Hostname: page-component-7c8c6479df-nwzlb Total loading time: 0 Render date: 2024-03-26T13:52:06.973Z Has data issue: false hasContentIssue false

The Chronology of the Iron Age in Bantu Africa

Published online by Cambridge University Press:  22 January 2009

D. W. Phillipson
Affiliation:
British Institute in Eastern Africa

Extract

Mrs Barbara Ottaway has proposed that the interquartile range of the radiocarbon dates available for a given industry will provide an estimate of the floruit of that industry. The present paper examines some four hundred radiocarbon dates which have so far been processed from Iron Age sites in Bantu Africa, and concludes that the intersextile range provides a more accurate approximation for the floruit of individual industries than does the interquartile range. This approach permits the presentation of a more precise inter-regional synthesis of Iron Age chronology than has previously been attempted. It is demonstrated that the earliest manifestations of the Early Iron Age Industrial Complex are in East Africa. Further to the south, two distinct streams of the Early Iron Age are recognized: that in the eastern part of the sub-continent had penetrated as far to the south as the Transvaal some centuries before the western stream spread into what is now Zambia. The inception of the later Iron Age, around the eleventh century A.D., is shown to have been at least as rapid as was that of the Early Iron Age.

In view of the greater precision which the above methodology imparts to the radiocarbon-based chronology, a brief discussion is presented of the calibrations which dendrochronological studies provide between radiocarbon ages and calendar ages. It is concluded that, for the period of the sub-Saharan African Iron Age, the corrections which are called for are minor in comparison with the standard errors of individual radiocarbon dates; and the application of such calibrations to Iron Age dates is not recommended.

Type
Research Article
Copyright
Copyright © Cambridge University Press 1975

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Ottaway, B., ‘Dispersion diagrams: a new approach to the display of carbon-14 dates’, Archaeometry, xv (1973), 512.CrossRefGoogle Scholar The concept of quartiles, like those of sextiles and octiles which will be introduced later in this paper, is an extension of that or the median. The median is ‘that quantity which is so related to the quantities occurring in a given set of instances that exactly as many exceed as fall short of it’ (Shorter Oxford English Dictionary). The quartiles each comprehend one quarter of the quantities when the set of instances is divided on the same principle into fourths; the sextiles and octiles similarly each comprise one-sixth and one-eighth of the quantities. The interquartile range will include one half of the observations in a sample, centred on the median; while the intersextile and interoctile ranges respectively include two-thirds and three-quarters of the measurements, likewise centred on the median. Expressed differently, the interquartile, intersextile and interoctile ranges may be regarded as respectively excluding from the total range of observations both the highest and the lowest quarters, sixths and eighths of the observations. For definitions of the terms ‘industrial complex’, ‘industry’ and ‘phase’ as used in this paper, the reader is referred to pp. 892–4 of Background to Evolution in Africa, ed. Bishop, W. W. and Clark, J. D. (Chicago, 1967).Google Scholar

2 Ottaway, B., ‘Estimating the duration of cultures’, Antiquity, XLVII (1973), 231–2.Google Scholar

3 Indeed, anomalous dates should not be excluded since their impact on the interquartile range is minimal, and to exclude them would result in too small an interquartile range and an underestimate of the duration of the industry under investigation. Some anomalous dates are to be expected, due to the statistical uncertainties of the radiocarbon dating technique; to exclude them and then to calculate the interquartile range would have the effect of making double allowance for them.

4 See p. 339, below.

5 Ottaway, , ‘Dispersion diagrams’, 10.Google Scholar

6 Diem, K. (ed.), Documenta Geigy Scientific Tables, 6th ed. (Basle, 1962), 128.Google Scholar Cf. also Ottaway, , ‘Estimating the duration of cultures’, 231.Google Scholar

7 This assumption may be rendered invalid by the tendency of archaeologists to concentrate their attentions on periods of industrial change; but this tendency is probably counteracted to a certain extent by the difficulty of locating sites which display the very first manifestations of a particular industry.

8 When more radiocarbon dates have been processed it may prove possible to refine the present methodology by omitting from our calculations multiple analyses relating to the same phase of occupation at the same site. This will also help to alleviate the difficulty raised in note 7, above. On a smaller scale of investigation than that here discussed, the intersextile, etc., ranges of radiocarbon dates for a single phase of occupation could be used to provide an estimate for the duration of that phase.

9 This work is described mainly in the following publications: Vogel, J. O., Kamangoza: an Introduction to the Iron Age Cultures of the Victoria Falh Region (Zambia Museum Papers, 2), Lusaka, 1971Google Scholar; idem, Kumadzulo: an Early Iron Age Village Site in Southern Zambia (Zambia Museum Papers, 3), Lusaka, 1971Google Scholar; idem, ‘The Kalomo culture of southern Zambia’, Zambia Museum Journal, 1 (1970), 7788Google Scholar; idem, ‘The Shongwe tradition’, Zambia Museum Journal, III (1972), 2734Google Scholar; idem, ‘Some Early Iron Age sites in southern and western Zambia’, Azania, VIII (1973), 2554.Google Scholar See also Daniels, S. G. H. and Phillipson, D. W., ‘The Early Iron Age site at Dambwa near Livingstone’, in Fagan, B. M., Phillipson, D. W. and Daniels, S. G. H., Iron Age Cultures in Zambia, 1 (London, 1967), 154.Google Scholar

10 In view of the very large number of radiocarbon dates considered in this paper, it has been found impracticable to do more than list the laboratory reference numbers of the individual dates upon which the following calculations are based. Only published dates have been used; and all have been, or will be, included in the regular summaries of newly released dates which appear in this Journal, i.e. II (1961), 137–9Google Scholar; IV (1963), 127–8; VI (1965), 107–16; VII (1966), 495–506; VIII (1967), 513–27; X (1969), 149–69; XI (1970), 1–15; XIII (1972), 1–24; XV (1974), 175–92, to which the reader is referred for more comprehensive details.

11 Unless it is specifically stated to the contrary, it may be assumed that all dates and calculations given in this paper are based on uncorrected age determinations using the 5,568-year half-life of radiocarbon. The relationship between these ages and true ‘calendar’ ages is discussed in a later section (pp. 339–342).

12 Vogel, , ‘Kalomo culture’, 78–9.Google Scholar

13 The ‘maximum limits’ quoted in the text of this paper are those between which, at 90 per cent probability based on the sample of dates at present available, half of the eventual dates for the industry under consideration may be expected to fall. For convenience, all dates given in this paper have been rounded off to the nearest ten years.

14 Vogel, , ‘Kalomo culture’, 81.Google Scholar

15 In chronological order, the dates on which these calculations are based are: N-408, C-829, N-409, N-1140, N-411, N-410, SR-106, N-412, SR-62, N-413, SR-110, N-1138, N-414, SR-97, N-735, N-1137, SR-98, N-1139, SR-96, N-455, N-416, N-415, N-1143, N-1141, N-1142, N-1144, N-668.

16 N-420, N-425, N-424, N-426, N-746, N-739, N-742, N-423, N-457, N-743, N-419, N-740, N-421, N-748, N-741, N-678, N-679.

17 N-571, N-565, N-564, N-563, N-734, N-748, N-745, N-572, N-573, N-744, N-738, N-737, N-733, N-732, SR-99, N-736, N-731.

18 Vogel, , ‘Kalomo culture’, 83.Google Scholar See also Phillipson, D. W., ‘Iron Age history and archaeology in Zambia’, J. Afr. Hist., xv (1974), 125.CrossRefGoogle Scholar

19 Vogel, , ‘Kalomo culture’, 81.Google Scholar

20 SR-65, GX-1114, SR-123, GX-1425, GX-1115, N-585, N-595, N-586, N-594, N-428, SR-41, N-597, N-430, N-431, N-596, SR-66, SR-139, N-429.

21 SR-19, UCLA-1632A, R-874, N-427, UCLA-723, UCLA-177, SR-30, SR-20, SR-31, UCLA-176, SR-74, N-433, UCLA-1632B, UCLA-175.

22 Fagan, B. M., Iron Age Cultures in Zambia, 1 (London, 1967)Google Scholar; unpublished excavations by B. M. Fagan at Gundu and Ndonde in 1967; Vogel, , KamangozaGoogle Scholar; idem, ‘Kalomo culture’.

23 Phillipson, , ‘Iron Age history and archaeology’.Google Scholar

24 Cf. Ottaway, , ‘Dispersion diagrams’, 8Google Scholar: ‘… if the contemporaneity of the two cultures is only apparent during the appropriate outer quartile ranges of each sample, i.e. if the inter-quartile ranges do not overlap, then the intensity of contact between them will not have been great enough for significant cultural transmission to have occurred’.

25 With samples of more than about forty dates the relative positions are reversed.

26 Soper, R., ‘A general review of the Early Iron Age in the Southern half of Africa’, Azania, VI (1971), 537CrossRefGoogle Scholar; Huffman, T. N., ‘The Early Iron Age and the spread of the Bantu’, S. Afr. archaeol. Bull., xxv (1970), 321CrossRefGoogle Scholar; idem, ‘A guide to the Iron Age of Mashonaland’, Occ. Pap. Nat. Mus. Rhod., IV, 1 (1971), 2044Google Scholar; Phillipson, D. W., ‘The Early Iron Age in Zambia—regional variants and some tentative conclusions’, J. Afr. Hist., IX (1968), 191211.CrossRefGoogle Scholar

27 This hypothesis is propounded and discussed in a forthcoming work by the present writer, provisionally entitled Later Prehistory of Eastern and Southern Africa.

28 Sutton, J. E. G., ‘New radiocarbon dates for eastern and southern Africa’, J. Afr. Hist., XIII (1972), 124CrossRefGoogle Scholar; the quotation is from p. 10.

29 For the distinction between the Early Iron Age and the later Iron Age, see Phillipson, , ‘Iron Age history’, 1.Google ScholarSutton, , ‘New radiocarbon dates’, 10Google Scholar,—and elsewhere—has introduced the term ‘Middle Iron Age’; so far as I am aware no definition of this term has been offered and neither its validity nor its utility has been demonstrated. The material attributed by Sutton to his ‘Middle Iron Age’ appears all to be such that the present writer would subsume it within the earlier stages of the later Iron Age.

30 Sutton, J. E. G. and Roberts, A. D., ‘Uvinza and its salt industry’, Azania, III (1968) 4586.CrossRefGoogle Scholar

31 Phillipson, , ‘Iron Age history’, 25.Google Scholar

32 Garlake, P. S., Great Zimbabwe (London, 1973), 182.Google Scholar

33 Robinson, K. R., ‘The pottery sequence of Malawi briefly compared with that already established south of the Zambezi’, Arnoldia (Rhod.), VI (1973), no. 18.Google Scholar

34 Ibid. 10.

35 e.g. Mason, R. J. et al. , ‘Early Iron Age settlement of southern Africa’, S.A. Journal of Science, LXIX (1973), 324–6Google Scholar; Klapwijk, M., ‘A preliminary report on pottery from the north-eastern Transvaal, South Africa’, S. AJr. archaeol. Bull., XXIX (1974), 1923CrossRefGoogle Scholar; Mason, R J., ‘Background to the Transvaal Iron Age—new discoveries at Olifantspoort and Broederstoom’, J. S. Afr. Institute of Mining and Metallurgy, LXXIV (1974), 211–16.Google Scholar

36 The most recent statement to this effect is that of Soper, ‘General review of the Early Iron Age’, 10Google Scholar, who emphasizes the absence of a definitive description of Uitkomst pottery. So far as I have been able to ascertain, the main basis of this attribution is the cautious view of Mason, R. J., Prehistory of the Transvaal (Johannesburg, 1962), 430Google Scholar, which was presented thus: ‘There are some similarities between the Rhodesian stamped pottery industries and our Uitkomst Culture wares, but many differences, especially in vessel and rim shape. In addition incised and burnished decoration was used in making some Uitkomst Culture pottery, but not by the early Rhodesian cultures. Though the evidence is very imperfect, there may be some basic connexion between the Uitkomst Culture and the early Rhodesian Iron Age, with substantial additions and modifications affecting the Uitkomst Culture people after they settled in the far South’.

37 This evaluation is based, in the absence of a full quantified analysis of any Uitkomst pottery assemblage, on a subjective study of published illustrations and descriptions and on collections housed at the University of the Witwatersrand, Johannesburg, which I was able to examine in 1965 by kind permission of Dr Mason.

38 Similarly, the published data strongly indicate that the ‘Early Iron Age’ pottery from Moor Park, Natal, should not be attributed to the Early Iron Age Industrial Complex. Davies, O., ‘Excavations at the walled Early Iron Age site in Moor Park near Estcourt, Natal’, Ann. Natal Mus., XXII (1974), 289323.Google Scholar

39 As will be apparent from table 3, modern political boundaries have generally been used. Only two adjustments have been made on archaeological grounds. Eastern Zambia has been included with Malawi since the closely inter-related Early Iron Age industries of both regions are attributed to the eastern stream, in contrast with the predominantly western stream settlements found to the west of the Luangwa. The dates for the western stream Early Iron Age of northwestern Mashonaland have been subsumed with those for related material in adjacent regions of Zambia, rather than with the eastern stream dates from elsewhere in Rhodesia.

40 N-647, N-894, N-895, N-890, GrN-5752, N-902, N-891, N-892, N-483, N-484, N-898, N-347, N-883, B-755, N-292, N-291, N-435, N-784, GrN-5753, B-758, N-436, GX-1186, N-437, N-463, N-648, N-465, N-901, N-646, SR-78, N-900, SR-76, N-464.

41 N-408, N-1501, C-829, SR-65, GX-1328, GrN-4646, N-1286, GX-1013b, GX-1013a, GrN-4647, N-409, GX-1114, SR-123, GX-1012, N-998, GX-1425, SR-126, N-1140, GX-1115, N-585, L-395C, N1498, N-411, N-410, N-1500, N-1499, SR-106, N-412, SR-62, N-413, N-1283, SR-118, SR-110, N-1138, N-414, N-595, N-1284, N-428, N-586, SR-97, N-594, N-735, N-1137, GX-1327, GX-535, SR-98, N-1139, SR-41, N-997, N-597, GX-1009, Pta-606, N-430, SR-96, L-395B, N-455, N-431, SR-211, GX-1010, N-596, N-416, N-415, GX-1329, L-395A, GrN-3580, SR-245, N-1143, N-1141, GX-662, SR-139, N-429, N-1142, SR-248, N-1144, GrN-3189, N-668.

42 SR-148, GX-1551, SR-226, SR-128, N-907, SR-174, SR-209, Pta-798, SR-161, SR-147, Pta-795, SR-207, SR-175, SR-237, N-906, SR-206, SR-242, SR-230, N-905, UCLA-1289, UCLA-1242, SR-239.

43 UCLA-929, SR-43, SR-17, M-913, SR-119, SR-26, SR-117, SR-79, SR-225, N-1275, N-978, SR-15, SR-73, I-4862, SR-33, SR-32, N-1148, SR-38, N-979, B-233, Tx-228, SR-163, SR-169.

44 Pta-901, Pta-914, Y-1212, GrN-5315, GrN-5022, RL-351, UCLA-1791B, Pta-328, UCLA-1791A, Y-1995.

45 GX-348, N-349, GX-347, N-257, N-448, N-649, N-1316, SR-114, HEL-224, HEL-223, HEL-222, N-256, N-462, N-275, N-882, N-274, HEL-225, HEL-226, N-273, GX-519, N-272, B-756, N-885, GX-900, GX-517, GX-247, M-1892, GX-516, N-290, N-1317, GrN-6111, GX-520, N-896, M-1893, N-893, N-461, M-1894, GrN-6109, N-1236, GX-743, B-257, M-1114, GrN-6108, GX-742.

46 R-908, GX-1422, GX-270, SR-23, GX-1423, SR-22, N-583, R-720, N-584, N-587, GX-767, N-593, N-1443, N-591, N-592, GX-1424, N-590, N-1603, N-589, GX-1331, N-571, SR-28, N-565, GX-1330, N-1606, N-1605, N-564, N-432, N-581, GX-109, N-563, SR-86, N-734, N-448, N-745, N-588, N-582, N-572, N-573 GX-1368, N-1602, N-744, SR-87, SR-42, GX-1369, N-1441, N-738, N-737, N-733, N-732, N-1281, SR-99, Y-618, N-1282, SR-89, Y-796, SR-54, GX-1332, N-736, N-731.

47 UCLA-1290, SR-235, SR-230, UCLA-1329, SR-236, SR-231, UCLA-1243, SR-221, UCLA-1244, SR-232, SR-220, UCLA-1236, SR-178, SR-233, UCLA-1245, UCLA-1246, UCLA-1238, N-1389, SR-212, SR-238.

48 C-613, C-917, SR-55, GL-19, SR-224, SR-68, N-772, SR-216, SR-136, SR-217, SR-218, M-914, SR-134, SR-37, SR-122, SR-53, SR-145, N-1147, SR-36, GrN-2341, SR-100, SR-115, Y-722, SR-101, SR-25, SR-58, SR-109, SR-44, SR-70, SR-47, SR-108, SR-219, SR-180, M-915, SR-94, SR-120, SR-196, N-1146, SR-124, SR-71, SR-177, N-1145, SR-179, SR-144, SR-35, SR-241, SR-195, Pta-605, SR-210, SR-162, SR-214, N-773, SR-213, SR-228, SR-153, SR-154, SR-140, SR-215, SR-229.

49 Y-1636, Y-1638, Y-1635, Y-1637, Pta-162, Y-135.17, Y-1338, Pta-911, Y-1639, Y-1662, Pta-849, Y-1828, Pta-850, Pta-853, Y-135.9, Y-135.14, Y-1657, GX-1014, Pta-703, GX-1015, Pta-715, Pta-777, GX-1462, Y-1323B, Y-1766, Y-1658, GrN-4797, GX-1463, GrN-4796, GrN-4727, Y-1767, GrN-4215, Pta-134, GrN-4928, GrN-4796, Y-1769, Y-1768, Y-1660, Pta-133, Y-1661, GrN-4929, Y-1997.

50 These dates are listed in notes 16 and 21, above.

51 N-894 N-895, N-890. GrN-5752, N-902, N-891, N-892, N-898, B-755, N-435, N-784, GrN-5753, B-758, N-436, GX-1186, N-437, N-901, N-900.

52 N-647, N-483, N-484, N-883, N-292, N-291, N-648, N-646.

53 The sample is here too small to permit the calculation of meaningful maximum limits.

54 Urewe ware may be regarded as at least partly ancestral to the eastern stream, but should probably not be subsumed therein. A more detailed discussion of this point is in preparation (see note 27).

55 Because of the small number of radiocarbon dates so far available from Early Iron Age sites south of the Limpopo, the relevant figures given in table 3 should be regarded as being liable to substantial revision.

56 Vogel, , ‘The Shongwe tradition’, 30.Google Scholar

57 N-1501, SR-65, GX-1328, N-1286, GX-1013b, GX-1013a, GX-1114, SR-123, GX-1012, GX-1425, N-998, SR-126, GX-1115, N-585, N-1498, N-1500, N-1499, N-1283, SR-118, N-595, B-263, N-586, N-594, N-1284, UCLA-716, GX-1327, SR-41, N-597, N-997, GX-1009, Pta-606, Pta-234, N-430, SR-211, N-431, GX-1010, N-596, B-264, SR-66, GX-1329, SR-245, GX-662, SR-139, N-429, SR-248, B-760.

58 See Phillipson, , ‘The Early Iron Age in Zambia’.Google Scholar

59 GX-1328, GX-1013b, GX-1013a, GX-1012, N-998, SR-126, SR-118, GX-1327, N-997, GX-1009, Pta-666, GX-1010, GX-1329, SR-245, GX-662, SR-248.

60 These dates are listed in note 22, above.

61 The early western stream of the Early Iron Age may prove to have been the source from which pastoralism and knowledge of pot-making techniques spread to the ‘Late Stone Age’ peoples of Namibia and the western Cape. Some support for such a hypothesis is provided by linguistic data cited in Ehret, C., ‘Patterns of Bantu and Central Sudanic settlement in central and southern Africa’, Transafr. J. Hist., III (1973), 171.Google Scholar For this to have been the case, it would be necessary for the Early Iron Age to have been established in more westerly regions, such as central Angola, by around the beginning of the Christian era. It is unfortunate that relevant archaeological data from this area are almost totally lacking.

62 Oliver, R. A., ‘The problem of the Bantu expansion’, J. Afr. Hist. VII (1966), 361–76.CrossRefGoogle Scholar

63 The late Professor Guthrie's concept of a ‘Bantu nucleus’ in this area has, of course, been widely criticized on linguistic grounds, notably by Greenberg, J. H., ‘Linguistic evidence regarding Bantu origins’, J. Afr. Hist., XIII (1972), 189216CrossRefGoogle Scholar, and Ehret, C., ‘Bantu origins and history: critique and interpretation’, Transafr. J. Hist., II (1972), 19.Google Scholar I shall return to the problem of correlation between archaeological and linguistic evidence in a paper presently in preparation.

64 Phillipson, , ‘Iron Age history’.Google Scholar

65 This rapidity may be, to a certain extent, illusory, in view of the broad geographical zones under which the data are, perforce, presented. As noted above, the later Iron Age may have begun in parts of Rhodesia—and perhaps elsewhere—at a date somewhat earlier than that here indicated.

66 Phillipson, , ‘Iron Age history’, 7, 1516.Google Scholar

67 Ibid., 7.

68 Phillipson, D. W., ‘Early Iron Age sites on the Zambian Copperbelt’, Azania, VII (1973), 119Google Scholar; idem, ‘Iron Age history’, 9.Google Scholar

69 For example, C. Renfrew has drawn attention to the major revisions made necessary to our view of European prehistory through the realization that radiocarbon-dated monuments and cultures such as Stonehenge or the Maltese megaliths are in fact older than Egyptian phenomena to which they were formerly believed to be related. Renfrew, C., ‘New configurations in Old World archaeology’, World Archaeology II (1970), 119211.Google Scholar

70 Burleigh, R., ‘The New Zealand Radiocarbon Conference’, Antiquity, XLVII (1973), 54–6.Google Scholar See also Switsur, V. R., ‘The radiocarbon calendar recalibrated’Google Scholar, ibid. 131–7.

71 Ralph, E. K., Michael, H. N. and Han, M. C., ‘Radiocarbon dates and reality’, M.A.S.C.A. Newsletter, IX (1973), no. 1.Google Scholar

72 Reliable data relating to the period before c. 5000 B.C. are not yet available.

73 I have prepared a table showing, for each decade between A.D. 1800 and 300 B.C., the relationship between radiocarbon dates (based on the 5,568-year half-life) and true ages according to the methods of Switsur (note 70 above) and Ralph et al. (note 71). A limited number of copies of this table are available and may be sent to interested individuals on request.

74 Although for all other periods the 5,730-year half-life will give a closer approximation to the true age, it seems preferable to avoid confusion by continuing to quote radiocarbon dates in terms of the 5,568-year version. This continues to be the practice in almost all date-lists and in journals such as Radiocarbon, and is in accordance with the resolutions of the last Radiocarbon Conference, noted by Burleigh, , ‘The New Zealand Conference’, 55.Google Scholar It is therefore confusing that the published correlation tables noted above have been drawn up on the basis of dates calculated on the 5,730-year half-life of radiocarbon.