11 Dakṣiṇa Kosala

Observations on the History and Culture of

Dakṣiṇa Kosala∗

Fifth to Seventh Centuries ad

Introduction

The historiography of the region called Dakṣiṇa Kosala, nowadays generally known as Chhattisgarh, is beset with difficulties of a predominantly chrono logical nature. Apart from quite a number of inscriptions, we do not possess written sources that can help us to unravel its early history. The chronological problems are due to the fact that, with one isolated exception, the charters of its kings are dated in regnal years. Moot questions, such as the dating of the kings of Sarabhapur, the relation, if any, between the Pān ´ . ḍava dynasties of Mekalā and Kosala, or the date of King Tīvaradeva, have been discussed again and again by a number of scholars during the last fifty years, a debate that has been dominated by three eminent Indian epigraphists, V.V. Mirashi, ḌC. Sircar, and A.ṀShastri. A first reading of this facinating corpus of learned articles gives the uncomfortable feeling that these three scholars dis agree among themselves on almost every issue. Only laborious study makes one realize that in this debate a large body of historical evidence has been dis closed and evaluated, as a result of which we know to date considerably more about the history and culture of this region than half a century ago. Still, many inscriptions await publication and this is, unfortunately, in particular true for those on stone. Unlike copperplate charters, stone epigraphs often inform us about particular historical circumstances and details beyond the official royal records. They are less formal and regulated, but they are usually more difficult to decipher and this may have hampered their publication.

It seems a whim of fate that the only inscription belonging to this period and said to be found in this area that is dated in a known era, sciḷthe Gupta Era, has no apparent connection with any of the other data. I refer to the

∗ The first version of this article was published in Balbir, Nalini & Joachim K. Bautze (eds.), Festschrift Klaus Bruhn zur Vollendung des 65. Lebensjahres, dargebracht von Schülern, Freunden und KollegeṇInge Wezler Verlag für orientalistische Fachpublika tionen, Reinbek 1994. Studien zur Indologie und Iranistik. pp. 1–66.

235

Araṅg Copper-Plate of Bhīmasena II ¯ . Though in this case the era is known, there is uncertainty with regard to the reading of the year. Hiralal, who edited the inscription (EI IX, 342–345), read 282, which would mean ad 601/2. This would make it the latest record dated in the Gupta Era in an area in which Gupta influence had virtually disappeared one century earlier. 1 Mirashi has vigorously argued against this interpretation, which was accepted by Sircar and Shastri, and has made it plausible that we should read 182 instead of 282. 2 This would place Bhīmasena of the Sūra family in that phase of the history of ´ Dakṣiṇa Kosala, in which it emerged as a politically and culturally autonomous state. Bhīmasena may therefore be seen as a figure who, in the turmoil of the age, hoped, by pledging allegiance to the once mighty Guptas, to secure a dominant position in the political arena of his days. However, Gupta support no longer amounted to much, and Bhīmasena and his dynasty disappeared from the scene without leaving a trace. 3 His inscription can therefore not be used for absolute dating of the dynastic history of Kosala, 4 and hence we have left it out of account in the following reconstruction.

There is another important source of the history of Kosala, viz. the archae ological materials in which this region abounds. Major contributions to their systematic and coherent treatment were made by Donald Stadtner in his the sis Sirpur to Rajim: the Art of Kosala During the Seventh Century,5 and by Krishna Deva in the Encyclopedia of Indian Temple Architecture (II.1, Chap ter 18, 1988). But whereas the archaeologists and historians of art concentrate on the abundance of material remains, and the historians and epigraphists on a sizeable corpus of inscriptions, very few attempts have been made to an integral treatment. Although, in my opinion, only such a study can really deepen our understanding of the cultural history of this region, the present article does

1 ‘The Sumandal copperplate inscription of King Pr̥thivīvigraha who was the ruler of Kaliṅga, refers to the sovereignty of the Guptas in the year 250, i.e. A.D. 569/70. This was perhaps the last flicker of an extinguishing lamp.’ (Agrawal 1989, 269). Cf. Bakker 2014, 60.

2 EI XXVI, 227; see also many of Mirashi’s later publications. Although the Plate is reported to have been found in Arang, this seems by no means certaiṇThe place of issue, Suvarṇanadī, is identified by Sircar as the River Son, though the Mahānadī might be another possibility. The characters of the inscription, according to Sircar (op. cit. 342), ‘belong to the Northern class of alphabets’.

3 A king named Harirāja, son of Niṣṭhurarāja, grandson of Bhāgraharāja (?), said to belong to the Sūra family, is known from a copperplate charter found in Vārān ´ . asī. The open nail-headed letters of this inscription point to Central India (possibly Dakṣiṇa Kosala), which led Michael Willis (2014, 109) to conclude that ‘this plate, although belonging to the fifth century, was found outside of its first geographical context’. King Harirāja and his Queen Anantamahādevī may have belonged to the same vaṁśa as Bhīmasena, though the latter’s inscription does not mention Harirāja and his ancestors in its pedigree. Neither the place of issue, Sāntanapura, nor the village in which land ´ was donated, Ambrakanagara, have been identified. See Bhattacharya 1945, 167–73. ¯ Also Chhabra 1949, 47–48 and Plates XXI–XXII.

4 In this article ‘Kosala’ is shorthand for Dakṣiṇa Kosala.

5 Unpublished PhD dissertation, University of California, Berkeley 1976.

12 / History and Culture of Dakṣiṇa Kosala 237

not have that pretentioṇIt was written after a short visit to Chhattisgarh and touches only on a fraction of its historical wealth. What it does hope to illustrate, however, is that the methodology employed, consisting in utilization of sources of various sorts with the aim of cross-fertilisation, is in principle a sound and fruitful one. Whether this is borne out by the following essay is up to the reader to decide.

Figure 2

Historical map of Mekalā and Dakṣiṇa Kosala

The Paṉ .ḍavas of Mekalā

This lineage of Pāṇḍavas is known from two inscriptions, both to be ascribed to the last recorded member of the dynasty, Sūrabala, alias Udīrn ´ . avaira, dating from the 2nd and 8th years of his reign. 6

6 Bamhanī Plates of Pāṇḍava king Bharatabala: year 2 (EI XXVII, 132–145; also in CII V, 82–88). This charter, found in the village Bamhani in Shahdol District, is probably

Figure 3

Dynasties of Mekalā and Dakṣiṇa Kosala

Figure 4

Dynastic capitals of Mekalā and Dakṣiṇa Kosala

to be ascribed to Bharatabala’s son Sūrabala, as a comparison with the latter’s ‘Malhar ´ Plates’ (JESI III, 183–193) suggests. A third, incomplete inscription in nail-headed characters, is reported by K.d. Bajpai 1977-78, 433–37. It concerns the second of three plates, found in Būrhīkhar (near Malhar). It contains a prose passage, followed by two verses in praise of King Nāgabala and one incomplete verse referring to his Queen Indrabhaṭṭārikā.

12 / History and Culture of Dakṣiṇa Kosala 239

A descendant of the legendary Pāṇḍu, this king considers himself to belong to the Lunar Race (somavaṁśa). The first two kings in the genealogy according to both inscriptions are styled rāja and nr̥pati, the later kings mahārāja. Son of Jayabala, founder of the dynasty, is Vatsarāja, from whose marriage with Droṇabhaṭṭārakā mahārāja Nāgabala is borṇFrom the latter’s marriage with Indrabhaṭṭārakā is born mahārāja Bharatabala, who is eulogized as a great warrior. 7 Bharatabala married a daughter of a family descending from Amara (amarajakula), a fact that is particularly emphasized. This princess, designated as ‘Lokaprakāśā’, is said to hail from Kosalā. 8 The accentuation of her origin seems to indicate that this relationship of the Pāṇḍavas of Mekalā with a lead ing family in Dakṣiṇa Kosala was of some importance to them. According to the inscription of the 8th regnal year, the son born out of this marriage, Sūrabala ´ Udīrṇavaira, allowed a donation to be made by one Narasiṅgha, son of Boṭa, of a village at a confluence (saṁgama), situated in the southern province of Mekalā (mekalāyāṁdakṣiṇarāṣṭre), to a temple of Jayeśvarabhaṭṭāraka.9 The Saiva leanings of this king not only appear from the recipient of the endow- ´ ment, Jayeśvarabhaṭṭāraka (i.e. Siva), whose eulogy opens the Malhar Plates, ´ 10 but also from the seal attached to them, which ‘bears in relief the figure of a

7 Of the four verses eulogizing King Bharatabala in the Bamhanī Plates (EI XXVII, 141, vv. 5–8), two (7–8) are omitted, one (6) replaced by another, and one (5) retained in the Malhar Plates (JESI III, 190).

8 Of the two verses (9–10) eulogizing this princess in the Bamhanī Plates (EI XXVII, 141) the first one (9) has been substituted by a prose passage in the Malhar Plates saying that she was the mother of mahārāja Sūrabala (JESI III, 190 f.). The two verses are ´ presented below, n. 23 on p. 242. The Malhar Plates read jātā yā kosalāyām, and, in my opinion, the Bamhanī Plates read the same (cf. Mirashi in CII V, 85 f.). The feminine form ‘Kosalā’ could mean ‘town of Kosala’ (cf. Kosalā = Ayodhyā, MBh 3.75.3; Bakker 1986 I, 6,7,9), but it seems equally possible that the actual name of the area was Kosalā (see EI XXXI, 221 ll. 8–9 kosalāmaṇḍala; CII V, 81 l. 27 and VSMA 1971/72, 75 l. 22 kosalā–mekalā; however CII V, 108 l. 14 reads kosala-trikūṭa). There is an old village Kosalā at the Kamji River, 25 km east of Malhar. ‘The antiquity of this large site, which has several mounds and the remains of a moat all round, go [sic] back at least to the Maurya period. Apart from early historical pottery, a copper coin of the Kushāṇa emperor Vima-Kadphises was acquired by us from the site.’ (Bajpai & Pandey 1977, 21).

9 JESI III, 191 f. (text corrected):

tato mekalāyāṁdakṣiṇarāṣṭre saṁgamagrāmake grāmakūṭapramukhaṁnāyakapra mukhaṁca grāmaṁsamājñāpayati | viditam astu vo ’smābhir ayaṁgrāmaḥso draṅgah. [. . . ] mātāpitror ātmanaś ca puṇyābhivr̥ddhaye yatra vaṇikamanoratha pautrasya boṭaputrasya narasiṅghasya prasādīkr̥tas tad anenāpy asmadanumatyā bhagavataḥśrījayeśvarabhaṭṭārakasya pratipāditaḥk

prasādīkr̥tas seems to imply that the village was actually held in fief by Narasi˙ngha, who therefore needed the permission of Sūrabala for the donation (cf. Sircar in ´ Indian Epigraphical Glossary ṣv. pasāita).

10 JESI III, 189 (text corrected):

aparimitaguṇasamudayasya bhagavatas triśūlapāṇer varavr̥ṣabhavāhasya sphuradbhu jagaparikarasyānaṅgāṅgavidhvaṁsinaḥtridaśapatinutacaraṇābjayugalasya śrījaye- śvarabhaṭṭārakasyedaṁśāsanaṁlikhyate śryudīrṇavairānumatyā k

couchant bull (Nandin)’. 11 The findspot of the inscription, Malhar, 12 suggests that the mentioned dakṣiṇarāṣṭra actually bordered on the northern parts of Kosala, which may have come under the sway of the house of Mekalā thanks to the matrimonial alliance contracted by Bharatabala. It would seem that in the 6th century the course of the rivers Sevnāth and Mahānadī running west to east was considered to be the border between Mekalā and Kosalā (see Figure

2).

The characters of Sūrabala’s inscriptions, showing a mixture of square- ´

headed (especially the letter ma) and triangular-headed letters, are compared with the grants of Prabhāvatī Guptā, 13 but Sitaraman & Sharma (JESI III, 184) rightly observe that they may rather be compared with those of the Mal har Plates of Vyāghrarāja and the Malgā Plates of Sāmanta Indrarāja. 14 In order to specify the dynasty’s links with Kosala we shall have a closer look at the evidence regarding a family (kula) descending (ja) from Amara.

The family descending from Amara in Kosala

This family is known from an inscription of Vyāghrarāja, which was also found in Malhar. 15 The characters of the inscription are, in the words of d. C. Sircar, ‘nail-headed and the triangular mark forming the top of the letters is of the linear or hollow type; [. . . ] (they) may be assigned to the 6th century’ (EI XXXIV, 45).

11 JESI III, 184. The seal is damaged; the upper half contains the image of a bull, ‘the lower half bears a legend in one line consisting of four letters which may be read as śrīpuruṣah.’.

12 As reported above, n. 6 on p. 238, another incomplete inscription of this dynasty is said to have been found in Būrhīkhar, adjacent to Malhar; it is in the possession of an inhabitant of Malhar (Bajpai & Pandey 1977, 23). In their excavation report these authors argue in favour of the identity of Malhar and Sarabhapur, and in my view they have established ´ that of all possible candidates Malhar holds the best claim of being the ancient capital of the Sarabhapurīya kings, by virtue of its being the major findspot of inscriptions and ´ its possession of a wealth of archaeological material, only comparable to that of Sirpur (op. cit. 26 ff.).

13 Chhabra in EI XXVII, 132. Mirashi (CII V, 82) remarks about the inscription of the 2nd regnal year: ‘They resemble in some respects those of the Poona Plates of Prabhāvatīguptā, which also are of the same nail-headed variety, but unlike the latter, they do not contain any admixture of the northern letters.’

14 Bosma 2018, 22 ff. argues that Bharatabala and the sāmanta Indrabala are one and the same persoṇI consider this a plausible hypothesis. It would make Sūrabala the fourth, ´ missing, son of Indrabala. That Indrabala had a fourth son, in addition to Nannarāja, Īśānadeva and Bhavadeva, seems to follow from the Mālinī verse 19 of the Ara˙ng Stone ¯ Inscription of Bhavadeva Raṇakesarin and Nannarāja (JRAS 1905, 626; Shastri 1995 II, 97), which is only partly leggible:

suvihitavr̥ṣa ˘ ˘ , ˘˘˘˘˘˘ ˘ ˘ |

˘˘˘˘˘˘ ˘ lokopakārī, bhava iva bhavadevas tasya putras turīyaḥk 19 k 15 Mallar Plates of Vyaghraraja (EI XXXIV, 45–50).

12 / History and Culture of Dakṣiṇa Kosala 241

Vyāghrarāja is the younger brother of a chieftain (manujapati) Pravarabha ṭṭāraka, a son of Jayabhaṭṭāraka, who is without title except for the honorific śrī. The family to which they belong is said to be that of Amarārya, 16 on which Sircar passes the following remark,

The expression Amar-ārya looks like a Brahmanical personal name and names of the same type are often met with in South Indian records. It may be supposed that Jayarāja’s [i.e. Jayabhaṭṭāraka’s, ḤṬB.] queen who gave birth to Pravara I [i.e. Pravarabhaṭṭāraka, ḤṬB] and Vyāghra was the daughter of a person named Amar-ārya or was born in a family of which a person of that name was believed to have been the progenitor, since South Indian rulers sometimes represented them selves as belonging to the family from which their mother sprang. (EI XXXIV, 48)

If Sircar is right, the Amarārya family was partly of South Indian origin; ap parently through affiliation with a family of Kosala, they came to belong to the aristocratic circles of that country. In this way they attained to the status of local chieftains in the ‘umbrageous’ residence Prasannapur on the bank of the River Niḍilā, ‘the waters of which were divided through being struck by the round hips of the concubines of Pravara (/ of distinguished lovely women)’. 17 Town and river are as yet not been identified. 18

The charter of Vyāghrarāja, provided with his own seal, 19 reading śrī vyāghrarājaḥ, is stylistically very different from the inscriptions of the so-called Sarabhapurīyas. Despite these differences Sircar argued in favour of an identity ´

16 amarāryakulāmbaraśaśinah. [. . . ] śrīpravarabhaṭṭārakasya. In view of the comparatively few pretentions voiced in this charter, it seems rather unlikely that a ‘divine’ (amara) origin of the family was envisaged.

17 EI XXXIV, 49: prasannapurād upavanavanarājirājitād amarapurakīrtivijayinaḥprava rakāminīnitambabimbābhighātabhinnāmbasā ca srotasvatyā niḍilayā pavitrīkr̥tād [. . . ] 18 Bajpai & Pandey 1977, 28 identify the Niḍilā river with ‘Līlāgar and the Pūrva-Rāṣṭra was the region across the left bank of this river’. I fail to see the specific connection between Niḍilā (Skt. Niṭala/niṭāla, ‘forehead’ ?) and Līlāgar. (‘play-ditch’ ?). Another possibility would be the site of Kosalā at the Kaṁjī Nadī (Brown River) 18 km eastwards (below, n. 23 on p. 242). ‘According to Sircar and Bhattacharya, it is not improbable that Prasannapura was situated in the neighbourhood of Srīpura’ (Shastri 1995 II, 67). ´ Mahajan 2000, 92 identifies Prasannapura with ‘Parasvani, Mahasamund tehsil, Raipur district’.

19 Sircar & Bhattacharya in EI XXXIV, 45:

The surface of the seal, which is 1.9 inches in diameter and is much corroded, has a thick line dividing it into two halves. The section above the line exhibits three symbols, viz. the side view of a cakra in the left, the head of an animal (probably a lion) to front in the middle, and a conch-shell in the right. The legend below the line, written in Southern characters similar to those employed in writing the text of the document on the plates, reads śrīvyāghrarājah.. There is another symbol below the legend, which is difficult to identify, though it may be the head of an elephant to front. It will be seen that Vyāghrarāja’s seal is totally unlike the seal of the Sarabhapurīya ´ kings [. . . ].

Shastri (JESI IX, 40) sees a ‘bird (probably garuḍa)’ instead of a lion, and a lotus flower instead of an elephant.

of the two dynasties. Ajay Mitra Shastri (1987, 179 ff.), on the other hand, who convincingly refuted Sircar’s hypothesis, thinks the Amarārya family to be one of independent rulers ‘contemporaneously with the later kings of the Sarabhapur family in a part of Kosala ´ not included in the dominions of the Sarabhapurīyas ´ ’. 20 The last clause, however, is unlikely to be true. There is little in Vyāghrarāja’s inscription that points to a sovereign royal dynasty. Admittedly, the charter does not refer explicitly to an overlord, but we should keep in mind that it was not issued by the manujapati himself, of whom we do not possess any inscription, whereas Vyāghrarāja may have thought it sufficient to mention only his direct superior, his elder brother. Anyhow, they must have ruled very near to the kings of Sarabhapur, whose charters are also ´ found in Malhar.

The name of the residence, Prasannapur, seems to indicate that this town be longed to the dominions of the Sarabhapurīya king Prasannamātra, an issuer of ´ repouss´e gold coins, or his successor Jayarāja. Shastri himself rightly remarks, ‘that Prasannamātra was a powerful ruler [who] had thrown off the yoke of the Guptas and started ruling as a sovereign monarch for all intents and purposes’ (Shastri 1987, 183). Consequently it does not seem plausible that Prasanna mātra or his successor would have tolerated an independent ruler in the close vicinity of his own capital. It may even cautiously be suggested that the ‘right honourable royal officers of the king’, 21 in whose presence Vyāghrarāja’s grant was declared, and the year 41 of the ‘reign in which victory ever increases’ 22 should be understood as referring to the ruling king of Sarabhapur. ´

Whatever the exact relationship of the Amarāryakula with the rulers of Sarabhapur might have been, the marriage of the king of Mekalā, Bharatabala, ´ with a princess of the former family seems to have furnished the hill people of the Maikala Range with a foothold in the fertile plains of Kosala, of which they were proud enough to have it explicitly mentioned in their charters. 23

20 Shastri 1987, 181; italics mine.

21 EI XXXIV, 49 ll. 9–10: rājñaḥsumānyarājapuruṣān.

22 EI XXXIV, 50: iti pravardhamānavijayarājyasaṁvat 40 1 pauṣa-di 20 [7]. 23 Sanskrit verses (Mālinī and Sragdharā) of the Bamhanī Plates praising the princess of the family descending from Amara, according to Mirashi’s edition in CII V, 85 f. ll. 25–31: ekaiva [i.e. Queen Lokaprakāśā],

sphaṭikavimalaśubhraṁbibhratī śīlatoyam. ,

yamaniyamataṭāntaprāntaśuddhapravāham |

praśamaguṇagaṇormir yā janaṁpāvayantī,

svayam iha suralokād āgatā jāhnavīva k 9 k

śrīmaccāndrāṁśukīrter bharatabalanr̥pasyottamā rājapatnī,

jātā yā kosalāyām amarajakulajāṁkīrtim uccair dadhānā |

śaśvaddharmārthakāmaprativihitatamātīva lokaprakāśā,

yātā pautraiḥprapautrair nayavinayaratai rājasiṁhaiḥpratiṣṭhām k 10 k

In 10b I read ◦kulajām. : ◦kulajā Mirashi (the anusvāra is clearly visible in the Bamhanī Plate). In 10c I read with Mirashi ◦prativihita-tamˆatīva (double sandhi ṁc.); an alter native would be to read a superlative: prativihitatamā atīva, ‘deeply devoted (?) to a high degree’ (cf. Shastri 1995 II, 75). Translation:

12 / History and Culture of Dakṣiṇa Kosala 243

However, it did not yet bring them sovereignty over Kosala, which for the greater part remained in the hands of the rulers of Sarabhapur for at least one ´ more generation. 24

The rulers of Sarabhapur ´

The kings who issued their charters from Sarabhapur, generally designated ´ as Sarabhapurīyas in want of any other dynastic nomination, may be divided ´ into two groups, which are linked by similarity in the style of their seals and inscriptions as well as by their capital. 25 Kings of both groups call themselves paramabhāgavatas, i.e. worshippers of Viṣṇu. To the earlier group belongs mahārāja Narendra, son of Sarabha, of whom we possess three epigraphical ´ testimonies, written in 5th-century square-headed characters which we think very similar to the ones used in the charters of the Vākāṭaka kings. 26 Also to be included in this group is a king called Mahendra or Mahendrāditya, who is known from repouss´e gold coins of the Khairtal Hoard, 27 and, possibly, from a

Peerless, the chief queen of King Bharatabala, whose fame resembles the illustrious rays of the moon—she, being as it were the river Ga˙ngā herself descended to earth from the world of the gods, who purifies the people, carrying the waters of her good character, bright and spotless like cristal, her pure stream contained within the two banks formed by self-control and good conduct, and her waves being a mass of virtues and tranquility—she, being born in Kosalā and holding high the fame of the family descending from Amara, and being very much the ‘Lustre of the World’ (Lokaprakāśā), by whom darkness has been counteracted by means of eternal dharma, artha and kāma, she has attained preeminence (owing to) her grandsons and great-grandsons, princes like lions, devoted to good policy and conduct.

Cf. EI XXVII, 141; JESI III, 190 f.

24 Bajpai & Pandey 1977, 23 f. seem to argue in a quite confusing and incoherent way that Sarabhapur was the capital of the Pān ´ . ḍava dynasty of Mekalā, in spite of the fact that the name of that town does not feature in their inscriptions.

25 On Sarabhapur see above, n. 12 on p. 240. ´

26 Pīparḍūlā Copper-Plate Inscription of King Narendra of Sarabhapur ´ (IHQ XIX (1943), 139–146), dating from regnal year 3; Kurud Plates of Narendra, year 24 (EI XXXI, 263–268); Rawan Plates of Maharaja Narendra (JESI VI (1979), 44 f.). Sircar (SI I, 488) describes the chararacters of the Pīparḍūlā Plates as ‘box-headed variety of the Central Indian Alphabet of the 5th or 6th century A.D’. In IHQ XIX, 140 the same author remarked ‘Medial ī is of the older type and is not represented by a dot in the circle (which indicates medial i) as the Araṅg grant of Jayarāja and the T ¯ . hākurdiyā grant of Pravararāja.’

27 V. P. Rode in JNSI X, 137–142:

They are all single-die coins with the device and legend embossed on the obverse. The reverse is blank. They measure .78 to .87 inches in diameter and weigh from 19 to 20.2 grains each. They bear on the obverse, inside the circle of dots along the edge, Garuḍa standing on a horizontal line with wings spread out. To his proper right are the Crescent-Moon and a Chakra encircled by dots and to his proper left the Sun symbol and Saṅkha. Below the line is the legend “ ´ Srī Mahendrāditya” in the box- ´

sealing found in the excavation in Malhar. 28

In his Kurud Plates Narendra reconfirms an earlier grant of the paramabha ṭṭāraka and the ensuing merit is also assigned to the latter. Sircar (EI XXXI, 267) argues convincingly that the respectful manner in which the paramabha ṭṭāraka is mentioned shows that Narendra ‘still considered himself, howsoever nominally, a subordinate of the Imperial Guptas’. Another argument in favour of an allegiance of Sarabha and Narendra to the Imperial Guptas pursued by ´ A. ṀShastri, who follows Sircar in this respect, has been contested by Mirashi in his later work. According to Sircar, Shastri and others, Narendra’s father Sarabha is ‘identical with the maternal grandfather of Goparāja who died in ´ fighting on behalf of the Gupta monarch Bhānugupta at Eran [. . . ] in the Gupta year 191 (= ad 510)’. 29 In his Indological Research Papers I (1982, 187 f.) Mirashi gives up his earlier view and argues against this identification, though, in our view, he missed the most plausible argument, which was formulated by J. Williams as follows: ‘Sarabha cannot be the maternal grandfather of Goparāja ´ of Eran, called śarabharāja-dauhittrah. . That term is used only when a male heir is lacking, and Sarabha of Kosala had a son’ (scil. Narendra). ´ 30

When the ‘Goparāja argument’ has thus proven to be invalid, there seems to be no longer any reason not to date King Sarabha around the middle of ´ the 5th century, as one is inclined to do on the basis of the palaeography of the inscriptions of his soṇThe paramabhaṭṭāraka to whom Narendra refers might therefore have been Kumāragupta I himself, whose biruda and coinage the kings of Kosala, in particular Mahendra/Mahendrāditya, may have sought to imitate. 31 During the second half of the 5th century, when Mahendra and Narendra ruled—Narendra at least for 24 years—Gupta power was certainly on the wane, a fact of which the Vākāṭaka kings Hariṣeṇa and Narendrasena may have temporarily taken advantage, to judge by their claims of suzerainty over Kosala/Kosalā and Mekalā. 32 However, a disruption of the rule of Sarabhapur ´

headed characters of the 5th–6th century A.d. Below the legend are a cluster of seven dots and a letter ‘ru’ (?); in one case the letter looks like u or d. [. . . ] The legend on these coins is in box-headed characters which resemble those of the Pīparḍūlā grant of King Narendra of Sarabhapur family.’ ( ´ op. cit. 137 f.)

Another hoard found in Bhandara contains one coin of Srī-Mahendrāditya and eleven ´ coins of Srī-Prasannamātra (JNSI XVI, 216). ´

28 Bajpai & Pandey 1977, 24. The excavators ‘feel inclined to identify the Mahendra of the sealing with his namesake mentioned in the Allahabad pillar inscription of Samu dragupta’ (kausalakamahendra◦, CII III (1888), 7 l. 19). If this were correct, the owner of this seal can not have been the issuer of the gold coins.

29 Sircar in EI XXXI, 267; cf. CII III (1888), 91–93; Shastri 1987, 175; Mirashi in EI XXII, 17.

30 Williams 1982, 25; cf. Kane III, 715 f.

31 Sircar EI XXXI, 268; cf. Altekar’s note at Rode’s publication of the Khairtal Hoard (JNSI X, 142): ‘All things considered I am at present inclined to think that the present coins were not issued by any Gupta emperor, but were the issues of some ruler in Mahākośala, who had adopted the biruda of Mahendrāditya. It is not improbable that he may have borrowed the device from the Gupta coinage.’

32 CII V, 108 l. 14; CII V, 81 ll. 27 f.; VSMA 1971/72, 75 l. 22.

12 / History and Culture of Dakṣiṇa Kosala 245

did not come from the side of the Vākāṭakas, as it would seem, but may have been brought about by their enemies, the Nala kings of Puṣkarī from the Andhra country to the southeast of Kosala, when they overran the kingdom of the Vākāṭakas in the last decade of the 5th century. 33

In fact the end of the 5th century marked the end of an era; when Prasanna mātra succeeded in reorganizing Sarabhapur’s power and started issuing gold ´ coins in his own name, he might have done so temporarily from another resi dence, viz. Prasannapur, which may have afterwards been turned over to the Amarārya family. The beginnings of the 6th century may have also offered enough freedom to the kings of Mekalā to call themselves mahārāja, beginning with Nāgabala. The new ´elan of the Sarabhapurīya rulers clearly speaks from ´ the first Malhar Plates of Prasannamātra’s son Jayarāja, which dates from his 5th regnal year. Grafted on earlier formal phrases, the epigraphs contain from this time onwards high flown eulogies, not yet heard in Kosala and Mekalā, such as,

The illustrious great Jayarāja, giver of riches, land and cattle, who made the women of his enemies tear out their parted hair, whose feet are washed by the sprinkling water that is the brilliance of the crest-jewels of his feudatories who have been brought into submission by his prowess, great devotee of the Lord, favoured by his venerable father and mother. 34

The same ´elan gave rise to the construction of stone temples, the oldest of which may date back to the first half of the 6th century. In the remains of some of them Vākāṭaka influences are still distinct, as in, for instance, the door-frame and some pillars preserved in the (later) Rāmacandra Temple in Rajim or almost identical pillars found in Turturiya. 35

33 Bakker & Isaacson 1993, 68 (above, n. 67 on p. 142); Bakker 1997, 53–57. This would explain from where the Nalas got the idea of issuing their own gold coins, which were devised after those of Mahendrāditya (Mirashi in JNSI XI, 109 f.).

34 EI XXXIII, 157 ll. 1–2:

vikramopanatasāmantacūḍāmaṇiprabhāprasekāmbubhir dhautapādayugalo ripuvilāsi nīsīmantoddharaṇahetur vasuvasudhāgopradaḥparamabhāgavato mātāpitr̥pādānu dhyātaḥśrīmahājayarājah.. 35 Stadtner 1980, 47. Krishna Deva in EITA II.1, 225: ‘Of two early pillars in the Rāmacandra Temple, one is Rucaka, the other octagonal with a circular fluted laśuna and crowning members. Similar pillars are known from Turturiā. These Mahākōsala art forms, which reveal the diffused impact of Vākāṭaka prototypes, can be dated c. ad 525–550.’ For photographs of the pillars in the Rāmacandra Temple see EITA II.1. nos. 439–443; for those in Turturiya see ASI XIII (1882), Plate no. XVII. For another pillar, found at the entrance of the temple at Turturiya, see below, Plate 17. Krishna Deva’s description of the Rucaka pillar to the effect that ‘the lower two-fifths [. . . ] is plain; the upper three-fifths is carved in three zones’ is probably based on the photo graph II.D 2. No. 443, since actually the lower two-fifths consists of two panels in low relief (below, Plate 16). Other (?) pillars of this temple are reported by Cunningham to have been brought here ‘about 250 years ago (some say 400) by Govind Sˆah, Kamˆasdˆar of Raypur’ from the Lakṣmaṇa Temple in Sirpur ‘in boats’ (ASI XVII, 28). See also Viennot 1958, 138–140; Dikshit 1960, 31 f.

Plate 16 Plate 17

Rucaka pillar (north side) Pillar at the entrance of

Rāmacandra Temple in Rajim temple in Turturiya

After the collapse of the Vākāṭaka empire artisans may have travelled and helped to create new styles to the west (e.g. Jogeśvarī, Elephanta) 36 and to

36 EITA II.1, 87 ff.

12 / History and Culture of Dakṣiṇa Kosala 247

the east (e.g. Kosala) of the old culture area. However, there seems to have been enough self-confidence to allow for revolutionary new experiments of an astonishing richness and beauty, such as are witnessed in the temples in Tālā (see below). They lay the foundation of the emerging Dakṣiṇa Kosala style of sculpture and architecture of the second half of the 6th and 7th centuries.

Jayarāja was, in all likelihood, succeeded by his nephew Sudevarāja, who soon inaugurated a second centre from where the kingdom was governed, a place named after the goddess who adorned the seals of the dynasty, Srī (Laks ´ .mī), viz. Srīpura (modern Sirpur) on the right bank of the Great River (Mahānadī). ´ The residence of the king remained in Sarabhapur, as is evident from the fact ´ that the charters issued from there were direct orders of the king himself (sva mukhājñayā), 37 whereas the charters from Srīpura were issued by his proxy ( ´ dū taka), the illustrious great feudatory (śrīmahāsāmanta) Indrabalarāja, whose office is described as that of chief minister (sarvādhikārādhikr̥ta). 38

There can be little doubt that Indrabala is the same as the king of the Pāṇḍava lineage from whom the later King Tīvaradeva traced his descent (see below). 39 The ending of the name ◦bala as well as the fact that he belonged to a Pāṇḍava lineage are strong indications that Indrabala was somehow related to the Somavaṁśa dynasty of Mekalā, whose last ruling king, Sūrabala, was ´ born from a princess of Kosalā and who probably ruled as a feudatory of the Sarabhapurīyas over parts of southern Mekalā and/or northern Kosala at the ´ times of Jayarāja and/or Sudevarāja. 40

Sudevarāja’s successor, Pravararāja—probably his younger brother—may have had a rather short reigṇBoth of his charters date from his 3rd regnal year. 41 They were issued from Srīpura, this time, however, without media- ´ tion of Indrabala or another sāmanta. It appears not implausible that the death of Sudevarāja was followed by a power struggle between the old family of Sarabhapur and the feudatories of the newly founded ´ Srīpura. Within the ´ family of the Pāṇḍuvaṁśins of Mekalā this may also have led to a conflict of interests, in particular between Sūrabala, reigning in Mekalā, and Nannarāja, ´ son of Indrabala, reigning in Kosala; according to Bosma’s hypothesis they were brothers (above, n. 14 on p. 240; see Figure 3). The outcome of this struggle we know: the Indrabala–Nannarāja–Tīvaradeva/Candragupta lineage of the Pāṇḍavas of Kosala emerged as the new monarchs of Srīpura and under their ´ rule this place evolved into an imposing temple town.

37 EI XXXI, 108 l. 24; EI XXIII, 22 ll. 25 f.

38 EI XXXI, 316 l. 23 f.; JESI V, 96 l. 26 (the letters -ś cātra dūtakaḥseem to have been forgotten by the engraver).

39 CII III (1888), 295 l. 16; EI XXXIV, 115 l. 16; EI VII, 104 l. 18.

40 The identity of Bharatabala and Indrabala has been argued by Natasja Bosma 2018 (above, n. 14 on p. 240).

41 EI XXII, 15–23; EI XXXIV, 51–52.

The Paṉ .ḍuvaṁśa of Sr´ īpura

As has been noted above, scholars nowadays generally agree that sāmanta Indrabalarāja, who ruled over Srīpura during Sudevarāja’s reign, was in some ´ way related to the Pāṇḍavas of Mekalā, but members of this lineage are never mentioned in the inscriptions of the later Pāṇḍavas of Kosala (cf. Shastri 1987, 204). The latter’s ancestor is said to have been Udayana, who is mentioned in the so-called Buddhist Inscription of Bhavadeva Ranakesarin (which is probably to be ascribed to Nannarāja I) and in an inscription of the time of Mahāśiva gupta. 42 This Udayana may be identical to the Pāṇḍava Udayana featuring in the Kālañjara inscription of the 8th or 9th century, in which he is said to have founded there a beautiful brick temple of Bhadreśvara. 43 If indeed these two Udayanas were the same, the Kālañjara inscription might indicate that the original homeland of this branch may have been the region around Rewa, but this issue has been the subject of discussion in my later work. 44

42 Ara˙ng Stone Inscription of Bhavadeva Raṅakesarin and Nannarāja (JRAS 1905, 626; Shastri 1995 II, 97):

gacchati bhūyasi kāle bhūmipatiḥkṣapitasakalaripupakṣaḥ|

pāṇḍavavaṁśād guṇavān udayananāmā samutpannaḥk 16 k

… ◦sya tanūjanmā k 17 k

The exact relationship between this Udayana (in verse 16) and subsequent kings is uncertain, since the Aryā verse 17 is lost, apart from the last five syllables featuring the ¯ word ‘son’. The actual findspot of this inscription is much disputed (see e.g. Mirashi in EI XXIII, 116 f. and Notes of Mirashi and Sircar in EI XXXIII, 251–256). Shastri 1987, 229 n. 88, following Mirashi, thinks that ‘the inscribed slab hails from Arang’. Actually, there might be another inscription of Bhavadeva/Nannarāja I in the Mahāmāyi Temple in Arang (see Hiralal 1932, 110 (No. 183); Shastri 1995 I, 136).

Sirpur Stone Inscription of Sivagupta Bālārjuna ´ (in the Gandheśvara Temple) IA XVIII (1889), 180 l. 2; Shastri 1995 II, 150:

āsīd udayano nāma nr̥paḥśaśadharānvitaḥ|

abhūd balabhidā tulyas tasmād indrabalo balī k 2 k

For a photo-zincograph based on a rubbing see ASI XVII, Plate XVIII A.

43 ASI XXI, 40 Plate IX L; quoted in JRAS (1905), 621. This inscription has been reex amined by Peter Bisschop and Hans Bakker (see Bakker 2014, 201):

udayana iti rājā yaḥkule pāṇḍavānām. ,

sakalabhuvananāthasyāsya bhadreśvarasya |

pavanalulitacihnaṁramyakāntīṣṭakābhir,

gr̥havaram atibhaktyā kāritaṁtena pūrvam k

44 I still think that both Udayana’s are the same figure, but I have given up the view that conceives of this king as a historical persoṇHe features in these pedigrees to forge a link with Pāṇḍu and the legendary Somavaṁśa dynasty. I wrote the following in Bakker 2014, 203:

This suggestion [of an historic Udayana] now seems to me naive. The Kālañjara inscription and the two Pāṇḍuvaṁśin inscriptions from Kosala refer to one and the same legendary Pāṇḍava king Udayana of Vatsa, son of Satānīka, descendant of Bhara- ´ ta through Pāṇḍu, Arjuna, Abhimanyu, Parīkṣit, Janamejaya [. . . ] and Satānīka II. ´

Telling in this respect is that in the inscriptions of Sūrabala, which give the pedigree of ´ the Pāṇḍavas of Mekalā, the second king of the lineage Vatsarāja is only compared to the famous king of Vatsa, i.e. Udayana (Shasti 1995 II, 74, 81 (v. 2).

12 / History and Culture of Dakṣiṇa Kosala 249

It is conceivable, nevertheless, that the territory over which King Indra(bala) and the Pāṇḍuvaṁśins of Mekalā ruled may have included the region of the present-day Shahdol District. This rule is apparently attested by an inscription found in Malgā (Shahdol District) of a sāmanta Indrarāja, dating from his first year, issued from Maṇḍaka (EI XXXIII, 209–214). The father of this sāmanta is only designated as ‘king’ (kṣitipati). About this inscription A. ṀShastri 1987, 201 observed:

As we have pointed out, in respect of general appearance, formal features, draft ing, the list of addresses of the royal order concerning the grant, the privileges bestowed on the grantee, the mode of recording the date as well as other matters, the Malgā Plates and the extant records of the Pāṇḍavas of Mekalā resemble each other very closely and, what is still more instructive, the scribes of these records were related to another as brothers [scil. Droṇāka (Malgā Plates) and Mihiraka (Bamhanī Plates), both sons of the goldsmith Īśvara, ḤṬB.].

If the hypotheses of the identity of, on the one hand, samānta Indrarāja and mahāsamānta Indrabala, and of Indrabala and Bharatabala, on the other (above, n. 14 on p. 240), are correct, the following career could be envisaged: the young sāmanta Indrarāja of the Malgā Plates, feudatory chief in (northern) Mekalā, through his marriage with Lokaprakāśā, a princess from the house of Amara, feudatory chiefs in Kosalā (Prasannapur), worked himself up to the position of chief minister (sarvādhikārādhikr̥ta) of the dominant regional ruler of his days, Sudevarāja of Sarabhapur, who entrusted him with the administra- ´ tion of the newly founded Srīpura. ´ 45 The mahāsāmanta of Srīpura had himself ´ called ‘Indrabalarāja’ in the Dhamatari Plates of Sudevarāja, Year 3 and in the Kauvatal Plates of Sudevarāja, Year 7.46

However, the Pāṇḍava family settling in Srīpura soon proved to be a cuckoo ´ in the nest, since either Indrabala himself, or his eldest son, Nannarāja I, usurped the throne of Kosala. 47 That the Pāṇḍuvaṁśa did not completely

45 Cf. Mirashi in Studies I, 240.

46 JESI V, 96 l. 26; Shastri 1995 II, 33; EI XXXI, 316 l. 23; Shastri 1995 II, 44.

47 In the as yet not published stone inscription in the Lakhneśvara (Lakṣmaṇeśvara) Tem ple at Kharod of his son Īśānadeva (see below, n. 51 on p. 250), Indrabala is described as sovereign monarch. Thanks to the good services of Natasja Bosma, who took pho tos of this inscription, I could decipher the following in lines 22–23 (four pādas of a Vasantatilaka and the first pāda of a Sragdharā):

sarvāvanīśvaraśiromaṇirājirājat-,

pādāmbujaḥśaśikulāmbarapūrṇacandraḥ|

āsīd bhuvaḥpatir avārita ī(ya) [22] ˘ ,

(a)pāstaripur indrabalābhidhānaḥk

śrīmān īśānadevaḥkṣitipatitilakas tasya putrottamo ’bhūd,

…

There was a lord of the earth named Indrabala, who had thrown off his enemies by . . . of his unimpeded . . . , who was the full moon in the sky that is the Lunar Dynasty, and whose lotus feet were shining due to a line of jewels in the crests of all the kings on earth (who prostrated before him). He had a most excellent son, an ornament of the kings, the illustrious Īśānadeva, . . .

give up their control over parts of the Vindhya (Mekalā) after they had seized sovereignty over Kosala ensues from the Senakapāt. Inscription of the time of Mahāśivagupta Bālārjuna, in which a certain brahmin Devarakṣita is said to have ‘obtained, apparently from King Nannarāja, the governorship of the Vindhya territory (vindhyadhūrdharatva)’. 48

Nannarāja I had three brothers, the youngest of whom, Bhavadeva (see above, n. 14 on p. 240), was glorified in the above-mentioned inscription, as he was responsible for the repair of an old Buddhist monastery (vihāra) originally built by one Sūryaghoṣa. 49 The inscription informs us that Bhavadeva was a great warrior, hence his birudas Raṇakesarin and Cintādurga (vv. 20, 32), but he does not seem to have been (sovereign) king (pr̥thivīpati) himself, if we read one of the verses in his praise correctly,

(Though) he has not taken (her) by the hand (/ has not taken tax), and with out walking around (the fire) (/ marching the surrounding (countries)), he, (al ready) being Lord of Lakṣmī, became an unparallelled supporter (/ husband) of the earth. 50

Another brother of Bhavadeva and Nannarāja I, Īśānadeva, is reported to have built a temple in Kharod. 51 The Pāṇḍavas of Kosala appear to have been great

Cf. Mirashi, Studies in Indology I (1960), 241 n. 1; Shastri 1995 II, 375 f.

Nannarāja had himself (probably) proclaimed rājādhirāja in the Ara˙ng Stone Inscription ¯ of Bhavadeva Raṇakesarin and Nannarāja (JRAS (1905), 629 v. 40; Shastri 1995 II, 99). The redrafting of the dynastic Pāṇḍava lineage, substituting Udayana for the Bala kings of Mekalā, gives the impression that the Pāṇḍava family, after the coup d’´etat in Srīpura, ´ no longer wished to be reminded of its tribal roots in Mekalā. A similar redrafting of a pedigree took place when the Maukharis came to power in Kanauj in about the same period (see Bakker 2014, 42).

48 EI XXXI, 32, 35 v. 7. The reading and interpretation of this verse are uncertain and disputed by Mirashi and Sircar in EI XXXIII, 251–256. Sanderson 2013, 237 f. accepts Mirashi’s reading (though not his interpretation): ‘. . . Devarakṣita, who had become a close confidant of King Nannarāja, had been appointed [by him] to govern the Vindhya region of the kingdom and had received various districts as the reward of his services.’ Cf. Bosma 2018, 34 ff., who argues that the Nannarāja mentioned may have been Nanna rāja I.

49 JRAS (1905), 628 vv. 35–36; Shastri 1995 II, 98.

50 JRAS (1905), 627 v. 22 (Shastri 1995 II, 97): karagraham akr̥tvāpi maṇḍalabhramaṇād vinā | apūrvo yaḥkṣiter bhartā jāto lakṣmīpatir bhavan k 22 k

51 Hiralal 1932, 125 (see also above, n. 47 on p. 249):

This inscription is in the Lakhneśvara Temple at Kharod. It was found plastered over. The pujārī endeavoured to take off the covering, but in doing so he damaged the record so badly, that it is now almost illegible. However, the names of the two kings, Indrabala and his son Īśānadeva of the Lunar Dynasty, are visible and there is mention of a village which appears to be Ghoṭapadrakagrāmaḥ, apparently granted for the maintenance of the temple.

Shastri 1995 II, 375 reads the name of the village as Sāṭṭhapadraka-grāma and adds: ‘It also refers to a place called Indrapura, evidently named after Indrabala, and mentions a district called Kośīra–Nandapura–viṣaya, obviously named after the twin localities of Kośīra and Nandapur, and a number of villages including Mekalapadraka, which was apparently named after the Mekalā country.’ A donation and repairs were (later) made

12 / History and Culture of Dakṣiṇa Kosala 251

temple builders. Many of the grants of this royal family concern the erection and maintenance of temples; Krishna Deva was right when he observed that ‘if the Pāṇḍuvaṁśīs did not found Srīpura, to them is mainly due its abundant ´ archaeological wealth’. 52

Nannarāja I was succeeded by his son Tīvaradeva, who acquired the sove reignty of the whole of Kosala. 53 Connected with Tīvaradeva is the conundrum of the absolute chronology of the history of ancient Kosala. Almost everything written about this king concerns his dating. 54

The date of the Pāṇḍuvaṁśa dynasty of Srīpura ´

Two dating criteria are thought to be relevant: 1) his capital (obviously Srī- ´ pura, but possibly also called ‘Trivaranagara’) might have been alluded to in an inscription of Mādhavavarman, the Viṣṇukuṇḍin king; 2) he might have been a contemporary of the (Maukhari) prince Sūryavarman, whose daughter Vāsaṭā was wedded to his nephew Harṣagupta (EI XI, 191 vv. 15–16).

With regard to the first criterium A. ṀShastri has convincingly argued, on the strength of the Indrapālanagara (Tummalagudem) charter of the Viṣṇu kuṇḍin king Vikramendravarman II, which is dated in the expired Saka year 488 ´ (= ad 566), that the latter’s great-grandfather, Mādhavavarman II Janāśraya, must have ruled between the termini post and ad quem of ad 470 and 528 (Shastri 1987, 122–131). This Mādhavavarman II, son of Govindavarman and married to a Vākāṭaka princess, is said in his Ipur Plates to have ‘rejoiced the hearts of the young ladies in the palaces of the town of Trivara’. 55 In view of the highly improbable, very early date that has to be assumed for Tīvaradeva in

by Mahāśivagupta (EI XXVII, 324) to a temple of Īśāneśvarabhaṭṭāraka in the township (pattana) Khadirapadra. The identification of this place is uncertain, but the identity of the temple in Kharod called today ‘Lakhneśvara’ with this Īśāneśvara Temple seems likely.

52 EITA II.1, 224. Cf. Beglar in ASI VII, 168–193; Cunningham in ASI XVII, 23–31; Srīvāstava 1984. Many of the temples of South Kosala have been constructed and ´ reconstructed during several generations. Among the oldest remains (apart from Tālā, see below) Krishna Deva reckons the Rājīvalocana Temple at Rajim, which ‘should be at least a generation earlier than the Lakṣmaṇa Temple at Sirpur and is assignable on ground of style to c. ad 600 [. . . ]’ (EITA II.1, 231 f.; cf. Meister 1984a, 121). In Sirpur itself the oldest surviving construction appears to be the Buddhist vihāra (see IAR 1954–55, 24–26; IAR 1955–56, 26 f.; EITA II.1, 232 f.). The so-called Lakṣmaṇa Temple at Sirpur was, according to a large stone inscription recovered from it, built by the queen mother of Mahāśivagupta in memory of her deceased husband Harṣagupta. Krishna Deva observes: ‘Paleographically, this epigraph belongs to c. ad 625–650 and the temple is also assignable to the same period.’ (EITA II.1, 235 f.; cf. Stadtner 1980, 39: ‘ca. A.D. 595–605’).

53 EI VII, 105 l. 19: prāptasakalakosalādhipatyah..

54 E.g., Mirashi in EI XXII, 19; EI XXVI, 229; Indological Research Papers I (1982), 179–190; Sircar in EI XXXIV, 112; Shastri 1987, 211–230.

55 EI XVII, 336 ll. 4–5: trivaranagarabhavanagatayuvatīhr̥dayanandanah.. Cf. the Pola muru Plates (Journal of the Depṭof Letters, University of Calcutta XI (1924), 59 ll. 8–9).

order to make his residence the scene of Mādhavavarman’s conquests, Shastri concludes that, whatever might have been meant by the expression, it had nothing to do with the Pāṇḍava king Tīvaradeva and his capital Srīpura. ´ 56

Regarding the second criterium, A. ṀShastri, following Sankaranarayanan 1977, argued that the father of Tīvaradeva’s nephew’s wife Vāsaṭā, Prince (nr̥pa) Sūryavarman, who is said to belong to a family of Varmans that had attained sovereignty over Magadhā, 57 has nothing to do with the homonymous Maukhari prince, son of Īśānavarman, who is known from the Haraha Inscrip tion (EI XIV, 110–120). Shastri’s arguments (1987, 215–217) fail to carry convictioṇThey had already all been cogently discussed by B. P. Sinha (1977, 116–119) and shown to be invalid. Shastri’s assertion to the effect that ‘history knows of several other dynasties like the Pallavas, all the members of which had names ending with varman; but that was never employed as a dynastic name’ 58 and that, consequently, no dynasty, including the Maukharis, was ever designated as ‘Varman’, is ipso facto refuted by the Sirpur inscription, unless we accept that only completely unknown ‘Varman’ families could be denoted varmaṇāṁkulam. Secondly, Shastri’s assessment that the Maukharis at the time of Īśānavarman were not in actual control of Magadhā is beside the point. They were certainly claiming to be sovereigns of this province and were about to realize it; epigraphical praśastis serve to promulgate this sort of claims, not to report truthfully about the often less rosy historical reality. Moreover, when Vāsaṭā/Mahāśivagupta made this claim, about half a century had passed since Īśānavarman.

Sinha’s conclusion that ‘ “the Varman dynasty famous for its supremacy over Magadha” on strong circumstantial evidence can be no other than the impe rial Maukhari dynasty which ruled over Magadha for many generations’ should be accepted as the most plausible interpretation of the evidence furnished by the Sirpur stone inscription (Sinha 1977, 118). Because the Haraha inscrip tion of Sūryavarman has been shown to date from ad 554, 59 we may safely conclude that his contemporary, Tīvaradeva, ruled in the middle of the 6th century, a conclusion that is fully consistent with the archaeological dating of the monuments in Sirpur that were built by Vāsaṭā and her son Mahāśivagupta Bālārjuna two generations later (ad 600–650). 60

56 Shastri 1975-76; Shastri 1987, 117–220. Cf. Sankaranarayanan 1977, 224–230: Therefore it is certain that Trivaranagara was the capital of the Viṣṇukuṇḍins. We have also seen earlier how both Mādhavavarmans of the Ipur (I) and Polamuru (I) plates brought prosperity to this capital Trivaranagara. We have identified this Trivaranagara with the modern Tiruvuru (17◦ 10´ North and 30◦ 35´ East), the headquarters of the taluk of that name in the Krishna District, Andhra Pradesh. (op. cit. 229)

57 EI XI, 191 v. 16: magadhādhipatya.

58 Shastri 1987, 216; Cf. Sankaranarayanan 1977, 227 f.

59 Haraha Inscription of the Reign of Īsānavarman [VS] 611, EI XIV, 113.

60 See above, n. 52 on p. 251.

12 / History and Culture of Dakṣiṇa Kosala 253

Tīvaradeva

To judge by his inscriptions, Tīvaradeva was a great warrior. In his military campaigns he may have been supported by the alliance which had evidently been forged between the Maukharis and the rulers of Dakṣiṇa Kosala.

There was enough geographical proximity for the Maukharis and Somavaṁśīs to come into contact. Sarvavarman [reigning brother of Sūryavarman, ḤṬB.] was ´ overlord of Kalanjar region and probably held the Vindhyan fort Asigarh; 61 and the Somavaṁśī kings of Mahākosala were associated with Mekala regioṇIt is likely that [the Maukhari king] Īśānavarman was helped by the Somavaṁśī king in his campaign against the Viṣṇukuṇḍins of Andhra. ¯ 62 Tīvaradeva who was master of the entire Mahākosala region must have facilitated Īśānavarman’s march into the interior of Orissa and Andhra. ¯ 63 (Sinha 1977, 118 f.)

Tīvaradeva may have actively participated in the exploits of his northern ally, since his successor, Nannarāja II, credits him with sovereignty over Orissa (Utkala) etc. 64

Irrespective of how much these campaigns (or raids) may have contributed to the wealth of Kosala, they did not result in a lasting expansion of the king dom. Nannarāja II, who in his own charter calls himself ‘son’, 65 but who might actually have been the son-in-law of Tīvaradeva, 66 only claims the sovereignty of the Kosalāmaṇḍala (EI XXXI, 221 ll. 8–9). This ‘modesty’ of Nannarāja II, also apparent from the comparison of his ‘father’ with Viṣṇu (Kaiṭabhāri), himself with Pradyumna (EI XXXI, 221 l. 7), and the fact that the charter does not feature a regnal year, may all point to the comparative insignificance of his reigṇWhether brought about by force or not, after him the line of succession switched again to the collateral branch. It must also remain uncer tain whether Candragupta, brother of Tīvaradeva, 67 and his son Harṣagupta, who was married to Vāsaṭā, ever really ruled, as long as we do not possess their inscriptions. Anyway, the exceptionally long reign (c. ad 590–650) of Harṣagupta’s son Mahāśivagupta (at least 57 years) indicates that he ascended the throne as a boy or young man. 68

61 Cf. CII III (1888), 219–221; HCI III, 69 n. 3.

62 Cf. Bakker 2014, 56.

63 Haraha Inscription (EI XIV, 117 v. 13). Bajpai & Pandey 1977, 22: ‘The ancient main route joining north India with south-eastern sea-coast passed from Kauśāmbī via Barhut through the present districts of Satna, Shahdol, Bilaspur and Raipur. [. . . ] The region of South Kosala had thus close contacts with Kali˙nga from very early times.’

64 EI XXXI, 221 l. 6: kosalotkalādimaṇḍalādhipatyaprāpta◦.

65 EI XXXI, 221 ll. 7–8: ātmaja.

66 EI VII, 105 ll. 25–26: priyajāmātr̥śrīnannarāja.

67 In the Sirpur Stone Inscription of the Time of Mahāśivagupta (i.e. Vāsaṭā’s inscription in the Lakṣmaṇa Temple), Tīvaradeva is only anonymously referred to by the words rājā dhikāradhavalaḥsabalo [. . . ] agrajo (sciḷcandraguptarājasya); EI XI, 190 v. 6 (Shastri 1995 II, 142).

68 On the basis of the Bonda Plates of Mahaśivagupta, year 22 (EI XXXV, 60–65) Mirashi and Pandeya argue that this ‘king came to the throne in 595 A.D’. The editor, d. C.

Mahāśivagupta Bālārjuna

Under Mahāśivagupta Srīpura evolved into ´

a centre of culture which deserves a mono

graph of its own in order to do it full jus

tice. 69 In the present context we must con

fine ourselves to only the most outstand

ing achievements. Queen Mother Vāsaṭā

built the splendid ‘Lakṣmaṇa Temple’: ‘She

caused this eternal abode of Lord Hari to

be constructed, so that it resembles ex

actly (the abode) in which he lives in per

petual adoration of the imperishable one

(i.e. Viṣṇu).’ 70 This temple was proba bly dedicated to Viṣṇu–Narasiṁha as the

first three verses of the inscription suggest.

Narasiṁha,

[. . . ] who, bending his eyes eagerly to his

nails—thinking that they had not (yet)

taken hold of the enemy to tear up—

caught sight of the demon that was hid

ing out of fear in the opening of the fis

sures (underneath) those (nails), and who

contemptuously bursting into laughter,

all of a sudden tore (him) up with the

Plate 18

Narasiṁha in Site Museum Lakṣmaṇa Temple Sirpur

nails of his other hand and cast him away in rage, as if (he were just) dirt under them (i.e. nails). 71

An image of Narasiṁha tearing the demon Hiraṇyakaśipu apart is preserved in the museum next to the temple (Plate 18).

Sircar, however, sees a flaw in the reasoning, as he usually does if Mirashi is involved (EI XXXV, 61 n. 7). Cf. Meister 1984a, 140 n. 6.

69 Maheśacandra Srīvāstava’s ´ Sirpur, Bhopal 1984, needs reinforcement in this respect. The work of Natasja Bosma (2018) fills this desideratum. The reports on his excavations in Sirpur by A.K. Sharma, published in Purāmanthana, are entirely unreliable, since this archaeologist abuses his exclusive rights in Sirpur to build an archaeological theme park.

70 EI XI, 192 v. 20: tayā nijaḥpretya patir yathāvidhe vasaty asau nityam upāsitācyutaḥ| prakāśitaṁtādr̥śam eva kāritaṁvibhor idaṁdhāma hareḥsanātanam k For a descrip tion of this temple see EITA II.1, 233–236. Cf. Krishna Deva 1960. 71 EI XI, 190 v. 2:

labdho nirbhettum ebhir na ripur iti rasād dattacakṣur nakheṣu,

trāsāt tatkroḍarandhrodarakuharadarīm eva līnaṁvilokya |

hāsollāsāvahelaṁtaditarakarajāgreṇa nirbhidya sadyaḥ,

krośāc cikṣepa tajjaṁmalam iva danujaṁyaḥsa vo ’vyān nr̥siṁhaḥk 2 k

12 / History and Culture of Dakṣiṇa Kosala 255

Mahāśivagupta, who, unlike his ancestors who were paramavaiṣṇavas, was a paramamāheśvara (EI XXVII, 323 l. 5), evidently pursued a policy of active patronage of various religious currents of his day, as is apparent from his grant to a Buddhist monastery in Taraḍaṁśaka (at the request of his maternal uncle Bhāskaravarman), 72 and, indirectly, from the Buddhist monasteries a little south of Srīpura. The inscription found in one of them records the donation ´ by a monk called Anandaprabha of a ‘free-feeding establishment for the ¯ yatis or monks’ under the reign of Mahāśivagupta. 73

Among the major constructions of this period, however, is the Gandheśvara Temple on the bank of the Mahānadī. According to Hiralal this temple was,

Repaired by the Bhonsalas who took all kinds of old material from the ruins of other temples and used it in making the mahāmaṇḍapa. We therefore find here a number of inscriptions, some of which do not really belong to this temple. [. . . ] Altogether there are six inscriptions, of which perhaps 2 or 3 only may be said to belong to Gandheśvara Temple. The one which specifically mentions the name Gandharveśvara is built into the plinth (on the right side as you enter). It records the arrangement made for the offerings of flowers for the pūjā of Gandh[arv]eśvara by one Jejuraka, a subject of prince Siva Gupta in whose kingdom pious people ´ lived. 74

The prospering capital of Kosala under the reign of Mahāśivagupta was visited by the Chinese pilgrim Xuanzang, who gave the following description,

The capital is about 40 li round; the soil is rich and fertile, and yields abundant crops. [. . . ] The population is very dense. The men are tall and black com plexioned. The disposition of the people is hard and violent; they are brave and impetuous. There are both heretics and believers here. They are earnest in study and of a high intelligence. The king is of the Kshattriya race; he greatly honours the law of Buddha, and his virtue and love are far renowned. There are about one hundred saṅghˆarˆamas, and somewhat less than 10,000 priests: they all alike study the teaching of the Great Vehicle. There are about seventy Deva temples, frequented by heretics of different persuasions. Not far to the south of the city is an old saṅghˆarˆama, by the side of which is a stūpa that was built by Aśˆokarˆaja. (Beal 1884 II, 209 f.)

Mahāśivagupta’s reign marks the acme of the early history of South Kosala. No inscriptions of his successors have come to light and we may conveniently end our survey at this point. It is meant to provide some sort of dynastic and chronological framework, in which the cultural history of South Kosala has

72 EI XXIII, 120 f. ll. 11–13.

73 EI XXXI, 197 n. 2.

74 Hiralal 1932, 98. Cf. Beglar in ASI VII, 168–170; Cunningham in ASI XVII, 23–31; Shastri 1995 II, 152 f., 160f., 382. One of these inscriptions has been edited in IA XVIII (1889), 179–181; two more have been published by A. ṀShastri in Vishveshvaranand Indological Journal Vol. XVII (1979), 196–202 (not available to me). For the others see ASI XVII, 25–27, Plates XVIII A, B and XIX C, D and XX E. Cf. Hiralal 1932, 97–99.

to fit. 75 To conclude we may select just one of its highlights for a tentative examination, the temples of Tālā.

Taḻ ā

Tālā is a tiny village in the Bilaspur District, situated a little north of the con fluence of the Maniārī and the Sevnāth rivers; the junction of the latter with the Mahānadī is c. 50 km downstream. The hamlet lies at a distance of 25 km to the west of Malhar as the crow flies. Near the village, on the left bank of the Maniārī are two stone temples known under the names Jiṭhānī and Devarānī. Bajpai suggested that one of these temples was the Jayeśvarabhaṭṭāraka Tem ple, which was endowed with the village at the confluence by one Narasiṅgha and King Sūrabala Udīrn ´ . avaira (see above, p. 239). 76

At the time of the grant this ‘confluence-hamlet’ (saṁgamagrāmaka) be longed to the southern province (dakṣiṇarāṣṭra) of the kingdom of the Pāṇḍavas of Mekalā. As we have argued above, this dynasty may have extended its ter ritory to the north of Kosala since the times of Sūrabala’s father, who had ´ married a princess of Kosalā. This princess, as has been shown, probably came from the Amarārya family, who were petty chiefs in Kosala. Their progenitor was Jayabhaṭṭāraka. In view of the common practice to name temples after their builders, we would venture the hypothesis, that this Jayeśvarabhaṭṭāraka Temple was founded by Jayabhaṭṭāraka of the Amarāryakula. From this it would follow that Sūrabala made a grant to the temple built by either his ma- ´ ternal grandfather or, less likely, his maternal uncle (depending on whether the princess ‘Lokaprakāśā’ was a daughter or (younger) sister of Jayabhaṭṭāraka). And this would place this temple around 500 ad, according to the dynastic history of Kosala that has been reconstructed above on the basis of epigraph ical testimony. Pending new epigraphical discoveries that throw fresh light on this issue, we shall now have a look at the archaeological evidence of the two temples themselves.

75 In addition to A.ṀShastri, Thomas E. Donaldson may be mentioned as an art-historian who has argued in favour of a much later historical framework for the Kosala temples (Hindu Temple Art of Orissa I (1985), 188 ff.). One of his main arguments is the dynastic gap or dark period of the region that follows the reign of Mahāśivagupta Bālārjuna. Apart from the fact that shifting the whole dynastic and art history with 200 years to fill this gap only results in another gap, beginning at the time that Vākāṭaka and Gupta hegenomy ended (c. 500 ad), it may be assumed that, after the Pāṇḍuvaṁśa, a re-emerged Nala dynasty ruled over parts of Kosala including Rajim, where Vilāsatuṅga reconstructed the Rājīvalocana Temple (EI XXVI, 49–58; cf. Krishna Deva in EITA II.1, 224 f.).

76 Bajpai & Pandey 1977,23: ‘I have identified the village saṁgamagrāma with the village Tālā (also called Saṅgama) near the confluence of the rivers Maniārī and Sivanātha in ´ the Bilaspur district.’ Cf. Risbud 1984, 60 f.

12 / History and Culture of Dakṣiṇa Kosala 257

Plate 19

Devarānī Temple at Tālā seen from the south

Plate 20

Jiṭhānī Temple at Tālā seen from the south

With regard to the relative chronology of the two temples, it would seem that the better preserved Devarānī is somewhat later than the Jiṭhānī Temple, by virtue of the former’s more refined carving and more conventional structure (Plate 19). The enormous slabs of stone of the Jiṭhānī ruin give the impression that this temple was a not wholly successful experimental construction that finally collapsed under its own load (Plate 20). 77 The brick buttresses that are found at the base along the west, north and east sides and the square brick prop in the centre of the southern stairs could have been meant to prevent the temple from falling down (Plate 21). 78 In the building of the Devarānī Temple these constructional errors were evidently remedied.

Plate 21

Brick buttresses at the western side of the Jiṭhānī Temple at Tālā

77 Cf. Williams 1982, 125: ‘The temple known locally as the Jeṭhānī is in ruins, per haps because it had an unstable stone superstructure, indicated by fragments of large āmalakas.’

78 This has first been suggested to me by the two excavators G. ḶRaykwar and R̥K. Singh. These brick supports may, possibly, date from the beginning of the 7th century when, as e.g. the Lakṣmaṇa Temple of Sirpur shows, brick had become the building material in Kosala.

12 / History and Culture of Dakṣiṇa Kosala 259

So far it is mainly the Devarānī Temple that has been discussed in the Indo logical literature. Donald Stadtner was the first to propose a date for it, viz. ad 525–550 (Stadtner 1980, 47). Two years later Joanna Williams observed: ‘To establish a precise date for temples of Tala is impossible, for nothing in the area is preserved to which they can be significantly related. [. . . ] The position of about 480 to 530 is a reasoned guess’ (op. cit. 128). Krishna Deva, finally, dates the Devarānī Temple to c. ad 550–575 (EITA II.1, 229 f.).

The dynastic chronology that we have tried to establish seems to support the early dating of Joanna Williams. Viewed from the dynastic perspective and considering the relative chronology of the two temples we are inclined to assign the Jiṭhānī Temple to the last decade of the 5th century. Keeping in view a margin of uncertainty in the dating of at least 25 years, however, it is virtually impossible to determine with certainty whether Jayeśvarabhaṭṭāraka of the Malhar Plates of Sūrabala refers to the deity of the Jit ´ .hānī or to that of the Devarānī Temple, if indeed Jayabhaṭṭāraka lent his name to one of these temples.

Williams (1982, 127) observes that the motifs found on both temples are on the whole ‘basically orthodox Gupta’. This observation should now, in view of the sculptures that have been discovered in Vidarbha since 1982 (Mandhal, Ramtek, Nagardhan, Mansar) be qualified to the effect that the artefacts of Tālā, especially the gaṇa type images, bear just as much or more relation to the Vākāṭaka as to the Gupta idiom. 79 However, it has been noted that the basal wall-mouldings and niched walls of the Devarānī Temple in particular have a definite South Indian flavour. 80 Krishna Deva (EITA II.1, 229) speaks of ‘some features reminiscent of Calukya and Pallava temples’. This theme had earlier been elaborated by Stadtner (1980, 39–41) who observes about the Devarānī Temple that,

Vertical exterior niches alternating recessed and projecting and with or without images, are a characteristic of early temples of the Deccan and of South Indian architecture in general. [. . . ] Certain elements of the base (adhiṣṭhāna) moldings of the Devarani also indicate a relationship between the temple and South Indian forms. [. . . ] The presence of these distinctly southern features in Kosala should probably be interpreted as a borrowing of architectural forms that probably existed in the sixth century in the Deccan and in South India but that appeared on temples constructed of perishable materials and therefore have not survived. It is highly unlikely that these forms originated in Kosala during the sixth century in light of Kosala’s rather provincial and isolated position.

79 Cf. K.K. Chakravarty in the Abstract of his dissertation (see below, n. 84 on p. 261): ‘As such, I have attributed the continuing volume of Tala art to the influence of the vigorous style of neighbouring Vidarbha and detected seeds of the medieval rarefication and complexity in Tala style.’

80 Stadtner 1980, 38: ‘The magnificent sculptures surviving at Tala bear strong affinities with Gupta art of the north; at the same time architectural and iconographic features of the temple reveal important connections with so-called Dravidian, or south Indian forms.’

Joanna Williams (1982, 26), though agreeing with Stadtner that the architec ture of Tālā may go back to wooden archetypes and that ‘these forms are in a general way akin to slightly later buildings to the south, both Calukya and Pallava’, disagrees with Stadtner’s emphasis of the southern origin of these forms.

Stadtner describes these (sciḷbase moldings) as forming a typically Dravidian base of the type known as pratibandha, citing the authority of ṀA. Dhaky. [. . . ] The makara toraṇas of the exterior, another seemingly south Indian element, are also found at Ajaṇṭā (Cave 6, lower). I would prefer to think of such forms as indigenous to Kosala in wood, rather than as resulting from influence from the South, where such forms are not known at this point. (op. cit. 125, n. 81)

We are not in the position to make a final assessment on this issue, which should be settled by historians of Indian architecture. However, the hypothesis advanced above, making a ruler of the Amarārya dynasty primarily responsible for the origin of these monuments, could provide an historical explanation of some South Indian influence in the architecture of Tālā in virtue of the possible South Indian descent of this family.

The Jiṭhānī Temple

Bajpai & Pandey 1977 are again the first to report on this temple. They noticed ‘the colossal nature of the stone-sculptures lying around the ruined Jethani Temple. [. . . ] Some of the sculptures have a height of 4 m to 5 m. The images of Narsimha, Buddha and a few other deities have been identified. The temple may have been a Daśāvatāra shrine. In point of time this temple is earlier than the Devrani Temple. It can be assigned to the 4th century A.D.’ 81

The archaeological exploration of the Jiṭhānī Temple mound was initiated in 1986 according to the Indian Archaeological Review of that year (IAR 1985–86, 48 (published in 1990)). Here it is said that ‘the unearthed temple has a large, simple but typical plaṇAmongst antiquities mention may be made of silver coin of Prasannamātra Sarabhpuriya king of Dakshina Kosala [. . . ].’ 82 This seems to confirm a dating of the site to the end of the 5th, beginning of the 6th

81 Op. cit. 37. Cf. Stadtner 1980, 46 f.:

The style of the sculptures surrounding the Jithani conforms to that of the Devarani, and it can be concluded that both temples were constructed at the same time. How ever, the fragments of architectural members reveal that the Jithani probably differed in construction from the Devarani. The most important of these are at least three large stone amalakas which, if they were used in the superstructure, may indicate that the temple adopted a form more consistent with north Indian standards.

Risbud 1984, 61:

Jiṭhānī maṁdir kī bhārī bharkam pratimāem. Kuṣāṇakālīn pratīt hotā haim. . Saṁbhavataḥunke lakṣaṇmeṁśilpakār ke sāmne Kuṣāṇa kālīn yakṣa-pratimāoṁkā svarūp rahā hogā. Is ādhār par Devarānī maṁdir kī tithi īsvī pāṁcvīṁchaṭhīṁśatī evam. Jiṭhānī maṁdir kī tithi īsvī cauthī śatī nirdhārit kī jā saktī hai. 82 I am not fully certain whether the section that begins with ‘Amongst antiquities . . . ’ really refers to Tālā.

12 / History and Culture of Dakṣiṇa Kosala 261

century. The final clearance of the mound was conducted during three seasons from 1985–86 under the supervision of G.ḶRaykwar and Rahul Kumar Singh (Nigam 2000, 46). R̥K. Singh published a short note on the uncovered temple in Purātan vol. 6 (1989), 169 in which he observes that, though the image of the deity enshrined in the garbhagr̥ha seems lacking, the whole collection of images definitely prove the temple to be of Saiva stamp. Among the free- ´ standing images found at the site mention is made of that of Kārttikeya, a Siva ´ head, a torso of Ardhanārīśvara, a head of an emaciated devotee, a head of Nandikeśvara and a stele of Gaurī. 83 As part of the architecture Singh lists images of Gaṇeśa, Nandin and Sivagan ´ . as.

K. K. Chakravarty devoted his PhD thesis to the subject: The temples at Tala and the art of Dakṣiṇa Kośala.84 Unfortunately this thesis was not avail able to me. I had the privilege of visiting the site in November 1992 in the company of the two excavators R̥K. Singh and G.ḶRaykwar along with Dr ḶṢNigam of Raipur University. To all three of them I owe many valuable observations.

First of all I fully agree with R̥K. Singh’s observation that the temple is decidedly Saiva although the overall scheme of the monument has yet to be ´ established. Bajpai and Pandey’s conjecture to the effect that the temple was a Daśāvatāra shrine is based on very shallow evidence, made obsolete by the clearance of the mound. The large slabs of stone lying to the west of the temple are too much damaged (they look as if their front sites are sliced off by a razor, probably caused by fractures in the stone) to allow identification as Narasiṁha and the Buddha, though one of them (measuring 110 x 265 cm) does seem to have an uṣṇīṣa (Plate 22). A smaller image of a seated Narasiṁha is found in low relief on a stone architrave lying to the southwest of the mound (Plate 23).

One of the remarkable aspects of this temple is its lay-out. There appear to have been three entrances reached by steps. The major entrance is a flight of steps, more than 7 m in width, at the bottom of which four huge pillars (measuring between 70 to 80 cm on each side, one still standing) must have supported a roof like projection (Plate 20). The space between these pillars is 114, 206 and 108 cm. This propylaeum faces south. The two pillars in the middle of the frontside rest on plinths that are bolstered by atlantid gaṇas (bhārarakṣakas) (Plate 24). The plinths of the two outward pillars are formed by lions (Plate 25). The bases of the two central pillars are formed by kumbhas on four sides, out of each of which two horses jump (Plate 26). The broadness

83 Some of these images are now in the Bilaspur Museum (State Archaeology in Bilaspur). For photos see Bosma 2018, Plates 3.17, 3.26, 3.25, 3.29.

84 Harvard 1992. For an abstract of this thesis see Dissertation Abstracts International. A: The Humanities and Social Sciences Vol. 53 No. 5, November 1992, 1297-A. Here Chakravarty states: ‘The coexistence of classical and medieval features, volumetric and linear styles in Tala art stamps it as one of the most important missing links in the transition from classical to medieval art.’ Publication of this thesis (Chakravarty 2018) came too late to be of use for the present volume.

of the steps and their direction to the south suggest that this temple faced the nearby waters of the confluence, which at those days may have been closer to the temple than it is to date.

The other two entrances are at the rear of the temple, facing east and west respectively (Plate 21). Near the steps of the western entrance lies a pillar that is deliciously carved with oblique bands of floral and foliage patterns (Plate 27). On these bands a mouse or rat is sculptured as if running down the pillar (Plate 28). A similar motif recurs in the Devarānī Temple. Along the north and east sides of the moulded base of that temple a snake emerges repeatedly out of the stone in chase of a mouse. The distance between snake and mouse gradually decreases until the mouse is caught.

Several ‘sliced off’ slabs of stone flank the eastern steps (Plate 29). A stand ing pillar, which seems to have split along its vertical axis, has different bands of ornamentatioṇThe upper portion consists of a ‘convex-lobed āmalaka’ that rests upon a vase decorated with festoons hanging from lions (cf. Williams 1982, 127 Plate 197). At the rear side of this pillar a figure with bird-like legs and ram-like head is carved standing against the vase (Plate 30). The sculp ture work evokes a strong flavour of wood carving. At the bottom of the steps stands a torso clad in a tiger skin covering a pronounced scrotum and erect penis (Plate 31). A claw and the face of the tiger is sculptured on the right thigh next to what seems to be a trident. Along the left thigh the tail of the animal runs downwards starting next to the phallus. Two huge elephant heads buttress the temple platform on the northern side (Plate 32). To compare, an image of Gaṇeśa with trunk to the left and holding a danta in his right hand is found at the northern base (i.e. right side) of the steps leading up to the Devarānī Temple (Plate 33).

The mound of the ruined temple is strewn with very lively carved images of makaras and gaṇas, which once supported pillars and other figures (Plates 34–39). These jolly, delicately sculptured, pot-bellied figures, all markedly different, are in a way reminiscent of Vākāṭaka and Gupta examples, yet they definitely breathe an atmosphere of their owṇLuxuriant patterns of flowers and foliage decorate the faces of the stone remains as far as these are not damaged and ‘sliced off’.

Two fragmentary images merit particular attentioṇOne is a complex, bro ken image, about 1.50 m as it stands now, fixed in concrete and resting against the brick prop at the middle of the southern staircase (Plate 40). It shows a group of three figures; the main one is a male deity broken just below the waist, against whose breast two makaras rise up (the right one has lost the upper portion of his trunk-like jaw). 85 The left arm of the main deity is bro ken off, but part of what seems to be his underarm in an upwards position is

85 Under the left makara a fan-shaped fold protrudes, which in a way resembles the sole of a left foot. If indeed it is that it must, in view of its position, have been attached to the image by those who are responsible for fixing the sculpture in concrete. Or it may belong to the makara itself which, anyway, is not complete as we see it now.

12 / History and Culture of Dakṣiṇa Kosala 263

preserved under his shoulder. The right arm is broken off at the elbow. The fracture above the bent elbow indicates that originally the right underarm also was directed upwards. This right underarm is preserved on a slab of stone that is placed to the left of the image (Plate 41). The right hand holds braided curls of hair, also seen at other images found in Tālā, which may have belonged to the female figure that rises above the hood-like halo of the main image. This hood consists of layers of petals, on the outer rim of which (lotus) buds are placed in a row. Altogether I count eleven buds, two preserved on the slab that contains the right underarm. The head of the female figure is missing. To her right, placed a little below her, stands a male figure, apparently making an añjali.

The head of the main deity is slightly slanted to the left. His round face, with eyes half closed as he looks downwards, with rather flat nose and thick lips, seem to express forbearance. He wears a broad necklace (graiveyaka) adorned with pendants. The female figure rising above his head wears a simple necklace and bangles on her right wrist.

The group of figures as it stands now is flanked by two round pillars, profusely decorated with scroll-work. The preserved upper portion of the right pillar consists of abacus, āmalaka, a vase and (damaged) capital, a sequence that recurs in other fragmentary pillars preserved at the site.

The interpretation of this sculpture poses many problems. There are few, if any, clues that connect it with other known iconographic material. In dis cussions with the excavators the suggestion arose that the image could be that of Varuṇa, i.e. the ocean, out of which Srī arises. The apparently invariable ´ attribute of Varuṇa, the noose (pāśa), is absent. It can not be excluded that it was held in the left hand which is missing. The makaras (the traditional vāhana of Varuṇa), and possibly, the scroll-work on the pillars (waterplants / waves) may point to the ocean. 86 His expression of forbearance would be appropriate if we assume that the goddess Srī arises out of him while he is being churned. ´ The halo of lotuses encircling his head serves the goddess above as pīṭha, the traditional one of Srī/Laks ´ .mī.

The other sculpture is also incomplete and apparently composed of at least three figures. It is hewn out of a large slab of stone that lies on top of the Jiṭhānī Temple mound (Plate 42). The main figure is a large decapitated male deity wearing a belt, keyūra adorned with lion’s head on his left arm, a kaṅkana, and a graiveyaka. His left arm seems to hold at the waist another, smaller male figure, who is positioned on the main image’s right side and of whom only the upper part of the head and left arm are preserved. The head is deliciously sculpted showing a youthful face, almond eyes, snail-like curls, and a cranial bumb (Plate 43). At both sides of the cranial bumb the feet are preserved of a third figure that must have stood on top of his head. The left

86 For the iconography of Varuṇa see Sahai 1975, 45–52.

arm, wearing bangles at the wrist, holds this third figure at his/her left ankle. 87 The identification of this group is hampered by the fact that, in the state it is in now, it lacks all significant attributes.

To conclude I would venture a tentative interpretation of this image group by directing attention to a passage from the Mahābhārata that describes the god Skanda and his offspring.

Deserted by the gods Sakra then hurled (his) thunderbolt ( ´ vajra) at Skanda. That (thunderbolt), being hurled, directly hit the right side of Skanda and split open the side of that noble one, O great king. Through the stroke of the thunderbolt another person (puruṣa) was born from Skanda, youthful, with golden armour, wielding a spear and wearing divine earrings; he became Viśākha, because he was born from the entering (viśanāt) of the thunderbolt. Then, when he saw (still) another one emerging, whose splendour resembled doomsday-fire, Indra folded his hands and took refuge with Skanda out of fear. Skanda put him and his army at ease, O foremost of men, and thereupon the Thirty (Gods) rejoiced and sounded their instruments. (MBh 3.216.12–15)

I would conjecture that the main figure is Skanda/Kumāra, the smaller deity at his right side Viśākha. 88 MBh 3.217 continues by describing the other deities that were born ‘from the thunderbolt’s impact on Skanda’, maidens (kanyās) and Kumārakas, whose father is Viśākha. MBh 1.60.23 lists three sons of Kumāra: ‘Sākha, Viśākha, and Naigameśa, born from his back’. ´ 89 Whether Naigameṣa (Naigameśa or Naigameya), the ram-headed deity, also formed part of this group is impossible to determine. That his existence was well-known to the sculptors, however, is proved by the image of Naigameṣa that is found at the left side of the steps leading up to the Devarānī Temple (Plate 44). 90 An early parallel for this sculpture is found among the Vākāṭaka images discovered in Mandhal (Plate 45). 91

Another origin myth relating to the various forms of Skanda/Kumāra, pos sibly relevant to the Jiṭhānī Temple, is told in MBh 9.43–46. In the course of his pilgrimage Baladeva (Balarāma) arrived at Sthāṇutīrtha at the Saras vatī (MBh 9.41.4). It is told how once the consecration of Kumāra as general (senāpati) took place here (kumārasyābhiṣekam, MBh 9.43.1). The curious birth of Kumāra is briefly related. Then it is said that Kumāra, in order to venerate his four parents (Rudra, Devī, Agni and Gaṅgā) splits himself up into four figures. 92

87 Below the gripping hand an anklet seems to be preserved, which would make the figure most probably a female one.

88 Cf. Mahābhāṣya ad P 5.3.99.

89 MBh 1.60.22cd–23ab: agneḥputraḥkumāras tu śrīmāñ śaravaṇālayaḥk 22 ktasya śākho viśākhaś ca naigameśaś ca pr̥ṣṭhajaḥ| Van Buitenen translates pr̥ṣṭhajaḥwith ‘as the last born’. Some mss read, however, pr̥ṣṭhataḥand this is also the reading of MBh 9.43.37, where they are said to be simultaneous (kṣaṇena) manifestations (see below, n. 92 on p. 264). See also Purāṇapañcalakṣaṇa (PPL p. 212, v. 27).

90 Cf. EITA II.1, Plate no. 446.

91 See Shastri 1977-78, 147 f.; Bakker 1997, 110 f.

92 MBh 9.43.37: tato ’bhavac caturmūrtiḥkṣaṇena bhagavān prabhuḥ|skandaḥśākho viśā khaś ca naigameṣaś ca pr̥ṣṭhataḥk 37 k

12 / History and Culture of Dakṣiṇa Kosala 265

All gods come together for Kumāra’s consecration, bringing with them all nec essary requisites (saṁbhāra). The text continues by saying that earlier at this site the consecration of Varuṇa as ‘Lord of the Waters’ (jaleśvara, MBh 9.44.20) had taken place, as is indeed described in MBh 9.46.5–11. Kumāra is conse crated and he is equipped with an army:

And the gods gave him the army, which thronged together in the southwest (?), to destroy the enemies of the gods, invincible, containing all sorts of beings. Then all the gods and descendants of Vasu, Gandharvas, Yakṣas, Rakṣas, sages and ancestors exclaimed the word ‘Jaya!’ (‘Victory!’). 93

They each present him with two servants and then the text embarks on a lengthy description of the soldiers making up Skanda’s army (MBh 9.44.51– 108), among whom we encounter all sorts of gaṇas (a summary is given by Dhaky 1984, 246 ff.). After also the host of mothers (mātr̥gaṇa), which sur rounds Kumāra (MBh 9.45.1–40), is described, the major gods present him with special gifts, a ceremony concluded by Varuṇa and Brahmā.

And King Varuṇa gave the noose provided with strength and courage, and Lord Brahmā gave Brahmaṇya (i.e. Kumāra) the skin of a black antelope, and the Creator of the World also gave victory (jaya) in battle. 94

Thus equipped Kumāra/Skanda gained victory over the enemy of the gods. We have rendered this passage from the Mahābhārata in some detail be cause it seems to fit with the materials we find in and around the ruins of the Jiṭhānī Temple. Two divine figures play a major part in the Māhātmya of Sthāṇutīrtha, Skanda/Kumāra and Varuṇa, and the two major images found at the Jiṭhānī mound could, tentatively, it is true, be identified with these two gods. 95 And to this could be added the recurrent motif of makaras and the abundance of gaṇa images at the site, the latter concurring with the promi nence of these figures in the epic story. Further research is needed to support the correspondence found. For the present purpose, the introduction of Tālā in its cultural-historical setting, may it suffice to end with a speculation, viz. that the name ‘Jayeśvara’, referring simultaneously to Lord Siva as well as to ´ the great victory obtained by Kumāra in his stead, could have been thought to be an appropriate name for the ‘Jiṭhānī Temple’ at Tālā.

93 MBh 9.44.25-26:

tathā devā dadus tasmai senāṁnairr̥tasaṁkulām |

devaśatrukṣayakarīṁajayyāṁviśvarūpiṇīm k 25 k

jayaśabdaṁtataś cakrur devāḥsarve savāsavāḥ|

gandharvayakṣarakṣāṁsi munayaḥpitaras tathā k 26 k

94 Mbh.9.45.47:

pāśaṁtu varuṇo rājā balavīryasamanvitam |

kr̥ṣṇājinaṁtathā brahmā brahmaṇyāya dadau prabhuḥ|

samareṣu jayaṁcaiva pradadau lokabhāvanaḥk 47 k

95 This is not to suggest, of course, that the Sthāṇutīrtha of the Mahābhārata is to be identified with Tālā. For the Sthāṇutīrtha and its mythology see below, p. 535 and Bakker 2014, 160 ff.

Plate 22

Image (with uṣṇīṣa ?) lying west of the Jiṭhānī Temple

Plate 23

Image of Narasiṁha on architrave lying southwest of the Jiṭhānī Temple

12 / History and Culture of Dakṣiṇa Kosala 267

Plate 24

Bhārarakṣakas supporting central entrance pillar of the Jiṭhānī Temple

Plate 25

Lions supporting outer entrance pillar of Jiṭhānī Temple

Plate 26

Base of central entrance pillar of Jiṭhānī Temple

12 / History and Culture of Dakṣiṇa Kosala 269

Plate 27

Pillar lying at the western entrance of the Jiṭhānī Temple

Plate 28

Mouse/rat on pillar at the western entrance of the Jiṭhānī Temple

Plate 29

Eastern side of the Jiṭhānī Temple

Plate 30 Plate 31

Rear side of pillar at eastern Figure at eastern entrance entrance of Jiṭhānī Temple of Jiṭhānī Temple

12 / History and Culture of Dakṣiṇa Kosala 271

Plate 32

Elephant’s head at the northern side of the Jiṭhānī Temple

Plate 33

Gaṇeśa at the northern side of the entrance of the Devarānī Temple at Tālā

12 / History and Culture of Dakṣiṇa Kosala 273

Plate 34

Supporting makara head and gaṇa (Jiṭhānī Temple)

Plate 35

Supporting makara (Jiṭhānī Temple)

Plate 36

Supporting gaṇa (Jiṭhānī Temple)

12 / History and Culture of Dakṣiṇa Kosala 275

Plate 37

Supporting gaṇas (Jiṭhānī Temple)

Plate 38

Gaṇa supporting standing image (Jiṭhānī Temple)

Plate 39

Supporting gaṇa (Jiṭhānī Temple)

12 / History and Culture of Dakṣiṇa Kosala 277

Plate 40

Image standing at the southern staircase of the Jiṭhānī Temple

Plate 41

Broken right underarm of image at southern entrance of the Jiṭhānī Temple

12 / History and Culture of Dakṣiṇa Kosala 279

Plate 42

Sculpture on top of the Jiṭhānī Temple mound

Plate 43

Profile of the head of the minor figure of the sculpture on top of the Jiṭhānī Temple mound

12 / History and Culture of Dakṣiṇa Kosala 281

Plate 44

Naigameṣa standing to the left (south) of the entrance of the Devarānī Temple

Plate 45

Naigameṣa image found in Mandhal (Vidarbha)