Chapter XXII. Origin and Development of Purāna Literature

Volume V: Part II pp. 815–886

Chapter XXII. Origin and Development of Purāna Literature

Origin and development of Purāṇa literature

The History of Dharmaśāstra (vol. I pp. 160–167) has a brief chapter on the Purāṇaṣ It has been shown therein how the Taittirīya Āraṇyaka, the Chāndogya and Bṛhadāraṇyaka Upaniṣads mention Itihāsa and Purāṇa (sometimes collectively as ‘Itihāsapurāṇam’ and sometimes separately as ‘Itihāsaḥ Purāṇam’) and how some of the extant Purāṇas are much earlier than the 6th century A.D. It is further pointed out there that the number of the principal Purāṇas has been traditionally handed down as eighteen,¹ that some Purāṇas such as the Matsya, the Viṣṇu, the Vāyu and Bhaviṣya contain much Dharma-śāstra material, that the Garuḍa-purāṇa and the Agnipurāṇa contain several hundred verses each that are identical with verses of the Yājñavalkya smṛti, that there is great divergence as to the extent of almost all Purāṇas, that some of the Purāṇas them-selves enumerate minor works called Upapurāṇas, that the Purāṇas are divided into three groups, viz. sāttvika, rājasa and tāmasa (as done by Garuḍa I. 223. 17–20, and Padma VI. 263. 81–84). Details were also furnished as to the chapters of the Purāṇas in which the several topics of Dharmaśāstra (such as ācāra, āhnika, dāna, rājadharma, śrāddha, tīrtha) were dealt with.

In the present section it is proposed to trace the great trans-formation that took place in the ideas, ideals and practices of the ancient Indian people owing to the influence of the Purāṇas in the first few centuries of the Christian era.

Before proceeding further several preliminary matters have to be dealt with. The mention of Purāṇas as a class of literature goes back much farther than was pointed out in the H. of DH. vol. I. p. 160. The Atharvaveda ² mentions Purāṇa (in the singular) in XI. 7. 24 and XV. 6. 10–11. ‘The ṛk and sāman verses, the chandas, the Purāṇa along with the Yajus formulæ, all sprang from the remainder of sacrificial food, (as also) the gods that resort to heaveṇ He changed his place and went over to great direction; and Itihāsa and Purāṇa, gāthās, verses in praise of heroes followed in going over.’ The Śatapatha brāhmaṇa also ( XI. 5. 6. 8 ) mentions ‘Itihāsapurāṇam’ (as one compound word) and states that on the 9th day of the Pāriplava the hotṛ priest instructs among other matters as follows: ‘The Purāṇa is the Veda; this it is; thus saying let him narrate some Purāṇa’ ( XIII. 4. 3. 13.). ³ The Śāṅkhāyana ⁵ Śrautasūtra ( XVI. 2. 27 ) and the Āśvalāyana Śrauta ( X. 7 ) remark that on two days of the Pāriplava the Itihāsaveda and Purāṇaveda were to be reciteḍ But the two sūtras (though affiliated to the Ṛgveda) differ as to the day on which they were to be reciteḍ It is difficult to say whether the Atharvaveda, the Śatapatha Brāhmaṇa and the Upaniṣads knew several works called Purāṇa or whether there was only a single work called Purāṇa known to theṃ But from the fact that the Tai. Ar. (II. 10) speaks of Itihāsas and Purāṇas (in the plural) it would not be unreasonable to suppose that in the later Vedic period at least some works (three or more) called Purāṇas existed and were studied and recited by those that were engaged in solemn sacrifices like the Aśvamedha. It is not unlikely that, where the singular 'Purāṇam' was employed in the Vedic texts, a class of works was meant. The facts that 'Itihāsa-Purāṇa' is called the fifth Veda in the Upaniṣads and that the Śatapatha employs 'Itihāsa-Purāṇam' as one compound word lead to the inference that the two categories 'Itihāsa' and 'Purāṇa' were similar in contents at least in several respectṣ The Āpastamba-Dharmasūtra ⁶ quotes two verses each in two places from a Purāṇa (in the singular), once cites the view of the Bhaviṣyat-purāṇa and in another place gives the summary of a Purāṇa passage or verse which says that, when a person attacks another with the object of harming him and the person attacked kills the attacker, no sin is incurreḍ From these passages it is clear that Āpastamba had before him a Purāṇa called Bhavisyat and also that the Purāṇa or Purāṇas which he knew contained rules about the food to be accepted for eating, about the stages of householder and the perpetual student, about resisting an ātātāyin even unto death and about creation and re-creation after dissolutioṇ These matters fall within the province of Smṛtis and Purāṇaṣ As the word 'purāṇa' means 'ancient', the word Bhavisyat-purāṇa is a contradiction in termṣ Long before Āpastamba the word Purāṇa had come to mean a work dealing with ancient tales &c; several such works must have been composed and they probably began to incorporate contemporary events and wrote about such events in a prophetic veiṇ Hence arose the name Bhavisyat-purāṇa.⁷ As Āpastamba mentions a Bhavisyat-purāṇa and also Purāṇa, it follows that before 500 B. C. several Purāṇas existed one of which was called the Bhavisyat and the Purāṇas then known contained the topics of sarga, pratisarga and smṛti matterṣ

This conclusion is strongly corroborated by other factṣ The Gautama-dharmasūtra provides that a learned (bahuśruta) brāhmaṇa is one who knows peoples' usages, the Veda, the aṅgas (auxiliary lores), Vākovākya (dialogues), Itihāsa and Purāṇa and that the king in administering the affairs of his kingdom and justice has to depend upon the Veda, Dharmaśāstra, the (six) aṅgas of the Veda, the (four) Upavedas and Purāṇa.⁸

From the above discussion it appears that, though we are not in a position to make definite statements about the contents of the Purāṇa or Purāṇas referred to in the Atharva, Śat. Br., Tai. Ār. and the Upaniṣads, by the time of Āp. and Gautama, Purāṇas approaching in contents to some extent at least some of the extant Purāṇas had come into existence. The Arthaśāstra of Kauṭilya states that 'trayi' means the three Vedas, viz. Samaveda, Ṛgveda, Yajurveda and that the Atharva- veda and Itihasaveda are (also) vedaṣ⁹ It follows that in Kauṭilya's time itihāsa was a definite work like the three vedaṣ In another place Kauṭilya provides 'a minister proficient in Arthasastras and bent on the good of the king should by means of itivrtta (history or historical occurrences) and purāṇas, admonish (and bring to the right path) a king led astray by other guideṣ' In laying down a time-table for the king's daily routine of work, ¹⁰ Kauṭilya provides that in the latter part of the day the king should listen to itihāsa and defines the latter as comprehending purāṇa, itivrtta, akhyayika (narrations), udaharana (heroic examples), dharmaśāstra and Arthaśāstra (science of government and statecraft). It appears that Kauṭilya meant by 'itihāsa' a Mahābhārata more or less very like the extant one, which describes ¹¹ itself as the best of itihasas, as a dharmaśāstra, Arthaśāstra, Kamasastra and as Kavyaveda. Among the officials ¹² to be maintained by the king with salaries are mentioned astrologers, persons conversant with portents and auspicious times, Pauranika, Suta and Magadha, who were to be paid 1000 (panas) as salary. The Daksmrti ¹³ which is com- paratively an early smṛti prescribes the reading of itihāsa and purāṇa for all dvija house-holders in the 6th and 7th parts of the day (divided into eight parts). The Ansanasna-smṛti prescribes that the Vedangas and purāṇas should be studied in the dark half of the months after Utsarjana (vide Jiv. part I p. 515). When the Manusmṛti¹⁴ provides that in a śrāddha rite the brāhmaṇas invited for dinner should recite the Vedas, the Dharmaśāstras, stories, Itihāsas, Purāṇas and khila hymns (such as Śrīśūkta), it should be taken as referring to Purāṇa works closely resembling the extant oneṣ The Yājñavalkya-smṛti¹⁵ enumerates the fourteen Vidyāsthānas (branches of knowledge) and sources of dharma as Purāṇa, Nyāya (logic), Mīmāṁsā (rules of Vedic interpretation), Dharmaśāstras, the aṅgas of the Veda and the (four) Vedaṣ These fourteen are probably arranged in the order of importance and authoritativeness at the time of the Yājñavalkya-smṛti. Yāj. refers to sages that expounded or promulgated the Vedas, Purāṇas, the vidyās (six aṅgas), the Upaniṣads, the ślokas (itihāsa?), aphoristic works (like those of Jaimini or on nyāya), the bhāṣyas and whatever other literature existṣ In another place Yāj. recommends that a householder after a bath in the morning and worship of gods and pitṛs should engage in japa-yajña, in which he should mutter according to his ability parts of the Veda, of Atharvaveda, Itihāsa and Purāṇas and philosophical textṣ These passages from Yājñavalkya establish that Itihāsa and Purāṇa went together, that both were works distinct from Vedic Literature and that both had attained a status of sacredness and authoritativeness in later than 3rd century (A. D.). A Vārtika mentioned by the Mahābhāṣya on Pāṇ IV. 2. 59-60 provides for the formation of words with the affix 'thak' (ika) in the case of ākhyāna (such as Yāvakrītika, Yāyatika), ākhyāyikā (such as Vāsavadattika, Saumanottarika), itihāsa (aitihāsika), Purāṇa (paurāṇika) in the sense of one who studies or knows that.¹⁶ In numerous passages the Mahābhārata mentions Purāṇa in the singular. (as in Ādi 5. 2, 31. 3–4, 51. 6, 65. 52, Udyoga 178. 47–48, Karṇa 34. 44, Śānti 208. 5, Anuśāsana 22. 12, 102. 21) and sometimes Purāṇas in the plural (as in Ādi 109. 20, Virāṭa 51.10, Strīparva 13. 2, Śānti 339. 106, Svargārohaṇa 5. 46–47 (which refers to Purāṇas as 18).¹⁷ Further, the Matsya-purāṇa and a Purāṇa proclaimed by Vāyu are mentioned in the Vanaparva.¹⁸ It is impossible to hold that all the numerous references to Purāṇas are later interpolations, though a few may be so. It is not correct to say that no Purāṇa resembling the extant ones existed before the time when Purāṇa stories were collected in the Mahābhārata. There is hardly any evidence to support any such view.

Early Sanskrit authors like Bāṇa (first half of 7th century A. D.) and commentators like Śabara (not later than between 200–400 A. D.), Kumārila (7th century A. D.), Śaṅkarācārya (between 650–800 A. D.) and Viśvarūpa (800–850 A. D.) leave us in no doubt that in their times Purāṇas existed, the contents of which were just like those of the extant Purāṇaṣ Śabara in his bhāṣya¹⁹ on Jaimini X. 4. 23, while discussing the question as to what is meant by ‘devatā’ in relation to sacrifices, states that one view was that they are Agni and others, that are described in the Itihāsa and Purāṇas as dwelling in heaveṇ Bāṇa in his Kādambarī and Harṣacarita frequently refers to the Mahābhārata and Purāṇas, but two passages from the Kādambarī and one from the Harṣacarita are very interesting. While describing the hermitage of the sage Jābāli Bāṇa employs a śleṣa 'there was Vāyupralāpita (proclaiming by god Vāyu; babbling due to the windy humour) in Purāṇa (but no babbling in the hermitage)'.²⁰ Similarly, in the des- cription of the palace of Tārāpīḍa Bāṇa employs a śleṣa (double extiendre) comparing it with Purāṇa (the two meanings being 'where the accumulated wealth of the world was arranged in appropriate groups', 'in which there is a description of the whole sphere of worlds each part of which was assigned a proper division'). In the Uttarabhāga (by the son of Bāṇa) of the Kādambarī it is stated that in all āgamas (sacred works tradition- ally handed down) such as Purāṇas, Rāmāyaṇa and Bhārata there are many stories about curseṣ The placing of Purāṇas first shows that they were probably more honoured or popular than the Rāmāyaṇa and the Bhārata. In the Harṣacarita²¹ it is stated that the book-reader Sudṛṣṭi treated Bāṇa and his relatives and friends to a musical recitation of the Purāṇa promulgated by Vāyu, that was sung by the sage (Vyāsa), that is very extensive, that is world-wide (i.e. known everywhere), that is holy and that is not different from the career of Harṣa (to which also all the adjectives applied to the Purāṇa are applicable). Here it appears that the Vāyu is expressly mentioned (as Pava- manaprokta and Pāvana) and it is further stated that the Purāṇas contained a description of the several divisions of the worlḍ This description applies to such Purāṇas as the Vāyu- Matsya (chapters 114-128), Brahmāṇḍa (II. 15 ff). It may be argued that the Purāṇa mentioned by Bāṇa may be the Brahmāṇḍa since that Purāṇa says²¹ in the beginning as well as at the end (IV. 4. 58ff) that Brahmā imparted it to Vāyu, from whom it passed on to several divine and semi-divine personages and ultimately sūta received it from Vyāsa. This latter argu- ment is not acceptable, since there was nothing to prevent Bāṇa from expressly saying that the Brahmāṇḍa was the Purāṇa recited by Sudṛṣṭi. Kumārilabhaṭṭa in his Tantravārtika refers in several places to the Purāṇas and their contentṣ A few interesting passages are set out here: On Jaimini I. 3. 1 Kumārila says²² 'Therefore the authoritativeness of all smṛtis is established by the purpose which they serve; whatever therein (in smṛtis) is connected with dharma and mokṣa (is authoritative), because it springs from (is based on) the Veda; whatever concerns wealth (artha) and pleasures is based on the usages of the people. In this way a distinction is to be made. This very reasoning applies to hortatory passages in the Itihāsa and Purāṇaṣ The Upākhyānas may be explained by the reasoning applicable to arthavādas (i. e. they have a purpose and authoritativeness just like the commendatory passages of the Veda). Narration of the divisions of the earth serves the purpose of distinguishing the regions for undergoing the consequences of the (actions that are the) means of dharma and adharma and are partly based on personal experience and partly on the Veda. The orderly presentation of vaṁśas (dynasties) in the Purāṇas is intended to facilitate the knowledge of the brāhmaṇa and kṣatriya castes and their gotras and is based on actual experi-ence and traditional knowledge; countries and measures of time are intended to facilitate worldly transactions and astronomical needs and are based on actual perception, mathematics, tradition and inference. The narration of what will happen in the future is based on the Veda, since it conveys knowledge of the variety of the ripening of the consequences of (complying with) dharma and of adharma and knowledge of the characteristics of yugas that have been in operation from time immemorial.' It is clear from this passage that the Itihāsa and the Purāṇas that Kumārila knew contained stories, geography of the earth, dynastic lists, measures of time and description of what will happen in the future. These are matters dealt with in the extant Purāṇaṣ On Jaimini I. 3. 7 Kumārila observes : 'In the Purāṇas ²³ it is narrated that in the Kali age there will arise Śākya (Gautama Buddha) and others that will cause confusion about dharma; who would listen to their words? This shows that before the 7th century A. D. the Purāṇas contained descrip- tions of the nature of Kaliyuga and that the Purāṇas Kumārila knew did not regard Buddha as an avatāra of Viṣṇu, but rather condemned hiṃ

From the facts that Kṣemendra composed his Daśāvatāracarita in 1066 A. D., that Aparārka (on p. 338) quotes a long passage from the Matsya chapter 285, verse seven of which enumerates the ten avatāras of Viṣṇu (including Buddha), and that the Gītagovinda of Jayadeva regards Buddha as an avatāra, it follows that before 1000 A. D. Buddha had come to be regarded as an avatāra of Viṣṇu, though before the 7th century some Purāṇas at least had condemned hiṃ

In the discussion of the meaning of the word ‘svarqa’ ²⁴ Kumārila asks: does it mean the region of the stars or does it mean the top of the Meru (mountain) in conformity with Itihāsa and Purāṇas or does it mean only a state of happiness? This indicates that in Kumārila’s day Purāṇas contained descriptions of the top of Meru as ‘heaveṇ’

Śaṅkarācārya in his bhāṣya on the Vedāntasūtra frequently mentions the contents and characteristic features of Purāṇas that are the same as those of the extant ones, though he actually names none of theṃ For example, on V. S. II. 1. 36 he remarks that it has been established in the Purāṇa that there is no limit to the number of past and future ²⁵ kalpaṣ On V. S. I. 3.30. the Ācārya quotes two verses that he designates as smṛti, that are but are not found in ancient smṛtis like those of Manu or Yāj. On V. S. III. 1. 15 ( api ca sapta ) the bhāṣya remarks that those who have studied or who know the Purāṇas declare that there are seven hells, Raurava and others, that are regarded as places reached for undergoing the consequences of evil deeds committed by sinnerṣ The Viṣṇupurāṇa enumerates seven hells viz. Tāmisra, Raurava and others, as the places for those who abuse the Veda, who cause obstacles to sacrifices and who give up their proper dharma. Even Manu ( IV. 87-90 ), Yāj. ( III. 222-224 ), Viṣṇu Dh. S. 43. 2-22 enumerate 21 hells and almost all Purāṇas mention 21 or more hellṣ Vide H. of Dh. vol. IV. pp. 162-164. On V. S. I. 3. 26 and 33 the bhāṣya says that one can see from Vedic mantras, arthavāda passages, Itihāsa and Purāṇa and popular belief that gods have bodieṣ On V. S. II. 1. 1. Śañkarācārya quotes a verse which is the same in Vāyupurāṇa and on I. 3. 30 quotes in all five verses as smṛti, four of which are the same as Vāyu 9. 57-58 and 64-65. Viśvarūpa in his commentary on the Yājñavalkyasmṛti has two interesting notes on Purāṇaṣ On Yāj. III. 170 which describes the Sāñkhya theory of the evolution of the world, Viśvarūpa¹³⁸ remarks that this theory concerning the creation and dissolution of the world is found everywhere in the Purāṇaṣ On Yāj. III. 175 which states that the path to the world of the pitṛs lies between the star Agastya (Canopus) and the Ajāvīthī, Viśvarūpa remarks that in the Purāṇa several vīthis (i. e. courses) of the Sun in the sky are found and that Ajāvīthī is immediately contiguous to Agastya.

From the foregoing discussion about the references to Purāṇas contained in writers from Śabara to Viśvarūpa it follows that between about the 2nd century AD. and the 6th or 7th century AD. the Purāṇas exhibited the same contents and characteristics as many of the extant Purāṇas do.

Before proceeding further, it is necessary to say here something about the Yuga-purāṇa, a part of the Gārgi-saṃhitā, and one of the earliest extant works bearing the title Purāṇaṣ Kern in his Introduction to the Bṛhatsaṁhitā (pp. 32-40) brought this rare Purāṇa containing valuable historical data to the notice of scholars from a fragmentary mṣ Later on K. P. Jayaswal²⁷ published a text of the purely historical material of the Yugapurāṇa in 115 half lines in Anuṣṭubh metre based on the fragment that Kern had got and two more msṣ that were secured by him and added a translation and noteṣ Later on Jayaswal obtained a copy made by Prof. Lévi from a mṣ (in Bengali script) in the Bibliothèque Nationale of Paris and published in JBORS. (vol. XV pp. 129-133) a table comparing the readings of that mṣ with the text he published in JBORS. vol. XIV. Prof. K. H. Dhruva contributed a paper on the text of the Yugapurāṇa to the JBORS. vol. XVI. pp. 18-66, wherein he altered the proper names and gave free scope to his conjectures and inferences with the result that succeeding scholars have not attached any weight to his emendations and remarkṣ Prof. D. K. Mankad²⁹ published a monograph in which he gave the text of the historical portion based on the msṣ utilized by Jayaswal and a fresh mṣ that he obtained in Saurāṣṭra, corrects some of the readings accepted by Jayaswal and considerably differs from some of Jayaswal’s interpretations and added a few noteṣ

The Yugapurāṇa is the 113th chapter of Gārgīsaṁhitā and is called Skandapurāṇa in the mṣ secured by Prof. Mankad, probably because the Purāṇa starts with a question by Skanda to Śiva about the characteristics of the different yugaṣ The characteristics of Kṛta, Tretā and Dvāpara are respectively described in lines 11–28, 29–45 and 46–74 of Prof. Mankad’s text; and lines 75–235 (of Prof. Mankad’s text) and lines 1–115 of Jayaswal’s text (in JBORS. vol. XIV. pp. 400–408) describe the characteristics of the Kaliyuga, and the political, social and economic history of a few centuries before the work was composeḍ The characteristics of the Kaliyuga in the Yugapurāṇa bear a very close resemblance to the description of Kaliyuga in the Vanaparva (chap. 188. 30–64), one half verse being the same in both.³⁰

The important points that emerge are briefly these:—I omit the discussion of different interpretations of Jayaswal, Prof. Mankad and Prof. Naraiṇ Kaliyuga started after Draupadī dieḍ In the beginning of Kaliyuga Janamejaya, son of Parīkṣit, would be a famous king but will come in conflict³² with brāhmaṇaṣ In Kaliyuga, Udayī son of Śiśunāga, will found the city of Pāṭaliputra on the south bank of the Ganges, which will come to be called Pusapapura and which will last for five thousand, five hundred and five years, five months, five days and five muhūrtaṣ In that Pusapapura there will be a deluded and wicked king called Śāliśūka who will establish at Saketa his virtuous elder brother named Vijaya. Then the valiant Yavanas, Pañcalas and Māthuras will attack Sāketa and will conquer Kusumapura which had a mud fortificatioṇ All countries will be disturbed (by this onslaught of the Yavanas). Then non-Aryans will follow the practices of the Aryaṣ At the end of Kali age brāhmaṇas, kṣatriyas, vaiśyas will dress alike and have the same practiceṣ People will join heretical sects and will make friends for (seducing their) wiveṣ Śūdras will offer oblations into fire with the syllable 'om' and they will address (others) with the word 'bhoh' and brāhmaṇas will address others with the words 'O āryas'. The Yavanas will establish five kings in the city (of Puṣpapura). The Yavanas will not stay long in Madhyadeśa. When the Yavanas will vanish there will be seven powerful kings at Sāketa; in Madhyadeśa there would be bloody warṣ All the Agniveśya kings will perish by war and so will the people.

Thereafter the greedy Śaka king will undergo destruction against Sāta, the king of Kalinga, and the earth will be devast-ated and Puṣpapura will be a wildernesṣ Amlāṭa called ‘red-eyed’ will secure Puṣpapura. The Mleccha king Amlāta will destroy the helpless people and the four varṇaṣ Amlāta with his relatives will perish and then there will be a king called Gopāla who will rule one year and then die. There will then be the just king Puṣyaka who will rule only one year. After two more kings, Agnimitra will be king who will wage a terrible battle with brāhmaṇas for a girl. After him his son will rule for 20 yearṣ The condition of the people will be very bad because of his fight with Śabaraṣ Then Sātu king will rule. Then there will be depredations of Śakas who will destroy one-fourth of the population and make the people demoraliseḍ In this way the Yugapurāṇa ends on a dismal note.^{1348 e}

As the Purāṇa stops with the Śakas and does not dilate on the dynasties of the Andhras, Abhīras and Guptas, it must be placed earlier than all the known Purāṇas which deal with these dynastieṣ Jayaswal places the Yugapurāṇa in the latter half of the first century B. C. In the opinion of the present author he is right. A very recent work 'The Indo-Greeks' by Prof. A. K. Narain (Oxford, 1957) has an illuminating note (pp. 174-179) on several difficult passages of the Yugapurāṇa³².

Several Purāṇas such as Matsya 53. 3-11, Vāyu I 60-61, Brahmāṇḍa I. 1.40-41, Liṅga I. 2.2, Nāradīya I 92. 32-26, Padma V. 1. 45-52 state that Purāṇa³³ was originally only one, that first, that after that the Vedas proceeded from Brahmā’s lips, that originally it had the extent of one hundred crores of ślokas and that the essence thereof to the extent of four lakhs of verses was declared in each Dvāpara age by Vyāsa. It is impossible to say whether the writers of the extant Purāṇas had any ancient tradition about this before them or whether all this about the original existence of a single Purāṇa was purely imaginary. The author holds that the latter view is more likely to be the correct one. It has been shown above that as early as the Taittirīya Āraṇyaka Purāṇas are mentioned in the plural. Therefore, the extant Purāṇas are only the successors of the ancient Purāṇas, about which, it must be admitted, we know very little.

The number of Purāṇas (in later times and by some of the Purāṇas distinguished as Mahāpurāṇas) has been traditionally handed down as eighteeṇ They are enumerated in several Purāṇas, such as Viṣṇu III. 6. 21-23, Varāha 112. 69-72 (verses 74-77 in Venk. eḍ), Liṅga I. 39. 61-63, Matsya 53.11 ff, Padma IV. 100. 51-54, Bhaviṣya I 1. 61-64, Mārkaṇḍeya 134. 7-11, Agni 272, Bhāgavata³⁴ XII. 13. 4-8, Vāyu³⁵ 104. 2-10, Skanda (Prabhāśakhanda 2. 5-7). There is some difference about the 18 names and great difference as to their extent and contentṣ The Matsya (53. 18-19), Agni 272. 4-5, Nāradīya I. 92. 26-28 enumerate the Vāyu as the 4th among the 18 Purāṇas, while most of the others substitute Śivapurāṇa as the 4th in place of the Vāyu. The Skanda (in Prabhāśakhanda 2. 5 and 7) puts Śaiva as the 4th in place of Vāyu and Vāyavīya as the last (probably meaning Brahmāṇḍa thereby). The Devibhāgavata³⁶ contains a verse naming the 18 Purāṇas by their first letters in which the Śivapurāṇa is omitteḍ The Saurapurāṇa (chap. 9. 5-12) enumerates the 18 Purāṇas, places Vāyu as the 4th (and not Śiva) and Brahmāṇḍa as the last. The Sūtasamhitā (I. 1. 7-11) names the 18 Purāṇas omitting Vāyu and putting Śiva-purāṇa in its place. The Dānasāgara in its Introductory verses (11-12 on pp. 2-3) mentions both Vāyaviya and Śaiva separately. In the verses of the Kālikā-purāṇa quoted by Hemādri on dāna I. p. 531, Śiva, Kālikā, Saura and Vahnija (Āgneya, the genuine one) are included among eighteen principal Purāṇaṣ On the whole I agree with Dr. A. D. Pusalkar³⁷ that it is the Vāyu that is entitled to be regarded as one of the 18 principal Purāṇas and not the Śivapurāṇa. Al-beruni in his work on India (com- posed in 1030 A. D.) sets out one list of Purāṇas as read out to him from Viṣṇupurāṇa, which is the same as set out below, except that the Śaiva-purāṇa is put in the 4th place instead of Vāyu (Sachan's tr. vol. I. p. 131, eḍ of 1888). It is clear, therefore, that the list of 18 principal Purāṇas had been completed long before 1000 A. D. and was introduced in the Viṣṇupurāṇa many years before that date. Al-beruni furnishes (ibid p. 130) another list of the names of 18 Purāṇas which he had heard, viz. Ādi, Matsya, Kūrma, Varāha, Narasimha, Vāmana, Vāyu, Nanda, Skanda, Āditya, Soma, Sāmba, Brahmāṇḍa, Mārkaṇḍeya, Tārkṣya (i. e. Garuda), Viṣṇu, Brahma, Bhavisya. It should be noticed that in this list Vāyu is included (and not Śaiva), that some purāṇas described in other works as Upapurāṇas are included (viz. Ādi, Narasimha, Nanda, Āditya, Soma and Sāmba) and some works almost unanimously declared to be Mahāpurāṇas (such as Padma, Bhāgavata, Nārada, Agni, Linga and Brahmavaivarta) are omitteḍ It follows that some upapurāṇas such as Ādi, Narasimha, Āditya, Sāmba, Nanda (Nandi ?) had been composed at least some time before 1000 A. D. Balambhaṭṭa (latter half of 18th century A. D.) states in his commentary on the Mitākṣarā (on Yaj. I. 3) that the Vayaviya-purāṇa is also called Śaiva.

The following is a table of the 18 principal Purāṇas together with information about the number of ślokas in each.

| Order | Name of Purāṇa | Number of verses acc. to Matsya, Vāyu 104 and some others | Number of verses acc. to some other Purāṇas and remarks | | :---- | :------------- | :---------------------------------------------------- | :----------------------------------------------------------- | | 1 | Brahma | 10000 (acc. to Nār-ada 92.31 and Bhāgavata XII. 13.4) | 25000 acc. to Agni 272. 1 | | 2 | Padma | 55,000 | | | 3 | Viṣṇu | 23,000 | The number of verses is given from 6 to 24 thousand in various workṣ | | 4 | Vāyu | 24,000 | 14000 acc. to Agni (272. 4-5), 24600 acc. to Devibhāgavata I. 3. 7. | | 5 | Bhāgavata | 18,000 | | 6. Nāradīya: 25,000 verseṣ 7. Mārkaṇḍeya: 9,000 verseṣ Other accounts: 6900 acc. to Mārkaṇḍeya itself (134. 39), 9000 acc. to Nārada I. 98.2, Vāyu 104. 4. 8. Agni: 16,000 verseṣ Other accounts: 15400 acc. to Bhāgavata XII. 13.5, 12000 acc. to Agni (272. 10-11). 9. Bhaviṣya: 14,500 verseṣ Other accounts: 14000 acc. to Agni (272. 12). 10. Brahma-vaivarta: 18,000 verseṣ 11. Liṅga: 11,000 verseṣ 12. Varāha: 24,000 verseṣ 13. Skanda: 81,100 verseṣ Other accounts: 84000 acc. to Agni (272. 17), vide below under Sknda. 14. Vāmana: 10,000 verseṣ 15. Kūrma: 18,000 verseṣ Other accounts: 17000 acc. to Nārada I. 106. 3 and Bhāgavata XII. 13. 8; 8000 acc. to Agni 272. 19, 13000 acc. to Agni 272. 20-21. 16. Matsya: 14,000 verseṣ 17. Garuḍa: 18,000 verseṣ Other accounts: 19000 acc. to Bhāgavata XII. 13.8 and Devībhāgavata I. 3; 8000 acc. to Agni 272.21. 18. Brahmāṇḍa: 12,200 verses acc. to Matsya 53. 54. Other accounts: 12000 acc. to Bhāgavata (XII. 13-8 and Agni 272.23).

It will be noticed from the table that the total number of verses in the 18 Purāṇas according to the figures furnished by most Purāṇas comes to 400600 ślokaṣ This closely agrees with the total of four lakhs assigned to the eighteen Purāṇas in some of the Purāṇas (vide note 1349 above). But several of the extant Purāṇas contain much smaller numbers of ślokas than are assigned to theṃ For example, it appears from the commentaries called Viṣṇucitti³⁸ and Vaiṣṇavākūṭacandrikā on Viṣṇupurāṇa III. 6. 23 that the estimates of the extent of the Viṣṇupurāṇa varied from 6, 8, 9, 10, 22, 23 to 24 thousand ślokas and that both the commentaries comment on a text of the Viṣṇupurāṇa that has only 6000 ślokaṣ Similarly, the extent of the Kūrma is said to be 17000 or 18000 ślokas by most of the Purāṇas that furnish the extent of Purāṇas, but the extant Kūrma has hardly 6000 ślokaṣ The Brahma contains 10000 ślokas according to the Nāradīya and 25000 according to the Agni, but the printed Brahma (of the Anandaśrama eḍ) contains about 14000 verseṣ On the other hand, the Skanda (that has two recensions at present) is said to contain 81000 ślokas, but the Skanda printed by the Venk. Press contains several thousands more. The Bhaviṣya (Brāhmaparva)³⁹ states that all Purāṇas had originally 12000 verses each, but their extent increased by the addition of tales, so that the Skanda was inflated to the extent of one lakh of ślokas and the Bhaviṣya to 50000 ślokaṣ The order in which the Purāṇas are enumerated is not quite uniforṃ Most Purāṇas put the Brāhma in⁴⁰ the first place and set out the order as in the table above, yet the Vāyu (104. 3) and the Devībhāgavata (I. 3. 3) begin the list of Purāṇas with Matsya. The Skanda (Prabhāsa-khaṇḍa 2. 8-9) puts the Brahmāṇḍa in the first place. The Bhāgavata (XII. 7. 23-24) enumerates the eighteen Purāṇas in a somewhat different order. The Vāmana⁴¹ regards the Matsya as the foremost among Purāṇaṣ Notices of the contents of all the 18 Purāṇas occur in Matsya (chap. 53), Agni (chap. 272), Skanda (Prabhāsakhaṇḍa 2. 28-76), Nāradīya (which devotes 18 chapters from I 92. 30-43 to I. 109 to the contents of 18 Purāṇas from Brahma to Brahmāṇḍa ). There is almost complete agreement among the Purāṇas as to the names of the 18 principal Purāṇas, except as to the Vāyupurāṇa.

The question about the Purāṇas is further complicated by the fact that some of the Purāṇas themselves mention the names of a number of Upapurāṇas, though others ignore theṃ For example, the Matsya ( 53. 59-62 ) names the Nārasimha, the Nandī, the Āditya and Sāmba as Upapurāṇas and appears to say that the Nārasimha extended to 18000 ślokas and elaborated the description of the Man-Lion avatāra declared in the Padma-purāṇa. The Kūrma ( I. 1. 16-20 ), Padma (IV. 111. 95-98 ), the Devībhāgavata (I. 3. 13-16) set out the names of eighteen Upapurāṇaṣ Some of the Upapurāṇas bear the same names as those of the principal purāṇas, viz. Skanda, Vāmana, Brahmāṇḍa and Nāradīya. Prof. Hazra¹³⁵⁸ states that the number of Upa-purāṇas is 100. Only a few of the Upapurāṇas have been published and those published do not differ much in the matter of subjects from some of the principal purāṇas and almost all belie the definition of Purāṇa as ' pañcalakṣaṇa '. It has been seen above ( note 1349 ) that the number of ślokas in the 18 principal Purāṇas is said to be four lakhṣ It must be remembered that this total does not include the number of the ślokas assigned to the Upapurāṇas, and no Purāṇa includes the verses of the Upapurāṇas in the total of four lakhṣ Further, no one should forget or ignore the significance of the remarks of the Matsya and Kūrma about the Upapurāṇaṣ The Matsya⁴³ speaks of the Upapurāṇas as sub-sections (upabhedas) of the principal eighteen Purāṇas and emphatically asserts ‘know that whatever is declared as distinct from the eighteen Purāṇas came forth ( or issued from ) these (18)’. The Kūrma is no less clear. It states that⁴⁴ the Upapurāṇas are the summaries or abridgments of the 18 ( principal ) Purāṇas made by sages after studying theṃ The lists of Upapurāṇas given by several works, most of which are set out by Prof. Hazra in his paper on Upapurāṇas (in ABORI vol. XXI, at pp. 40–43) and in his Studies (pp. 4–13) diverge a good deal from each other. As the Matsya mentions by name only four Upapurāṇas it is not unreasonable to hold that more than four were not in existence at the time when this passage of the Matsya was inserted in it or at least that more than four had not been accepted as Upapurāṇas at that time. A good many of the Upapurāṇas are of late date. Only a few of the rest such as the Narasimha, the Viṣṇu-

Upapurāṇas are not included in the 4 lakhs of verses of all the Purāṇas together; fourthly, early commentators and nibandha writers like the Mitākṣarā and the Kṛtyakalpataru either do not mention any of the Upapurāṇas or at the most only about half a dozen or so and that too rather sparingly; and lastly, as Prof. Hazra himself asserts ('Studies', vol. I, p. 23) adherents of various sects such as Śāktas, Sauras, Pāñcarātras interpolated chapters in the Purāṇas of the established group and in some cases wrote new and independent works to propagate their own ideas and styled them Purāṇaṣ

Early commentaries and digests of Dharmaśāstra very rarely refer to any of the well-known Upapurāṇaṣ The Mitākṣarā, though it names the Brāhma (on Yaj. 1. 3 and 45), quotes passages from the Matsya (profusely), Viṣṇu (on Yaj. III. 6), Skanda (on Yaj. III. 290), Bhavisya (on Yaj. I. 236, 254, III. 19, 287, 289) and Brahmāṇḍa (on Yaj. III. 30), hardly refers, so far as I know, to any Upapurāṇa. The Kṛtyakalpataru of Lakṣmīdhara (composed about 1110-1130 A. D.), twelve parts of which published in the G. O. S cover several thousand pages, profusely quotes many of the Mahāpurāṇas, and only six of the Upapurāṇas by name, viz. Ādi (only twice on śuddhi), Nandi (profusely only on four profusely on several topics). Aparārka (first half of 12th century A. D.) who quotes profusely from Brahma, Brahmāṇḍa, Bhaviṣyat, Mārkaṇḍeya, Vāyu, Viṣṇu and Matsya quotes by name only the following Upapurāṇas and only sparingly viz. Ādi, Āditya, Kālikā, Devī, Nandi, Nṛsiṁha, Viṣṇudharmottara (seven times), Viṣṇurahasya (once), and Śīvadharmottara (once ). The Danasagara^{1, 62} (written in 1169) states ‘Upapurāṇas have been promulgated, which clearly set out the procedure of gifts' and it expressly mentions as Upapurāṇas Ādya (Ādi or Brahma?), Āditya, Kālikā, Mārkaṇḍeya, Viṣṇudharmottara and Śāmba and remarks that the Viṣṇurahasya and Śīvarahasya are merely compilations (saṅgrahārūpa). It is not necessary to refer to writers later than about 1170 A. D. in the matter of Upapurāṇaṣ

The references to eighteen Purāṇas in about a dozen principal purāṇas and the description of their contents in some of them naturally lead to the inference that these passages were added some time after all the eighteen Purāṇas assumed a complete forṃ It is not possible to hold that all the main purāṇas were composed at the same time by one individual author or even by several contemporary writerṣ Besides, most of the editions of Purāṇas are based either on a single mṣ or on few msṣ collected at random and are not critically edited as is the case with the critical edition of the Mahābhārata published in Poona by the B. O. R. I. Many conclusions therefore, drawn from the current printed editions of the Purāṇas or from msṣ of the Purāṇas, must be regarded as merely tentative and as likely to turn out to be wrong. What Winternitz said in his History of Indian Literature (English translation, Calcutta, vol. I. p. 469) viz. 'that the date of each section, nay, sometimes of each single verse in the Mahābhārata must be determined separately' applies with equal (or perhaps greater) force to the Purāṇas, particularly when one wants to use the section or stanza for historical or comparative purposeṣ⁴⁶

It may be conceded that the Purāṇas and some of the avail- able Upapurāṇas also contain many ancient legends and tradi- tions; but these have been so much tempered with and inflated by additions intended to bolster up particular forms of worship and particular tenets that great caution is required before one can recognize them as genuine and reliable representatives for ascertaining the general state of Indian society and beliefs in ancient and medieval timeṣ

In the opinion of the present writer there is no positive objective evidence for placing any of the upapurāṇas except the Viṣṇudharmottara before the 8th or 9th century A. D. Even in the case of Purāṇas, there have been large interpolations such as verses about the 18 purāṇas, their number and contentṣ But they contain much ancient material and are far more reliable than most of the Upapurāṇas, having been quoted by writers from 8th and 9th centuries A. D. or even earlier.

The Amarakosa⁴⁷ defines 'Itihāsa' as 'what happened in the past' and 'Purāṇa' as 'what has five signs or characteristics'. It is no doubt true that some purāṇas speak of Purāṇa as ‘pañcalakṣaṇa’ and set out the five characteristics⁴⁸ as `sarga` (creation), `pratisarga` (re-creation after dissolution of the world), `vaṁśa` (dynasties of gods, the Sun and the Moon, and the patriarchs), `manvantara` (the vast periods of time that are so called), `vaṁśānucarita` or `vaṁśyānucarita` (deeds and history of the descendants of the solar, lunar and other dynasties). The Bhāgavata⁴⁹ states that ten topics are dealt with by Purāṇas and that some say that they are only five. The ten topics of the Bhāgavata are `sarga`, `visarga` (dissolution or creation after destruction), `vṛtti` (modes of subsistence, natural or prescribed for all men by śāstra), `rakṣā` (protection i. e. `avatāras` destroying those that hate the Vedas), `antarāṇi` (i. e. `manvantaras`), `vaṁśa`, `vaṁśyānucarita`, `saṁsthā` (four kinds of `laya`), `hetu` (the cause of creation viz. the soul that is subject to `avidyā` and collects `karma`), `apāśraya` (the refuge of individual souls, viz. `brahman`). The Matsyapurāṇa also mentions other characteristics of Purāṇas viz. in all Purāṇas the four goals of man (`puruṣārthas`) are described, also the consequences of doing what is opposed to `dharma`, division of Purāṇas into `sāttvika`, `rājasa`, `tāmasa`, the `sāttvika` and `rājasa` Purāṇas respectively extolling the greatness of Hari and Brahmā, `tāmasa` Purāṇas extolling Agni and Śiva, the mixed ones extolling the greatness of Sarasvatī and `pitṛs`. The questions addressed by Manu to Keśava (in Matsya 2. 22–24) indicate the subjects that would be dealt with by that Purāṇa viz. creation and dissolution, the `vaṁśas`. manvantaras, vaṁśyācarita, the expanse of the world, rules about dānas, śrāddha, varṇas and āśramas, iṣṭa and pūrta, establishment of images of gods, and everything else.

It is not quite clear why the Amarakosa seized upon the above mentioned five topics as characteristics of Purāṇaṣ Amarakosa⁵⁰ cannot be placed later than the 5th century A. D. It is likely that before that time the number of Purāṇas was not large, that they had not become very inflated, and that, as Itihāsa and Purāṇa were often lumped together as the 5th Veda in the Upaniṣads, they both had certain matters in commoṇ Itihāsa did not probably deal with creation, dissolution and manvantaras, but contented itself with the dynasties of kings and with the deeds and legends about the heroes of the past. Rarely, Itihāsa (Mahābhārata) is called Purāṇa and some extant Purāṇas describe themselves as Itihāsa. For example, the Vāyu⁵¹ calls itself in the same context both itihāsa and purāṇa. The Brahmapurāṇa calls itself Purāṇa as well as ākhyāna (245. 27 and 30). The Mahābhārata⁵² though generally called itihāsa by itself (as in Ādi 1. 19, 26, 54) or the best of itihāsas, still applies to itself the wordṣ ‘ākhyāna’⁵³ (as in Ādi 2. 388-89), Kāvyā (Ādi 2. 390), Kārṣṇaveda (Ādi 1. 264) and Purāṇa (Ādi 1. 17). From this it appears that originally the line of demarcation between the two was rather thiṇ In defining Purāṇa as ‘pañcalakṣaṇa’ the Amarakosa and some of the Purāṇas seized upon such topics as distinguished Purāṇas from Itihāsa and other branches of Sanskrit literature. It has been shown above that the Purāṇa and Bhaviṣyatpurāṇa that existed before Āpastamba contained not only sarga and pratisarga but also some smṛti materials as well. From the definition given in the Purāṇas and the Amarakōśa it does not at all follow that those five topics alone were the constituent parts of the very ancient Purāṇas, as Kirfel holds;⁵⁴ What could be said at the most is that five were the distinctive topics that differentiated the Purāṇa class of works from other branches of literature and probably cognate works called Itihāsa or it may be that the five are generalized as an ideal for Purāṇas and that the very early representatives of this class ( that existed before Āp. Dh. S. ) had not these five as characteristicṣ

The extant Purāṇas contain far more subjects than the five. Some Purāṇas barely touch these five and deal at great length with altogether different topicṣ Only a few of the extant Purāṇas the extant Viṣṇu alone closely agrees with the definition of Purāṇa as 'pañcalakṣaṇa,' though it also contains a good many other topicṣ On a modest calculation the four subjects of vrata, śrāddha, tīrtha and dāna cover at least one hundred thousand ślokas in the extant eighteen main Purāṇaṣ Several of them have identical chapters on several topics (e.g. Matsya⁵⁵ and Padma, Vāyu and Brahmāṇḍa contain long passages that agree verbatim with one another in many matters and details). It is probable, therefore, that the extant principal Purāṇas are partial and gradually inflated representatives of an earlier group of Purāṇas (not necessarily 18 in number) that existed before Yājñavalkya. In the present state of our knowledge it is almost an insoluble problem to find out what these Purāṇas were or containeḍ The number 18 was probably due to the fact that the number is prominent in several connections as regards the Mahābhārata. The Bhārata war was fought for 18 days, the total of the vast armies engaged in the conflict came to 18 akṣauhiṇīs, the epic has 18 parvans, the Gītā also has 18 chapterṣ⁵⁶

The Purāṇas may be classified into several categories, viz. (1) encyclopaedic like the Agni, Garuḍa and Nāradīya, (2) those mainly dealing with tīrthas such as Padma, Skanda and Bhavisya, (3) sectarian, such as Liṅga, Vāmana, Mārkaṇḍeya, (4) Historical such as Vāyu and Brahmāṇḍa. The Vāyu, Brahmāṇḍa, Matsya and Viṣṇu are probably the oldest among extant Purāṇas, though they too have received substantial additions from time to time.

There are seven Purāṇas that contain historical material, viz. the ancient dynasties down to the time of the Bhārata war and from the Bhārata war to the downfall of the Andhras and the rise of the Guptas,⁵⁷ viz. Vāyu 99. 250-435, Viṣṇu IV. 20. 12 to IV. 24. 44, Brahmāṇḍa III. 74. 104-248, Bhāgavata IX. 12. 9-16, IX. 22. 34-49 and XII. 17, Garuḍa 140 and 141. 1-12, Bhaviṣya III. 3 and 4 (this account is practically worthless). Matsya contains the fullest list of Āndhra kings and states (273. 16-17) that 29 Āndhra kings will rule for 460 years, while the Vāyu (99. 357-358) states that 30 Āndhra kings will rule⁵⁸ for 456 (406?) yearṣ Both Vāyu (99. 355) and Matsya (373. 16) ('Pulova' in Vāyu) make Puloma as the last king of the Āndhraṣ Ptolemy who published his 'Geography of India' about 150 A.D. states that Ptolemaios was king of Baithana (Paithana) in his days (vide J.I.H. vol. 22,1943, at p. 84 in an article on 'Apostles of Kalyan'. So that these historical references to Āndhras must be a good deal later than 150 A. D. Only four Purāṇas viz. Vāyu, Brahmāṇḍa, Bhāgavata and Viṣṇu mention in a general way that kings of the Gupta dynasty will rule along the Ganges up to Prayāga, Sāketa (Ayodhyā) and Magadha, but no names of Gupta kings are specifieḍ The passages referring to the Guptas are rather corrupt. It is argued by Pargiter (in 'Dynasties of the Kali age' p. XII.) and others that Samudragupta was a great conqueror as his Allahabad Praśasti shows (Fleet's 'Gupta Inscriptions' No. 1). Most scholars are agreed that the Gupta dynasty began to rule about 320 A. D. It is argued that, if the reviser or revisers of the Purāṇas had known of the brilliant campaigns of Samudragupta they would have named him at least and that therefore the revision of the Vāyu took place between 320–335 A. D.

There is a large mass of Literature dealing with several questions relating to Purāṇaṣ For those who desire to make a close study of them and wish to be acquainted with the controversies relating to Purāṇas, some of the more important works and papers on the Purāṇas may be mentioned here: H. H. Wilson's Introduction to the English tr. of the Viṣṇupurāṇa, vol. 1. (1854); F. E. Pargiter's 'Purāṇa texts of the dynasties of the Kali age' (1913), 'Ancient Indian genealogies' in Sir R. G. Bhandarkar Presentation volume pp. 107-113, 'Ancient Indian Historical Tradition' (Oxford, 1922); W. Kirfel's 'das Purāṇa Pañcalakṣana' (Bonn, 1927), 'die Cosmographie der Inder' (1920), 'Bharatavarṣa' (Stuttgart, 1931); Vries on Purāṇa studies' in Pavry commemoration vol. pp. 482–487 (applies Kirfel's method to the subject of śrāddha in the Brahmāṇḍa, Harivaṁśa, Matsya, Padma and Vāyu); Haraprasad Shastri's descriptive cat. of Msṣ at the Asiatic Society of Bengal, vol. V. Preface pp. LXXXII-CCXXV and his paper on 'Mahā-purāṇas" in J. B. O. R. S. vol. XV. p. 323–340; Prof. B. C. Majumdar's paper in Sir Asutosh Mookerji Silver Jubilee vol. III, Orientalia, part 2 pp. 9–30; Dr. A. Banerji Śāstri's paper on⁵⁷. 'Ancient Indian Historical Tradition' in J. B. O. R. S. vol. XIII pp. 62–79 ( supplies a useful corrective to many sweeping assertions of such scholars as Macdonell, Pargiter and others); Cambridge History of India, vol I pp. 296–318; Winternitz's 'History of Indian Literature,' English Tr. vol I. pp. 296–318; Prof. H. C. Hazra's 'Studies in the Purāṇic Records of Hindu rites and customs' (Dacca 1940), papers on ‘Purāṇas in the History of smṛti' in 'Indian Culture,' vol. I pp. 587–614, 'Mahā-purāṇas' in Dacca University Studies' vol. II, pp. 62–69; 'Smṛti chapters in Purāṇas (I. H. Q. vol XI pp. 108–130), 'Pre-Purāṇic Hindu Society before 200 A.D.' (I. H. Q. vol. XV. pp. 403–431), 'Purāṇic rites and customs influenced by the economic and social views of the sacerdotal class' (in Dacca University Studies' vol. XII, pp. 91–101), 'Influence of Tantra on Smṛtinibandhas' (in A. B. O. R. I. vol. XV. pp. 220–235 and vol XVI pp. 202–211), 'the Upapurāṇas' (in A. B. O. R. I. vol. XXI. pp. 38–62); 'Purāṇa Literature as known to Ballālasena' (in the J. O. R., Madras, vol. XII. pp. 129–146), 'Some Minor Purāṇas' in A. B. O. R. I. vol. XIX. pp. 69–79, 'the Aśvamedha, the common source of origin of the Purāṇa Pañcalakṣaṇa and the Mahābhārata' A. B. O. R. I. vol. 36 (1955 pp. 190–203); 'Some lost Upapurāṇas' in J. A. S., Calcutta, vol. 20 pp. 15–38; and many other papers on individual Purāṇas specified at the end of this part; Das-Gupta's 'Indian Philosophy' vol. III. pp. 496–511 on 'Philosophical speculations of some Purāṇas'; Dr. D. R. Patil's paper on 'Gupta Inscriptions and Purāṇic tradition' (in Bulletin of D. C. R. I., vol. II pp 2–58, comparing passages from Gupta Inscriptions and Purāṇas); Prof. V. R. Ramchandra Dikṣitar's 'The Purāṇas, a study' (in I. H. Q. vol. VIII. pp. 717–67) and 'Purāṇa Index' in three volumes; Dr. A. D. Pusalkar's paper in 'Progress of Indic Studies' (1917–1942) in Silver Jubilee Volume of BORI. pp. 139–152 and 'Studies in Epics and Purāṇas of India' (B. V. Bombay, 1955); Prof. D. R. Mankad's ⁵⁹ papers on 'Yugas' (in P. O. vol. VL part 3-4 pp. 6-10), on ‘Manvantaras’ (IHQ. vol. XVIII. pp. 208-230) and in B. V. vol. VI pp. 6-10; Dr. Ghurye's Presidential Address in the section on Ethnology and Folklore in Pro. of 9th A. I. O. C. (1937) pp. 911-954; Dr. A. S. Altekar's paper ‘Can we re-construct pre-Bhārata-war history?’ in J.B.H.U. vol. IV. pp. 183-223 (holding that the various pre-Bhārata-war dynasties mentioned in the Purāṇas are as historical and real as the dynasties of Mauryas and Andhras and the Paurāṇic genealogies really refer to kings who figure in the Vedic Literature also); Dr. Jadunath Sinha's ‘A History of Indian Philosophy’ vol. 1 pp. 125-177 on the philosophy of the Purāṇas’ (1956); two papers ‘on the ancient chronology of India’ by R. Martin Smith in J. A. O. S. vol. 77 No. 2 (April-June 1957) and No. 4 Dec. 1957 ( He follows Pargiter in his texts ). Some remarks on the important conclusions of Pargiter and Kirfel are necessary. Pargiter tries to construct history from the earliest times to the Bhārata war which he holds to have taken place about 950 B. C. (AIHT, chap. 15 p. 182). He holds that there were two traditions in ancient India, viz. the Kṣatriya tradition and the Brāhmaṇa tradition (AIHT, chap. 5 pp. 58-77). He harps dozens of times on the utter lack of the historic sense among the brāhmaṇas in his work (AIHT), holds that the Purāṇas represent the kṣatriya tradition, that there were three racial stocks, viz. the Mānavas (or Mānvas as he styles them), the Ailas and the Saudyumnas, that respectively represent the Dravidian, the Aryan and Munda (AIHT chap. 25 pp. 289-302 ), that the Purāṇas are Sanskritisations of works in Prakrit ( pp. X-XI ) of ‘Dynasties of the Kali age’. His war has not been accepted by later scholars, since his handling of that subject is not judicial, objective or straightforward, but relies too much on his own prejudices and on averageṣ In his ‘Purāṇa texts of the dynasties of the Kali age (pp. 58, 74) he appears to favour the view that the Bhārata war was fought 1050 years before the Nandas i. e. about 1475 B. C. The msṣ and the printed Purāṇas give four different periods between the birth of Parikṣit⁶⁰ and the crowning of Nanda, viz. 1015 years (Viṣṇu), 1050 years (Vāyu, Brahmāṇḍa and msṣ of Matsya), 1115 years (Bhāgavata), 1500 years (some msṣ of Viṣṇu and of Matsya). Pargiter himself argues forcefully for the trustworthiness of tradition and Paurāṇika genealogies in R.G. Bhandarkar Presentation volume pp. 107–113 and in AIHT chap. X. p. 119–125. Besides, there is the com- mon experience that the total number of years between one well- known event and another can be easily remembered and handed down even orally for hundreds of years, while handing down hundreds of royal names is a difficult matter and some names may easily drop out. Moreover, the Matsya, Brahmāṇḍa and Vāyu themselves say that they ⁶¹ mention only the prominent kings of the Ikṣvāku line and the Bhādradratha line ⁶² and even as to the Paurava ⁶³ line it is clear that that dynasty had many names not all of which are enumerateḍ There is therefore every possibility that some kings dropped out even in later genealogies also (as for example Matsya 213. 16, saying that Āndhras were 29 and Vāyu 99. 357 saying they were 30 ). Merely counting the total of the kings actually named in the Purāṇas would not convey a quite accurate idea of the total length of years during which that dynasty ruleḍ Bearing in mind the two matters (viz. trustworthiness of tradition and Paurāṇika genealogies and the ease of remembering the time distance between too well-known events) he should have endeavoured to find out the age of the Bhārata War. He brushes aside the statement of the period between Parikṣit and Nanda as unreli- able simply on the ground (which appears to me flimsy) that the figures (1015, 1050) are discrepant (A. I. H. T, p. 180). In almost all passages of the Purāṇas there is some discrepancy or other. Therefore, he should have made an endeavour to find out which one of the three periods (1015, 1050, 1500) is supported by the best and oldest msṣ and should have stuck to them, particularly when the Sanskrit equivalents of the three periods (pañcadaśa, pañcāśat and pañcaśata) are so much alike that scribes might easily have been confused and made their own readingṣ Even taking the least period of (1015) years, the Mahābhārata War would have to be placed at about 1440 B. C. (adding 1015 to the date of Nanda's enthronement, viz. about 425 B. C.). Most Western writers and Prof. S. N. Pradhan (in 'Chronology of ancient India', Calcutta 1927, pp. 249ff) find fault with Paurāṇika statements and brush them aside as practically worthlesṣ Prof. Pradhan takes the kings actually named in the three lines, holds that 28 years are the average reign period of each and multiplying the number of kings by 28 arrives at the conclusion that the Mahābhārata war was fought about 1150 B.C. It is not possible to deal at length with his argumentṣ But he ignores the express words of the Purāṇas that they enumerate only the principal or important kingṣ Besides, there are scholars like Pargiter who regard 17 or 18 years as the average reign of a king in other countries (and in India also). The author cannot accept Prof. Pradhan's reasoning. Most Western writers are cloth to assign old dates to matters Indiaṇ Pargiter is no exceptioṇ Instead of straightforwardly accepting one of the three periods that was strongly supported by mss, he indulges in some devices that appear like tricks of jugglery (A. I. H. T. pp. 180–183). His method requires some explanation and examinatioṇ

Vyāsa is said to have been alive when the Bhārata battle was fought at the end of the Dvāpara age and he is also held to be the author of the 18 Purāṇaṣ The kings before the Bhārata war, the Pāṇḍava heroes and a few descendants of them and of some contemporaries of them are treated by the Matsya, Vāyu, Brahmāṇḍa and others as past (atīta). Adhisomakṛṣṇa or Adhīsimakṛṣṇa, who ⁶⁵ was 6th in direct descent from Arjuna (excluding Arjuna) is said to have been alive when the Purāṇas were narrated to the sages at the sattra. Both Vāyu (99. 282) and Matsya (271. 5) state that in the Ikṣvāku line king Divākara, 6th (or 5th in Matsya) in descent from Bṛhadbala, was alive at the time when the Purāṇas were narrateḍ Then the same Purāṇas (Vāyu 99. 30, Matsya 271. 23 and Brahmāṇḍa III. 74. 113) state that in the line of Jarāsandha (ruler of Magadha), who was a contemporary of the Pāṇḍavas and whose son Sahadeva was killed in the Bhārata war, there was Senājit who was a contemporary of Adhisīma-kṛṣṇa and Divākara and who was 7th in descent from Sahadeva. These three are described as vartamāna kings in the Purāṇas and all those that came after these three are described as bhaviṣya. Pargiter first (AIHT p. 180) takes the total of the kings of the three dynasties, Aikṣvāka, Paurava and Māgadha that are actually named (ignoring what the Purāṇas say viz. that only the principal or important kings are named and not all) and the total of the reigns of all these (1408 years), finds that the average of the reign of each king in the three lines works out at such large figures as 47, 50, 31, which he regards as impossible when tested by real historical averageṣ But he forgets that the Purāṇas say that in the Aikṣvāka, Magadha and Paurava lines and also generally only the prominent kings are mentioned and also that the extant Purāṇas are only fragments left of the originals, since in the Brahmāṇḍa (III. 74) all Paurava and Aikṣvāka kings are altogether lost. Then he takes the total number of kings in ten kingdoms up to Mahāpadma and arrives at the average of 26 kings for each kingdom (AIHT p. 181). Then he says that the average of fourteen series of kings in Eastern and Western countries which he examined comes to 19 years for each reign and, holding that the average in eastern countries is less than in western countries, he arrives at the average of 18 years for each reign (pp. 181–182), which he regards as fair and rather liberal. He then multiplies the average 18 (of length of reigns) by 26 (the average number of kings in ten countries which he supposes to be the only kings in those ten) and arrives at the figure 468 yearṣ He adds these to the date which he assigns to Mahāpadma Nanda 382 B. C. and thus arrives at 850 B. C. as the mean date of the beginning of the reign of Adhisīmakṛṣṇa, Divākara and Senājit, who were vartamāna kingṣ Then he takes five as the average of the kings between the vartamāna kings and Yudhiṣṭhira and assigning about 100 years to these 5 kings, arrives at the date 950 B. C. for the Bhārata war. He discards (AIHT p. 180 n 3) the astronomical evidence contained in the Purāṇas (and the Mahābhārata) about Bhārata war in a single sentence viz. that astronomical statements can have no scientific precision and can only have been formed by estimate at the close. The probable date of the Mahābhārata war has been discussed by the author at some length on materials supplied by the Mahābhārata, the Purāṇas, Varāhamihira, Āryabhaṭa and Inscriptions in vol. III. pp. 895-923 and there- fore he does not go into that question here. But he strongly disapproves of the methods of Pargiter and the date he deduceṣ Later scholars like Kirfel have not accepted the conclusions of Pargiter about two separate traditions and about the Purāṇas being Sanskritizations of originals written in Prakrit and Kharosṭhi script (p. XVI. of Intro. to ‘Purāṇa Texts &c.'). The importance of another independent source has not been taken proper notice of by Pargiter and even by Kirfel. It appears that about 300 B. C. Megasthenes was supplied with a list of kings from Bacchus to Alexander's time (153 or 154 in number) covering by their reigns a period of 6451 years and three monthṣ Vide 'Ancient India as described by Megasthenes and Arrian' by McCrindle (1877) p. 115 and Cambridge History of Ancient India, vol. I. (1922, p. 409). Even supposing for argument that the account of kings is not trustworthy the fact remains that about 300 B. C. Indians claimed that they had lists of kings that reigned before that date for thousands of years (and not for a few hundred years as Pargiter would have us believe).

It has been shown above that Āpastamba mentions a Bhāvi- ṣyatpurāṇa and quotes four verses from a Purāṇa (p. 817, note 1328). That Bhaviṣyat-purāṇa was probably so called because it contained in a prophetic vein the names and other details of kings that flourished after the Bhārata heroes and after a few generations of descendants of them and their contemporaries and probably purported to have been composed by some sage or by Vyāsa. As the Kali age is said to have started after the Bhārata war, as Parāśara, his son Vyāsa and Vyāsa's son Śuka were regarded as more or less contemporaries of the Pāṇḍavas that lived in the Dvāpara age and as all the 18 Purāṇas are deemed to have been composed by Vyāsa⁶⁶ in the Dvāpara age, the history of the kings of the Kali age from the descendants of Adhisimakṛṣṇa and his contemporaries downwards has been furnished by the Purāṇas in the form of a prophetic style. It has not been clearly noticed by both Pargiter and Kirfel that the so-called future kings are divided into two groups, viz. the kings of the Aila, Aikṣvāka and Māgadha lines from Adhisimakṛṣṇa, Divākara and Senājit to the last scions of these lines (such⁶⁷ as Sumitra in the Aikṣvāka line, Kṣemaka in the Aila line) form one group and later kings in the lines of Pradyotas, Śungas, Andhras, Śakas and others form another group and further that the first group was most probably dealt with in the ancient Bhaviṣya-purāṇa or some other Purāṇa if we rely on Āpastamba, but the other group not having been in existence when the Bhavisyat was composed (before 500-400 B. C.) was dealt with by the extant Purāṇas from information received by them apart from the ancient Bhavisyat. This is clear from the passages of the Matsya and Vāyu quoted⁷⁰ below. The Matsya says 'After this I shall proclaim those future kings in the Aida (Aila), Aikṣvāka and Paurava lines and those with whom those three benign (or virtuous) families will come to an end and I shall enumerate all of these kings that are narrated in the Bhaviṣya. Other kings different from the preceding that will arise, such as Kṣatras (? of the kṣatriya class), Pārāsavas (the Pārsī tribe or persons born of a śūdra father and brāhmaṇa mother?), Śūdras (as kings) and others that are foreigners, the Andhras, Śakas, Pulindas, Cūlikas, Yavanas, fishermen, Abhīras and Śabaras and others born of Mleccha (tribes)—these kings I shall proclaim one after another in order and by name. Out of these (two groups) the first is Adhisimakṛṣṇa who is now alive and I shall speak of the kings of his line that are narrated in the Bhaviṣya. This passage makes it perfectly clear that in the ancient Bhaviṣyat kings of the three lines of Aila, Aikṣvāka and Paurava to the last of them were enumerated (vide note 1382), but that the later kings like the Andhras and Śakas were not enumerated thereiṇ I agree with Pargiter (p. VIII in Intro. to 'Purāṇa Texts' &c.) that the words 'Bhaviṣye kathitān' in Matsya (50. 77) or 'Bhaviṣye pathitān' in Vāyu (in 99.292) refer to the descendants mentioned in the Bhaviṣya and that they do not simply mean 'mentioned in future.' I fail to understand, however, why he regards 'Bhaviṣyat' as a perversion of 'Bhaviṣya'. Bhaviṣyat is as good a word as Bhaviṣya, being employed in several passages such as Varāha (177. 34), Matsya (53. 62).

Pargiter probably wants to identify the Bhaviṣyat of Āpastamba with the Bhaviṣya of later timeṣ There is no evidence except the name to identify the two. It, therefore, appears that the extant Purāṇas base their narrative as to the three lines of Aila, Aikṣvāka and Paurava kings on the materials contained in the ancient Bhaviṣya and as to other lines and comparatively later kings they relied on other materials or oral traditions that they could collect. This inference receives support from other circumstanceṣ The extant Purāṇas quote verses called Anuvarśa ślokas or gāthās about ancient kings, such as Kārta-virya (in Vāyu 94. 20, Matsya 43. 24, Brahmāṇḍa III. 68-20, Brahma 13.17) and also about the last scions of the Aila and Aikṣvāka lines, viz. Sumitra and Kṣemaka. But so far as kings of comparatively later dynasties such as those of the Andhras, Śuṅgas and others are concerned, no such gāthās or ślokas are quoted in the Purāṇaṣ There is absolutely no evidence to hold, as Pargiter does (p. XIII of 'Purāṇa Texts' &c.), that the ancient Bhaviṣya contained a reference to the Guptaṣ The ancient Bhaviṣya was composed before Āpastamba (i. e. before the 4th or 5th B. C.) and hence originally could have contained no reference to the Guptas (whose rule began about 320 A. D.). The Matsya does not refer to the Guptas and mentions only the

says that this is scarcely more than an arbitrary assumption (p. XLVIII of Kirfel's Intro. and J. V. O. I. vol. VIII, p. 31). The present author may tentatively accept most of the conclusions of Kirfel, but he differs as stated above from Kirfel's view that the five characteristic topics (sarga &c.) are the oldest constituent parts of the whole Purāṇa Literature. A lengthy discussion of the age of the Purāṇas is not very relevant to the subject of this sectioṇ But it would not be entirely out of place if the author said a few words thereoṇ The author's position about the Purāṇas is as follows: We know hardly anything about the Purāṇa mentioned in Atharvaveda, the Śatapatha and the ancient Upaniṣads; but this much is clear that Purāṇa had attained a status of sacredness like the Vedas and was closely associated with Itihāsa even in Vedic timeṣ This is the first stage in the evolution of Purāṇa Literature, but we know nothing about the contents of the Purāṇa in those ancient timeṣ The Tai. Ā. mentions Purāṇāni; therefore in its time there must have been three Purāṇas at least. As Āp. Dh. S. quotes four verses from a Purāṇa and expressly names Bhaviṣyatpurāṇa, it follows that by the 5th or 4th century B. C. at the latest, there was in existence a Bhaviṣyat-purāṇa and other Purāṇas or a Purāṇa, that contained sarga and pratisarga and some Smṛti material. This we may regard as the 2nd definite stage of Purāṇa Literature, of the contents of which we have some traces at least. The Mahābhārata quotes hundreds of verses (called ślokas, gāthās, anuvamśa ślokas), some of which have a bearing upon Paurāṇika subjects and have a Paurāṇika ring. Some examples may be citeḍ The Vanaparva ⁷¹ quotes two verses about the spiritual prowess of Viśvāmitra and about his assertion that he was a brāhmaṇa. The Anuśāsanaparva ⁷² quotes certain gāthās said to have been sung by the pitṛs about the importance of a son or sons, which agree in letter and spirit with verses on the same subject in the Purāṇaṣ In the Udyoga-parva⁷³ Bhīṣma is said to have addressed a verse to Paraśurāma that was sung by Marutta and declared in anuvaṃśaślokas sung by people described as Paurāṇika (in Vāyu 70. 76, 88. 114–116, 88. 168–169, or as purāvidāḥ or purāṇajñāḥ in Brahmāṇḍa III. 63. 171). Vāyu (93. 94-101) mentions several gāthās as sung by Yayāti, most of which occur in the Ādiparva 75. 50–53 and 85. 12–15, Brahmāṇḍa III. 68. 96–103 and in other purāṇas also. It is quite possible that these gāthās and ślokas said to have been declared by those who knew Purāṇas were taken from the Purāṇa or Purāṇas known to Āpastamba. As Yāj. I. 3 regards Purāṇa as one of the sources of dharma, it follows that some Purāṇas containing smṛti material must have been composed a good deal before that smṛti i. e. before the 2nd or 3rd century A. D. at the latest. This is the third stage in the evolution of Purāṇaṣ It is difficult to say when the extant Matsya was originally composed but it was revised about the middle (or close) of the 3rd century A D., since it speaks of the downfall of the Āndhra dynasty, but does not refer to the Guptaṣ But it is possible that the original kernel of the Matsya may be earlier than this by a few centurieṣ The same applies to Vāyu and Brahmāṇḍa. The Vāyu and Brahmāṇḍa also were compiled or added to about 320–335 A. D., since they refer to the Guptas but do not name any Gupta king. These two in their present form may also be referred to this third stage. Most of the Mahāpurāṇas were composed or completed in the period from the 5th or 6th century A. D. to the 9th century A. D. This represents the 4th stage in the evolution of Purāṇa literature. The Upa-purāṇas began to be compiled from about the 7th or 8th century A. D. and their numbers went on increasing till about the 13th century A. D. or even later. This is the last phase. Thus there is enough evidence to hold that the Purāṇas began to influence Hindu society a few centuries before Christ, that their influence continued in full force till the 17th or 18th century A. D. and that it continues to some extent even now. After the 9th century no further Mahāpurāṇas appear, but additional matter appears to have been unscrupulously inserted in several Purāṇas, the worst example of the kind being the third part of the Bhaviṣya, which contains stories of Adam and Eve, of Pṛthvirāj and Jaicandra, Taimur, Akbar, Caitanya, Bhaṭṭoji, Nadirshah and so oṇ

The word ‘purāṇa’ occurs over a dozen times in the Ṛgveda, is an adjective and means ‘ancient, old’. The Nighaṇṭu (III. 27) mentions six Vedic words as having the sense of ‘purāṇa’ viz. pratnam, prādīvāḥ, pravayāḥ, sanemi, pūrvyam, ahnāya. Yaska (Nirukta III. 19) derives the word ‘purāṇa’ as ‘purā navam bhavati’ (what was new in former times). The Ṛgveda does not contain the word ‘purātana’ (ancient). Purāṇa may be a very old form of ‘purātana’ through the intermediate form ‘purāṇa’. From meaning ancient the word ‘purāṇa’ came to mean a work dealing with ancient tales; it became a noun and was applied in the times of the Atharvaveda, the Śatapatha and the Upaniṣads to a class of works containing ancient taleṣ When purāṇa came to mean a work dealing with ancient tales, to speak of a Bhaviṣyat-purāṇa was apparently a contradiction in termṣ That contradiction was probably not minded or was ignored by the thought that works that narrated old tales gradually came to include comparatively recent ones and had therefore to adopt a prophetic style of composition with reference to the latter.

The Vāyu ⁷⁵ derives the word ‘Purāṇa’ from ‘purā’ (in ancient times, formerly) and the root ‘an’ (to breathe or live), and therefore according to it the word literally means ‘that which lives in the past’ or ‘that which breathes ancient times’. The Padmapurāṇa propounds a slightly different etymology, viz. "it is called Purāṇa because it desires or likes the past" from 'purā' and the root 'vaś' (to desire or to like).

The question why the extant Purāṇas do not narrate the traditions about the dynasties of the Guptas and their successors cannot be satisfactorily answereḍ One reason may be that the original kernel of some Purāṇas like the extant Matsya were compiled before the Guptas rose to power, while others like the Vāyu and Brahmāṇḍa were compiled while the Gupta rule was in its infancy. Another reason may have been that in the 5th and following centuries when many of the extant Purāṇas were compiled northern India was very much disturbed by the invasions of foreigners like the Hūṇas, ⁷⁶ Toramāṇa and Mihirakula, numerous sects and schisms had arisen, Buddhism had become powerful and therefore the first task of the intelligent and devout followers of the Veda was held to be to wean the common people away from schisms like that of Buddha, to lay down the foundations of a new ideology among the masses and to emphasize and assimilate as many of the doctrines of the sects and schisms as possible with their ancient traditions and practiceṣ The intelligent classes, therefore, emphasized the importance of such virtues as ahiṁsā, satya, bhakti and of vratas, pilgrimages, śrāddhas and dānas and were probably not in a mood to record the names of foreign conquerors or of small chieftains fighting with each other and unable to repel the cruel invaderṣ For the absence of references to the dynasties of the Guptas and their successors, Pargiter blames the brāhmaṇas in the following characteristic passage (AIHT chap. 4 p. 57) 'the absolute dearth of traditional history after that stage is quite intelligible, both because the compilation of the Purāṇa had set a seal of tradition and because the Purāṇa soon passed into the hands of brāhmaṇas who preserved what they had received, but with the brahmanic lack of the historic sense, added nothing about the later kings'. Supposing for argument that brāhmaṇas lacked the historic sense, Pargiter's opinion appears to be entirely one-sideḍ He assigns no reason why the sūtas, whose business was to record and preserve historic tradition (as he himself says on p. 58 of AIHT), did not stick to their business and did not continue to compose genealogies of later kings and to add further items of history to already existing recorded tradition, nor does he explain how the sūtas could be ousted or allowed themselves to be ousted from their age-old avocation by the brāhmaṇaṣ It is probable that foreign dynasties like that of Kaniska and the Hūṇas did not encourage the sūtas who had sunk low in the social scale and the sūtas probably became Buddhists, as Buddhism with its Jātaka stories gave to all persons following a bardic profession sufficient scope for earning their livelihooḍ

The legends about Vyāsa and Sūta may be briefly considereḍ The Purāṇas declare that Vyāsa was the son of Parāśara, was also called Kṛṣṇa Dvaipāyana and was an incarnation of Viṣṇu ( of Brahmā also in Vāyu 77. 74-75 and of Śiva in the Kūrma II. 11. 136 ). He was called Dvaipāyana⁷⁷ because he was born on an island ( dvīpa ) in the Yamunā river and Kṛṣṇa because he was of dark complexioṇ His mother was Satyavatī and son was Śuka. He was called Vyāsa because he is supposed to have divided or arranged the one Veda into four parts ( from the root ‘as’ 4th conjugation ‘to throw’ with the upasarga ‘vi’ ). He instructed four disciples in the four Vedas, viz. Paila, Vaiśampāyana, Jaimini and Sumantu respectively in Ṛgveda, Yajurveda, Sāmaveda and Atharvaveda. His 5th disciple was Sūta Romaharṣaṇa to whom was imparted Itihāsa-Purāṇa. The son of Sūta was Sauti who narrated the Mahābhārata to Śaunaka and other sages in the Naimiṣa forest. It was believed that whenever dharma and Veda declined Vyāsas were born for the benefit of men ( Brahma 158. 34 ). The Kūrma ( I. 52. 1-9 ) sets out 27 names of different Vyāsas, while Vāyu ( 23. 115-219 ), Brahmāṇḍa ( II. 35. 116-125 ), Viṣṇu⁷⁸ ( III. 3. 11-19 ) enumerate the names of the 28 Vyāsas of the 28 Dvāpara ages of the Vaivasvata Manvantara (which is the current one). How Vyāsa put together the Purāṇas is described in several Purāṇaṣ He who was an adept in the meaning of Purāṇa composed a Purāṇa-saṃhitā from (the material supplied by) tales, episodes, gāthās (stanzas) and correct ascertainment or descriptions of Kalpas⁷⁹. This shows that, while the Vedic texts were preserved with unparalleled care by the brāhmaṇas, the very ancient Itihāsa-purāṇa, though called the 5th Veda, was not kept intact with care similar to that bestowed on the four Vedas and that this fifth Veda was allowed to be inflated by fresh additions from time to time.

In connection with Vyāsa’s legendary role as an arranger of the Veda, Pargiter has a theory of his own which must be briefly noticed and examineḍ He develops that theory on pp. 9-10 of A. I. H. T. He refers to the Ṛgveda as the greatest brahmanical book, says that it is a compilation of hymns composed by many authors and is arranged according to certain principleṣ His words are ‘It (Ṛgveda) must manifestly have been compiled and arranged by some one or more persons, yet Vedic Literature says absolutely nothing about thiṣ The brāhmaṇas cannot have been ignorant about it, for they professes to know and declares the names of the authors of nearly all the hymns and even of single verses, yet it ignores all knowledge of the person or persons who afterwards compiled and arranged these hymnṣ To suppose that when it preserved the earlier information it was ignorant of the later work in so vital a matter is ridiculouṣ’ From this silence in the Vedic Literature about the persons that compiled and arranged the Ṛgveda, Pargiter at once jumps to the positive and emphatic conclusion, as is usual with many western writers on Sanskrit Literature and Indology to argue from silence, that ‘Vedic Literature has deliberately suppressed all information on these matters’ (AIHT p. 9 ). He refers to the fact that the Mahābhārata and Purāṇas are full of Vyāsa and repeatedly declare that the Veda was arranged by Vyāsa and points out that Vedic Literature is remarkably reticent about Vyāsa Parāśarya ( who is mentioned as a pupil of Viśvaksena in the Vamśa list at the end of the Sāmavidhāna Brāhmaṇa and in the Taittirīya āraṇyaka ) and then he reiterates his charge of the conspiracy of silence about Vyāsa ( AIHT, p. 10 ). Pargiter is ready with a reason for the supposed conspiracy of silence viz. ‘ the brāhmaṇas put forward the doctrine that the Veda existed from everlasting; hence to admit that any one had compiled or even arranged it struck at the root of their doctrine and was in common parlance to give their whole case away ’ ( ibid. p. 10 ).

Several objections can be raised against the positive assertion of deliberate and fraudulent suppression inferred from mere silence. In the first place, Pargiter is very loose in his statement of factṣ Pargiter totally ignores that even in the Ṛgveda itself, r̥k verses, yajus texts and sāman chants are differentiateḍ To give only a few references, vide for R̥ks ( Rg. II. 35. 12, V. 6. 5, V. 27. 4, V. 44. 14–15 ), in both verses of the last r̥k and sāmans being separately mentioned; for yajus vide Rg. V. 62. 5, X. 181. 3 ); for sāman chants vide Rg. II. 43. 2 ( udgāleva śakune sāma gāyasi ), VIII. 81. 5 ( śravat sāma gīyamānam ), VIII. 95. 7 ( śuddhena sāmna ).

The epic and purāṇic texts indicate that Veda was thought to be originally one, but was arranged into four groups, that the four groups of texts were entrusted for preservation and propagation to four different disciples of Vyāsa. The Ṛgveda has two arrangements, one into maṇḍalas and sūktas and the other into aṣṭakas, adhyāyas and vargaṣ The Taittirīyasamhitā and Atharvaveda are arranged into kāṇḍaṣ Not a word is said in any of these accounts to which Pargiter refers about picking up hymns already existing or their being arranged in maṇḍalas or adhyāyas or kāṇḍas by Vyāsa. Further, the reason assigned for the supposed deliberate suppression of the name of the arranger of the Ṛgveda is quite flimsy, not to say ridiculouṣ Every hymn of the Ṛgveda or every mantra has a r̥ṣi who by the ancient Indian tradition was not the author (as Pargiter puts it) but only the seer. It is clear from Brāhmaṇa texts, Upaniṣads and smṛtis that from very⁸⁰ ancient times it was a very strict principle that no one should teach or repeat in japa or employ a mantra in a sacrifice without knowing the ṛṣi, the metre, the deity and the use (viniyoga) of it and dire consequences were declared to follow for him who was remiss in these matterṣ Hymns and mantras were arranged in different groups for different religious rites and solemn sacrifices or for other purposes (such as śāntis). It is not necessary to remember who arranged the required mantras for rites, sacrifices and other purposeṣ The Brāhmaṇa texts and śrauta sūtras prescribe the manner of the employment of the same mantras for various purposes and the Anukramaṇīs contain the names of the seers, metres and the deities of hymns and individual mantras also. As every mantra of the Veda was supposed to be only seen by a sage and as eternal, the mere compilation of them in one or more series or the mere arrangement of mantras or hymns in different groups for different purposes did not at all affect or interfere with the eternity of mantras or hymnṣ Pargiter's so-called reason for suppressing the name of the arranger of the Veda is simply no reason at all.

Pargiter did not stop to consider possible explanationṣ One of the most plausible is now put forth. The Mahābhārata and the Purāṇas (a very extensive literature) were attributed to Vyāsa, who, as shown above in ṇ 1390, had come to be looked upon as Viṣṇu or as an avatāra of Viṣṇu. The four Vedas and the several different śākhās (recensions) of each Veda were well knowṇ By a sort of post fact explanation, the distribution of the Veda into four main groups was claimed to have been brought about by the divinely inspired Vyāsa, whose Purāṇas are, as will be apparent from note 1349 above and as will be shown in more detail later, glorified as even anterior and superior to the Veda. The eternality of the Veda had to be maintained and at the same time Vyāsa was to be glorifieḍ The easiest way was to proclaim that Vyāsa, the author of the great Epic and of the distribution of Purāṇas into eighteen, was also responsible for the division or arrangement of the Veda. If all this glorification of Vyāsa occurred in the centuries immediately preceding and following Christ, how could the supposed arranger of the Veda be mentioned in the early Vedic Literature, which, most scholars agree, was closed some centuries before Buddha (i.e. before the 6th century B.C.)? No body claims that the arrangement into maṇḍalas or aṣṭakas or kāṇḍas is eternal. It is only the hymns or mantras that are claimed to be eternal. Even the padapāṭha of the Ṛgveda is declared to be non-eternal and is ascribed to Śākalya whom the Nirukta criticises in VI. 28. Viśvarūpa on Yaj. III. 242 expressly states that the pada and krama arrangements of the Veda are due to human authorṣ⁸² This theory explains all matters and has far better claims to be accepted than Pargiter's bold ascription of fraudulent suppression inferred from (a supposed) silence.

The Purāṇas do not speak with one voice about their own origin and transmissioṇ After declaring that Vyāsa entrusted the preservation and propagation of the Purāṇas to Sūta, the Vāyu and other Purāṇas contain a somewhat different versioṇ The Vāyu narrates (61. 55–61): Sūta had six disciples, viz. Sunīti, Ātreya, Akṛtavraṇa Kāśyapa, Agnivarcas Bhārdvāja, Mitrayu Vasiṣṭha, Sāvarṇi Saumadatti, Suśarman Sāṁśapāyana. Three of these, viz. Kāśyapa, Sāvarṇi and Sāṁśapāyana prepared new Purāṇa-saṁhitās and Sūta's own was the 4th and the original one. All were divided into four kāṇḍas, contained the same sense (matter), but differed in their readings as the śākhās of the Veda differ. All had four thousand verses except the saṃhitā of Sāṁśapāyana. These four are said to be the basic saṁhitās (in Brahmāṇḍa II. 35. 66) or original saṁhitās (in Vāyu 61. 58 pūrvasaṁhitās). The Brahmāṇḍa (II. 35. 63–70) has the same account in almost in the same wordṣ Viṣṇu (III. 6. 16–17), Agni 271 (11–12) are briefer but agree in the main with Vāyu. The Bhāgavata (XII. 7. 4–7) differs from all these to some extent. That there is some substance in this story appears from stray passages in several chapters of the Vāyu (56. 1, 60. 33–34, 62. 1, 89. 16) and the Brahmāṇḍa (II. 34, 34, II. 36. 1 &c.), where Śaimśampāyana is the inquirer and Sūta replieṣ The personality of Sūta is somewhat of an enigma in the Mahābhārata and the Purāṇaṣ Sūta is called Romaharṣaṇa⁸³ or Lomaharṣaṇa because he made the hair (roman or loman) of his audience bristle or stand erect by his touching and eloquent speech. In the Skanda it is stated that he was so called because and when he was being instructed by Dvaipāyana. One meaning of the word sūta is 'chrioteer' and another meaning is 'a person of a mixed pratiloma caste born of the union of a brāhmaṇa woman with a kṣatriya male' and the cognate word 'Māgadha' means 'one who is born of a pratiloma union between a vaiśya male and a kṣatriya female' (vide Manu X. 71, Yāj. I. 93-94). The Arthaśāstra of Kauṭilya says⁸⁴ the same thing about sūta and māgadha, but adds that 'the sūta and māgadha mentioned in the Purāṇas are different from these, because he (the sūta) is distinguished from (ordinary) brāhmaṇas and kṣatriyas'. Kauṭilya means that in his days sūta and māgadha were pratiloma castes, but the sūta and māgadha mentioned as the first reciters of the Purāṇas are a category apart, that they do not belong to the pratiloma castes and are both distinguished from brāhmaṇas and kṣatriyas (i. e. divine person). The Vāyu (1. 26-33 and 62-147 ff), Padma (II. 27. 65-87, V. 1. 29-32), Brahmāṇḍa (II. 36. 158-173), the Skanda (Prabhāsakhaṇḍa 1. 8) say⁸⁵ that in the sacrifice of Pitāmaha (i. e. Brahmā) Sūta sprang up as a partial avatāra of Viṣṇu oṇ The day when soma juice was extracted and Māgadha also thus arose. It is further stated in the same Purāṇas that the havis (offering) meant for Indra (symbol of the kṣatriya class) got mixed with that meant for Bṛhaspati (symbol of brāhmaṇa class and learning) and that the sūta was born just at the time when the mixed-up offering was presented to the godṣ From this the sūta (in later times) had duties similar to those of the original sūta and it was said that the sūta is offspring of the mixed union of brāhmaṇa (woman) with kṣatriya (male). Then another story is grafted on to this (in Vāyu 62. 147 ff, Brahmāṇḍa II. 36. 170-173 and in others) that the original ⁸⁶ Sūta and Māgadha sang the praises of king Pṛthu, son of Vena, who being pleased, made a present of the country of Anūpa to Sūta and the country of Māgadha to Māgadha and since that time sūtas and māgadhas sing the praises of kings and awaken them in the morning with blessingṣ The Vāyu itself, ⁸⁷ however, says (1. 33-34) that sūta was born at the time of the extracting of soma juice in the sacrifice of Pṛthu Vainya.

The authors of the extant Vāyu and other Purāṇas are conscious that the sūta and māgadha in their times had no adhikāra for Veda, that the business of the sūta was to note the dynasties of gods, sages and kings that are found in the Itihāsa and Purāṇaṣ They felt scandalized by the fact that great sages like Śaunaka were said to have learnt Purāṇas from the sūta who in their times belonged to a pratiloma caste, about which Gautama, Viṣṇudharmasūtra and Kauṭilya himself lay down that pratilomas are like śūdras, and are condemned by āryas and are beyond the pale of the usual duties of brāhmaṇas and kṣatriyas ( such as upanayana, Veda study, teaching &c. ). Learning by a brāhmaṇa from a kṣatriya was considered even in Upaniṣad times as contrary to the natural order of thingṣ Vide the words of king Ajātaśatru to Gārgya Bālāki⁸⁸ quoted below. Therefore, to account for the position of Sūta as instructor in Itihāsa and Purāṇa of great sages like Śaunaka, the story of Sūta's birth was invented and he was placed in a separate category by himself. This must have occurred some centuries before Kauṭilya who was aware of the low position of sūta and māgadha and differentiates the Paurāṇika sūta from the pratiloma sūta and māgadha. One need not accept the divine character of Sūta, one has only to understand that in very ancient times brāhmaṇas could, without any qualms and without loss of prestige, learn about legends from a sūta, but that in the times when the extant purāṇas were compiled the position had entirely changeḍ

Next to Pargiter and Kirfel, one must mention the name of Prof. R. C. Hazra who has bestowed much labour and thought on the purāṇas in general and on individual purāṇaṣ One cannot but feel high admiration for Prof. Hazra's industry, patience and enthusiasṃ One regrets, however, to find that he has developed a tendency to assign rather more ancient dates to the extant Purāṇas and Upapurāṇas than the available evidence would warrant. Besides, he has been so much engrossed in the study of Purāṇas that he sometimes scents a reference to Purāṇas where none in fact existṣ For example, in ‘Purānic records on Hindu rites and customs’ ( p. 6 ) Prof. Hazra observes that Vijñāneśvara tells us of Hārīta’s reference to the opinion of Purāṇas in prescribing penance in normal circumstances to those who eat the food dedicated to the patriarchṣ The words of the Mitākṣarā⁸⁹ of Vijñāneśvara are quoted in the note below and clearly show that there the word ‘purāṇeṣu’ means ‘Śrāddhas called purāṇa’ and has nothing to do with Purāṇa workṣ Another defect noticed in Prof. Hazra’s writings is that he sees too much meaning in simple words and phrases and is not cautious in his conclusions as a scholar of his standing and experience should be. In a recent paper on ‘the Aśvamedha, the common source of origin of the Purāṇa Pañcalakṣaṇa and the Mahābhārata’ in ABORI, vol. 36 (1956) pp. 190-203, he cites the Atharvaveda verse quoted above (p. 816, n 1325 ) in which ṛk and sāma verses are mentioned separately and ‘purāṇa’ (purāṇaṁ yajuṣā saha) is associated with yajus, and says this collocation seemed to him highly significant and that he felt fully convinced that the Purāṇa pañcalakṣaṇa and the Mahābhārata owe their origin to the Aśvamedha sacrifice and especially to its Pāriplava ākhyānaṣ Reasons of available space preclude a detailed examination of this paper. But a few fundamental objections and matters must be mentioneḍ The words ‘purāṇaṁ yajuṣā saha’ should ordinarily mean (as in passages like Devadattaḥ saputra āgataḥ ) no more than ‘Purāṇa and Yajus’. Yaj. 1. 101 provides¹⁴⁰² that after the daily bath, a vedic householder should undertake everyday the japa of portions of the (three) Vedas, the Atharvaveda, Purāṇas together with Itihāsa and of adhyātmiki-vidyā (Upanisads). There is no special meaning here in the association ‘purāṇāni setihāsāni’ beyond ‘Purāṇas and Itihāsa’. One, therefore, fails to understand how the words ‘purāṇaṁ yajuṣā saha’ are highly significant for arriving at the conviction that Aśvamedha is the origin of Purāṇa and Mahābhārata. Then on p.202 of the paper In question Prof. Hazra quotes a passage from ⁹² Śaṅkarā- cārya's bhāṣya on Chāndogya III. 4. 1-2 and completely misun- derstands the great ācārya when he observes "Śaṅkara's use of the word 'rātri' in the plural (in 'rātrīṣu') shows that in his opinion the Itihāsa and Purāṇa were employed every night during the Pāriplava and not merely on the 8th and 9th nights respectively, as the Śatapatha Br. and Śāṅkhāyanasrautasūtra say" (Italics author's). The Aśvamedha sacrifice lasted for a year and listening to the Pāriplava went on for a year, each Pāriplava being a cycle of ten days (or rather nights as the recitation by the hotṛ priest was to take place after the morning, mid-day and evening iṣṭis to Savitṛ were finished). The texts to be recited and the nature of the legends to be narrated on each of the cycle of ten days are fixed and Itihāsa and Purāṇa are to be recited only on the 8th and 9th nightṣ As each cycle was of ten days, there would be 36 cycles of Pāriplava in a year and Itihāsa would be recited on 36 nights in the year and Purāṇa also would be recited on 36 nights in the year. It is for this reason that Śaṅkarācārya speaks of 'Pāriplavāsu rātrīṣu' in the plural, but he does not say that Itihāsa and Purāṇa were to be recited on 'all' nights ( sarvāsu rātrīṣu), while Prof. Hazra represents him as so saying. There is absolutely no warrant in the ancient texts for saying that on each day (or night) of the Pāriplava Itihāsa or Purāṇa was to be recited or that Śaṅkarācārya said anything of the kinḍ The testimony of the Vedāntasūtra ⁹³ is completely against this view of Prof. Hazra. Vedāntasūtra III. 4. 23 refers to certain stories mentioned in the Upaniṣads such as 'Yājñavalkya had two wives, Maitreyī and Kātyāyanī' (Br. Up. IV. 5. 1), 'Pratardana, son of Divodāsa, went to Indra's abode' (Kauṣī- takī Up. III. 1), 'Jānaśruti Pautrāyaṇa was a pious donor giving much wealth to the people and keeping an open house for distributing food' (Chāṇ Up. IV. 1. 1) and remarks that such stories were not to be recited in Pāriplava, since the stories to be recited therein are expressly specified beginning with the story of king Manu Vaivasvata (which was to be recited on the first night of the Pariplava).

Prof. Hazra has recently published (in 1958) ‘Studies in the Upapurāṇas’ vol. I pp. 1-400 on Saura and Vaiṣṇava Upa- purāṇas (in the Calcutta Sanskrit College Research Series 1958). This would be briefly dealt with a little later.

Prof. Ramchandra Dikshitar also has written a good deal on the Purāṇaṣ His writings are beset by the same infirmity that attaches to Prof. Hazra’s work to some extent. For example, in a paper published in the Proceedings of the 13th Indian History Congress (pp. 46-50) on the Viṣṇupurāṇa he first states (p. 46) that he is more concerned with the extant Viṣṇupurāṇa and after pointing out that the topics of vratas, of fasts, of tīrthas, are absent from the extant Viṣṇupurāṇa he concludes that the extant Viṣṇupurāṇa can safely be placed in the 6th or 7th century B. C. Hardly any modern and critical scholar would accept such a date for the extant Viṣṇupurāṇa. Instead of relying on the absence of certain topics he should rather have relied on what it actually contains to find out the probable date of the extant Viṣṇupurāṇa.

In connection with the Purāṇas the author must refer at some length to the Introductory remarks of Ballālasena, king of Bengal, in his Dānasāgara, edited in the B. I. series (1953-1956) by Mr. Bhabatosh Bhaṭṭācārya (three parts of text pp. 1-722 and 4th part an Introduction in English with Indices). Those remarks evince a bold critical faculty rare in our medieval Sanskrit writerṣ He mentions, besides the Gopathabrāhmaṇa, the Rāmāyaṇa, the Mahābhārata, the smṛtis and dharmaśāstras of Gautama, Manu, Yājñavalkya, (count- ing Śaṅkha Likhita as two), Dāna-Bṛhaspati and Bṛhaspati (as separate), Vasiṣṭha and others (in all 28), the Chandoga-pari- śiṣṭa of Kātyāyana, thirteen principal Purāṇas viz. Brāhma, Varāha, Āgneya, Bhāviṣya, Mātsya, Vāmana, Vāyavīya, Mārkaṇḍeya, Vāmanīya, Śaiva, Skānda, Padma and Kaurma and the Upapurāṇas named in Kūrma and Ādi Purāṇas as contain- ing the procedure of (various) dānas, viz. Ādya, Sāmba, Kālikā, Nanda (v. 1. Nandin), Āditya, Nārasiṁha, Viṣṇudharmottara (declared by Mārkaṇḍeya) and the śāstra called Viṣṇudharma (eight in all). He mentions that he drew upon all these for the production of his work on dānas, 1375 in number (verses 11-20 pp. 2-3). Then he mentions certain Purāṇas and Upapurāṇas which he discarded in his work on dānas for various reasonṣ

Some of these remarks are very important and the original verses are set out in the note below.⁹⁴ He states that he did not draw upon the Bhāgavata, the Brahmāṇḍa and the Nāradīya as dānas are absent therefroṃ He did not rely in his work upon the Liṅgapurāṇa, though it is a large work, because its essence, he decides, is the same as the treatment of Mahā-dānas declared in the Matsyapurāṇa; the Bhaviṣyapurāṇa has been assiduously utilised by him only up to the (vratavidhis of ) 7th tithi, but he discards the procedures of the 8th and 9th tithis (of the Bhaviṣyapurāṇa), since they are overwhelmed (tainted) with (the doctrines of )⁹⁵ heretical sects (Tāntrikas, Bauddhas, &c.); both the Viṣṇurahasya and Śīvarahasya, though well-known among people, have not been accepted in this work, since they are considered to be mere compilations; the Bhaviṣyottara (Purāṇa) which is followed in peoples' practices and is not in conflict (with orthodox views) has been excluded from this work, since no indications (evidence) of its authoritativeness could be found; the following are ignored in the Dānaśāgara for reasons stated: three khaṇḍas, viz. those concerned with the tales of Pauṇḍra, Revā and Avanti of the Skanda apart from a part of it that is prevalent (among people), the Tārkṣya (i. e. Garuḍa) purāṇa, another Brahmapurāṇa, another Āgneya (i. e. Agnipurāṇa), a Viṣṇupurāṇa containing 23000 verses, another Liṅgapurāṇa containing six thousand verses; all these have been discarded for various reasons such as the procedure of dīkṣā (initiation of a disciple by a guru in a cult like the Tāntrika or Pāñcarātra or Pāśupata) or of the establishment of an image, heretical reasoning, testing of jewels, stories of the doings of (persons of) false genealogies, treating of such matters as dictionaries and grammar, containing incoherent tales and contradictions, because they lead to the misleading of people by the description of or reference to Love affairs, to those who are buffoons, or are heretics or make their livelihood by displaying some sign (such as matted hair); the Devīpurāṇa has not been utilized in this work, because it is not included in the enumeration of the number of Purāṇas and Upapurāṇas (in various works), because it contains delusive acts ⁹⁶ and because it approves of heretic śāstraṣ Some important conclusions can be drawn from the above mentioned remarks of Ballālasena in his Dānasāgara. Next to the Mitākṣarā, the Kṛtyakalpataru and Aparārka's commentary, the Dānasāgara is among the earliest extant nibandha works the dates of which are nearly certaiṇ⁹⁷ If does not mention the Mitākṣarā, nor Kṛtyakalpataru nor Aparārka. Mr. Bhabatosh Bhaṭṭāchārya is right in his contention that the words 'Kalpa-drumo jaṅgamaḥ' in the 3rd opening verse have nothing to do with Kalpataru of Lakṣmīdhara and that Prof. Raṅgaswamy Aiyangar is wrong in thinking that the verse refers to the Kalpataru (vide Mr. Bhaṭṭāchārya's Introduction to Dānasāgara p. XVIII and note 1).

The principal points that emerge from Ballālasena's remarks on Purāṇas are that he included both Vāyu and Śiva among the principal Purāṇas (sometimes called Mahāpurāṇas), that there were two Purāṇas called Liṅga, Brāhma, Āgneya and Viṣṇu, that the four counterparts bearing these names were not treated as authoritative (one pseudo-Liṅgapurāṇa having 6000 verses and one Viṣṇu having also 23000 verses being unautho-ritative), that he abhorred Tāntrik rites and therefore totally discarded the Devīpurāṇa and parts of Bhaviṣya, that he did not utilize three named sections of the Skanda, that he did not regard the Garuḍa as authoritative. It may be stated here that acc. to the (printed) Kūrma I. 1. 17-20 some Upapurāṇas such as Skanda, Vāmana, Brahmāṇḍa and Nāradīya bear the same names as the Mahāpurāṇaṣ Prof. Hazra relies upon a passage quoted from the Bhaviṣyapurāṇa⁹⁸ by Kalpataru (Brahmacāri- kāṇḍa p. 25) wherein it is said that the appellation Jaya is applied to 18 Purāṇas, Rāmāyaṇa, Viṣṇudharmasāstras, Śivadharma, Mahābhārata, Sauradharmas and Mānavadharmas (Manusmṛti?). I shall discuss the question of Viṣṇudharma-purāṇa later. But I have serious objections against the antiquity and authenticity of this passage. Being quoted in the Kalpataru it may be earlier than 1050 A. D. That is all. In order to glorify the 18 Purāṇas the meaning of Jaya is extendeḍ Jaya is applied only to the Mahābhārata in the latter e. g. Udyogaparva 136. 18-19 and Svargārohaṇika 5. 49 and 51 quoted in note 1369 above. Hence this passage was inserted very late after all 18 Purāṇas had been composed i. e. after the 9th century A. D. Besides the plural ‘Viṣṇudharmādiśāstrāṇi’ shows that many works are meant and not one, i. e. the meaning is that all śāstras dealing with Viṣṇudharmas and the like. If a single work were meant, one expects ‘Viṣṇudharmādiśāstram ca’ and that would not have spoilt the metre. Besides, the Kalpataru itself indicates that this verse about ‘Jaya’ was cited by some authorities as ‘Smṛti’. Therefore, it is doubtful if it is a genuine Bhaviṣya passage. Ballalasena mentions only eight Upapurāṇas on dānas by name (including the four mentioned by the Matsya).

In spite of the very admirable and praiseworthy efforts of Dr. Hazra in the matter of the place of the Upapurāṇas, their contents, the search among numerous msṣ to find out what the text of the several upapurāṇas has been, the present writer must say here once for all that he does not at all agree with most of the dates that he assign to the Sāmba, the Viṣṇudharma, the Viṣṇudharmottara, the Narasiṃhapurāṇa, which are the principal Upapurāṇas he has dealt with in the first volume of ‘Studies in Upapurāṇas’. His dates for the four Upapurāṇas are: Samba between 500 and 800 A. D. (p. 91), the Viṣṇudharma- purāṇa between 200-300 A. D. (p. 143), the Viṣṇudharmottara- purāṇa between 400 to 500 A. D. (p. 212) and the present Narasiṃha-purāṇa between 400 to 500 A. D. To examine all his reasons the present writer would have to write another book. Therefore, he proposes to give only a few illustrations of the way in which Dr. Hazra arrives at his dateṣ But before this is done it is better to mention some of his own findings on the Purāṇas and the four Upapurāṇas mentioned above. On p. 27 he observes that the text of the extant Mahāpurāṇas which are the results of innumerable changes, modifications and interpola- tions made at different times and by different sects is scarcely reliable and can be used with great caution and careful discriminatioṇ I agree with hiṃ But the same or perhaps worse is the case with the Upapurāṇaṣ Prof. Hazra himself says ('Studies' vol. I, p. 23) that after the group of 18 principal Purāṇas had been compiled many sub-systems and sects like the Śāktas and Sauras came into prominence and their adherents interpolated chapters in the 18 established Purāṇas and wrote new and independent works styled Purāṇas in order to propagate their own ideas and that some of these latter came to be called Upapurāṇaṣ The result is that, unless we have critical editions of the critical edition of the Mahābhārata at the BORI in Poona, all chapters and often single verses are suspect. But the task of preparing critical editions of even the principal Purāṇas and some of the Upapurāṇas based upon ancient and medieval msṣ collected from all parts of India would be far more colossal and costly than even the critical edition of the Mahābhārata. Therefore, most chronological conclusions about the dates of Purāṇas and Upapurāṇas and about the borrowings of one Purāṇa from another are just tentative at the most and likely to be set aside by new evidence as long as critical editions of Purāṇas and Upapurāṇas are not available.

Let us now turn to the four Upapurāṇas dealt with at length by Prof. Hazra. About the Sāmba (which is one of the four Upapurāṇas expressly named in the Matsyapurāṇa 53. 60-63) Prof. Hazra observes ('Studies' vol. I. p. 68) that the present Sāmba-purāṇa consists of different units mostly belong- ing to different countries and ages and after analysing in his own way the several chapters of the Sāmba he arrives at the conclusion (on p. 93) that chapters 17, 22 and 23 of the printed edition were added later than 950 A. D., that chap. 44-45 were inserted between 950 and 1050 A. D. and chap. 39-43 and 47-83 were added between 1250 and 1500 A. D. There is at present only one printed eḍ of the Sāmba viz. that of the Venk. Press in 84 chapters based probably only on one mṣ Out of these 84 chapters Prof. Hazra himself finds that 47 chapters are later than 950 A. D., of which 42 were added between 1250-1500 A. D. Prof. Hazra has himself examined several mss of the Sāmba, but they do not come from all over India, many being from Bengal and almost all seem to be late ones (p. 33 last line), being copied in śaka 1764 i. e. 1842 A. D. When more than half of this Purāṇa ranges between 950 to 1500 A. D. according to Prof. Hazra, how can it be usefully employed for chronological purposes ? Nobody can say when the verses about the four Upapurāṇas were inserted in the Matsya, but one can affirm that it was done about the 9th century A. D. or even later. Two dates about Upapurāṇas are certain, viz. that Sāmba is mentioned by Alberuni (Sachau, I, p. 130) who wrote in 1030 A. D. and that the Dānasāgara (verses 13–15 on p. 3), composed in 1069 A.D. mentions eight Upapurāṇas on dānas of which four viz. Sāmba, Narasiṁha, Nandi and Āditya are the same as are mentioned in the Matsya. Therefore, an Upapurāṇa called Sāmba must have been composed a century or two earlier than 1000 A. D. On p. 91 he holds that the Sāmba cannot be dated later than 800 A. D. It is difficult to fall in line with all the assumptions on pp. 90-91 for arriving at this date. What the Sāmba named by Matsya contained beyond the words 'story of Sāmba' is not at all knowṇ Prof. Hazra himself has given up at least half of the printed Sāmba as later than 950 A. D. and there is absolutely no reliable evidence to hold that the remaining portion of the Sāmba is earlier than 800 A. D. or even earlier than 950 A. D.

The next Upapurāṇa is Viṣṇudharma dealt with by Prof. Hazra in ‘Studies’ vol. I. pp. 118–155. There is no printed editioṇ Prof. Hazra (p. 119) refers to six msṣ but he has chiefly used only one mṣ viz. Bengal Asiatic Society's mṣ No. 1670. This Purāṇa has 105 chapters and over 4000 verseṣ Prof. Hazra admits (p. 119) that it has very little of the principal characteristics of a Purāṇa and deals exclusively with the religious rites and duties of Vaiṣṇavaṣ Alberuni names Viṣṇudharma but the verses that he quotes therefrom are found in the Viṣṇu- dharmottara as Bühler pointed out long ago (I. A. vol. XIX. pp. 381–410 and vide table given by Prof. Hazra on p. 208 comparing the two). Prof. Hazra holds (p. 116) relying on two verses quot- H. D. 110. ed below from Viṣṇudharmottara ⁹⁸ that the Viṣṇudharmottara is only the latter part of Viṣṇudharma and that, as he holds, the date between 400-500 A. D. the date between 200 and 300 A. D. (p. 143). The other arguments that he advances are practically worthlesṣ In the present author's opinion those verses can be interpreted in two other but different ways, viz. that the first section of the present Viṣṇudharmottara is so called because it was composed after the Viṣṇupurāṇa, which contains the greatness of Viṣṇu and the dharmas of Vaiṣṇavaṣ He often trots out the theory that, if a work is free from Tantric elements, it must be an early work belonging to the 3rd or 4th century A.D. (see p. 142). The Sarvada- rśanasaṁgraha of Mādhavācārya (which was composed in the 14th century) makes no reference to the cult of Śakti or the system of Cārvāka (a thorough-going atheist), Bauddhas and Jainaṣ The Sarvasiddhāntasaṅgraha also does not refer to the Śaktas or the Tantraṣ But no one can argue that Mādhava is earlier than 4th or 5th century A.D. There may be various reasons for silence. One may abhor a thing and may not refer to it or use it at all as the Dānāsāgara says about the Devipurāṇa. The argument from silence is a slippery one. I challenge the whole basis of Prof. Hazra's date for Viṣṇudharma, viz. the date of the Viṣṇu- dharmottara. This last is a vast work in three sectionṣ It would be proved later that portions of the latter Purāṇa dealing with prognostieations from dreams cannot be placed earlier than about 600 to 650 A. D. For the Viṣṇudharma we have to rely on what Prof. Hazra quoteṣ Chap. 66 quotes the famous words of the Gītā 'whenever there is decline of dharma &c. ' (p. 143 ṇ 94 ) and then the same chap. mentions the incarna- tions of Viṣṇu including Buddha (p. 125 ). Therefore, the mention of the ten avatāras in the context of the words of the Gītā is natural and should not be regarded as spurious simply because it comes in the way of one's pet theorieṣ On p. 144 he quotes ten verses from chapter 66 in which Buddha is described as son of Śuddhodana and his doctrines are stateḍ Prof. Hazra (on pp 145–146) gives four reasons which are quite unconvincing. The Purāṇas mention the avatāras of Viṣṇu in several placeṣ The Bhāgavata in I. 3 names 22 avatāraṣ On p. 150 Prof. Hazra himself quotes a long passage about the evils of the Kali age from the Viṣṇudharma in which occurs a significant half verse ' utkocāḥ saugatāś-caiva Mahāyānanaratās-tathā '. Here not only are the followers of Buddha mentioned but also those who are of the Mahāyāna persuasioṇ On p. 124 Prof. Hazra states that the Viṣṇudharma (p. 124 ṇ 45 ) mentions by name 33 authors of Dharmaśāstras, besides the Saptarṣis and otherṣ Yāj. (I, 4–5 ) mentions only 19 promulgators of Dharmaśāstra ( including himself and holding Sāṅkha-likhita as one ). Both the Viṣṇu- dharma and the Viṣṇudharmottara are not mentioned by the Matsya. Therefore, it must be held that they were not reco- gnized as Upapurāṇas at the time when the verses about Upa- purāṇas were interpolated in the Matsya and were not so recog- nized till at least the 8th or 9th century A. D. The Viṣṇudharma is opposed to what is stated by all writers from the Gṛhya and Dharmaśūtras, by Manu ( III, 128–186 ), by Matsya, Kūrma and other Purāṇas about the qualifications of the brāhmaṇas to be invited at a śrāddha dinner ( vide H. of Dh. vol. IV pp. 384–387 ). It expressly says that squint-eyed, hunch-backed, impotent, poor and diseased brāhmaṇas should be mixed up at a śrāddha along with those who are deeply versed in the Veda.¹⁰⁰ This does not bespeak an early date. On p. 138 Prof. Hazra refers to several nibandhakāras as quoting from Viṣṇudharma such as Gadādhara up to Ballālasena, Aparārka and Kṛtyakalpataru. These do not carry the date beyond 1000 A. D. and besides an examination of quotations in only one work ( Aparārka ) will show that the Viṣṇudharma was an hotch-potch bringing together passages from several sources (vide note ), ¹⁴¹⁰ Prof. Hazra himself holds that the original Viṣṇudharma was appropriated and recast by Bhāgavatas and that many verses quoted from Viṣṇudharma by Raghunandana, Govindānanda and the Madanapārijāta do not occur in the present work (pp. 154-155 ).

Then we come to the Viṣṇudharmottara. The only printed edition is that of the Venk. Presṣ It is a vast work divided into three sectionṣ The first deals with geography, solar and lunar dynasties, astronomy and astrology, gotra and pravara, śrāddhas, Manvantaras, Bharata's fight with Gandharvas and Śatrughna's with Lavana. The 2nd section deals with various aspects of Rājadharma and the third section of the Purāṇa deals with Citrasūtravidhāna and contains several topics such as painting, dancing, music, song, rasas, riddles, dramaturgy, metrics, figures of speech, approaching death, gifts of various kinds, law and justice, hermits and sannyāsinṣ On p. 212 Dr. Hazra places this Purāṇa between 400-500 A. D. Reasons of limitations of space preclude any detailed criticism of Prof. Hazra's treatment. In III. 351.54 Buddha is mentioned as an avatāra which Prof. Hazra says is ‘most probably spurious’ (p. 212) and advances no reasons for this summary judgment. He puts in his usual argument of its non-Tantric ¹⁴¹¹ character and frequent use of the word prādurbhāva and not 'avatāra' which latter, he says, occurs in two places only. Prof. Hazra often attaches undue importance to insignificant detailṣ Whether the word 'prādurbhāva' is used or 'avatāra' is used matters little. The same Purāṇa uses both words promiscuously e. g. Narasiṁha (36. 1, 'avatārāņhaṁ vaksye') begins with the word 'avatāra' and in 39. 1 employs the word 'prādur- bhāva' (ataḥ paraṁ Hareḥ puṇyaṁ prādurbhāvaṁ) and in the colophons of all chapters from 36 to 53 the word 'prādurbhāva' is useḍ The Matsya 247. 1 starts with 'prādurbhāvān purā- ṇeṣu Visņorām itatejasaḥ'. In 247. 19-21 the Matsya employs the word 'avatīrṇaḥ' with regard to Vāmana and Nṛsiṁha and in 246. 4 'avatīrņo jagadyoniḥ ··· Vāmaneneha rūpeņa &c.'. Padma V. 13. 182 speaks of twelve avatāras; Viṣṇu V. 16 also uses the word 'avatāra'. On p. 199 he gives a table of passages that are common to Manu, Yāj., Nārada and Viṣṇudharmottara, on p. 200 another table of passages common to Bharata's Nātya- śāstra and Viṣṇudharmottara, and on page 202 a table of passages common to Matsya and Viṣṇudharmottara and holds that the Matsya borrowṣ This is a most astounding propositioṇ The Matsya does not even mention the Viṣṇudharmottara as an Upapurāṇa; the only sure and earliest date is that the latter is mentioned by Alberuni. That does not carry the matter beyond 900-1000 A. D. The Matsyapurāṇa might have been tampered with. The reasons assigned for his opinion by Prof. Hazra are to say the least, flimsy and unconvincing. It is further to be noted that there are twelve verses that state that some phenomena are not to be held to be utpātas, that are common to Matsya 229. 14-25, Viṣṇudharmottara (II. 134. 15-26) and the Bṛhatsaṁhitā (45. 83-94), which are quoted by the Adbhutasāgara pp. 743-744 as occurring in all these three and in Bārhaspatya. I have dealt with this matter above on p. 768 n 1240. Varāha states that these verses are the summaries ( rṣiputrakṛtaiḥ ślokair-vidyād-etat samāsoktaiḥ ) of the ślokas of Rṣiputra. Therefore, the view that they are borrowed from the Matsya or Viṣṇudharmottara would be out of questioṇ The two Purāṇas do not say whence they have taken theṃ Therefore, it would not be unjustifiable to hold that in both Purāṇas the verses of the Bṛhat-saṃhitā were inserteḍ There is another circumstance that strengthens this last view. The Adbhutasāgara on pp. 493-494 quotes seven verses from the Bṛhad-yātrā of Varāha alone. These verses are quoted by Utpala on Br. S. 47. 22 (where Varāha says ‘sadasat-svapnanimittaṁ yātrāyām svapnavidhiḥ-uktaḥ ) from ‘Yātrā’. These verses (8 in all) are Varāha’s own (as he says) and not taken from any other source. Three of these verses on dreams occur in the Viṣṇudharmottara (II. 176. 9-11 ). They are quoted above on p. 776 note 1254. Prof. Hazra does not appear to have carefully gone into the several works of Varāhamihira and is not right when he says ‘ the Viṣṇudharmottara does not refer to or utilize the works of Varāhamihira ’. The Viṣṇu-dharmottara is an encyclopaedia, while Varāhamihira was a great astronomer and astrologer. He would turn to his predecessors in his own subject if he wanted to borrow and he expressly mentions numerous predecessors (vide pp. 591-594 above) and not to an Upapurāṇa (which in my view did not exist or at least was not recognised as an authority when Varāha wrote in the first half of the 6th century A. D.) Besides, the presumption to be drawn from the nature of the Upapurāṇa is in favour of holding that it borrowṣ

Prof. Hazra deals with the Narasimha-purāṇa on pp. 219-266 of his ‘Studies’ vol. I. The only printed edition is that published in 1911 by Gopal Narayan and Co. (Bombay) in 68 chapters based on three msṣ about which no information is giveṇ This Purāṇa appears to have been composed solely for the glorification of Nara-siṁha identified with Nārāyaṇa. Prof. Hazra has used several msṣ, two, the oldest, being dated in śaka 1567 (i.e. 1645 A. D.), some of the others are not dated and a few are dated so late as 1793 A. D. and 1810 A. D.; and some are written in Bengali script. On p. 322 Prof. Hazra says about one mṣ in Eggeling’s cat. of India Office msṣ that it was copied about 1500–1600 A.D. and its last five folios were supplied in 1789 A. D. No reasons are given why this mṣ should be regarded as being copied between 1500– 1600. This is probably a guess, so all the msṣ consulted are not earlier than the 17th and following centurieṣ Most msṣ do not agree with the printed edition in the number of chapters and also lack certain chapters of the printed editioṇ Prof. Hazra holds (p. 242) on various grounds, all of which cannot be examined here, that the ‘ present Narasimhapuraṇa is to be placed between 400-500 A. D.’ One or two of his characteristic arguments will be noticed here. In chap. 36. 9 (kalau prāpte yathā Buddho bhāven–Nārāyaṇah prabhuh) Buddha is mentioned; Mārkan- deya promises in chap. 36 to narrate stories about eleven avatāras (among whom Balarāma, Kṛṣṇa and Buddha are included) and in chapters 37–54, the stories of all avatāras except Buddha are given and it is added ¹⁰⁴ in chap. of the ten incarnations of Viṣṇu. The devotee of Nṛsimha who always listens to these attains Viṣṇu’. It may be noted that Balabhadra is mentioned in a half verse (36. 8) which is found only in mṣ ‘ga’ of the three msṣ of the printed editioṇ In chap. 53 a few exploits of Balarāma and Kṛṣṇa are narrated in the same chapter and so the half line was inserted later. As the story of Buddha is not given Prof. Hazra holds that the verse in 36. 9 is undoubtedly spurious (pp. 230 and 249). It never occurred to Prof. Hazra that the sectarian zeal of a bigoted Vaiṣṇava might have never cared to give the life of Buddha who was not concerned with (but was antagonistic to) the varṇa system and the Vedas or might have omitted the story of the life of Buddha even if it occurred in the mṣ from which he copieḍ The present author thinks that 36.9 is a genuine verse and enumerates Buddha as an avatāra following the universal belief held in India at least from the 9th or 10th century A. D. that Buddha was an avatāra, but probably bigotry led to the omission of the details of Buddha’s life. Besides, the verse is capable of another interpretatioṇ It is only the stories of ten avatāras (excluding Buddha) that lead a devotee to Viṣṇu. A devotee may worship any one of Nṛsimha, Rāma or Kṛṣṇa, and reach Viṣṇu, but hardly any orthodox work has ever stated that worship of Buddha alone (an avatāra) will lead to attainment of Viṣṇu. Therefore, it is proper to hold that the story of Buddha’s life might not have been given at all or was purposely omitteḍ On, p. 239 Prof. Hazra puts forward a peculiar argument. According to the Narasimha-purāṇa chap. 53. 31, says Prof. Hazra, Kṛṣṇa embodies only a part of Viṣṇu’s śakti ¹⁰⁵ and therefore ‘the Narasimha is to be dated earlier than the present Bhāgavata in which Kṛṣṇa is called bhagavān himself (Bhāgavata I. 3. 28). As the present Bhāgavata is to be dated in the 6th century A. D. the Narasimha is to be placed not later than 500 A. D. Great controversies have raged over the date of the Bhāgavata. The present author holds that the Bhāgavata-purāṇa does not belong to the 6th century A. D. It will be shown below that while Rāmānuja quotes more than a hundred verses from the Viṣṇupurāṇa, which he regarded as one of supreme authority over other Purāṇas (in Vedārthasaṅgraha pp. 141-142 of D. C. edition) he does not quote the Bhāgavata at all in his bhāṣya on the Vedāntasūtra. A bigoted devotee of Narasimha may say that Narasimha avatāra is the perfect avatāra of Viṣṇu, while Kṛṣṇa is only a partial one. But surely that cannot determine the dates of the respective purāṇaṣ What bigoted devotees are capable of saying may be seen from Brahmāṇḍa III. 36. 18-20 quoted later in this section, where it is asserted that the reward of repeating thrice the thousand names of Viṣṇu is yielded by repeating one name of Kṛṣṇa once. The strongest argument against the authenticity and early date of the present Narasimhapuraṇa has not been properly and adequately emphasized or appreciated by Prof. Hazra. The Matsya (53. 60-63) gives meagre details about the four Upapurāṇas expressly named by it, but about the Narasimha it makes one definite statement that it contained 18000 verses (53. 60). The present Narasimha contains only about 3400 verseṣ¹⁰⁶ Therefore, the old Narasimhapuraṇa exists no more and in its place a new one has been substituted, probably containing some of the topics and chapters contained in the old one. We do not know the extent of the Narasimha mentioned by Alberuni. For ought we know he might have referred to the old extensive purāṇa or the new stripling substituted in its place. Even some of the msṣ copied in comparatively recent times drop several chapters of the printed purāṇa; on p. 249 Prof. Hazra holds that chapters 34 verses 43-55, and chapters 61, 65 (on 68 holy places for Vaiṣṇavas), 66, 67 are genuine parts, though dropped in some msṣ Prof. Hazra’s judgments on various matters are most liable to be challenged; but considerations of space prevent any detailed treatment. On p. 252 he holds that in chap. 6 the story of the birth of Vasiṣṭha and Agastya as the sons of Mitra and Varuṇa, the story of Yāma and Yamī are comparatively late additionṣ The story of the birth of Vasiṣṭha from Urvaśī and Mitra and Varuṇa is alluded to even in the Ṛgveda (VII. 33. 11) and in the previous verse (10) of the same hymn Vasiṣṭha, Agastya and Mitrāvaruṇau are brought together. The story of Yāma and Yamī occurs in Ṛg. X. 10. The Narasimha itself states (in chap. 14. 1) that the story of Yāma and Yamī is Vaidiki. The Purāṇa may have added some fringes and embellishments, but the main theme is the same in both. On p. 254-255 and note 330 Prof. Hazra was constrained to admit that the Narasimha- purāṇa had been revised more than once, remarks that Hemādri had a more extensive purāṇa before him than the printed one and that a comparison of the verses quoted from the Narasimha- purāṇa by the Madanapārijāta pp. 301-302, Smṛtitattva on āhnika (vol. I. p. 411) and Nityācāpradipā (vol. I. p. 617) show that the text of the Narasimha used by Raghunandana and Narasimha-vājapeyin agrees more with the printed text than the text used by Madanapāla (about 1375 A. D.), less than 200 years before Raghunandana.

The extant Narasimhāpurāṇa is an insignificant one as compared with the principal Purāṇas like Vāyu, Matsya, Viṣṇu; the Sāmba is not quoted at all by Aparārka, the Viṣṇudharma only twice and even the Viṣṇudharmottara is quoted by Aparārka only 7 times for about 25 verses in all. The very extensive digest Kalpataru (first half of 12th century A. D.) contains hardly any quotation from Viṣṇudharmottara on vratas (though the latter devotes about 1600 verses to vratas), nor is it quoted in the sections on brahmacāri and other kāṇḍas, while hundreds of verses are quoted from Matsya, Brahma, Bhaviṣyā, Mārkaṇḍeya and other Purāṇaṣ Aparārka in his huge digest of 1252 printed pages quotes the Narasimhāpurāṇa only 9 times and in all about 30 verses only, of which about half are in respect of sannyāsa, while the Brahmapurāṇa is quoted 68 times for hundreds of verses, the Matsya is quoted over 50 times for several hundred verseṣ The Tulasī plant is now sacred to the devotees of Viṣṇu, but it does not figure in early Paurāṇika or other literature. It is mentioned in chapters 31. 87 ( in a prose passage ) and 34 ( 19 and 23 ) of the printed Narasimhapurāṇa. As usual Prof. Hazra holds these references are due to later revisions and should not be held to suggest a late date for the extant Narasimhamāhātmya is very much in evidence in Padma e. g. in IV. 94. 4–10, V. 58. 109ff and V. 59, VI, 24–43. The very late part of Padma VI ( chap. 98–107 ) contains the story of Jalandhara, whose wife Vrndā is said to have become Tulasi. Prof. Hazra fails to realize the full force of his own admission that the Narasimha was revised several times and that writers of the 16th century had a version before them different from that before the writers of the 13th or 14th century. In the author's opinion the whole Purāṇa becomes suspect and that by mere diplomatic criticism one would not be able to support beyond doubt any chronological conclusion derived from the present truncated and garbled version of the original Narasimha.

So much space had to be devoted to an examination of Prof. Hazra's work on the Upapurāṇas for cogent reasonṣ His is the latest and most elaborate treatment of Purāṇas and particularly of Upapurāṇas, on which he has bestowed immense labour. His opinions, particularly about the dates of the Upapurāṇas, are likely to be taken as the last word on the subject and even his unfounded surmises are likely to be taken as well-established conclusionṣ Modern Sanskrit scholars have sad experience about such matterṣ Max Muller, working back from the date of Buddha and assigning arbitrarily two hundred years each for three periods viz. the upaniṣad period, the Brāhmaṇa period and the saṃhitā period, inferred that the composition of the Vedas was to be assigned to a period about 1200 B. C. He no doubt said that his periods were the minimum and that his was a pure surmise. Yet most writers who have not themselves made a deep study of the Veda assign the Vedas to about 1200 to 1400 B. C. even after the lapse of about one hundred years from the time when Max Muller launched his surmiseṣ For example, Prof. Toynbee in the table of the time spans of the growth phases of civilizations on p. 758 of volume IX (1954) gives the dates of Epiphany and of breakdown of the Indic civilization as 1375 B. C. and 725 B. C. respectively. The present author does not at all agree with the dates assigned to the Upapurāṇas by Prof. Hazra and with the methods and reasoning adopted for arriving at those dateṣ He had to perform this painful duty even in spite of his regard for Prof. Hazra. There are numerous works, translations and papers on Purāṇaṣ A few alone are mentioned here. Eugene Burnouf translated the Bhāgavata Purāṇa in French in five volumes in 1840 and later yearṣ The translations of Viṣṇu and Mārkaṇḍeya were made respectively by H. H. Wilson and Pargiter; Prof. Kirfel’s Introduction to Purāṇa pañcalakṣaṇa (1927, Bonn) was translated in the Journal of the Shri Veṅkaṭeśa Institute, vol. VII pp. 81-121 and vol. VIII pp. 9-33; Kirfel in Festschrift Jacobi pp. 298-316; K. P. Jayaswal on ‘Chronological tables in Purāṇic chronicles’ in J. B. O. R. S, vol. III pp. 246-262; ‘Purāṇas and Indus Āryas’ and ‘Study of ancient Geography in Agnipurāṇa’ in I. H. Q. (1933), vol. 18 pp. 461 and 470; Prof. Ramchandra Dikshitar’s studies on Vāyu and Matsya and Index to (five) Purāṇas in three volumes; J. R. A. S. 1941 pp. 247-256 and pp. 337-350 by W. Ruben for ‘Puranic line of heroes’; J. A. S. B. for 1938, vol. IV. Article 15 pp. 393 ff; ‘Purāṇas on Guptas’ in I. H. Q. vol. 21 pp. 141 ff and ‘Gupta Inscriptions and Purāṇas’ by Dr. D. R. Patil in B.D.C.R.I. vol. II. pp. 148-165; H. C Raychaudhuri in a paper in the Pro. of the 10th Oriental Conference, pp. 390 ff for discrepancies of Puranic accounts with epigraphic records; paper by Mr. B. C. Majmudar on ‘Origin and character of Purāṇa literature’ in Sir Asutosh Mukharji Silver Jubilee vol. III, Orientalia, part 2 pp. 9-30; a verse from p. 30 of ‘Inscriptiones Sanskrites de Champa et du Cambodge’ refers to the provision made by a king of the 6th century A. D. for the daily reading of the Bhārata, Rāmāyaṇa and Purāṇas;¹⁰⁷ the Penukonda plates of the early Gaṅga king Mādhava II, ( E. I. vol. 14 p. 338 ) speaks of him as knowing ‘the essence of many śāstras, Itihāsa and Purāṇas’ (probably in 6th century A. D.); Purāṇas are mentioned as studied in 578 A. D. ( in E. I. vol. 28 p. 59 ).

In the following notes on individual Purāṇas and Upapurāṇas the author adds a few notes based on his own study of the Purāṇas and (printed) digests ( nibandhas ) of an early age. He desires to emphasize that the earliest works that can be called digests and are available in print cannot be placed earlier than about 1100 A.D. Though there is difference of views among scholars about the exact dates of composition, the Mitākṣarā, the Kṛtya-kalpataru ( which is a regular Digest of various kinds of materials on Dharmaśāstra ) and Aparārka’s work ( though in form a commentary on Yājñavalkya-smṛti is still in the nature of a Digest) are three printed works more or less contemporary and composed between about 1100 to 1140 A.D. The Kṛtyakalpataru on Vyavahāra mentions by name Prakāśa, Hālakalpataru and Pārijāta. Besides, Kṛtyakalpataru (Niyata p. 280) cites the explanation of Parārka (in Ap. Dh. S. I. 17. 26) by the Smṛtimañjarī (of Govindarāja) and also on śrāddha (pp. 46 and 259). It has been shown in H. of Dh. vol. I. that Prakāśa (pp. 306-308), Pārijāta (pp. 308-309), Smṛtimañjarī (pp. 312-314) were works of the digest genre. The Kāmadhenu of Gopāla also appears to have been a digest (H. of Dh. vol. I. pp. 293 ff.). The author Gopāla was a friend of Lakṣmīdhara (vide Introḍ to Dānakāṇḍa p. 49), but, since the latter employs the past tense (cakre) with reference to Gopāla’s work and the present tense (tanyate kalpavṛkṣaḥ) as regards his own work it follows that the Kāmadhenu was composed at least a few years before the Kalpataru. As no msṣ of the Prakāśa, Pārijāta and Kāmadhenu are available it is impossible to say anything about their exhaustiveness or otherwise. But from the summary at the end of the mṣ on the Prāyaścitta section of the Smṛtimañjarī (H. of Dh. vol. I. p. 312 note 714) it appears that it must have been fairly large and was composed on the same lines as the later Kṛtyakalpataru, since it began with Paribhāṣākāṇḍa and Brahmacārisection, then dealt with Gṛhasthadharmas, Dāna, Śuddhi and Āśauca, Śrāddha, then with Vānaprastha and Pravrajyā (corresponding to the mokṣakāṇḍa of Kalpataru) and ended with Prāyaścittaṣ These predecessors of the Kalpataru were superseded by the extensive work of Lakṣmīdhara, who himself receded into the background when the works of Hemādri, Cāṇḍeśvara, Madanaratna, the Viramitrodaya, the Mayūkhas of Nīlakaṇṭha became popular. Even before the Kāmadhenu and probably the Smṛtimañjarī also, Bhoja (about 2nd quarter of 11th century A. D.) composed (or patronised the composition of) several works such as the Bhujabala and Rājamārtaṇḍa which exhaustively dealt with the astrological requirements of saṁskāras from puṁsavana to marriage and also of vratas, yātrā, śāntis, pratiṣṭhā (vide the author’s paper on ‘King Bhoja and his works on Dharmaśāstra and astrology’ in JOR, (Madras), vol. XXIII for 1953-54 pp. 94-127 for five works of Bhoja). So there was nothing very novel in the Kṛtyakalpataru except its exhaustive and logical treatment of all topics and profuse quotations from the Epics and Purāṇaṣ The Mitākṣarā does not quote much from the Purāṇas, but Aparārka and Kalpataru quote them profusely. The Kalpataru quotes about 600 verses from Devīpurāṇa, over 200 verses each from Kālikā, Āditya- purāṇa, Nandipurāṇa and Narasiṃhapurāṇa (all Upapurāṇas) while it quotes none at all from Viṣṇudharmottara. The Kalpataru did not probably regard it as authoritative, though Aparārka and the Dānasāgara utilized it to some extent. The learned editor of the extensive Kalpataru has made great efforts to identify the quotations therein from the Purāṇas, and laid all workers and scholars under great obliga- tionṣ But many quotations have eluded him as will be shown later oṇ He has been assiduous in pointing out Hemādri, Caṇḍeśvara and Mitramiśra have copied wholesale from the Kalpataru. It is not impossible that even the Kalpataru might have done the same to some extent as regards its predecessors such as the Pārijāta, Prakāśa, Smṛtimañjarī, and Kāmadhenu. But as those works are not yet available, no positive conclusion can be drawn about its borrowing for the present

The present author edited 286 verses from the Rājamārtaṇḍa ( containing 1462 verses) on tithis, vratas and utsavas in ABORI vol 36, parts III-IV, 1956, pp. 306-339,). It describes several vratas and utsavas like Indradhvajotthāpana and the work is about 75 years older than the Kalpataru. The Kalpataru is studiously silent as regards Bhoja, though it mentions Kāma- dhenu, Govindarāja, Prakāśa and Halāyudha and it hardly any- where mentions the vratas described by the Rājamārtaṇḍa. The Mitākṣarā does mention Dhāreśvara Bhoja. Probably Lakṣmīdhara did not like that a comparison should be instituted between his treatment of vratas and Bhoja'ṣ

Alberuni's work on India translated by Sachau furnishes us with some data about the chronology of Purāṇaṣ On p. 130 (of Sachau's tr. ) he says he heard the following as the Purāṇas viz. Ādi, Matsya, Kūrma, Varāha, Narasiṃha, Vāmana, Vāyu, Nanda, Skanda, Āditya, Soma, Sāmba, Brahmāṇḍa, Mārkaṇḍeya, Tārkṣya (i. e. Garuḍa ), Viṣṇu, Brahma, Bhaviṣya. It will be noted that he mixes in this list Purāṇas and Upapurāṇaṣ He further states that he had seen only portions of Matsya, Āditya and Vāyu. Then on p. 131 ( of Sachau's tr.) there is another list which was read to him from the Viṣṇu (viz. the 18 principal Purāṇas, Śaiva being substituted for Vāyu ). Then, on p. 229 ( of Sachau's tr. ) he sets out from Āditya some details about the regions below the earth and shows how Vāyu differs from it and on p. 248 he sets out details about Meru from Viṣṇu, Vāyu and Aditya. As he wrote his work in 1030 A. D. one has to conclude that Purāṇas having the names cited by him existed some time before 1000 A. D. at the latest.

Some of the papers of Prof. Hazra and others on the princi- pal Purāṇas and on some of the Upapurāṇas are brought together in one place by Dr. Pusalkar in his work 'Studies in Epics and Purāṇas' pp. 218-225; a few of them are mentioned here. About sixteen papers of Prof. Hazra contributed to various Journals from time to time were included in his work 'Studies in Purāṇik Records of Hindu Rites and customs', which is referred to as PRHR here and H. stands for Prof. Hazra and U. for Upa- purāṇa.