Excavated sites related to a hunting economy cover a wide environmental range. Among the earliest such sites may be Birbhanpur in Bengal, some sites near Allahabad, and several coastal locations in Tamil Nadu. Birbhanpur has quartz micro- liths appearing together with faint traces of shelters, and soil studies suggest that the habitation may belong to a period when the climate was drier than at present. Though some archaeolo- gists believe that these microliths antedate the more westerly microlith sites, the climatic evidence in favor of such a position is hardly conclusive (cf. text for plate I.C.1). The finds at Sarai Nahar Rai, near Allahabad, associate microliths and crude pottery with the remains of fish and of deer and other large ungulates. A ¹⁴C date of c. 8000 B.C. (uncorrected) makes a Homo sapiens find there the oldest hominid find in South Asia. The fossil sand dunes (teri) on which the coastal microliths at the tip of the peninsula are located also provide them with a claim to considerable antiquity. Bifacially worked points and transverse arrowheads occur there that are other- wise rarely found, although blades and crescents remain in the majority. As in the previous period, the teri sites and the sites near Bombay suggest that marine life was exploited as a food source, but the artifacts themselves give no clear indication of this, for a similar industry occurs at Jalahalli, near Bangalore, where the open plateau presents a quite different environment.

Langhnaj in Gujarat and Bagor in Rajasthan show that an initial prepottery microlith phase with a hunting economy is in time replaced by pastoralism. Langhnaj, like many Gujarat sites, is on a loess dune at the edge of a waterhole. The earliest levels there show a microlithic industry and include humans buried in a flexed position along with shells that were brought from the coast. Bagor, a riverside site, for which the ¹⁴C dates range between c. 3000 and 4500 B.C., shows substantially the same cultural traits. At both places the early strata grade into a later horizon (for which ¹⁴C dates are lacking) characterized by the presence of pottery, iron arrowheads, and ring stones (possibly used to weight digging sticks). At Bagor glass beads now appear, and there is also evidence for a human cremation. Mixed with the latter remains was a calcined feline tooth, per- haps worn by the cremated person. Other animal remains from this late microlithic phase were those of deer, boar, wolf, rhi- noceros, sheep, goat, cattle, and buffalo. At Adamgarh, men- tioned previously for its Stone Age sequence, the microlith deposits reach to a depth of 150 cm below the present surface. From the upper layers, dog, humped cattle, buffalo, goat, sheep, and pig remains occur in proportions equal to those of feral animals such as deer and porcupine. The ¹⁴C dating here of c. 7450 B.P (corrected)¹ appears early in comparison with another date of c. 3660 B.P (uncorrected) obtained for com- parable levels at the Jambudip and Dorothy Deep sites. At Bagor these levels have not been dated by the ¹⁴C method, but the association with iron can hardly make them earlier than the first third of the last millennium B.C.

In eastern India, microliths and pottery of uncertain age are reported from Selbalgiri in the Garo Hills, from parts of Orissa, and from the Godavari River delta. Now for the first time Sri Lanka (Ceylon) enters the archaeological picture. Three stone axes found there are said to be in the Acheulian tradition. Should this claim be substantiated, revisions in the present understanding of Sri Lanka's early cultural development will be necessary. Microlithic sites there mainly cluster about Ban- darawela, where they occur near caves, but in the vicinity of Jaffna a few microlithic sites were located in the open plain. Typologically the tools from Sri Lanka are related to those of south India, whence they might well derive. From one of the southern Sri Lanka sites, Bellanbandi Palassa, come a dozen skeletons of Homo sapiens. Their morphology closely ap- proaches that of the primitive Veddas, but they are evidently of recent times, since ¹⁴C dating places them in the 1st century B.C. (uncorrected). Further discussion of microliths in an agri- cultural context is best deferred until the peninsular farming communities are considered.

Sources (in addition to those in the General Bibliography)

B. Allchin (1963); K. W. Butzer (1971); J.-M. Casal (1970-71); A. H. Dani, ed. (1967); S. Deraniyagala (1971a), (1971b); H. de Terra and T. T. Paterson (1939); L. Dupree ¹ In speaking of "corrected" ¹⁴C dates throughout this section, we assume a half-life for radiocarbon of 5,730 ±40 years, rather than the previuosly used figure of 5,568 ±30 years. The ¹⁴C dates of the graph on plate II.2 and dates provided on maps throughout section II are also based on the 5,730 ±40 years half-life. Recent research reported on by Dales (1973) indicates, however, that the half-life of ¹⁴C has varied over time and that our "corrected" figures are in need of further adjustment, the age to be added generally increasing with time. For example, it is suggested that our ¹⁴C dates 1551-1750 B.C. should be some 200 years older than we have cited, while those from 2651-2850 B.C. should be 650 years older. But in our judgment the art of dating currently appears to be in sufficient flux not to warrant additional "corrections" being made in the data prepared for this section. (1973); R. V. Joshi (1968); V. D. Krishnaswamy (1949); B. B. Lal (1958); L. S. Leshnik (1968a), (1968b), (1972); T. T. Paterson and H. F. H. Drummond (1962); H. D. Sankalia (1965); M. Seshadri (1956); K. R. U. Todd (1950); F. E. Zeuner and B. Allchin (1956).

General

The text relating to the maps on plates II.2–5 is arranged to follow regional developments, rather than cultural stages, throughout the subcontinent. Although the latter method is commonly used to organize the presentation of archaeological data and is followed generally in the order of presentation of our maps, in the case of these plates it imposes a strain on chronological coherence. The different rates of cultural evo- lution in the various regions of the subcontinent would require great time leaps if similar stages were to be treated under a single rubric (see chart on plate II.2). Hence, the method adopted for the following text is to treat continuously develop- ments in the northwestern part of the subcontinent from the beginnings of settled agriculture, through the rise of civiliza- tion in the Indus Plain, and up to the "dark ages" following its decline. Thereafter, the geographical focus of the text moves to northern and central and then south India and the advent of farming in those regions.

Early Village Farming Communities of the Northwest

The advent of sedentary agriculture on the northwestern fringes of South Asia in the late 4th millennium B.C. is a con- sequence of the flow of ideas and the movements of peoples from the Iranian Plateau eastward (see plate II.3, maps (a) and (b)). Its practice took hold in the northwest of the Indian subcontinent and in the peripheral highlands of what are now Pakistan and Afghanistan a thousand years or more earlier than it did in the Gangetic Plain and peninsular India. (These latter regions are treated on plate II.2 "Neolithic and Chalco- lithic Cultures," which typologically follows those illustrated on II.1, relating to the Stone Age.)

Before the development of fully settled life in the northwest, the meager evidence from that general region points to an era of seminomadism based on the herding of sheep, goats, and Indian cattle, together with rudimentary agriculture. This mode of life is difficult to document archaeologically; yet the circum- stance that it has not been recognized either in the hills of southern Baluchistan or along the Makran coast may be more than mere chance. The sites of this period are all in north Baluchistan and the plain of Kandahar—that is, close to the traditional routes that lead from the Indus Plain to Iran.

The earliest levels at Mundigak, a continuously inhabited site hidden in a mountain valley roughly fifty miles northwest of Kandahar, provide the best evidence for this stage, called the Pre-Indus Period. The first signs of human presence there are an assemblage of pottery, steatite bowls, lapis lazuli beads, a copper pin, and some bone awls and stone blades, Signifi- cantly, there are no structural remains of any kind in this ear- liest phase. Thereafter, however, mud brick walls appear, and a quantity of small terra-cotta models of humped bulls add a new feature to the group of small finds. Although specific re- semblances are few and the ceramic fashions tend to be diver- gent, comparable assemblages occur in the Zhob River valley (Rana Ghundai), in the Quetta region (Kile Ghul Moham- mad), along the Gomal River (Gumla) and in central Balu- chistan (Anjira). Farther east this phase has been uncovered in Sind (Amri) and the Punjab (Sarai Khola). An unusual feature of the stone industry at Sarai Khola is that, in addition to flint blades, stone celts are present, which are otherwise little known beyond peninsular India. A central dating for this pe- riod falls in the middle of the 3d millennium B.C., and, as the steatite and lapis lazuli objects reveal, these settlements al- ready participated in a network of long-distance trade. From Badakhshan lapis lazuli found its way to the Iranian Plateau, and the track of distinctively carved steatite vessels has been traced beyond Sind into Mesopotamia. But for all this, the subsistence base remained mixed farming. Wheat and millet were the chief grains, and cattle, sheep, and goats provided the animal products. A clue to the farming techniques of the period is supplied by the abundant remains of dry stone walls throughout northern and central Baluchistan. These structures were erected across the mouths of narrow hill valleys, thus closing them and preventing soil runoff during the short sum- mer rains. The accumulated soil at the bottom of the valleys then provided fertile ground for grainfields.

As the next period dawned, settlements at a few of the older sites (Mundigak, Gumla, Amri, Sarai Khola) merged into it ² The terms "Harappan Civilization" and "Indus Civilization" or "Culture" are used interchangably by many scholars. Since the sites known to have been associated with that culture extend well beyond the Indus region, we generally prefer the term Harappan, relating to the locality at which the distinct nature of the culture was first recognized. gradually without perceptible discontinuities (plate II.3.b). A ¹⁴C dating from Gumla puts this time at c. 2800 B.C. (cor- rected). But at other places, such as Rana Ghundai II, a new, finely painted pottery marks an abrupt break with the former traditions, and contemporary nearby Periano Ghundai I intro- duces typical Zhob "mother godess" figurines in terra-cotta alongside bronze leaf arrowheads and terra-cotta bull models. From Seistan to the Indus Plains a considerable number of new settlements were founded at this time. The evidence shows that the newcomers had strong links with the west, and in Baluchistan foreign and old local traditions mix to give rise to a new cultural synthesis. One of the major archaeological achievements of the last two decades has been to establish that the new western influence originated in southern Turkmeni- stan.

By the middle of the 3d millennium B.C. farmers of the Jeitun (Namazga III) region had already developed a complex irrigation network and inhabited planned towns. The geometric designs they painted on their pottery unmistakably relate it to ceramics found in the Quetta valley at Damb Sadaat and at many other mounds. The character of such long-range affin- ities is being given close attention. The emergent pattern is one of a movement southward out of southern Turkmenistan to the Seistan Basin, thence eastward along a former course of the Helmand River, to the plain of Kandahar through the val- leys of the Sulaiman mountains, and into Sind. This route is well marked by related settlements, the largest of which is the seventy-five-hectare Shahr-i-Sokhta in Seistan. The central lo- cation and extent of this site justify the claim that it was the hub of an important lapis lazuli trade route from Badakhshan to Mesopotamia. A hoard of golden vessels found at Khosh in western Badakhshan strongly supports other lines of evidence, since the style and subjects of their designs link them with Mesopotamia. Mundigak III, a station along this route, has in its assemblage circular and rectangular flat seals, wheel-turned polychrome and red pottery with black painted designs, and a number of bronze and copper implements, the most prominent of which is a shafthole ax. A cemetery attached to the settle- ment at first contained only flexed burials but later included communal ossuaries. Contemporary settlements connected with the lapis lazuli trade on the Pakistan side of the border were established at Sarai Khola, Jalilpur, and Pandi Wahi.

In Sind and Rajasthan the larger sites were protected by massive walls of mud brick. Of these, Kot Diji and Kalibangan have been extensively excavated. In the lowest levels they re- veal a number of cultural traits once thought unique to the mature Harappan Civilization. In the settlement pattern a cita- del/lower township division is already evident, and the do- mestic architecture made use of multiroomed houses. The early pottery, though generally different from that which fol- lowed, was painted with the fish-scale pattern that dominates the mature phase, and in other small finds, such as terra-cotta cart models, flat triangular votive cakes, and copper celts, there is undoubted continuity. At Kalibangan there are clay ovens much like those in use today, and lined water-storage pits. Saddle querns for grinding grains and spices are also found. The remarkable discovery of a furrow-marked field belonging to this Early Indus period, as it is called, lays to rest the ques- tion whether the plough was known in the later, mature civili- zation. Yet the seals and writing system typical of the next period are wholly lacking.

The archaeological record in the mid-3d millennium B.C. shows an increase in characteristically Indian motifs and a concomitant decline of Iranian ones in the Baluchistan settle- ments. This is especially evident in the painted pottery tradi- tions, where the ibex is replaced by the humped bull and pipal leaves also become common. The times were unsettled, with evident wide abandonment of both large and small sites. Shahr- i-Sokhta IV came to an abrupt end and was never rebuilt. Mas- sive defense walls were erected at Mundigak IV; Rana Gundai II and Kot Diji II ended in conflagrations; Jalilpur became de- serted; and Kalibangan's early culture also met with catastro- phe. What caused these changes is unknown, but it is thought that there was an earthquake at Kalibangan. When this site and Kot Diji were resettled, after an interval of undeter- mined duration, it was by carriers of the urban Harappan Civilization.

The Harappan Civilization

The discovery of the Harappan Civilization in 1922 is the most glorious achievement of South Asia's archaeology. It has fundamentally changed all previous concepts of the origins of Indian culture, whose roots hitherto were thought not to reach in any important way deeper than the advent of the Vedic Aryans. Indeed, so central was that idea that after the major cities of Mohenjo-daro and Harappa were excavated, several scholars attempted to establish these same Vedic warriors and pastoralists as their builders.

To date, nearly 300 large and small sites of the Harappan Civilization are known. They cover an area of half a million square miles, centering in the plain of the Indus River, whence the name often assigned to it, and reaching out to the Makran