Hostname: page-component-78c5997874-8bhkd Total loading time: 0 Render date: 2024-11-13T01:29:50.786Z Has data issue: false hasContentIssue false
  • English
  • Français

Some phantom dynasties of early and medieval India: epigraphic evidence and the abhorrence of a vacuum1

Published online by Cambridge University Press:  24 December 2009

David P. Henige
Get access Rights & Permissions [Opens in a new window]

Extract

Conventionally, the history of early and medieval India is viewed against a framework of a large number of dynasties, some local, others ‘imperial’, some ephemeral, others extending over several centuries.3 In this paper I argue that such a conceptualization is to some extent illusory, at least if a dynasty is defined as a series of genetically related rulers succeeding to a given office over time. The purpose of the paper is not to deny that many of these dynasties could have existed, or even that they did exist, but only to argue that the nature and extent of our evidence is not sufficient to permit many of the hypothesized reconstructions of them. In other words, the impression of a dynastic Indian past often has resulted from a maximum of inference based on a minimum of satisfactory data.

Type
Articles
Information
Bulletin of the School of Oriental and African Studies , Volume 38 , Issue 3 , October 1975 , pp. 525 - 549
Copyright
Copyright © School of Oriental and African Studies, University of London 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

2 Ibn Khaldū0n, Muqaddima, I, 14.

3 See, e.g., HCIP, iii-vi, where nearly 200 dynasties are surveyed; Karnataka through the ages. Bangalore, 1968;Google ScholarSharma, D. (ed.), Rajasthan through the ages. I. From the earliest time to 1316 A.D., Bikanir, 1966. The present paper concentrates on works which have appeared since the publication of the relevant volumes of HCIP since this series generally took cognizance of earlier work. However, when the argumentation was illuminating of itself I have included materials from pre-HCIP sources.Google Scholar

4 I have not considered the use of numismatic evidence here but it is obvious that data of this kind have even more restricted possibilities than epigraphic materials.

5 Henige, D. P., The chronology of oral tradition: quest for a chimera, Oxford, 1974.Google Scholar

6 ‘Ego autem coacervavi omne quod inveni’ Nennius in the preface to his Historia Brittonum. Lot, F., Nennius et l'Historia Brittonum; ètude critique, Paris, 1934, 147.Google Scholar

7 Ramesh, K. V., A history of south Kanara, Dharwar, 1970.Google Scholar

8 Ibid., xi.

9 Ibid., 43–136 passim.

10 Ibid., 96, 135–6, 149.

11 Ibid., 104–15. Ramesh is not alone in this type of assumption. See below, pp. 540–2.

12 Ibid., 116, 116, n.

13 Ibid., 54–5, 63–4, 114.

14 Ibid., 114.

15 For the last 150 years of Āḷupa history, for which the epigraphic data are more abundant, Ramesh posits no fewer than nine rulers, or an average of only 16–6 years per reign. Furthermore, as table 1 illustrates, several of these individuals seem to have ruled contemporaneously.

16 There may be less than 29 different names if certain identifications are accepted.

17 For a discussion of these rulers see Ramesh, South Kanara, 75–83, 125–9.

18 I propose to defer discussion of Ramesh's postulated genetic relationships among the Āḷupa rulers to section III of this paper, which discusses collateral suppression. See below, pp. 541–2.

19 Arokiaswami, M., ‘The Pallava Nāyakas of Kōrtāmpet’, Journal of the Asiatic Society of Bengal, Fourth Ser., v, 1–2, 1963, 28.Google Scholar

20 Ibid., 29. For another example of two rulers of unknown relationship being constructed into an embryonic ‘dynasty’ see HCIP, iv, 101.

21 See also Desai, P. B., ‘A new Kadamba family’, IHQ, xxxiv, 2, 1958, 67–9.Google Scholar

11 Swift, Gulliver's travels, pt. iii, ch. v.

23 I exclude the Sāntara inscription of four generations for the reasons discussed in the text.

24 Mahalingam, T. V., Kā***cīpuram in early South Indian history, London, 1969, 25.Google Scholar

25 See, for example, Sircar, D. C., The successors of the Sātavāhanas in the lower Deccan, Caloutta, 1939, 140–212;Google Scholaridem, The early Pallavas, Lahore, 1935; HCIP, iii, 255–6, 275–83;Google ScholarAruvamathan, T. G., ‘Early Pallavas and Kāñohi’, Transactions of the Archaeological Society of South India, xxv, 1960–1962, 51–84;Google ScholarMahalingam, T. V., ‘The early Pallava genealogy and chronology’, Proceedings of the 26th International Congress of Orientalists, New Delhi, 1964, iii, 2, Poona, 1970, 693–9.Google Scholar

26 At the risk of trespassing beyond my competence I would argue that this dichotomy may be too sharp. Language tends to change slowly and it is arguable that some or all of the Prakrit inscriptions could have been contemporaneous with some of the Sanskrit inscriptions.

27 HCIP, iii, 275–7; Mahalingam, Kāñcīpuram, 25–6. The seventeenth inscription is the Darśī grant which mentions the great-grandfather of an unnamed ruler.

28 Analogous to the Pallava case in this regard is the continuing debate over portions of the Hittite royal genealogy—a debate caused in no small measure by recurring royal names. See, inter alia, Goetze, H., ‘The predecessors of Šuppiluliumaš of Ḫatti’, Journal of the American Oriental Society, LXXII, 2, 1952, 67–72;CrossRefGoogle Scholaridem, ‘The problem of chronology and early Hittite history’, Bulletin of the American Schools of Oriental Research, 122, 1957, 18–20;Google ScholarGüterbock, H. G., ‘The predecessors of Šuppiluliuma again’, Journal of Near Eastern Studies, xxix, 2, 1970, 73–7CrossRefGoogle Scholar, and, most recently, idem, ‘Ḫattušili II once more’, Journal of Cuneiform Studies, xxv, 2, 1973, 100–4, and sources cited there.Google Scholar

29 This ruler is known only from an inscription of the Gupta ruler Samudragupta (c. 330/5-c. 376) and it is possible that he may not have been a Pallava. Nevertheless he is usually inserted in reconstructions of the Pallava dynasty in such a way as to date him as a contemporary of Samudragupta.

30 Ṥāntivarman Caṇḍadaṇḍa is known from a Kadamba inscription and, like Viṣṇugōpa, he may not have been a member of the Pallava ruling line. See HCIP, iii, 278.

31 Sircar in Ibid., iii, 282–3.

32 Ibid., 279, 283.

33 Mahalingam, Kāñcīpuram, 33–4.

34 Ibid., 32.

35 HCIP, iii, 277. Nor, and properly so, did Sircar structure the Prakrit group into a dynastic sequence.

36 Siṃhaviṣṇu, the last ruler in the sequences listed, has been described as ‘starting the Pallavas on their grand career of political and cultural achievement’, HCIP, iii, 258–9. Could this not have been made possible by his uniting several Pallava ruling lines ?

37 See below, p. 547, for a brief discussion of new epigraphic data for the Pallavas which illustrate the fragility of the models discussed here as well as of other reconstructions of the early Pallava genealogy and chronology.

38 HCIP provided two accounts of the Viṣṇukuṇḍin genealogy and chronology, as shown in table 3. Other recent attempts to dynasticize the ViṣŇukuṇḍins include Rao, M. Rama, ‘New light on the Viṣṇukuṇḍins’, PIHC, 27, 1965, 78–82;Google ScholarSundaram, K. S., ‘Further light on the genealogy and the chronology of the Vishnukundins’, Journal of the Andhra Historical Research Society, xxxi, 1965–1966, 39–44;Google Scholar and Sankaranarayanan, S., ‘New light on the genealogy and chronology of the Viṣṇukuṇḍins’, Journal of the Oriental Institute (Baroda), xvi, 4, 1967, 375–81, which provides a brief synopsis of earlier views and advances yet another model based on the Indravarman/Mādhavavarman/Govindavarman genealogy discussed in the text.Google Scholar

39 I have included here only the two hypotheses advanced in HCIP even though new data have undermined them, in order to display to best advantage the nature of the variants advanced. At the same time I would not like to suggest that any of the more recent theories mentioned in the preceding footnote are likely to prove more enduring than these two.

40 HCIP, iii, 207.

41 Ibid., 223.

42 Ibid.

43 Rama Rao, ‘New light’, 78–81.

44 Ibid., 79. Cf. HCIP, iii, 210–11.

45 Rama Rao, ‘New light’, 79.

46 Behera, S. C., ‘Māṭhara rule in Kaliṅga’, JIH, xlviii, 1, 1970, 119.Google Scholar See also Rajaguru, S. S., ‘The chronological account of the Māṭharas of Kaliṅga’, PIHC, 20, 1957, 94–100.Google Scholar

47 Behera, ‘Māṭhara rule’, 121–2.

48 Ibid., 122–3.

49 The thirteenth ruler is one Nandaprabhañjavarmā whom Behera places in ‘the next generation’ after Anantavarmā although he admits that ‘we have no evidence to show the relationship between the two’, Ibid., 123. Several authorities feel that many of the rulers counted as Māṭharas by Behera actually belonged to other families. See, e.g., HCIP, III, 212.

50 Gopal, B. R., ‘The Sindas of Reñjēru’, Journal of Ancient Indian History, ii, 1968–1969, 84–93.Google Scholar

51 Gopal argues that ‘the reigns of Bhairava and Hemma (Permāḍi II), respectively the son and grandson of Permāḍi I, covered a period of thirty years’, Ibid., 91. This interpretation, plausible in itself, requires accepting the unlikely twin premises that Permāḍi I died in the very year of his latest known inscription and that Barma, the ruler of the 1169 record, had acceded in that same year.

52 Ibid., 89–91. For more Permāḍis, this time belonging to a Sinda family at Yelburga, see Karnataka through the ages, 308–9.

53 For other Sinda families see Sircar, D. C., Studies in the society and administration of ancient and medieval India, i. Society, Calcutta, 1967, 137–8; HCIP, v, 175–88 passim.Google Scholar

54 Gopal, ‘Sindas of Reñjēṛu’, 92–3.

55 Sharma, Rajasthan, 224–5, 553–9. For the Tomara coins see Cunningham, A., Coins of mediaeval India from the seventh century to the Muhammadan conquest, London, 1894, 80–8.Google Scholar See also Sharma, D., ‘New light on the Tomaras of Delhi’, PIHC, 19, 1956, 150–2.Google Scholar

56 Sharma, Rajasthan, 553–9.

57 Ibid.

58 Ibid., 554–9.

59 Hiralal, R. B., ‘Four Bhañja copper-plate grants’, EI, xviii, 29, 1925–1926, 286–303.Google Scholar

60 Banerji, R. D., ‘The Tekkali plates of Ṥatrubhañja—v.s. 800’, Journal of the Bihar and Orissa Research Society, xviii, 3, 1932, 387–90.Google Scholar See also Sircar, B. C., ‘The Bhañjas of Khiñjali-Maṇḍala’, IHQ, xxviii, 3, 1952, 225–31.Google Scholar

61 By Sircar, art. cit. In fact the royal styles of the two pairs of rulers differ markedly.

62 See, e.g., HCIP, iv, 70; Acharya, P., ‘The Bhañja kings of Orissa’, Journal of the Bihar and Orissa Research Society, iii, 1, 1926, 92–102.Google Scholar

63 Sircar, ‘Bhañjas of Khiñjali-Maṇḍala’, 228.

64 The difficulties of accommodating all the available data to the concept of a single long-lived dynasty at Khiñjali can be discerned in the attempt to do so in HCIP, iv, 71–2.

65 For the other Bhañja ruling lines see HCIP, iv, 72–3; Misra, B., Dynasties of mediaeval Orissa, Calcutta, 1933;Google ScholarGupta, C. C. Das, ‘Some notes on the Ādi-Bhañjas of Khijjiṅga Kotta, earlier Bhañjas of Khiñjali-Maṇḍala, Bhañjas of Bauda and later Bhañjas of Khiñjali’, Annals of the Bhandarkar Oriental Research Institute, xii, 3, 1931, 231–45.Google Scholar

66 One example that compares in scope with those discussed here concerns the studies of the Gaṅga rulers of Śvetaka. From the inscriptions we again know of three groups of members of the Gaṅga family who ruled in the Śvetaka area. These three groups amount to 10 names and few genealogical relationships are known. The available evidence is not sufficient to justify the claim that ‘there is hardly any doubt that they [the Gaṅgas] constituted an independent line of rulers for a period extending over nearly 500 years’, Majumdar, R. C., ‘Phērava grant of Sāmantavarman, king of Kaliṅga, year 185’, EI, xxvii, 19, 1947–1948, 110.Google ScholarBehera, S. C., ‘Chronology of the Gaṅgas of Śvetaka’, Journal of the Oriental Institute (Baroda), xix, 4, 1970, 361–8,Google Scholar constructs a sequence of 10 rulers spanning without intermission the period from c. 720 to c. 925. For other examples of this process see Mishra, V. B., ‘The Gurjara-Pratīhāras in Bhṛgukaccha (Broach)’, Annals of the Bhandarkar Oriental Research Institute, xxxi, 1950, 250–7;Google ScholarGanguli, H. P., ‘The Śailodbhavas and their contemporaries of Kaṅgoda’, IHQ, xxxiii, 3, 1957, 208–21;Google ScholarMishra, P. L., ‘New light on the Nalas of Chattisgarh’, IHQ, xxxvi, 4, 1960, 247–59;Google ScholarBath, A. K., ‘A note on the Vigraha dynasty’, Orissa Historical Research Journal, xi, 1, 1962, 54–9;Google Scholar N. K. Sahu, ‘The Nalas’, Ibid., 95–102; Ganguly, D. K., ‘The genealogy and chronology of the Śailodbhava kings of Orissa’, PIHC, 26, 1964, 51–2;Google ScholarBajpai, K. D., ‘Fresh light on the post-Aśokan history of Kauśāmbī’, Journal of the Numismatic Society of India, xxvi, 1, 1964, 1–6;Google ScholarLahiri, B., ‘Early history of the Māghas of Kauśāmbi’, PIHO, 27. 1965, 102–7;Google Scholar D. K. Ganguly, ‘The genealogy and chronology of the Śailodbhava kings of Orissa’, Ibid., 82–92; Bamesh, K. V. and Katti, M. N., ‘The origins and early history of the Sāntaras of Sāntaḷige—1,000’, Quarterly Journal of the Mythic Society, lvii, 1966–1967, 25–36;Google ScholarVerma, T. P., ‘The line of Dhanabhūti’, Journal of the Numismatic Society of India, xxxi, 2, 1969, 143–50;Google ScholarKrishnamurthy, C., ‘Some minor dynasties of the Kolar district (1179–1338 A.D.), in Ganesan, S. (ed.), Professor K. A. N. Nilakanta Sastri felicitation volume, Madras, 1971, 72–83Google Scholar; O. Ramachandraiya, ‘ Muñjeru plates of eastern Ganga Avantavarma', Ibid., 119–20; M. D. Sompath, ‘ New light on the Nolambas’, Ibid., 165–79.

67 Sauvaget, J., ‘Les ēpitaphes royales de Gao’, Bulletin de l'Institut Français de l'Afrique Noire, XII, 2, 1950, 418–40.Google Scholar

68 This research is being conducted by J. O. Hunwick of the University of Ghana and P. F. de Moraes Farias of the University of Birmingham. I am grateful to Dr. Farias for advising me of some of the results of this research.

69 Tedeschi, S., ‘Note storiche sulle isole Dahlac’, Proceedings of the third International Conference of Ethiopian Studies, Addis Ababa, 1966, Addis Ababa, 1970, I, 58–66.Google Scholar

70 The number of works dealing with the chronology of ancient South Arabia is legion. Among the more important are Jamme, A., La dynastie de Sarahbi'il Yakuf et la documentation épigraphique sud-arabe, Istanbul, 1961;Google Scholaridem, The al-‘Uqlah texts, Washington, D.C., 1963;Google ScholarH StPlilby, J., The background of Islam, Alexandria, 1947;Google ScholarPirenne, Jacqueline, La royaume sub-arabe de Qataban et sa dotation, Louvain, 1961;Google ScholarRyckmans, J., L’institution monarchique en Arable mèridionale avant l'Islam, Louvain, 1951;Google Scholaridem, La chronologie des rois de Saba et du Raydan, Istanbul, 1964;Google ScholarWissmann, H. von, Zur Geschichte und Landeskunde von alt-Südarabien, Wiesbaden, 1953. Of these works Philby's, though the most dated, has the advantage of presenting the problem most clearly. Numerous articles on the subject have appeared in Bibliotheca Orientalis, Bulletin of the American Schools for Oriental Research, Bulletin of the School of Oriental and African Studies, and Le Musèon. See the bibliography in Wissmann, Geschichte, 9–23, and in recent numbers of BASOR.Google Scholar

71 Newton, Isaac, Chronology of ancient nations amended, London, 1728, 54.Google Scholar

72 Fox, J. J., ‘A Rotinese dynastic genealogy: structure and event’, in Beidelman, T. O. (ed.), The translation of culture, London, 1971, 43–4.Google Scholar

73 Thus it really is not ‘curious’ that the Gupta ruler Skandagupta (c. 455-c. 467) was omitted from ‘the official records of a later date’ and from the Gupta ‘ royal genealogical list’ since he was succeeded by his brother and the latter by his own son, HCIP, iii, 25. Skandagupta is known only from his own inscriptions. For similar cases of collateral suppression in Nepalese epigraphs see Regmi, D. R., Medieval Sepal, 2 vols., Calcutta, 1965–1967, n, 44–5, 206, 207, 214.Google Scholar

74 There are exceptions to this statement, including Basham, A. L., ‘The average length of generation and the reign in ancient India’, in idem, Studies in Indian history and culture, Calcutta, 1964, 83;Google Scholar the discussion of the later Pratīharas in HCIP, iv, 34; Sen, B. C., ‘Early Indian approach to history and some problems of its reconstruction’, Journal of the Bihar Research Society, LIV, 1968, 12–13.Google Scholar

75 See the genealogical tables in Sharma, Rajasthan, 274–6. See also HCIP, iv, 108–9; v, 89.

76 For this portion of Tod's history of Mewar see his Annals and antiquities of Rajasthan,3 vols., London, 1920, 1, 281–307.Google Scholar

77 For a discussion of the problems of extended father/son succession patterns in the Indian chronicles see my Chronology of oral tradition, 78–81.

78 See, for example, the Paramāras of Jalor in Bhatia, P., The Paramaras c. 800–1305 A.D., New Delhi, 1970, 186–9. See also the items listed at p. 541, n. 93.Google Scholar

79 Henige, Chronology of oral tradition, 72–3. Only nine instances of such a lengthy succession have been determined from a sample of over 10,000 documented successions. The longest of these covered only 15 generations.

80 It should be borne in mind that these are minimum figures which could increase as new data become available. In these calculations rulers of unknown relationship have been excluded. Although I did not rely on it for this table, a convenient tabulation of several medieval Indian dynasties can be found in Trautmann, T. R., ‘Length of generation and reign in ancient India’, Journal of the American Oriental Society, LXXXIX, 3, 1969, 564–77. In his calculations Trautmann used only the best-known portions of the most important medieval dynasties. Even so, his assertion that these ‘genealogies are quite accurately known from the epigraphs’ might be questioned. Some of his data, such as those for the Kadambas, are based on outmoded sources and this affects his argument on reign length although not necessarily that for generation length. Trautmann does not seem aware of the problem of collateral suppression in his calculations and this too could effect his reign-average argument.CrossRefGoogle Scholar

81 The Bhauma-Kara dynasty presents several striking features. Nearly all of the inscriptions for the dynasty are dated in the years of an era apparently of their own creation and this enables sequence to be established more closely than is normally possible. We have records of 17 rulers within a 200-year period. Of these no fewer than six were reigning queens. HCIP, iv, 62–8.

82 One of the collateral successions among the Candelas was that of Kīrtivarnian (c. 1060 to c. 1100), who succeeded his brother Devavarman (c. 1050 to c. 1060), thereby dropping the latter from the direct line. Devavarman is not mentioned in Kīrtivarman’s inscriptions, in which Kīrtivarman is called the son (which he was) and the successor (which he was not) of Vijayapāla (c. 1030 to c. 1050). Bose, N. S., History of the Candellas of Jejakabhukti, Calcutta, 1956, 71, observed that ‘but for the contemporary inscriptions of Devavarman himself…we should not have known that a king of that name had ruled at all’.Google Scholar

83 These figures do not include t h e last several Pratīhāra rulers, whose genealogical relationships to earlier rulers and to each other is not known. It has generally been agreed that Kakutstha, the second ruler of the dynasty, was the nephew of Nāgabhaṭṭa, the founder, but Katare, S. L., ‘The Pratīhara genealogy from Nāgabhaṭ ṭa I to Vatsarāja I’, JIH, xlviii, 2, 1970, 437–43, has argued that Kakutstha was the son of Nāgabhaṭ ṭa.Google Scholar

84 For the early Yādava genealogy see below, pp. 545–7.

85 cf. the genealogies of the Kadamba dynasties in Moraes, G. M., The Kadamba-kula: a history of ancient and mediaeval Karṇāṭaka, Bombay, 1931, 93,Google Scholar 167, and those in Gopal, B. R., ‘The Kadambas of Hāngal’, Journal of Karnalak University, iii, 1967, 99–115,Google Scholar and idem, ‘The Kadambas of Goa’, Journal of Karnalak University, iv, 1968, 164–77.Google Scholar

86 For the Śilāhāra dynasties see Altekar, A. S., ‘The Śilāhāras of western India’, Indian Culture, ii, 3, 1935, 394–434;Google ScholarKarnataka through the ages, 311–12; Maharashtra State gazetteers: Kolaba district, Bombay, 1968, 65–6.Google Scholar

87 Altekar, ‘Śilāhāras’, 398.

88 Ibid. Elsewhere Altekar admits the possibility that Raṭ ṭarāja may have had one successor; Ibid., 401.

89 Another inscription attributed to this dynasty has since come to light but it only provides a date for the father of Raṭṭaraja. See Karnataka through the ages, 311.

90 Venkataramanayya, N., The Chālukyas of L(V)emulavāḍa, Hyderabad, 1953.Google Scholar

91 Ibid., p. 36, n.

92 Ali, S. M., ‘Chandra kings of Pattikera and Arakan’, Journal of the Asiatic Society of Pakistan, vi, 1961, 269,Google Scholar and Sarma, I. K., ‘Select inscriptions of Ikshvāku rulers of Nāgārjunakoṇ ḍa‘, Journal of Oriental Research, xxix, 1959–1960, 44, used these terms to describe ascendant genealogies relating to their ruling lines.Google Scholar

93 Among others these would include Sircar, D. C., ‘Two grants of early Guhilas’, EI, xxxiv, 26, 1961–1962, 167–76;Google ScholarSastri, S. B., ‘A dynasty of Buddhist kings of Śrīkaṭ ṭa‘, Proceedings and transactions of the 23rd All-India Oriental Congress, Aligarh, 1966, 342–5;Google ScholarAli, , ‘Chandra kings’, 267–74;Google ScholarPandey, R. B., ‘Jabalpur plates of Mahārāja Hastin, G. E. 170’, EI, xxviii, 42, 1949–1950, 264–7;Google ScholarMajumdar, R. C., History of ancient Bengal, Calcutta, 1971, 282–4.Google Scholar

94 Ramesh, South Kanara, 50, 63, 68, 69, 75, 79, 81, 84, 92, 93–4, 100, 107.

95 HCIP, iv, 142–3; v, 205–9; vi, 361–5. The chronology and genealogy of early Gaṅga rulers in Kaliṅga is extremely obscure. See HCIP, iii, 215–17; iv, 139–40, and Sahu, N. K., ‘Chronology of the early Gañga kings of Kaliñga’, Orissa Historical Research Journal, vi, 4, 1958, 251–70. All of the proposed solutions involve some of the dynasticizing called into question here.Google Scholar

96 HCIP, v, 864; vi, 836.

97 Ibid., vi, 363.

98 The latest date for a member of the dynasty is 766. The most recent account of the Maitrakas is Gujarat District gazetteers: Bhavnagar, Ahmedabad, 1968, 48–55,Google Scholar while the fullest account remains Virji, K. J., Ancient history of Saurashtra, Bombay, 1952.Google Scholar See also HCIP, iii, 60–3, 147–50, and Ray, N. R., ‘The Maitrakas of Valabhi’, IHQ, iv, 3, 1928, 453–74.Google Scholar

99 Gujarat District gazetteers: Bhavnagar, 48–55; HCIP, iii, 711. The period after Ṛīaditya III (c. 660 to c. 685), when the epigraphic evidence is less abundant, is particularly doubtful. All the known rulers were named Ṛlāditya and each was the son of his predecessor. This pattern is much at variance with the earlier variety of royal names and fragmented succession patterns for the dynasty. For this later period see Virji, Saurashtra, 63–105.

100 Devahuti, D., Harsha: a political study, Oxford, 1970, 47, makes both these assumptions.Google Scholar

101 Hsüan-tsang, who visited India from 630 to 644, stated that a Śīlāditya had ruled Mo-la-po, generally thought to mean Malwa including SaurāṢṭra, 60 years before his time. This statement cannot be taken literally since we know that Dharasena was ruling at least as late as 589. This being so, neither can we use it to suggest a date for Śīlāditya's accession. See HCIP, 111, 63, 112, 607.

102 An instance from the Gāhaḍavāla dynasty may serve to complete this point. Govindacandra is known to have been’ruling as late as 1154 and in his inscriptions his son Asphoṭavarman is described as his yuvarāja and Vijayacandra as his second son. Dates of 1168 and 1169 exist for Vijayacandra but none for Asphoṭavarman as ruler. While I would not argue that because Asphoṭavarman was yuvarāja he eventually succeeded his father (see the next paragraph), the fact that none of his inscriptions has been recovered is not of itself sufficient reason to assert that ‘Asphoṭavarman seems to have predeceased his father and Vijayacandra succeeded to the throne sometime after 1154’, HCIP, iv, 54, 849. Trautmann, ‘Length of generations’, 575, dates Vijayacandra from 1155 to 1170. Surely the question of whether Asphoṭavarman actually reigned must be considered an open one?

103 HCIP, 111, 277.

104 Rao, B. V. Krishna, A history of the early dynasties of Andhradeśa, c. 200–625 A.D., Madras, 1942, 213–14. Mahalingam, Kāñcīpuram, 26–7, also excludes these two names from his reconstruction although he does assume that Buddhavarman was the son as well as the yuvarāja of Skandavarman.Google Scholar

105 For another view of this problem see Goody, Jack, ‘Sideways or downwards? Lateral and vertical succession, inheritance and descent in Africa and Eurasia’, Man, NS, v, 4, 1970, 627–38.CrossRefGoogle Scholar

106 On this see Henige, Chronology of oral tradition, 73–91 passim.

107 Lethbridge, R., The golden book of India, London, 1900,Google Scholar referring to Pal-Lahara, one of the Orissa Feudatory States. Similar remarks echo throughout the gazetteer accounts of the Native States of India and other works based on the testimony of the chronicles, e.g. H. Wilberforce- Bell, The history of Kathiawad, London, 1916, and J. Hutchison and Vogel, J. P., History of the Panjab hill states, 2 vols., Lahore, 1933.Google Scholar

108 Among instances where new discoveries have shown collateral successions we might cite Sircar, D. C., ‘The Guhilas of Kiṣkindhā’, Our Heritage, xi, 1, 1963, 56–7;Google ScholarDutta, K., ‘New light on the history of Assam’, PIHC, 12, 1949, 157–9;Google Scholar and Dikshit, D. P., ‘The early western Chālukya chronology reconsidered’, Orissa Historical Research Journal, xv, 3–4, 1967, 18–20.Google Scholar Pulakeśin [II] of the Cālukya dynasty was killed in 642 and the first known date of his son Vikramāditya [I] is 655. Although Pulakeśin is known to have had several other sons it has generally been argued that an interregnum intervened between the two rulers rather than that other sons of Pulakeśin might have ruled before Vikramāditya. The new evidence indicates that as many as four rulers intervened between Pulakeśin and Vikramāditya. For an Indo-Chinese example of this problem see Briggs, L. P., The ancient Khmer empire, Philadelphia, 1951, 205, and the sources cited there.Google Scholar

109 Dewey, J., Experience and nature, Chicago, 1925, 154.Google Scholar

110 Murthy, A. V. M., The Sevunas of Devagiri, Madras, 1971, 32.Google Scholar For a list of over 400 of these see Verma, O. P., The Yādavas and their times, Nagpur, 1970, 369–84. There are less than 10 inscriptions from the pre-1185 period under discussion.Google Scholar

111 For Hemādri and his writings see Kane, P. V., History of dharmaś āstra, 5 vols. in 7, Poona, 1930–1962, 1, 354–9,Google Scholar and Yazdani, G. (ed.), Early history of the Deccan, 2 vols., London, 1960, 11, 570.Google Scholar

112 Verma, Yādavas, 21–3, 42–9, 367–8.

113 Ibid., 21–49, 367–8.

114 Ibid., 367–8.

115 For Seuṇacandra III see p. 547, n. 122, below.

116 Verma, Yādavas, 368.

117 Ibid., 37.

118 On the elision of ‘usurpers’ see Henige, Chronology of oral tradition, 28–33.

119 The problem of the conflicting data for these rulers is discussed in Verma, Yādavas, 46–9, and idem ‘A discrepancy in the Yādava genealogy’, Nagpur University Journal (Humanities)’ xvi, 1, 1965, 32–6.Google Scholar

120 Sircar, D. C. and Sankaranarayanan, S., ‘Dharwar plates of the time of Siṁhaṇa’, EI, xxxiv, 7, 1961–1962, 34, suggested that Karṇa might be identified with Amaramallugi.Google Scholar

121 We would like to know more about Hemādri's sources for this earliest period. Presumably he consulted inscriptions that were available to him and many of these would not be available to more recent historians. The sequence of rulers for this period may have been known only from some later retrospective inscriptions; if so, Hemādri had no more to work with than contemporary historians. The fact that no inscriptions before the time of Bhillama [II], putatively the eighth ruler of the dynasty, have survived suggests either that the earliest names were not historical personages or that they were not reigning kings. The question must remain moot.

122 A certain Seuṇa left an inscription dated 1141 but he is not known either from the Vratakhaṇṭa or from later inscriptions. His inscription credited him with ‘the usual attributes of the Yādava kings’ but he may have been the member of an (otherwise unknown) collateral branch. Verma, Yādavas, 43, 368, includes him in the main line as Seuṇacandra III.

123 Murthy, Sevunas, differs considerably from Verma in his presentation of the Yādava genealogy. He omits Dhāḍiyasa and equates Siṃharāja with Verma's Seuṇacandra (see preceding footnote) but without citing any reasons. He sees the five rulers between Mallugi and Bhillama [V] as contenders with Bhillama [V] and not independent rulers in their own right. He notes that Bhillama [V]'s own inscriptions do not begin until 1187, but on the basis of a Kalacuri reference he considers that he was ruling at least as early as 1173 and dates his accession to that year. Murthy, Sevnnas, 34–6. On the face of it Verma seems to have taken less liberties with the evidence and his account, while not demonstrably correct, is more persuasive.

124 See above, pp. 529–32.

125 For this dating see, e.g., HCIP, 111, 258–60, 279–80; Mahalingam, Kāñcīpuram, 34.

126 I have based this paragraph on a brief notice in Weekly India News of 26 August 1971. It is possible that this preliminary interpretation may prove to be incorrect and, if so, the present analysis would of course also be nugatory.

127 An accession date of c. 540 for Siṃhavarman would be supported by the Viṣṇukuṇḍin inscription of 566/7 mentioned above (p. 533) since this inscription mentioned a Pallava ruler named Siṃha. Rama Rao, ‘New light on the Viṣṇnukuṇḍins’, 80–1, took this reference to be to Siṃhavarman s/o Viṣṇugõpa of the Ōṃgōḍu II inscription and alleges that this Siṃhavarman invaded Veṅgī in his eighth year and thus acceded in 556. This conclusion is based on a controversial identification of a toponym and on this Siṃha-Siṃhavarman s/o Viṣṇugōpa identification. It now appears more likely that it was to Siṃhavarman f/o Mahendravarman that this inscription refers.

128 See, e.g., Shastri, H. G. and Dholakia, P. V., ‘Āṁbaḷās plates of the Śaindhava king Āhivarman’, Journal of the Oriental Institute (Baroda), xix, 3, 1970, 279–85;Google ScholarSundaram, J., ‘Two Bhañja copper plates’, EI, XXXVII, 42, 1967–1968, 233–40;Google ScholarSircar, D. C. and Bhattacharya, G., ‘Mallār plates of Vyāghrarāja’, EI, xxxiv, 9, 1961–1962, 45–50;Google ScholarSircar, D. C., ‘Two grants of early Guhilas’, EI, xxxiv, 26, 1961–1962, 167–76, as well as the works cited at p. 544, n. 108.Google Scholar

129 Sircar, , Indian epigraphy, Delhi, 1965, 11–13.Google Scholar

130 This is a point to which I have not devoted specific attention but the practice can be seen in many of the examples discussed in the paper, particularly in table 1. Altekar, ‘Śilāhāras’, assigns to a certain Vijayāditya the dates 1140 to 1175 on the basis of three of his inscriptions ranging from 1143 to 1153, and the dates 1175 to 1215 to a certain Bhoja on the basis of inscriptions dated 1179 and 1205. No thought is given to the possibility of one or more rulers intervening between 1153 and 1179. This is but one of a very large number of examples.

131 There are probably not many examples of this practice since non-epigraphic data are so seldom available. In this regard see the discussion in HCIP, 111, 17–18, of the possible reign of Rāmagupta between Samudragupta and Candragupta [II] of the Gupta dynasty.

132 Sircar, Indian epigraphy, passim, esp. 23–30; Sen, ‘Early Indian approach to history’, 11–15; Basham, A. L., ‘Modern historians of ancient India’, in Philips, C. H. (ed.), Historians of India, Pakistan and Ceylon, London, 1961, 290–3.Google Scholar