Ancient India - Lecture Notes - Indian History - Dr. B. R. Ambedkar, Study notes of Indian History

Download Ancient India - Lecture Notes - Indian History - Dr. B. R. Ambedkar and more Study notes Indian History in PDF only on Docsity! SUBSCRIPTION RATES, POST FREE IN INDIA, FOR FOUR NUMBERS (TWO YEARS) Rs 7 or Subscriptions to THE MANAGER OF GOVERNMENT PUBLICATIONS, CIVUL LINES, DELHI Dr. B R AMBEDKAR THE Dnr ^ v OPEN UNIVERSITY 'AUTHOR __ ~->-" TITLE of the cultures and civilizations of India exclude all these studies Let us not forget the traditional curriculum, but let us constantly remember the need for enlarging its scope, for enquiring into the material things climate, geography, the changing courses of rivers, rocks and their minerals, the soils of varying fertility which have shaped the destiny of Man and must be understood if we are to understand him Archaeology in India is blest with a wealth and variety of material unsurpassed elsewhere m the world With a proper effort, it can rival that of any other country But a primary need is an enlargement of outlook, a fuller comprehension of the natural sciences as ancillanes to humanistic research a more sustained urge on the part of our students of archaeology and history to supplement the study of the great literature which they have inherited by exploring, at first hand, the Good Earth which is India and is a heritage no less relevant to their enquiry R E M W THE IMAGE OF MAHADEVA IN THE CAVE-TEMPLE ON ELEPHANTA ISLAND By STELLA KRAMRISCH The great rock-cut Siva temple on Elephanta Island is one of the best-known monuments of its kind in India, by reason partly of the grandeur of its sculptures and partly oj its proximity to Bombay Nevertheless the colossal carving of triple-headed Siva which looms out of the rock in the innermost recess of the cave has never been adequately illustrated The dim, unemphatic light which gives it an added quality of power ana mystery does not facilitate the task of the photographer Recently, however, Mr Pnthwish Neogy, one of Professor Stella Kramrisch's pupils in the University of Calcutta, experimented with sun- reflectors and artificial light and supervised the taking of a number of new photographs, some ofwhich are here reproduced (pis I-VII) The triune head which, with its shoulders, rises to a height of 17 feet 10 inches above a moulded base, itself nearly 3 feet high, faces the principal (northern) entrance of the temple at the end of a double row of seven pillars This vista in fact cuts across the major structural axis, which lies east and west, with the hnga-shnne standing free within the western end (fig 7) Nevertheless the huge Siva panel is the focus of the whole design Like the doorways ofthe shrine itself, it is guarded by d\drapdlas or doorkeepers, each some 13 feet high Behind, two pilasters flank a recess 10% feet deep, within which the heads emerge cliff-like from the native rock In the highest relief, they have at the same time a proper attachment to their material environment and something of the independence and imminent mobility of sculpture in the round This is not the conte\t in which to discuss the plan of the temple or the features of the other sculptures which adorn it Suffice it to say that, though not dated bv inscription, the type ofcolumn used, with amid capital and circularfluted shaft springingfrom a rectangular pier, is of a kind which was in use at Bdddmi in the latter part of the sixth century A D and was still in use at Ellora two centuries later On this evidence the cave is ascribed to the seventh or eighth century A D The more subjective evidence of sculptural stvle may be thought to point to the earlier of the two centuries Certainly the bold and vital handling of mass is consistent with the age which, alike in the south and the north of India, saw the first complete fulfilment of medieval art The image is that ofSiva in theform of Mahesa, the Great Lord In thefollowing article, Dr Kramnsch re-interprets the iconography of the work THE great sculpture of Mahadeva is an image of the fully manifest Supreme Siva In the middle is the face of Tatpurusha, the faces of Aghora and Vamadeva are collateral The breadth of the shoulders belongs to the central face , the chest, showing but the slightest modelling, is smooth and young It is as if breathing and holding the breath, as is shown by the ebbing curves of the necklace laid on its raised surface It is hemmed m on the right and the left by the hands , the right hand is raised it is damaged , the left rests on the base and holds a ripe fruit with its point up The shoulders are also those of the lateral faces These are turned against them, and it is on their backs that their hands come to rest The one on the spectator's left, belonging to the wrathful face, has a serpent rearing its head from between its fingers, while the one on the right, belonging to the blissful face, holds a lotus flower and is delicately poised on the shoulder Thus beset with emblems and hands, their fingers pointing upwards in the middle of the bust, THE IMAGE OF MAHADEVA IN THE CAVE-TEMPLE ON ELEPHANTA ISLAND closing m where they rest upon the shoulders, the broad body fills the width of the recess like an altar beset with offerings The middle boldly projecting, the frontal image surges upwards straight and strong as a pillar, with silence on its face and radiance on its crown The heads on the right and the left repeat the ascent of the central pillar, clinging to it collaterally with their high curving crowns, and forming with it the outline of a strong triple arch which firmly binds together the outline of the triple image LLEPHANTA ISLAND PLAN OF MAIN CAVE (T) MAHADEVA (D APDHANAQls'VARA - ^IVA (3) PACVATl IN ATTITUDE OF MANA @) QAVANA UNDER KAILASA () SIVA AS LAKULIS'A (g) NATARAJA - ^IVA (D ANDHAKA5UQAVADHAMUQTI - s'lVA (J) -SIVA SHRINE. () KALVANA3UNDARAMURTI - s'lVA <) GANGADHACA -SIVA FIG 1 The three faces are each steeped in its own mood, a closed world, each silent, unseeing, each turned away from the other, but each blossoming in generous curves round the stem whence they originate and derive their stability They are carved in wideJv sweeping surfaces which bind the crowned triune head in depth just as the triple 'arch' binds it vertically Hands and flowers, serpent, hair and jewels are laid against the smoothness of face and body Each face has its own physiognomy, each crown correspondingly its own ornaments, the hands their respective symbols, but they are upheld and comprised by the power and unity of the total image PLATE ft To face Plate III To face Plate II PLATE HI \liihadc\a of h.k phuuru Inland (Scale ot Uci) PLATE IV To face Plate V The MaluuU MI of I hphanta /\lnnJ To face Plate VI PLATE VII The \1uhath\tiolEItphantal\lanti I Aniadeiti tin (Scale ol teet) THE IMAGE OF MAHADEVA IN THE CAVE-TEMPLE ON ELEPHANTA ISLAND universe 1 , the rosary is a symbol of the re-mtegration of this world 2 In his crown of matted locks Tatpurusha wears the crescent of the moon ; it is its sixteenth digit (ama-kala\ symbol of perfectedness and the power of the Lord (aitvaryd) 3 Aghora is fierce, terrific, like Time (Kdla) itself, serpent and skull belong to him 4 Vamadeva is handsome, he is 'woman's delight' B The Vnhnudharmottara, moreover, knows the five faces of Siva under a double set of names The first series, Hana, Tatpurusha, Aghora, Vamadeva and Sadyojata, belongs to the Supreme Siva in his mental 'body of mantras* , they are the faces of the five-faced linga The second set of names is SadaSiva, Mahadeva, Bhairava, Uma and Nandm * Thus the face in the middle is that of Tatpurusha-Mahadeva , Aghora-Bhairava is on the one side and Vamadeva or Uma on the other The latter face, which is that of the Beauteous God (Vamadeva), is also that of the Goddess Uma, who is Siva's Sakti, inseparably part of his nature The second set of names denotes the faces of the Supreme Siva, fully manifest , he is the 'Great Lord \ Mahesa 7 Thus was carved the image of Mahadeva in Elephanta It is symbol and image, hn%a and murti, in one, the concrete form of the Supreme Siva, fully manifest Around this image, in separate panels, are carved his lila-miirtis, the various forms and actions in which is displayed his divinity But the image of Mahadeva dominates the whole assemblage It is the concrete form of Siva whose linga is worshipped in the adjacent shrine The plan of the rock-cut temple and not only the disposition of the iTla-murtn was thus adjust- ed to the image, the entire hall is laid out in front of it in its deep recess The stage is set for that form (riipa) of His who is beyond form, and who pervades the universe, \\hich thus is His visible body The central pillar, the mulastamhha, rises with the face of Tatpurusha His matted hair forms his cylindrical crown It is clasped by a diadem of wide curves clear cut The serpentine locks, the splendour of the jewel crests which has its symbol in the leonine Face of Glory (kirtnmukha) in front, the crescent moon on its right, the tender, folded awka leaves of early spring, and their full flower-cups all this precious, delicately caned, inti- mately agitated coiffure, full of meaning, closely adheres to the shape of the central pillar and none of its subtle detail is allowed to caress the august serenity of the face The nm of the diadem with its wide curves tightly fits its planes, no curl transgresses nothing disturbs its silence The long curves of the ears are set against locks, a closely \vo\en crop that frames each of the lateral faces, capricious, viciously encircled, tumescent serpentine on Aghora-Bhairava's face, with drooping elegance they touch the cheek of the face on the right Within their symmetry, each of these two crowns is truly a part of the nature of that god whose face bears his name and whose hands hold his insignia Obstinate and coagulate, skull-adorned, the crown of Aghora-Bhairava ascends, then slopes back in a 1 V Dh, III, xlviii, 12 2 V Dh , III, xlvi For its resorption by Time (K&la), VQchaspat)a t s v 1 V Dh , III, xlvin, 17 4 IP , TIF, xii, 36-17 , RUpamanfrna, IV, 5-12 6 IP III, xii, 18, strtMlOsa 8 V Dh , TIT, xlviii, 4-6 7 He is the immanent and primordial cause of the um\erse and its primordial substance (prakritt) The V Dh , ITT, xlviii, 19, says that Mahes"vara is white, as white as prakfin The image of MaheSvara (Mahadeva) must originally have been painted white 7 4NCIENJ INDIA, NO 2 is clear, meaning places where stones are planted or pitched Since the sites contain pottery urns, they are also described by the local people as places with madamadakkaitak, which is a corruption of tnudumakkal-tah, meaning the urns or receptacles (tali) in which the ancients or ancestors are buried This is the name found in early Tamil works from the third century B C to the twelfth century A D The other synonyms found in these are mudu-makkat-cadi (cadi or y<//-jar) imattali (funerary urn or receptacle) or simply tali The name tali itself, indicating a large receptacle for burial, is evidently derived from tal which means 'to lower into the earth' or k to bury', and the original sense in which it is used is of a vessel that is buried An early inscription from Tirukkattalai, near Kalasakkadu where these urns are in plenty, denotes a certain land as andarat tali punt e\ l dry land with buried funerary urns The stone circles are called karkidai (kidai* circle, of kaL stone) in a Tanjore inscription 8 which mentions them along with the burning-grounds of the Vellalar and Parai\ar in the village, and this is noticed by Mr K V Subrahmanya Aiyar ' The oldest extant Tamil work, Tolkdppnam (For 60). has nac/ukallhe stone planted over the grave, the 'menhir' The name pandavakkuji prevalent in some other Tamil districts may be a corruption of mandavar-kuli the burial-pit of the dead, or of bhandakkuli the pit in which the pots (bhanda, in Tamil pandam) are buried Or it may be a corruption of palndavakkuli the burial oi those who performed useless penance, as one of the beliefs in later times was that the Ajivakas or Jamas, whose penance was 'useless' in the eyes of the followers of the Vedic religion were buried m such pots The Todd name ularam means in Tamil the buridl circle (al-aram) The Kdnnadd ndme mornar-nume is difficult to explain Sewell 4 notices the Telugu names Rak\ha\a gullu or gdli the graves of the Rdkshasas and derives the name of Goli village from this Before we proceed to examine the literary references, it is of interest to mention a few facts about the word tali of the early Tamil inscriptions In the inscriptions of the early period, ranging from the seventh to eleventh century, when the Pallavas and Colas ruled the Tamil country and excavated and built stone temples the term tali always denotes the sanctum of the stone temples " Prior to this we ha\e literary and mscnptional evidences to show that temples were built of brick mortar and timber which perished The natural caverns in the hills, v\ith drip-ledges, beds and inscriptions, are the earliest monuments extant These religious resorts were not called tali but pajji or aman-pdli (Jama cave resorts) since they were mostly associated with the Jama ascetics An early inscription (ninth century AD) referring to the excavation of a rock-cut shrine, which is called tali, is published in Inscr of Pudukkottai State, No 18 The chief, who excavated this Siva cave temple, says that 'having excavated the Tiruvalattur hill m the jorm of a tali, he installed the god m it' The expression *m the form of a tali' is significant It evidently probably a derivation from padu, to sleep or die In early Tamil inscriptions of the ninth to tenth centuries the word is used in the same sense in pulhpptujiui, by which name shrines built over the graves of the C6}a kings are referred to Patfai also means the whole or part of the strueture or cdifiee buried or overground, patfatfui is used in the sense padu, lay to rest, or paffu, to die, and atfai (atjakkam or atfakkudal), burial 1 Ibid, No 18, p 18 2 South Indian Inscriptions, Archaeological Sur\ey of India, II (1891) Part 1, No 5, p 54 " K V S Aiyar, Historical Sketches of the Ancient Dekkhan (\9\7), p 359 4 R Sewell, List of Antiquarian Remains in the Presidency ofMadras (Madras, 1 882), I, Topographical List, 57-58 and 60 ' Sec my paper 'A note on Ta)i\ Archaeological Society ofS India, Madras, July 1944 f> M^rali western temple , ki["tali eastern temple vatfa-tafi - northern temple , karrajt (kal-taji) = a temple built of stones, etc 10 PHE MEGALITHK BURIALS AND URN-FIELDS OF SOUTH INDIA denotes that the structure, a cell, was fashioned m the form of a pre-existing type called tali or a rock-excavation This does not occur m &angam Tamil In our quest of the origin of this term, our attention is naturally drawn to the rock-cut tombs of Malabar, which happen to be the only rock-cut monuments prior to these early cave temples in the South The most famous among them is in Tah-paramba, which is the Malayalam equi- valent of the Tamil term tali-parambu. and means in both cases 'the hill containing tali or rock-cut cells' Thus we seem to get here what the term tali originally meant, and the force of the expression "in the form of a talC in the inscription quoted above becomes clear Incidentally we have got here the local Tamil or pre-Tamil name for this type of funerary monument, which is peculiar to the softer latente hills of the west coast Early literature The earliest extant works are of the 'Sangam epoch', which was the Augustan age of Tamil literature They are mainly collections of anthologies of different poets, and the two epics, Mammckhalat and Silappadikaram, arc said to belong to the close of the period, or to a period slightlv later The poems of the Sangam age are realistic and prnna facie trustworthy, they portray a civilization with advanced customs and manners, and relate anecdotes of the kings their wars and their patronage of literary men The most widely accepted date for this literature is the first three centuries of the Christian era, though some of the works or the literary tradition embodied in them may well go back to a century or two before Christ The sheet-anchor for this chronology is the synchronism of the Stf/Jtf<//H Cera king Senguttu\an with Gajabahu 1 of Ceylon who according to the \1ahd\am\a ruled between 1 13 and 135 A D or 173 and 195 A D ' There is perfect concord between the data relating to the Tamil kings and the life of the Tamils as depicted m the Sangam anthologies on one side and the writings of the classical authors of the early centuries of the Christian era (notably the compiler of the Periplus and Ptolemy ) and the finds of Roman coins of the early Imperial period on the other The earliest stratum of Tamil literature shows the influence of the growing religions of the North and the date of this active penetration of Brahmamcal, Buddhist and lama religions into the South may well be placed in the last three centuries before Christ We ha\e Jama caverns of this date m the Tamil country, literary and other eudences of Buddhistic migration to the South and Cevlon, and notices of South India and her trade in the 4rthaw\tra of Kautilya and the Indika of Megasthenes Thus we may take it that what the Sangam literature portravs is the culture which existed in the extreme South between the third century B C and the third century A D The Purananuru* an anthology of 400 verses by different poets composed in different periods of the early Sangam age, has many references to the burial-ground and to urn- burials and a few to cremation The general term used for the grave is kadu or pwangadu which means a waste-land set apart for burial in the wilds near the village It is also called mudukadu or mudiir which means the place for the ancients or old men after death - The term mudukaa'u has persisted m later works such as the Tc\aram (7731) and Nalvah and a Pudukkottai inscription, dated 1237 A D \ refers to the ancestors of the signatories as mudukkal 1 See K A N Sastn, The Cofa*. I, 68-70, The P&nfaan Kingdom, pp 16-24 and V R R Dikshitar, /t s in Tamil Lit and Histor\, pp 7^-74 2 Purum, verses 228, 256, Mammfthalai, VII, line 6^, Pufam, \crsc 156 * Op cit , No 317 11 ANCIENT INDIA, NO 2 It is only later that one finds the two differentiating names itfukdfa 'burial-ground', and tudukadu, 'cremation-ground' In some places the term imam or lmakk&(]iu is also used* where Imam refers to funeral rituals Later commentators on Tamil works invariably explain all these terms as the 'burning-ground' They were influenced by contemporary practices, whereas many of the contexts clearly show that /</wAJ</ alone is meant The puram gives word-pictures of the grave-yard in many contexts, 1 and all these verses may be summarized as follows The grave-yard (kadu\ the place of the ancients, on the outskirts of the town or on the skirts of the hilly ground, an expanse of brackish or alkaline soil overgrown with the tree-spurge (Euphorbia antiquontm or kalh) and other xerophytes and many thorny shrubs, full of buried urns, where, even during da>time, the owl from its hole m the old tree hoots, where the red-eared' cock and the paginal bird sit without fear on the conical heap (of stones) above the hds inverted over the burial-urns, where the vulture descends on its many foot-paths which cross each other and the strong-billed crow flies as it likes, v\here the carnon-eatmg jackals abound and the hordes of demons \\ith their teeth defiled by carrion, grasp the carcases of the dead and eat the white flesh till their mouths reek of the odour, and where he strewn fragments of white bones and numerous stones concealed by the over- growth of jungle ' As regards the actual methods of the disposal of the dead there are many references to burial and some to cremation The epic Mammekhalai* which belongs to the last period of the Sangam epoch, summarizes the customs, contemporary and traditional, as follows &uduvdr-idu\ or-todu kulippaduppdr Talvayinadaippor-tahyirkavippor (Ch 6,11,66-67) The context of these lines is the description of the great grave->ard of the famous Cola capital, Puhar or Kavinppumpattinam It was a cosmopolitan city in the early centuries of the Christian era, when in the Tamil country the Vedic, Jama, Buddhist and other cults had more or less taken root These two lines enumerate the different sets of people who came there for disposing of the dead, wz those who cremated (suduvor). those who cast away or exposed the dead to the elements or animals (iduvor), those \*ho laid the body in pits which they dug into the ground (todu-kuli-paduppor)* those who interred the dead body in subterranean cellars or vaults (tal~va\m-adaippor), and those who placed the body inside a burial-urn and inverted a lid over it (tdjiyir-kavippdr) The first two methods hardly need any explanation The third method refers to inhumation, the fourth refers to vaults or cellars (vaym) let into the ground (/J/), i e stone cists and the like, in which the body or the remains of cremation are interred , and the last method, which is brief and descriptive, refers to the placing of the body (or the remains of cremation) into burial urns (tali), the mouth of which was covered by inverting a lid (kavi) This is actually what one finds in the case of urn-burials, which represent by far the most dominant custom as described in the earlier Sangam works the Narnnai, Padirruppattu and Purananuru The poet Paranar describes the mental condition of a distracted mother whose daughter has run away with her lover The mother prefers death to a life exposed to the scandal of the neighbours and apostrophizes the God of Death as follows Mayirum-tah-kavippa-t Tannru-ka]ika-\er-koM-k-kurre (Narnnai, 271, lines 11-12) 1 Puram 225, lines 7-8, 218, lines 1 -5, 240, lines 7-9, 244, line 1, 217, line 13, 356, lines 1-4, 359, lines 1-8, 360, lines 15-16, 162, lines 12-21 , 361, line 10, 364, lines 10-13 12 THE MEGALITHIC BURIALS AND URN-FILLDS OF SOUTH INDIA or Ajlvakas whose system of penance was fruitless' m the eyes of the followers of the Vedic religion Tdliyir-pmangalum-talaippada-verum-tavap Paliyir-pmangalum-tuhppelap'paduttiyc (Tak 376 ) In his two other compositions, the Vikrama-Co}an- Via and Kulottunga-Colan-Ula* eulogistic poems on the two contemporary Cola monarchs, he mentions mudu-makkal-cadi, 'the jar for the ancients' In his Via on Vikrama one of the legendary progenitors of the Cola king was Tarapati, who designed the jar or urn for the ancients when the 'God of Death having become powerless to take any life (at his will) fled and hid himself from view' maralmdu- Odi-marah- \ olippa-mudumakkat Cadi-vakutta-Tarapativum (Vik Via , lines 14-16 ) In the other Via on Kulottunga one of the progenitors, the twelfth in the line from the Sun, is said to be the first to devise the *mudumakkat-cddi' for the old and feeble Killi- talaip-pandu-k d-n-alwn Padnmakkada vul-padaip-padai \ ak -A alia Mudunwkkatiadi-mudaldn (Kid Via , lines 22-24 ) The Sankara-Colan-Ula, another composition of the same class, gives a similar idea The Tiru\enkartupuranam too has a passage where an ancient king is described as one 'who designed sescral mitdu-nmkkai-(adt for the grand old men at a golden age, when the God of Death could not take away any life' &itta~niakilndu-mbamw a- <engd-nadatta-naman Ultamanam-enrannal-uvir-k odu-pokamai yindl Movtta-mudi \ orkku-mudumakkatcadi-pala Vaitta-kula-dipakane-nmnnako-mannako (Tinven San anal, 9 ) This accounts for the tradition in later times that very old people, who had lived their 'four-score and twentv' and were decrepit were placed inside large urns until their death, to avoid discomfort This accounts aKo for the quotation of a late commentator, Naccmarkmi>ar (c fourteenth century), in his commentary on the earliest extant Tamil work Tolkappnam. attributing to the Ajlvakas the practice of 'entering the tab for penance until death* Tah-kavippa-t-tavarn-ceyvar-mannaka Valiya-nonana-malvarai) e Conclusions If we accept the postulate that the literature of a people of a particular period not only portrays contemporary life and events but may also embody in it earlier traditions, and that the advanced state of civilization which we find in the Sangam period probably had its origin much earlier, we may place the earlier limits of the megahthic and urn-field culture, which seems to have been a dominant factor of early Tamil civilization, in the pre- Sangam epoch, i e earlier perhaps than the last three centuries B C and earlier, too, than effective 'Aryan* contact with South India The gradual percolation of 'Aryan' ideas into the Tamil culture is noticeable m the different strata of the Sangam literature, and as we approach the close of the period the influence of these ideas gets more and more marked until the two cultures become 15 ANCIENT INDIA, NO 2 thoroughly intermingled , when we hear of the Tamil kings performing and protecting Vedic sacrifices, and find cremation according to Vedic rites taking precedence over the more ancient customs of burial By about the fifth century A.D comes a dark chapter in Tamil history, synchronizing with the Kalabhra interregnum, and when we see light again in the sixth to seventh centuries A D we have a literature totally different from the Sangam works in vocabulary, diction and metre, and predominantly devotional in nature This, taken together with the misconceptions about the burial-urn enumerated above from the literature of the eleventh century and after, indicates the later limit of the megahthic and urn-field bunal customs in South India as the beginning of the fifth century A.D or earlier. 16 ANCIENT INDIA, NO 2 CONTENTS Page 1 The site and its historical background 1 8 2 The chronology of the site 22 3 The structures 24 Notes Dimensions of bricks 32 Use of tanks m the textile industry 34 4 The pottery A Wares imported from the Mediterranean (i) Arretme ware 34 (11) Imported amphorae 41 (m) Rouletted black ware 45 B Local wares 49 C Chinese celadon ware 91 D Addenda to the pottery (i) Comparison with North Indian wares 93 (n) Chemist's notes 93 5. Other small finds E Beads 95 F Gracco-Roman gem 101 G Roman lamp 101 H Roman glass bowls 102 I Terracotta objects 102 J Metal objects 103 K Stone objects 104 L Wooden objects and ropes 104 M Miscellanea ' 108 N Coins 108 Pictorial graffiti on potsherds 1 08 P Inscribed potsherds 109 Q Bones 1 14 6 Appendices I Roman coins found in India and Ceylon 1 16 H Semi-precious stones 121 III The ancient name of Ankamedu 124 1 THE SITE AND ITS HISTORICAL BACKGROUND THE classical geographers and the Tamil literature of the 'Sangam' age have familiarized historians with the outlines and some of the details of Indian trade with the West in and after the first century AD At its prime, the trade was extensive It included as Indian exports pepper, pearls, gem-stones, muslin, tortoise-shell, ivory and silk , and as imports from the West coral, lead, copper, tin, glass, vases, lamps, wine and, at first, coined money l By the latter part of the first century A D the literary evidence makes it clear that this trade was organized on lines not unlike those of the European * factories 1 established m India from the sixteenth century onwards The Penplus of the Erythraean Sea (c A D 1 The best reviews of Indian trade with the Roman Empire are by M P Charlesworth, Trade-routes and Commerce of the Roman Empire (Cambridge, 1926), and E H Warmington, The Commerce between the Roman Empire and India (Cambridge, 1928) See also H G Rawlmson, Intercourse between India and the Western World (Cambridge, 1916), K A Nilakanta Sastn, The Cdfas, I (University of Madras, 1935), and P T Srmivas lyengar, History of the Tamilsfrom the earliest times to 600 A D. (Madras, 1929). 18 To face page 18 PLATE IX Arikatnedu Northern Sector in middle distance, Southern Sector in background PLATE X To face plate XI \ -\nkanieihi from the vu\/ Southern Sutor in n^ht B Ankumcdu wall\ projecting from rner hank AR1KAMLDU 60-100) speaks of IpndQUx. v6fit/i(x. t the unqualified IpndQux. of Ptolemy (c A D 150), which may fairly be described as treaty-ports 1 That is to say, permanent lodges of Western traders * were settled in them under formal agreement with the appropriate Indian ruler, and were visited at the proper seasons by convoys of deep-sea merchantmen And just as the agents of the Portuguese, Dutch, Danish, French or British establishments in the days of the Companies travelled widely in the interior to circumvent the middlemen, so we may with probability recognize in some of the Roman coin-hoards far from the sea (fig 48) the penetration of earlier Western traffickers with similar intent On the scale indicated, this organized interchange implies a knowledge of the periodi- city of the monsjoons Without that knowledge, which, according to Pliny, 3 could restrict the crossing of the Indian Ocean to forty days, the laborious coastal voyage to India or the still more precarious overland routes must have prevented the development of regular and direct trade with southern or eastern India The establishment of permanent agencies in those parts must therefore have post-dated the discovery of the so-called 'Hippalus' or south-western monsoon , a discovery of unknown date but appreciably earlier than Pliny and the Penplus, and possibly, though not certainly, earlier also than r AD 21, when Strabo speaks of 120 ships sailing for India from Myos Hormos on the Red Sea 4 On the other hand, there is no hint that the monsoon was familiar to Mediterranean merchants before the time of the Roman pnncipate The unification of the western world under Augustus (23 B C -A D 14) and the recorded reception by him of at least tv\o Indian delegations (c 25 B C and 21 BC) 6 provide an obvious context, if not for the actual discovery, at least for its diffusion beyond the corporations of Arab sailors and other agents who had jealously monopolized the Indian traffic thitherto Further than this the literary evidence fails to carry us Warmmgton conjectures from it that the direct route from the Red Sea to the Malabar coast, i e the full use of the monsoon, was introduced k soon after 41 ' * Archaeology now indicates a rather earlier date Some considerable time prior to the abandonment of the manufacture of Italian red-glazed pottery (Arretme and related fabrics} an event which is unlikely to have been later than A D 50 a Roman emporium is now known to have been firmly established far up the eastern coast of India, near Pondicherry. We may infer that at least as early as c AD 30, and possibly before the death of Augustus, regular monsoon-trade had been established between the Mediterranean and western India, with a coastwise or overland ex- tension to the Coromandel coast Whether at this early date the route was carried to its logical conclusion in Malaya and China is at present unknown By the second century, at any rate, Antonine coins would appear to have penetrated to Long Shuyen and the peninsula of Cochin China, where they were dug up by Mr L Mallaret in 1945 7 Before we turn to the Pondicherry site, the literary picture of these trading-cities may be completed in outline by reference to one or two familiar passages in the Tamil literature This literature is very insecurely dated, but the relevant passages, which are numerous, may be ascribed on general grounds to the early centuries AD It was probably in the second century, for example, that the city of Puhar or Kaverippattmam, at the mouth of the Kaverl river 60 miles south of Pondicherry, was thus described in the epic Silappadikaram ('The Lay of the Anklet') 'The sun shone over the open terraces, over the warehouses near the harbour and over the turrets with windows like the eyes of deer 1 Warmmgton, p 107 2 At Muzirii (Cranganore, Cochin State) the Peutmger Table, second to third century A D , marks a 'temple of Augustus' 8 VI, 104 II, V, 12 6 The Monumentura Ancyranum states that such embassies came 'frequently' * Op cit.p 45 7 Information from Mr H Gcnais, of the 6cole Francaise d'Extrfime-Onent 19 3SLJ- N N ^,J PRINCIPAL PLACES MENTIONED IN THE ARIKAMEDU REPORT JCALf Of MILKS too so o IPO aoo 300 +00 HEtOHTS ABOVK /5OO fMKT V-^ HDNDAPURj& '^"V ?V^At^vU AMAR^VAT? ^ MASw /^Sr . jr *" X*" *** ' "Sf / v' ^* / 5 MAHAAAGIRJ^' i irf" ( O l^f^ir^ >\^O * & ^ /* >\,tt _ /^'S < /^ ^-A?3 "^ - * M̂A"*_,%%<) i>7 t VMARIVANAM (/'JUH.T <#V-T, ^w ^ V^ -'^ONDlCHeiRy "AA T\T* ^ ^' 'O (of">**s'fa *"< ycsX Cr.rcOIAATOMo JTRANQUEBAJl KA CHANGANOWI^^^"* J" Jo ( ^*}\'T&r, OU,AK^->J & ' AIKAMDU' auitjON^^^V * 7\^ TKORIOJJ %jk \J JfmLmtMi FJG. 1 20 ARIKAMEDU In different places of Puhar the onlooker's attention was arrested by the sight of the abodes of Yavanas [men from the Graeco-Roman world] whose prosperity never waned At the harbour were to be seen sailors from distant lands, but to ail appearance they lived as one community ... .' l In another poem, a Pandya prince is exhorted to drink the cool and fragrant wines brought by the Yavanas m their vessels 2 Again, 'agitating the white foam of the PSnyaru, the beautifully built ships of the Yavanas came with gold and returned with pepper, and Muzins resounded with the noise' 3 And Tamil rajas employed bodyguards of western mercenaries, 'the valiant-eyed Yavanas whose bodies were strong and of terrible aspect' and who, equipped with 'murderous swords', were 'excellent guardians of the gates of the fort-walls' 4 In this capacity they are said to have been employed at Madura Yavana craftsmen were also sought after in southern India, especially for the manufacture of siege-engines 6 In one way and another, the Yavana in partibus enjoyed a considerable prestige whether as trader or as settler******* Two miles south of the centre of Pondicherry, the capital of French India, a former outlet of the Gmgee or VarShanadi river forms a lagoon locked to day, save m full flood, by a sand-bar from the Bay of Bengal The lagoon, known as the 'Anyankuppam river' from the name of the largest adjacent village, is fringed by cocoanut-palms which, with a belt of sand-dunes, shelter it mostly from the sea In the vicinity of the little village of Kakkayantoppu, the eastern bank of the lagoon stands some 20 feet above the water and from the scarp project the jagged ends of successive brick buildings to which the mound or mcdu owes its being (pi XB) To the villagers the site is known as Ankamedu, and, though no more formal origin for the name has been discovered, it will serve as well as any other 8 Destruction of the buried buildings by the river in flood has been accompanied by extensive destruction from other causes Tree-planting and, above all, brick-robbing have combined m the work As far back as 1734 there is record of the latter process,7 and fragments of Chinese celadon ware of the ninth to twelfth centuries found m 'robber- trenches' throughout the excavated site indicate a similar process as early as the Middle Ages To-day, the nucleus of the site is protected by the French Government In the eighteenth century the visible remains were described by Le Gentil, 8 who incidentally remarks upon their destruction by the river But archaeological attention was first directed to the site by Mr G Jouveau-Dubreuil who, in and after 1937, made a collection of beads and gems found on the surface by the local children One of the gems is reported to bear a head of Augustus in intaglio, and this induced Mr Jouveau- Dubreuil to exclaim 'Nous avons la une veritable ville romaine' 9 In the sequel, excavations were carried out m and after 1941 by Brother L Faucheux of Pondicherry in collaboration with Mr R Surleau, then Chef des Travaux Publics The excavators ignored stratification 1 The &ilappadik&ram> trans V R Ramachandra Dikshitar (Oxford, 1939), p 110 2 K Pillai The Tamils eighteen hundred years ago, pp 12 and 37 , lyenger, op at , p 312 lyenger, p 312 * Ib , p 313, and Warmmgton, p 80 5 I>cnger, pp 313ff* fl Mr K R Snnivasan has suggested that the name may be Arukkume'a'u, meaning 'mound of ruins' (arukku=to destroy or rum), or it may be Arukume'a'u, meaning 'mound on a river bank' (aruku, according to old usage still surviving in Pondicherry see Tamil Lexicon under Aruku-Kollai and Arukadigam signifies the neighbourhood or bank of a river) 7 L Faucheux, Une vieille citt indienne (Pondicherry, 1945), p 2. Voyage dans les mers de VInde (Paris, 1779-81), I, 542-5 9 Bulletin de rfccole Franfaise d'Extrtme-Orient, XL (1941), 450 Mr Jouveau-Dubreuil's collection was sent to the French School museum at Hanoi, Indo-Chma, and I have not seen the Augustus gem 21 ANCIENT INDIA, NO. 2 Thus from a convergence of evidence it is here inferred that the riparian sites excavated in 1945 were first occupied at the end of the first century B C or beginning of the first century A D , with an inclination towards the later date The duration of the occupation is more in doubt. Almost all the buildings excavated were constructed after the cessation of the importation of Arretme ware The large building, here called a warehouse, m the Northern Sector was built on and into the main Arretine-producmg layers, and only a few stray sherds occurred in the contemporary 'make-up' or subsequent spoliation In the Southern Sector all structures, save two fragmentary walls, also post-date the cessation of Arretme In other words, the intensive structural development of these two areas belongs to the end of the 'Arretme period \ here ascribed to c A D 50 (before rather than after that date) This helps, incidentally, to drag down the initial phase into the first century A D , m preference to the preceding century Subsequently to our date A D 50 there were, in the Southern Sector, several successive stages of construction and reconstruction, accompanied by some modification of the associated Indian pottery These developments were controlled by two unifying factors, a general continuity m the main units of the plan, and the occurrence m all strata of sherds of Mediterranean amphora It is evident that the function and contacts of the site remained unchanged Within these general limits, the duration of the occupation of the site m years is anyone's guess A minimum of a century would appear to be required by the changes indicated, but there seems to be no good reason for allowing more than two centuries A terminal date m the neighbourhood of A D 200 would be a reasonable supposition The great 'warehouse' m the Northern Sector had vanished long before this date, and a pit, full of potsherds of the first to second centuries A D , had been cut into the remains of one of its walls The reason for the doom of this building is not far to seek , it lies at the lower end of the site, its foundations are at the present mean sea-level, and, although its floor was, from the outset, raised to a height of 4J feet above them, it must have been liable to flooding even on the necessary assumption that in the first century A D the relative level of land and sea here differed considerably from that of the present time It is noteworthy that no coins have been found in the ancient levels of the site Only m the extensive areas of late disturbance, due to systematic brick-robbing m medieval and modern times, do Cola coins occur with fragments of the Chinese celadon ware which is almost universal m Asia in and after the ninth century A D In summary, the site generally is assumed to have been occupied during the first two centuries A D , and to have been despoiled for its bricks in the Middle Ages and later 3. THE STRUCTURES NORTHERN SECTOR AK II, III, V, VI, VII, VIII AND X The 'warehouse* Of the two mam Sectors excavated, the Northern was an amplification of an area cleared by the French excavators in and after 1941 These earlier excavations had disclosed the south-eastern angle of a large brick building and the remains of a ring-well or soak-pit (R W. 1) built of potterv rings apparently at a somewhat earlier period. A second ring-well is said to have been found but has not been seen by the present excavators Two more (R.W. 2 and 3) were found to the north and west respectively of the building m 1945, and a few feet to the south or south-west of the former the French claim to have discovered a brick-lined well in 1944 24 ARIKAMEDU orel IS V ry A774 2g (asowse zon sontns Wenn) ANCIENT INDIA, NO. 2 As now revealed the building, largely destroyed by brick-robbers and curtailed by the river, was a featureless structure upwards of 150 feet long (pi XIII) Traces of a single partition-wall were found, but, apart from a side-chamber towards the east, the building was substantially an oblong shell, massively built and with all the appearance of a warehouse, which it is presumed to have been Its brickwork had been rendered externally by its builders with a tough plaster, apparently intended to keep out water on this low-lying site, at least, elsewhere at Ankamedu lime-mortar or plaster was only used when it was desired to waterproof walls or floor, as m Tank A on Site AK IV The bricks themselves were mostly held together with a mud-mortar devoid of lime The date of the building was reasonably clear and has been discussed above, p 24 It was approximately the middle and second half of the first century A D No later structure was found here or hereabouts It is evident that the 'warehouse' was built on a site which was as near to water-level as was feasible, and that the latter was already rising The main walls were built on rough foundations of bricks which had been sunk on a rubble bottoming into the soft sand, and the floor-level was 4 feet 6 inches above the lowest course of built brickwork For the most part the walls did not remain to a sufficient height to indicate doorways, but a former entrance was identified in the eastern wall of the side-chamber by a gap, at which a thick layer of concreted sand and broken bricks probably represented a street approaching from the east To the north of this gap, a rectangular mass of brickwork partly covered with concrete suggested the former existence of a ramp or staircase here Below sea-level, 1 abundant pottery and fragments of cut timber were found to a depth of no less than 10 feet, but these lower deposits, with two possible exceptions at the top, gave the impression of being estuarme slime into which the objects had sunk, rather than submerged occupation-layers In other words, the building was erected actually on the foreshore of the estuary, on which apparently timber buildings had already stood and into which the debris of the port had penetrated , it was floored at a level high enough to escape normal flooding, and its external walls were specifically reinforced to keep out water , but, after an occupation which does not appear to have been extensive, it \vas disused, its bricks were largely removed, and the water-logged site was abandoned, whilst the higher ground to the south (at and around AK IV) remained in use At a reasonable guess, the abandonment of the Northern Sector may be supposed to have occurred before the end of the first century A D , i e within little more than a generation of the date of construction Evidence of occupation in the 'warehouse' is slight, and, although superficial disturbance of the area robs this negative evidence of certainty, it is significant that Pit 1 of AK VIII, containing a great quantity of potsherds of first to second century types, was cut into the southern wall of the eastern side-chamber of the building after its demolition SOUTHERN SECTOR* AK IV The site known as AK IV lay along the edge of the river to the south of the second of the two mam areas already explored by the French Whereas in the Northern Sector 1 Our excavation in loose sand and clay down to a depth of 1 3 feet below sea-level was fraught with difficulties, which were overcome by fencing the area with planks, driven lower and lower as the excavation proceeded, and by pumping out water constantly with an excellent pump lent by the Travaux Publics of Pondicherry * This section of the report is contributed by Mr A Ghosh 26 ARlltAMEDU evidence of occupation penetrated to a depth of 9 feet to 10 feet 6 inches below sea-level, here the natural soil was found at an average height of 10 feet above sea-level On this higher site, less accessible for the work of the port but out of reach of flood, the evidence of pottery indicates that occupation began a little later than on the foreshore but continued after the foreshore was abandoned It falls into three mam Phases, Early, Middle and Late, with sub-periods, but it should be emphasized at once that there was no break in its continuity, unless between the Middle Phase and the very slight remains constituting the Late Phase Early Phase The Early Phase may be divided into three sub-periods Sub-period 1 The earliest structures were separated from the natural sand by a deposit of greenish sand 3 feet thick, rather poor in pottery and other small finds The remnants of this sub-period were very few and did not yield any structural plan They consisted of (1) a fragmentary wall carrying three square blocks of brickwork which doubtless supported wooden posts , (2) a ring-well (R W 1) near the southern end of the wall, and (3) a patch of brick floor (Floor A) near by, all the three were to the south of Tank A of the next sub- period , and (4) a wall in the open space between Tanks A and B, entirely over-built by a wall of the next sub-period An important feature of this wall was that it contained five irregular vertical chases in its eastern face, showing that it was built against wooden posts and roughly cut to fit them This method of reinforcing an essentially timber structure by a brick 'skin' was repeated m later buildings in the Sector At the end of this sub-period a shallow but wide pit was cut into the ground imme- diatelv to the west of the wall first mentioned above This pit yielded a very large quantity of pottery (Group A), which included several interesting sherds, notably a fragment of stamped Arretme several pieces of amphora, and a sherd with a Prakrit inscription (see below, pp 36, 43 and 111) Suh-penod 2 This sub-period witnessed brisk building activity The mam feature was the construction of two complexes of pavements called Tanks A and B each with an adjacent room of uncertain purpose Both continued to be repaired and restored till the last sub-period of the Middle Phase Throughout their existence they seem to have lain at the edge of two large and apparently unroofed quadrangles They had a close family- likeness to each other and must have been used for similar purposes The extensive use of drams or conduits and substantial pavements of large bricks apparently made for the purpose, and the absence of domestic features, indicate the probability that the site was used for industrial purposes involving a constant inflow and outflow of water It is likely enough, though not proved, that the tanks were used as cisterns or vats for dyeing the muslin which formed one of the most important exports of this part of India m ancient times 1 The yards may have been used for drying the cloth Tank A (fig 4 and pis XXA, XXIA) was originally paved with bricks four courses deep and plugged with clay and had a doorway in its eastern wall The portion of the wall to the north of this doorway had an irregular outer face, having been built against wooden posts resting on a single course of bricks The western and part of the southern walls had an irregular footing, indicating that a part of the tank had been built against the sides of an excavation below ground-level There is no evidence as to how water was drained out of the tank in this period i For the use of brick tanks m the present muslin industry m the Madras Presidency, see p 34 For Indian muslin exports in classical times, see Warmington, op cit , pp 210-12 27 ANCIENT INDIA, NO 2 AR.IKAMEDU SOUTHERN SECTOR. SITE AK IV TANK A SCALE OF METRES BRANCH CULVERT r~~l EARLY PHASE SUB PERIOD II MIDDLE PHASE SUB PERIOD I MIDDLE PHASE SUB PERIOD n H SECTION ON OH 5 I 194-5 FIG. 4 28 To face page \ PLATE XVI SUB-PERIOD 1 K SUB PERIOD U d- A S I >f ts e s te it i- is 1, g >f IS h r k h e ie n i- ie \- d Is >f ANCIENT INDIA, NO 2 Sub-period 1 This sub-period became known to the excavators as the 'Drain Period', since it was characterized by the construction of an elaborate system of drainage, sloping north towards the river In construction all the new drains or conduits, unlike Conduit A of the Early Phase, were constructed by means of corbelled walls (not single bricks on edge), floored and roofed with horizontal bricks Conduit B, built upon the ruined top of the eastern wall of the quadrangle of Tank A, first discharged itself into Tank A, whence the water was emptied by another conduit built in line with it through the opposite wall. The end of the inlet was rebated and its side- walls were provided with slits for a wooden sluice-gate (pi XXIA) From the extant fragments of the southern end of Conduit C it seems that it had first a northern course , then, after a short turn to the east, it turned again to the north and overlying the earlier Conduit A ran approximately along the median wall between Tanks A and B, which was built upon the earlier wall with chases for timbering and itself had the same characteristic (pi XXIIA) In fact, this wall and Conduit C were essentially restora- tions of the corresponding earlier structures Conduit D, originating like the others somewhere outside the excavated area, ran through the eastern portion of Tank B, utilizing the earlier floor of the tank as its floor and proceeded northwards till it reached the northern terminal wall of the quadrangle within which Tank B was situated Beyond this point it continued with a slightly different orientation (as Conduit E), till, crossing a probable street, it discharged itself into a wider cross-drain (Conduit F) which came from the east and probably formed part of the mam drainage-system of the town The point of junction with the mam dram was indicated by a gap in the latter (pi XIXB) It is likely that further west Conduits B and C similarly joined Conduit F The relation of Conduit B with Tank A has been described above The floor of the tank was now raised by 1 foot 4 inches over the earlier floor, the intervening space being filled up by debris The new floor was drained by a branch-culvert, the entrance to which was stopped by a small square brick plug The southern, western and northern walls of the tank were now rebuilt upon the corresponding earlier walls The outer face of the last wall contained two chases, indicating the position of posts against which it had been built The eastern wall was missing it is likely, however, that it too followed the alignment of the earlier wall In Tank B, the eastern wall was replaced by a new wall with practically the same alignment and the northern wall was thickened The partition-wall between the tank and the southern room continued without alteration Within the walls of the tank important changes were effected The partition-wall between the tank and the verandah was discarded, and a new east-to-west wall was erected, partitioning the room into two long strips Both strips were floored with bricks, the new floors being separated from their predecessors by a deposit of brick-bats and earth The floor of the northern strip was 8 inches higher, indicating that it now took the place of the earlier platform The northern strip itself was divided into two halves by a cross-wall, through which ran a subsidiary dram or conduit taking off from the western half and joining Conduit D through a hole made into one of its cap-bricks It holds the same position as the branch- culvert in Tank A The history of the two tanks is thus mutatis mutandis identical A paved floor outside the eastern wall of the tank, made of fragmentary bricks and edged with bricks on edge, also belongs to this sub-period Though the fragmentary remnants render the reconstruction of a complete plan of the sub-period impossible, the general lay-out is fairly clear Tank B seems to have formed part of a large quadrangle, the wall against which Conduit C was built forming its western 30 To face page SO PLATE XIV ARIKA^ SIT uill llMldJlil SUB-P SUB-F SUB-F JO /o METRES \ PLATE XV FER Hi To face page YER Scare a Scace of AST 1945  ARIKAMEDV I - 00 - j fM ' O O C i O JHSir_ I :' x | o { o 33 ARIKAMEDV A.D according to J H Iliffe, who has made a useful preliminary study of these provincial wares 1 At Ankamedu, however, we are concerned primarily with the highest category of metropolitan sigillata, namely, that which was produced in Italy, at Arezzo, Puteoli and elsewhere, and has been named 'Arretine' from the Latin name of Arezzo (Arretium). 2 This Arretme sigillata originated at an undetermined date probably in the first century B C By c A D 25 it was confronted in western Europe by rival potteries established in Gaul, in some cases by the Italian potters themselves, and, though it lingered on until after the Roman invasion of Britain in A D 43, it was driven from the western market by AD 50 In the East, where provincial rivalry may not have been at first so acute, it is possible (though not proved) that Arretine ware lasted a little longer It is conceivable (though again not proved) that, in its simpler forms such as are normal at Ankamedu, the ware may even have been retained specifically for export-trade, like the celadon of medieval China (see below, p 91) In the absence of alternative evidence, however, we must assume the general validity of the results of detailed study m western Europe, and must suppose that no Arretine pottery reached India after A D 50 The literary record makes it clear that Romano-Indian trade remained brisk until long after the middle of the first century A D There is furthermore no break m the occupation of Ankamedu at the moment when Arretine ware ceased to arnve there It is a fair inference therefore that the latest Arretme on the site belongs to the last phase of the production of the ware, i e c A D 45-50 , and m considering the initial moment of Roman contact, we may reasonably argue backwards from that date Such argument is necessanly somewhat subjective, but in the cutting where the evidence was clearest (AK II, see fig 2) only four layers, all of them of sandy estuanne mud likely to have been deposited quickly, contained Arretme sherds, and little more than thirty years might be regarded as sufficient for their accumulation So far as it goes, the character of the pottery itself is consistent with this Save for one tiny fragment, probably of Dragendorff's form 11, with leaf-pattern (fig 7, 32 and pi XXIV, 4), the sherds represent undecorated cups and dishes, in some cases even without the normal roulettmg These negative features may be regarded as symptoms of lateness , although it may also be supposed that only the cheaper products of the kilns were used for this overseas trade, particularly since the fabnc was in any case technically superior to that of the Indian pottery of the period and would therefore find a ready market I have proposed therefore A D 20-50 as the inclusive period for the Arretme pottery at Ankamedu The stratum which produced most of the sherds of Arretme on site AK II (thirteen out of a total of eighteen from the cutting) is shown as 4 7 Ar ' on the section, fig 2, and was approximately at the present mean sea-level Below it, a succession of sandy deposits contained pottery to a further depth of 10 feet, beyond which clean sand, devoid of relics, was encountered Of this further 10 feet, only the uppermost 2 feet (layers 8A, 8 green, and 1 1C) yielded sherds of Arretme, to the total number of four There were thus still 8 feet of accumulation, the pottery from which did not include Arretine 1 'Sigillata wares in the Near East', Quarterly of the Department of Antiquities in Palestine, VI (19^8), 4ff * The phrase vasa arretina (Arretine pottery) was retained by the Romans m the sense of pottery de luxe even after the actual potteries of Arretium had ceased production Standard references to this ware include H Dragendorff, Banner Jahrbuther, XCVI-XCVII (1895-6), S Loeschcke, 'Keramische Funde in Haltern', Mitteilungen der Altertwns-Kommission fur Westfalen, V (Munster i W , 1909), E Ritterlmg, 'Das frUhrOmische Lager bei Hofheim im Taunus', Annalen des Veremsfw nassauische Altertumskutule (Wiesbaden, 1913) , F Oswald and T Davis Prycc, Terra Stgillata (London, 1920) 35 ANCIENT INDIA, NO 2 Above stratum 7 Ar only one sherd of Arretme was found in an undisturbed deposit in stratum 5, which formed a part of the make-up of the floor-level of the * warehouse', a brick building constructed on and into stratum 7 Ar Otherwise, the most significant occurrence of sigillata was on the mam southern site, AK IV, where a stamped sherd immediately underlay a brick wall of the Early Phase By sites, the Arretine ware found during the present excavations was as follows (all from the Northern Sector except AK IV), together with some sherds from the previous excavations ARRETINE SHERDS, STRATIFIED Fig 5 AK II (a) From the make-up of the floor of the 'warehouse' 1 Fragment of flat base of dish (AK II, 589 ) (b) From stratum 7 Ar 2 Fragment of rouletted rim of Loeschcke type 2A (AK II, 342 ) 3 Flat base of dish possibly of similar type, stamped VIBIE (see p 39) Indeterminate graffiti on under side of base (AK II, 266) 4 Fragment of similar dish (AK II, 312 ) 5 Fragment of cannated side of very thin, fine fabric, variant of Dragendorff type 8 or Loeschckc type 1 5 Approximates to a stamped example of Claudian date (AD 41-54) from Mainz (Oswald and Pryce, op cit , pi XXXIX, 14) (AK II, 311) 6 Minute sherd with internal groove, possibly fragment of Loeschcke type 10A, but too small for reasonable conjecture (AK II, 344 ) 7 Fragment of flat base of dish (AK II, 343 ) 8 Fragment of rouletted nm of Loeschcke type 2A or similar (AK II, 314) 9. Fragment of base-ring of dish (AK II, 298 ) 10 Indeterminate fragment of ' marbled ' ware (yellow with red veins) This type of ware was being produced by Arretine potters at least as early as the beginning of the first century A D (see p 14) (AK II, 506) Not illustrated 1 1 Fragment of Dragendorff type 17 or Loeschcke type 3B (AK II, 587 ) 12 Fragment of flat base of dish with ring of fine rouletted pattern on upper surface (AK II, 587 ) 13 Fragment of variant of Loeschcke type 3B (AK II, 530 ) 14 Fragment of flat base of dish of 'marbled' ware See above, under No 10 (AKII.4IO) (c) Layer 8A (prior to 7 Ar) 15 Fragment of flat base of dish with incised line on interior (AK II, 397 ) (d) Layer 11C (cut into and therefore later than layer ' 8 green') 16 Fragment possibly of variant of Loeschcke or Dragendorff type 8 Orange-red glaze (AK II, 517 ) (e) Layer 8 green (lowest Arretine layer) 17 Fragment of flat base of dish (AK II, 550 ) Not illustrated AK W (f) Pottery Group A (earlier than the earliest, save one, of the brick structures on the site) 18 Base, form indeterminate, stamped ITTA (see p 40) Fragmentary graffito (svastika 7 ) on under side The glaze is of mediocre quality and is orange-red (AK IV, 1 54 ) AK y (g) Pit 7, high up outside S wall of warehouse 19 Fragment of dish-like bowl, variant of Dragendorff type 24 Red glaze on exterior of nm, yellow glaze on interior, and possibly on exterior below nm (AK V, 61 ) 36 To face plate XIX PLATE XX \ (A li lank \ \ho\\inz i\\o main puiod\ of fhni i B AK 7^ ' /tfwA, B E P = fiir/y P/JflVf , V^ P = A//J<//e Phase, LP = Late Phase PLATE XXI To fat e page 37 AR1KAMEDV 2TT '-- * 22 C L--"-==~.-J FIG 5 Arretme pottery from stratified layers, 1945. \ (potters' names m 3, 18 and 21, 37 PLATE XXIII A irutint Wt;/>//> I MIL or \ Hill To face page 39 B \tnitnt \nm,/> C l\fl HI C Arretine stamp ITT I Scale 2 1 ARIKAMEDU 29 Variant of Locschckc type 2 (AK X, 3 ) 30 Fragment of rouletted nm of Loeschcke type 8 or Ritterlmg type 5 (AK X, 4 ) 31 Rim of Ritterlmg type 5 No roulcttmg (AK X, 1 ) AKRBTINE POTTERY FROM THE PREVIOUS (FRENCH) EXCAVATIONS Fig 7 33 - 1 , 32 34 ' 35 39 .- A f"">E:j . '. : i 4 I IG 7 Arretine potter\ from the previous (French) excavations % These sherds are unstratified, but Nos 37 and 38 are said to have been found at a depth which would equate them with our Arretme-producing lajers on AK II 32 Fragment of hne ware with leaf-pattern impressed from a mould The sherd is too small to indicate form , possibly DragendortT type 1 1 PI XXIV, 4 33 Fragment of large dish, Locschtke type 2A 34 Fragment of dish, Loeschcke type 3B 35 Ditto PI XXIV, I 36 Base of dish 37 Base of cup, with indeterminate fragment of potter's stamp 38 Base of cup, probably Loeschcke type 8 or Ritterlmg type 5 On the under side of the base is a graffito k or, more probably, BrShml a PI XXIV, 3 39, 40 Cups of Loeschcke type 8 or Ritterlmg type 5 PI XXIV, 2, 5 POTTERS* STAMPS ON ARRETINE WARE (pi XXIII and fig 5) I VIBII {probably VIBIE, possibly VIBIF) On the interior of the flat base of a dish, form uncertain Found on Site AK II, layer 7 Ar There seem to have been two families of Arezzo potters, the Vibu and the Vibiem, to the latter of whom this stamp would appear to belong, but no attempt has been made to distinguish their respective work and date Miss M V Taylor, to whom I have referred the problem, has very kindly spent much time m looking up the appropriate literature at Oxford, and the following is a summary of her report The first and perhaps best account is by Gamurnm in Notizie degh Scavi, 1 883, pp 45 Iff He says that the pottenes of the Vibiem at Arezzo were near those of M Perenmus and P Cornelius by S Maria in Gradi, and worked both before and after them. On the death of C. Vibienus, his sons succeeded to the industry. 39 ANCIENT INDIA, NO. 2 What their relation was with the pottery of "Vibius" is not clear, or which employed the other, but the Vibieni and Vibii descended from an ancient Etrurian family, the Vibia, whose name is still preserved in Bibbiena. In any case, north of Arezzo, close to S Domemco, was the pottery of A(ulus) and C(aius) Vibius who, according to coins found, worked in the second and first centuries B.C. After the death of C Vibius, the industry was carried on by his workers, Athemus and others The names of workers with the Vibn include A Vibius, C Vibius and L. Vibius (Corpus Inscnptionwn Latmarum, XI, 6,700, 765ff ; XV, 5,750-60 , XIII, 10,009, 290-300) 'Loeschcke, m Mitteilungen der Altertums-Kommission fur Westfalen, V (1909), 186, says that, since the Vibn are found at Mont Beuvray, Neuss, Tnon, Vechten and Xanten but do not occur at Haltern (occupied 11 B C-A D 16), they evidently flourished before the latter dates He does not, however, distinguish between the two families of Vibii and Vibieni, and most of his evidence appears to relate to the Vibii T Kmpowitsch, Die Keramik romischer Zeit aus Olbia (1929), 9, No 3, Abb 2 and p 12, gives VIB1E (might be VIBIF), and, basing on Loeschcke, dates it to the last years of the first century B C The collection however has no independent documentation Ihffe (op cit on p 35, above) records VIBI and C VIBI on dishes from the agora at Athens, but again there is (at present) no independent dating On the other hand, F Oswald in his Indet of Stamps on Terra Sigillata ascribes C Vibius of Montans, in southern France, to the period Claudius- Vespasian (c A D 41-79) ' It would appear therefore that members of the Vibia family were producing Arretme or related pottery from the beginning of the first century B C until after the middle of the first century A D , though in the latter stages only denvative provincial wares are in question The evidence is consistent with an earlier initial date for the Ankamedu series than I have suggested above (p 22), but is far too confused to be determinate 2 CAMVRI On the interior of the flat base of a cup, probably Loeschcke type 8 or Ritterlmg type 5 Found on AK VII in Pit 2A, probably contemporary with the 'warehouse' The stamp is somewhat blurred, but the cross-stroke of the A appears to have been amalgamated with the right instead of the left half of the M The stamp of Gams Amunus occurs on Arretine at Olbia (Knipowitsch, as cited above), at Beisan m Palestine, at Alexandria, and m the Athens agora See Iliffe, as cited above In at least three instances it is recorded on dishes of Dragendorff's form 1 8 , this form is characteristic of the last phase of Arretine production, but the date of its introduction is unknown It occurs also on the dish known as 'Ritterlmg 5 ' which is found at Hofheim m the Claudian period (A D 41-54) 3 ITTA On the interior of the flat base of a dish, form uncertain, with an orange- red glaze Found on AK IV in Pottery Group A which just antedated the earliest, save two, of the brick walls on that site Possibly a retrograde stamp of the Augustan potter P Attms It is uncertain where he worked ( 9 Puteoh or Arezzo) Ilifle records his stamp from the Athens agora, and it has been found (ATTI) at Mainz on Arretine of Ritterlmg type 5 in 'marbled' ware (Behn, Romische Keramik, Abb 10, 1 , whence Oswald and Pryce, op cit , pi XXVIII, 1), and at Haltern (Loeschcke Nos 117 and 118) His wares were therefore in use within the period 1 1 B C -A D 16, the period of the occupation of Haltern COPIES OF ARRETINE FORMS, AND RELATED WARES (fig 8) Close copies of Arretine forms at Ankamedu are restricted to Dragendorff type 24/25, a small bowl with rouletted nm The copies are of a hard greyish buff ware with polished surface, and differ in fabric sufficiently from the local wares to make it reasonably certain 40 ARfKAMEDU (a) Stratified amphoraefrom the Northern Sector From pre-Arretine layers 46 From AK II, stratum 14 (the lowest save one , 6 feet below present sea-level) Rim of pinkish buff ware 47 Cannated shoulder, pink ware with yellow slip From AK V, stratum 11 A, underlying the wail of the 'warehouse 4 and equivalent to strata 13 and 14 on AK II 48 Rim of yellow-slipped buff ware, same type as No 46 From AK V, stratum 11, a pre-Arretine layer immediately below present sea-level From Arretine layers 49 Fragment of handle of gritty pinkish buff ware from AK II, stratum 8A, which also contained a sherd of Arretine (above, p 36) 50 Bluntly pointed base, of pink ware with yellow slip, from AK n, stratum 10A, which immediately overlay 8A 51 Part of straight handle from AK II, stratum 7 Ar, which yielded most of the Arrctine sherds from the site Yellow ware 52 From AK VHIA, stratum 7, which equates with the Arretme-produung layer 7 Ar of AK II Pinkish buff ware with yellow slip 53 Angular shoulder from AK VIII, stratum 5, which equates with the Arretine layer 7 Ar of AK II Pink ware From post-Arretine layers 54 Bluntly pointed base from Ak II, stratum 7H, contemporary with the construction of the warehouse' Pink ware, yellow slip 55 Rim and straight high-shouldered handle, of pink ware From AK II, stratum 9, immediately over- Ivmg 7 Ar and equivalent to 7H (No 54) The straight, high-shouldered handle is characteristic of the first half of the first century A D (Loeschcke type 66, Rittcrlmg tvpe 73) 56 Angular shoulder, of pink ware with yellow slip From the same stratum as No 55 57 Angular shoulder, pink ware with vellow slip, from AK II, stratum 7A, equivalent to that of Nos 55 and S6 58 Straight handle, pinkish ware with yellow slip, from Ak If, stratum 7D, make-up of floor of warehouse 59 Bluntly cannated shoulder, pinkish ware with yellow slip From \k II, stratum 6, make-up of floor of 'warehouse 60,61 Short handles, pink ware with yellow slip, from \k V, Pit 1, subsequent to destruction of warehouse This stump) type of handle is contemporary with the tall straight ivpe, No 5\ but also long out- lived it (b) Stratified amphoraefrom the Southern Sector (AK IV) 62 Rim, pink ware, from Pottery Group A (with sigillata base stamped IITA), immediately underlying the earliest-but-onc of the britk structures 63 Angular shoulder, fine pinkish buff ware, dating from the beginning of the mam Dram Period' 64 Stumpy handle, pinkish buff ware, from the same layer as No 63 65 Fragment of straight handle, from a slightly later structure of the same phase as Nos 6"M Pmk ware with yellow slip 66 Pointed base, pink ware with yellow slip Late Drain Period' 67. Fragment of handle of gritty pink ware Late Phase 68 Fragment of handle of fine yellow clay Late Phase 69 Fragment of straight handle, gritty pink ware Late Phase (c) From mi\ed deposits of the Northern Sector 70 Fragment of handle, pink ware with yellow slip (Ak II ) 71. Straight handle, pink ware (AK II ) 72 Rim, buff ware (AKII) 73 Part of straight handle, pink ware with buff slip (AK II ) 74. Rim, pink ware (AK II ) 43 ANCIENT INDIA, NO. 2 FIG 10 Imported amphorae 62-69, from stratified layers. Southern Sector , 70-84t from mixed layers 44 AR1KAMEDV 75 Curved handle, fine pink ware with yellow slip (AK VITI ) 76 Rim, pinkish ware with yellow slip (AK VIII ) 77 Bluntly pointed base, pink ware with yellow slip (AK VIII ) 78 Bluntly pointed base, pink ware (AK Vin ) 79 Straight handle, pinkish ware with buff slip (AK VII ) (d) from mi\eil deports of the Southern Sector 80 Fragment of handle, pink ware with yellow slip (AK IV ) 81 Sharply shouldered handle, fine pinkish ware with yellow slip This type of handle occurs at Haltern early in the first century A D and is found rarely at Hofheim towards the middle of the century It is exceptional after A D 50 (AK IV ) 82 Rim of pink ware, traces of yellow slip (AK IV ) (e) From the prewous (French) e\cavation<> 83-84 Necks of amphorae of light pink ware, 83 with buff slip Note an amphora from Ta \ila The only other Indian site known to have yielded an amphora of Mediterranean type is Taxila (Punjab), where an example (fig 1 1) was found at a depth of 7 feet 4 inches in the Parthian city of S'irkap, and is dated by Sir John Marshall to the first centuries B C -A D Now in the Taxila Museum Unlike the Ankamedu amphorae, which must represent maritime trade, the Taxila example is on general grounds more hkeh perhaps to have travelled overland from western Asia FIG 11 Imported amphora from Ta\ila t Punjab $ (in) Rouletted black ware (Ankamedu Type 1) l A characteristic pottery-type of Ankamedu is a dish (Type 1) sometimes more than 12 inches in diameter, with an incurved and beaked rim which usually has a facetted edge The ware has a remarkably smooth surface, is thin, brittle and well-burnt, and has an almost metallic ring The flat interior is normally decorated with two, occasionally three, concentric bands of rouletted pattern (fig 12 and pis XXIIB, XXV, XXVI) This pattern is not an Indian feature and may be regarded as an importation from the Mediterranean region, but it has not yet been possible to ascertain whether the type itself is of similar origin It may 1 This and the following sections on the pottery are contributed by Mr Krishna Deva 45 PLATE XXV To face plate XXVI A T\pe /, \anetit"* of rvuletnn% B T\pc /, varufies oj rouletting To face plate XXV PLATE XXVI A T\pe I \arieiu s of r B T\pe /, tourw rouletiing PLATE XXVII To face page 47 A Rouletteii \\ure /, from Cluindru\a{li, \f\sort' , 2 anil J, from 4marCi\iHi B. Base ofpre-Arretme pot sho\\ mg basket'impression ARIKAMEDU occupation From Amaravati are derived two sherds of rouletted ware of Ankamedu Type 1, now in the Government Museum, Madras (pi XXVIIA, 2 and 3), and rims of the same type have recently been picked up on the Amaravati town-site (fig 13, v and vi) In North India, in Jaipur State, a fragment of a dish with an mturned beaked nm generally approximating to the Ankamedu type and with a hard grey fabric and polished brown surface, also analogous, was found at Baira{ and is preserved in the Archaeological Museum at Jaipur (fig. 13, in) The fragment is too small to indicate whether the base was rouletted, and this feature must not be assumed The rim, unlike the Ankamedu type, is upturned, as is that of a further example of similar ware from Bairat (fig 13, iv) For other resemblances between northern wares and those of Ankamedu, see p 93 The following are selected examples of Type 1 and its variants (fig 12) 1 Polished grey ware of fine fabric with black slip inside and on outer base, and light brown on outer nm This is the arch-type The specimen illustrated has, however, individual features m an inconspicuous external cannation near the rim and a groove below, which runs less than half-way round the pot Like a majority of the type, it bears two rows of fine roulettmg From a pre-Arretine layer (AK V, 1 1 A) , early first century A D \a Polished grey ware with pink patches on exterior and black-slipped interior, and two rings of rouletted pattern Similar to 1 but with a more pronounced inward curve From Pottery Group A, Early Phase of AK TV, containing an Arretme sherd stamped ITTA, mid first century A D \b Similar to 1 but with ribbings on the exterior From the 'Arretme' horizon, AK V, 8N Ic Variety with unusually prominent beaked run projecting inward, and a series of external grooves Fabnc as I From a pre-Arretine layer (AK V, 1 1A) Id Smaller and thinner variety of Ic with a more incurved side Polished grey ware of fine fabric, greyish blue slip inside and on outer base, and light brown on exterior nm From a pre-Arretine layer (AK V, 1 1A) \e Rim-fragment of polished greyish brown ware of fine-grained fabric with a blunt but prominent notched beak and external groove From a pre-Arretme layer (AK II, 14) If Fine grey ware, polished black slip inside and brown outside From a layer containing Arretme ware (AK II, 1 1C), first half of first century A.D \g Fine grey ware with pink patches Shape similar to If Traces of greyish black slip A svastika cut on the outside after firing From the Late Phase of AK IV , probably second century A D l/i Similar but slightly inferior ware, black internally and brown externally, with blunt beak and high external grooving This example represents a small bowl rather than a dish, and ma) not have had internal roulettmg From Pottery Group A on AK IV , approximately mid first century A D U Similar ware to la Blunt beak and high external grooving, former presence of internal roulettmg uncertain From Southern Sector, 'Main Drain' Period, first-second century A D B LOCAL WARES Introductory note Previous sections of this report have dealt mainly with imported wares or motifs But by far the greater mass of the Ankamedu pottery consisted of local wares, to which the imported material now for the first time gives an approximate chronology With a few exceptions, the local pottery was turned on the wheel The exceptions include a class of portable ovens, rings of 'ring-wells ', some large troughs and storage jars, and conical jars of Type 75, the last being partly wheel-turned and partly hand-made The normal d6graissants are sand and grit, but mica and straw or husk are occasionally used for this purpose. The fabric was usually a porous clay containing a fair amount of sand, which burned pink or greyish red to light red under oxidizing conditions of firing and dull grey to greyish black under reducing conditions in the kiln For a special class of grey or greyish pink wares a superior quality of clay, free from grit and remarkably fine-grained, was employed. Slips are fairly common, and a fair percentage of the pottery was either 49 ANCIENT INDIA, NO. 2 salt-glazed (below, p 93) or treated with a burnished slip. The vast majority of the types are plain and utilitarian in character. Decorated types are rare, and decoration is generally simple and primitive in character, the commonest being finger-Up ornament A small percentage of Arikamedu pottery was subjected to inverted firing, i.e. was placed upside-down in the kiln with a fair quantity of carbonaceous matter stuffed inside As a result of this firing, the whole of the interior and the portion of the exterior (usually the nm) which came into direct contact with the fuel turned jet black, and the remaining outer surface became red under oxidizing conditions and grey under reducing conditions in the kiln For the sake of brevity we shall refer to the former as black-and-red ware and the latter as black-and-grey ware The effect was strikingly picturesque if the pots were treated with a slip and salt-glazed (below, p 93) in course of firing, a feature which is characteristic specially of the 'Arretine' period of the Northern Sector (first half of first century AD) For purposes of classification, the local pottery is divisible into two broad groups, the first coming from the Northern Sector and the second from the Southern The wares from the earliest layers of the Northern Sector are predominantly grey, with a sprinkling of red wares mostly painted with haematite or bright red slip In later deposits of this Sector there was a progressive increase in the quantity of red ware, until it assumes panty with the grey and finally outnumbers the latter as the top layers are reached In the Southern Sector, the occupation of which began later and lasted later, a crude type of pottery overwhelmingly red in colour is predominant Though many types of the Northern Sector are present here, some new varieties of familiar types and a number of altogether new types are in evidence , whilst the fabric, even in the earliest layers, is generally inferior to that of the Northern Sector, and becomes increasingly degenerate towards the top layers The conclusion is that the pottery of the Southern Sector represents in the main a continued process of devolution already incipient in the later layers of the Northern Sector The occasional introduction of new types in the Southern Sector seems to indicate, however, that other influences were at work alongside the normal processes of evolution or devolution For the convenience of study, the local pottery and other finds from Arikamedu may be divided into the following chronological classes, corresponding with the stratigraphical evidence A Northern Sector (first century AD) (1) 'Pre-Arretme', le finds from the layers underlying those which yielded Arretme pottery and ascnbable to the end of the first century B C. or the beginning of the first century A D , (2) 'Arretine', i e finds from the layers yielding Arretine pottery, and other equivalent layers of the first half of the first century A D (mainly c A D 20-50) , (3) 'Post-Arretine', i e finds from all the upper layers, which in the absence of any significant sub-division will be dealt with together They are of the mid or late first century A D B Southern Sector, mostly AK IV (mid first-second century AD) (1) 'Pre-structural', shortly before the middle of the first century A D , (2) 'Early Phase', mid or late first century A D , (3) 'Mam Dram Period', i e finds contemporary with the elaborate system of drains and connected walls of Middle Phase, sub-period 1; first-second century AD., (4; 'Late Dram Period ', i e. finds contemporary with the additions and restorations to the mam dram and other connected walls of Middle Phase, sub-periods 2 and 3 , second century AD., so To face page 50 PLATE XXVIII /)r/v<> \ho\\mg basket-impression and graffito decoration B Decorated sherdsfrom pre-Arretme lay ers PLATE XXXI To face plate XXXII Graffito So {them Set for, MttMlt Pha\t \ B Pre~ 4rrelme sherd, \ortlnrn Sector \ C Painted and incised sherds , Southern Sector, Earl \ Phase To face plate XXXI PLATE XXXII A Pu\l-4rrctini \heid, \itrtliern bettor /)<< orated \herd\ , Southern Set lor Laic Drain Pha\c AR1KAMEDU c 2b 3b 3c 3e 4b C 2c 5b Fio 14 Types 2-5 53 ANCIENT INDIA, NO. 2 6b : ] ed a ) J ARJKAMEDU It may be added that a number of superficial and disturbed layers yielded sherds of Chinese celadon ware, evidently deposited during the spoliation of the site in the middle ages (below, p 91) Types 2-148 Type 2 is a dish identical in fabric with Type 1 (p 46) but lacking the inward projection or beak at the rim Like Type 1, it has a sharply incurved side and the lip is usually facetted at the edge, but, unlike Type 1, it constitutes one of the minor types of the site The arth-type and the variant 2a with a slightly thicker nm arc occasionally found throughout the occupation of both Sectors, though they arc relatively more abundant in the pre-Arrctme and Arrctine layers of the Northern Sector Variant 26, with a thinner and more pronouncedly incurved side and a larger number of facets at the nm, is peculiar to the prc-Arretmc layers Though of a markedly coarser fabric, variant 2c shares all the other features of the type, including its polished surface , it is confined to the post-Arrctme strata of the Northern Sector Type 3 is a dish with a beaked rim similar to Type 1 but, like the majority of the local types, is of an inferior and coarser fabric , also it is invariably devoid of rouletted decoration This is another of the minor types of the site, though it is found in all periods of the two Sectors Variant 3a deviates from the mam type in the decoration of the inside of the base with rows of roughly incised concentric circles This feature is peculiar to the pre-AjTctmc and Arrctme layers Variant 36, with an almost cannated profile, is a unique specimen from an Arretine stratum (AK VIII, 5) Variants 3c, 3d and 3/, which are invariably grey wares, belong to the prc-Arretine strata , while 3c occurs occasionally from the pre-Arretme period to the Late Phase^of AK IV, exclusive!) in red ware Type 4 is a simple dish with an incurved rim and a rounded base and, amongst dish-forms, is next to Type 1 in frequency This type is of similar fabric to Type 3, and occurs in a variety of colours such as grey, red, black- and-red, or black-and-grev. due to differential firing Variant 4a, having an internally clubbed rim, is a unique pot from a mixed group (AK VIII, pit 1 ), while 46 with a sharply incurved side is remarkable as much for its shape as for its whitish fabric Only two examples of the latter are known, both from the Late Phase of the Southern Sector Type 5 is a carmatcd dish with an almost straight side and is of the same coarse fabric as Type 3 This is a very rare type and does not occur earlier than the post-Arretine period in the Northern Sector and the Early Phase in the Southern Variant 5a differs from Type 5 in having a slightly concave profile, while 56 is a dish of finer grey fabric peculiar to the Early Phase of AK IV, including Pottery Group A Type 6 is a rare dish with an incurved rim and flattened sagger base The arch-type comes from the pre- Arrctine layers and is only found in black-and-red ware of coarse fabric The specimen illustrated has, however, individual features in being decorated with mat-impressions (pi XXVIII \) over the exterior of the base, which is embellished with a large lotus flower incised after baking Variants 60-6 are grey wares occurring exclusively in the pre-Arreune layers and are even rarer than the main type Variant 6a f with a lip thinner than the arch- type, is incised with numerous concentric grooves on the ulterior of the base, while 66 is distinguished by a slightly concave base Variants 6c -d are unique miniature dishes of fine grey ware (the former from Pottery Group A', i e Early Phase, and the latter from the Main Dram Phase of the Southern Sector), which ha\e similar profile as 6a but show a pronouncedly concave base Variants 6f-/are thicker grey-ware dishes of fine-gramed fabric, and occur at all periods in the Northern Sector and in the Pro-structural and Early Phases of the Southern Sector Type 7, a dish with a rounded base, is distinguished by a clubbed and externall) grooved nm Variant la is larger in size and has a deeper groove than the mam type, while 76 has a smaller nm and more rounded profile The type is extremely rare and belongs to the post-Arrelme period, while its \anants come from the Mam Drain Phase of the Southern Sector The arch-type is of coarse greyish buff ware, black inside and on outer rim, and greyish buff on the exterior of the base Variant la is of black-and-red ware of coarse fabric Variant 76 is of greyish buff ware of a fine-grained fabric with grey core and smoothed buff-slipped surface inside and out Type 8 is a dish distinguished by a sagger base and constitutes one of the most characteristic types of Arikamedu It is of a coarse fabric and is generally grey in the pre-Arretme and Arretine layers and red in the post-Arretme period of the Northern and all periods of the Southern Sector Some of the grey examples are salt-glazed (see below, p 93) Though it was primarily a dish, some pots of this type may also have been used as lids The arch-type and all its variants save 8/-, 8y, 8/ and 80-z are found in abundance from the earliest to the latest occupation of both Sectors Variants $a-e and the miniature variety 8 have a similar base to the arch-type but differ from it in rim-form Variants 8/and 8/ are peculiar to the Arretine layers of the Northern Sector, while 8$ and 8; arc confined to the pre-Arretme layers Variants 8/1-1 form a class by themselves by reason of their distinctive sharpened nm-fonn Similarly variants 8m-/> constitute a homogeneous group in 55