^{page 316 note 1} I wish to thank Prof. I. Schapera and Prof. E. E. Evans-Pritchard for their encouragement and assistance.

^{page 316 note 2} Memo. XXIII of the International African Institute, Some Problems of Anthropological Research in Kenya Colony, by Schapera, I., 1949Google Scholar, contains a list of publications on the Nilo-Hamites of Kenya. For the people outside Kenya, the major works consulted are: Wright, A. C. A., ‘Notes on the Iteso Social Organisation’, Uganda Journal, ix, 1942, pp. 57–80Google ScholarC. G., and Seligman, B. Z., Pagan Tribes of the Sudan, London, 1932Google Scholar; Nalder, L. F., A Tribal of Mongolia Province, London, 1937.Google Scholar Since Schapera's memorandum was published, Fosbrooke's, H. A. ‘An Administrative Survey of the Masai Social System’ has appeared in Tanganyika Notes and Records, xxvi, 1948, pp. 1–50Google Scholar; as well A Preliminary Survey ofthe Turkana, by P. H. University of Cape Town, 1951Google Scholar, which I was able to consult before publication; and Peristiany, J. G., ‘The Age-set System of the Pastoral Pokot’, Africa, xxi, 1951, pp. 188–206CrossRefGoogle Scholar and 279–302. I should state that my analysis was already completed when Peristiany's paper appeared. On the whole the new evidence on the Pokot (also known as the West Suk) presented by him and his conclusions, as I read them, add weight to my argument. In particular I wish to stress his agreement with my views on the political function of the age-system among the Nilo-Hamites.

Though he does not specifically discuss this point, or the relationship of the (lineage-)clan system and of the territorial organization with the age-system, he says, in his final résumé (p. 301): ‘La société pastorale, à laquelle une structure territoriale ou de lignage fait défaut, est ainsi divisée en une hiérarchie de classes d'âge, qui lui sert d'armature politique.’ I should also mention here the recent publication by Huntingford, G. W. B., Nandi Work and Culture, Colonial Office, 1950Google Scholar. The author does not deal with the social and political organization of the Nandi, but gives a few references to the age-system and, at pp. 104-6, attempts an historical outline of the system. In so doing Huntingford assumes that ‘the age-sets of the Nandi and Masae are without doubt of Hamitic origin, and derived direct from the gada system of the Gala’. This assumption is frequently made, but has never been critically analysed. Recently Jansen, A. in ‘Elementi della Cultura Spirituale dei Conso’, Rassegna di Studi Etiopici, ii 1942, pp. 217–59Google Scholar, showed that it is highly probable that the gada system came to the Galla from the Conso. Jansen's opinion seems to me consonant with Baumann's theory that the age-class system in Africa originated in the paleonegritic culture, and not in the Hamitic culture. See Baumann, , Völkerkunde von Afrika, Essen, 1940Google Scholar(French translation, Les Peuples et les Civilisations de 1'Afrique, Paris, 1948, p. 261).Google Scholar

^{page 317 note 1} The endungore and the engibata are parts of a ceremony described by Whitehouse, L. E., ‘Masai Social Customs’, J.E.A.U.N.H.S., iii, pp. 146–8Google ScholarEngibata is the starting dance, very solemn. Then follows endungore, when the sponsors of the (age-unit) to be opened ceremonially break a stick and the legs of a small dead bird, and lay curse on a circumcision knife, to symbolize that it is no longer free to initiate candidates for the old poror which is thereby closed.

A third ceremony, eunoto, has also some structural significance. It is performed during each of the two age-unit periods, and at present marks the promotion of junior initiates to a senior rank. It seems that in the past it served to close the time for individual initiations. But nowadays individual initiations can still be performed after the eunoto, though not after endmgpre. It is at eunoto that the age-unit is given a special name.

^{page 317 note 2} The cyclical naming system is opposed to the linear system. In this system age-set names are taken from current events or from peculiar traits of the sets. None of the Nilo-Hamitic peoples follows either system exclusively. Thus the Masai keep a fixed series of names for their minimal age-divisions. The Kipsigis never use the same names for their minimal age-divisions. It seems to me that the difference between the two systems is one of stress. Where the linear system is followed the stress is on the changing character of the age-system; where the cyclical system is observed, the consideration of the system's permanent form prevails. See p. 321.

^{page 317 note 3} When Huntingford, The Nandi, Nairobi, 1944, p. 9, states that ‘every Nandi male is born into an age-set’, he is misled by the apparent predetermination of the cyclical system.

^{page 318 note 1} When describing this second part Peristiany emphasizes its military aspect at the expense of its structural significance. My interpretation is based on the reason for the reluctance of the Kipkoymet (the class to be closed) to consent to the request of the Chuma (the following class) for permission to hold initiation and then saket ap eito: ‘the Kipkoymet would refuse permission as long as possible, as, once this permission was granted, their generation [age-class in this paper's terminology] was “cut” and they had to retire as elders’. Peristiany, J. G., The Social Institutions of the Kipsigis, p. 31.Google Scholar

^{page 318 note 2} Rocks (nyetapa) and zebras (nyeru) are the two names. A man cannot be under the same name as his father.

^{page 316 note 3} Hulley, D. M., Short Notes on the Turkana, n.d. A MS. at the School of African Studies, University of Cape Town.Google Scholar

^{page 319 note 1} The ‘leopards’ and the ‘stones’ ‘are initiated in two groups at the same time but physically and socially apart’. Gulliver, , Preliminary Survey of the Turkana, p. 128.Google Scholar

^{page 320 note 1} Of the West Suk Peristiany states that ‘no Pochon can answer a question as to the duration of a sapana-set in terms of years’. Africa, xxi, p. 290.

^{page 320 note 2} For this reason the Kipsigis elders found it possible to curtail the class-period of the Maina class. See Peristiany, , The Social Institutions of the Kipsigis, pp. 43–44.Google Scholar

^{page 322 note 1} When the institutive ceremonies take place the movement from one stratum to another is automatic, but prior to those ceremonies the lower set brings pressure to bear on the classes above to hasten their performance. This pressure results in some rivalry between the lower classes, the senior classes acting as intermediaries between them. There are clear examples of this rivalry among the Masai, the Kipsigis, and others. Peristiany for the Pokot has stressed ‘the reluctance’ of the youngest colour section to recognize publicly the emergence of a new colour section. See Africa, xxi, p. 284.

^{page 323 note 1} Only the Teso of all the Nilo-Hamites are politically organized on a centralized chiefship system. But it is known that the system was imposed on them by the Ganda conquerors and left practically unaltered by the British when they took over from the Ganda. The scanty information on the earlier organization of the Teso shows that their society conformed to the decentralized type of the Nilo-Hamites.

^{page 323 note 2} I have borrowed this term for the chief medicine-man from Evans-Pritchard's analysis of the political system of the Nandi-speaking peoples. Africa, xiii, 1940, pp. 250–67.Google Scholar

^{page 323 note 3} On this point see Radcliffe-Brown, and Forde, , African Systems of Kinship and Marriagt, 1950, pp. 18–20.Google Scholar

^{page 324 note 1} It is significant in this context that the Dorobo tribes who live on the same land as the Masai and the Nandi-Suk are not considered enemies, although differing in culture and social structure.

^{page 324 note 2} Probably ecological reasons may account for the indefiniteness of the territorial units, especially where nomadism prevails, since it is particularly noticeable among such tribes as the Masai and the Turkana, who are almost entirely nomads. Gulliver describes the Turkana territorial organization as ‘a neutral feature of Turkana social organisation’. The absence of any centralized authority may be viewed as another reason for the phenomenon. New territorial units are easily formed without any special form of control and irrespective of previous territorial relations and clan allegiances. Fosbrooke has studied this among the Masai. People from any of the clans and territorial sections may be brought together by their transhumant wandering in search of water and good grazing. Residential neighbourhood and common interests foster a sense of solidarity which asserts itself in opposition to other similar groups and units. Finally, a claim to independence is made by holding certain tribal ceremonies of their own—Gulliver has recorded the tendency of the Turkana to join whatever initiation centre suits them best. It is at these centres that all social transactions of interest to those present are carried out. The site is usually a water-source or well.

^{page 324 note 3} With regard to the Nandi-speaking peoples I have accepted the conclusions of Evans-Pritchard's analysis, ‘The Political Structure of the Nandi-speaking Peoples of Kenya’, Africa, xiii, pp. 250–67, not, however, without checking all the original evidence.Google Scholar

^{page 325 note 1} The Nandi territorial unit corresponding to the Kipsigis province (emet), which has a political nature, is the pororiet. See Evans-Pritchard, , op. cit., p. 262.Google Scholar

^{page 326 note 1} ‘The Kipsigis consider the Nandi, the Keyo, the Tugen and the Suk as their brothers, but refuse this title to the Masai and the Turkana, whom they call punik, i.e. enemies.’ Peristiany, , The Social Institutions of the Kipsigis, p. 2.Google Scholar

^{page 328 note 1} The possibility of the juniors showing disrespect to the seniors is well illustrated by the reasons why the Kipsigis elders curtailed the class-period of Maina class: (1) the Maina had no proper respect for their kot-ap-thi (people of the house) and were rude in matters of kinship; (2) they stole cattle from their fellow Kipsigis and not only from punik or enemies; (3) they showed ignorance of and disrespect for the customs of their fathers. Sec Peristiany, , The Social Institutions of the Kipsigis, p. 44.Google Scholar

^{page 328 note 2} Gulliver, in Preliminary Survey of the Turkana, chap. 14, pp. 143Google Scholar ff., describes how this took place among the Turkana as follows: ‘The men would assemble by age-set and region, more or less under the supervision each of their senior man.’ Junior sets are in the van and senior sets in the rear. Merker suggests an organization of a similar kind for the Masai when he calk the senior classes ‘Reserve’, Merker, M., Die Masai, Berlin, 1904, p. 73.Google Scholar

Gulliver also describes the ways in which a raid was started: ‘firstly by an age set (or part of a set), secondly by a recognised war leader, and thirdly by a diviner.’ He states also that any individual setmember can summon his age-mates for a raid.

Fosbrooke says that a Masai prophet could initiate a raid by inciting the warriors to a prospective booty of cattle.

^{page 328 note 3} The early and recent history of the Masai is a very good illustration of this point. See Fosbrooke, , ‘An Administrative Survey’, p. 10.Google Scholar

^{page 328 note 4} Leakey, L. S. B., ‘Some Notes on the Masai of Kenya Colony’, J.R.A.I., xl, p. 200.Google Scholar But see the comment of Storrs-Fox, D., ‘Further Notes on the Masai of Kenya Colony’Google Scholar, ibid., p. 455.

^{page 329 note 1} I understand that these councils have been described for the Nandi by G. W. B. Huntingford in his book on Nandi social and political organization, not yet published.

^{page 329 note 2} The ingobir are singled out from the very begin ning of the class-period, so that usually they are the most senior of their age-mates, but any one distinguished for his talents is at any time considered engobiro. See Whitehouse, , ‘Masai Social Customs’, J.E.A.U.N.H.S. iii, pp. 146–8;Google Scholar and Fosbrooke, , ‘An Administrative Survey’, p. 34.Google Scholar

^{page 329 note 3} Huntingford, , The Nandi, Nairobi, 1944, p. 6.Google Scholar

^{page 329 note 4} Ibid., p. 11.

^{page 330 note 1} Cf. Peristiany, , ‘The Age-set System of the Pas- total Pokot’, p. 298.Google Scholar

^{page 330 note 2} A peculiar case, which I think should be related t o the structural status of the most senior elders, is reported by Peristiany in connexion with the religious leader of the Kipsigis, the Poyot ap Tumda, the elder of the ceremony. It is the duty of the Poyot ap Tumda to offer prayers to God for the candidates and the country at the ceremonies for individual initiation and the institution of the age-sets. His office is said to be hereditary. But when the Poyot ap Tumda dies, he is succeeded not by his son but by his own assistant. The reason given is that at the time of the Poyotap Tumda's death, his son is always ‘too young’ for the office. The proper age for it is over 55 or 60 years: a very advanced age for the Kipsigis. When the son of the Poyot ap Tumda comes to proper age, if he wants to pursue his father's activity, he must do so in a different place. In the light of the foregoing analysis the proper age for the functions of the Poyot ap Tumda is more satisfactorily explained in terms of a proper structural status giving authority for religious leadership. See Peristiany, , The Social Institutions of the Kipsigis, pp. 216–17.Google Scholar

^{page 331 note 1} See p. 530, n. 2, referring to the Poyot ap Tumda.

^{page 331 note 2} Fosbrooke, , ‘An Administrative Survey’, pp. 16–17.Google Scholar