woman.

¹⁶ A man caught in adultery with a married woman was obliged to pay her husband a cow—koi. The first calf from this cow would nevertheless belong to her.

¹⁷ This nephew, on inheriting his uncle's cattle, would have recourse to a ‘fortune stick’, the kalenda to prevent his dead uncle returning to kill or harm the cattle. This indicates an awareness that such an inheritance altered the traditional system for property transmission and threatened the ‘ancestor law’. Relations between uncle and nephew were in any case unfriendly considering that the nephew awaited his uncle's death to acquire cattle.

¹⁸ When Kwandu established a new household, his first wife lit a fire before the “ancestor altar” in the traditional way by rubbing sticks together. Only then could they sleep within the new house for without this ritual their animals would become barren.

¹⁹ When a widow re-married the brother of her dead husband had the right, as he had during the first Kwandu period, to receive the marriage compensation paid by his brother. If the widow was a second wife the compensation due was received by the woman's family.

²⁰ It was during this period that female initiation ceremonies stopped being presided over by the “chiefs of the matriclan”.

²¹ Frequently the chiefs of the matriclan belonged to different Herero ethnic sub-groups.

²² Men continued to choose and establish Kwandu localities; both girls and boys collected before the altar of patrilineal ancestors before their initiation rites; women as well as men knew more patrilineal than matrilineal antecedents; the only three cases of incest I could ascertain took place between matrilineal relatives.

²³ Fear, that ancestors might prevent entry into the world of the dead, or equally that victims might have recourse to witchcraft in revenge, clearly functioned as a deterrent to crime.

²⁴ Contact with the Portuguese dates back much further than this however, from the second half of the 18th Century at least. Colonial occupation of south-west Angola dates from the 19th Century.

²⁵ The first reference to Kwandu as workers in the European-owned fisheries appears in Vilela in 1923. Their women only began to accompany them and work for Europeans in the 'sixties.

²⁶ This tax was instrumental in forcing the Kwandu into wage-earning although the “marriage compensation” rule and the desire for more cattle were also cited as motives.

²⁷ When a portion of land was attributed to Salamao, an elder of the Dombe, for his own enclosure, the issue of land ownership became even more confused. At the time of my departure two Kwandu were making their own request to t he colonial administration for concessions of land to wire off.

²⁸ It is interesting to note that while the Kwandu related sheep to the world of men, the cock belonged to the world of women.