A final aim of this atlas is to contribute in modest measure to redressing a con- spicuous imbalance in the presentation of South Asian history, which, despite the recent growth and vigor of the historical profession within the region, remains ex- cessively preoccupied, in our judgment, with the impact of the West on South Asia and with the roles played by specific Westerners. As a simple illustration of a cor- rective, to be found nowhere else, we cite our map (on plate VIII.C.1) of "Cen- ters of South Asian Religious Movements Abroad." Maps of Western missionary activity in India, by contrast, can be found in abundance.

What we have been able to map, and how we have portrayed those subjects we did map, inevitably reflects the type of source materials available to us and the quantity and quality of the data they contain. Wherever practicable, we have relied on primary data: archaeological, epigraphic, numismatic, textual, statistical, and cartographic. In the case of the last three categories we refer, of course, to texts, statistics, and maps prepared in or shortly after the periods being considered. Orig- inal sources in Sanskrit, Pali, Persian, and Arabic have been used in abundance, as have their translations into English or other modern languages, both European and Indian. Sources in classical Western languages and in Chinese have been uti- lized in translation only. But, given the unavailability of many potentially useful sources and the limited time at our disposal, we frequently had to forgo the ideal of relying on primary sources. We then resorted to secondary works, which, as will be evident from the lists of sources following the various units of the text and from the atlas bibliography, exist in great number.

Though the several sections of the atlas text discuss sources in considerable de- tail, a few broad generalizations by period are in order here. For pre-and proto- historic times, we must rely almost wholly on the data of archaeology, supple- mented to a minor degree by independent deductions about physiographic and climatic history. For the historical period before the advent of Muslim rule, pri- mary reliance is on epigraphic and numismatic evidence; but inferences from nonhistorical writing, some of which contains abundant, even if incidental, refer- ences to contemporary peoples and places, are very important, especially for the pre-Mauryan era. There is as yet, except for Sri Lanka and Kashmir, little in the way of historical texts of South Asian provenance, though the historical and de- scriptive accounts of classical Western authors and the records of Chinese pilgrims are exceedingly valuable for the specific periods when they were written. But much of the record is uncorroborated, inconclusive, or one-sided, and enormous lacunae persist. Some of the gaps in the available record are surprising. Extant Greek and Latin sources, for example, completely ignore the great Mauryan, Aśoka, even though he named in his inscriptions five contemporary Hellenistic monarchs ruling in Western Asia and lands beyond as far as Epirus. On the other hand, the name of Alexander, whose Indian exploits are richly related in the history and legends of Europe, is not so much as mentioned in ancient Indian literature. During the period of the Delhi Sultanate and the Mughal Empire, the writing of history relat- ing to India becomes increasingly abundant, and such writing is supplemented by an ever-increasing volume of travel literature from Muslim and European sources. Finally, for the colonial and postindependence periods, we have a plenitude not only of historical accounts, but also of maps, statistics, official reports, and other source materials. In certain cases we have such an embarrassment of riches that our problem is not so much one of finding adequate data as of choosing wisely from what is available.

The sketchiness of the materials available to us for some of the more remote periods of history affords wide scope for individual interpretations of what few data we do have. The ten views of the limits of the Ku&stod;ā&ntod;a Empire (p. xxxiii) illustrate what varied conclusions can be reached by able and conscientious schol- ars utilizing essentially the same set of source materials, with due allowance for the varying dates of the studies. The nature of the source materials, excepting literary sources, is indicated on the plate itself. Coins are extremely important, not only for the material thereon, but also in terms of the distribution of their find-spots. In- scriptions, now numbering over one hundred, are widely scattered and not richly informative from a geographical standpoint, though they do provide useful evi- dence for determining frontiers; the brief Ara inscription, which we have illustrated, provides a typical example. Archaeological remains are concentrated in only a few widely dispersed localities, where there is little question about a major Ku&stod;ā&ntod;a presence.

In presenting the several views of the extent of the Ku&stod;ā&ntod;a Empire it is not our intention to cast aspersions on the scholarship underlying them. The problem of the Ku&stod;ā&ntod;as is particularly vexatious, and different lines of reasoning legitimately lead to different conclusions. A part of the problem lies in the fact that, notwith- standing the dated inscriptions and abundance of coins, the very dates of the prin- cipal Ku&stod;ā&ntod;a monarchs are subject to wide differences of interpretation. Estimates of the date of accession of the great Ku&stod;ā&ntod;a emperor Kani&stod;ka, for example, used to range from 58 B.C. to A.D. 288, but at present the opinions of authorities range between A.D. 78 and 244. With such an uncertain chronology, placing whatever evidence we have on the Ku&stod;ā&ntod;as into its appropriate historical-geographic con- text becomes exceedingly difficult; yet it is precisely in regard to context that evi- dence must properly be weighed. In these circumstances it would be folly to put forward our own view of the maximum limits of the Ku&stod;ā&ntod;a Empire (map (j)) with any sense of certitude. We do, of course, have definite reasons for portraying each portion of the Ku&stod;ā&ntod;a frontier in the particular position we have given it, and we have been more meticulous than other authors in making the frontiers conform logically to physiographic constraints to expansion in many sectors and in avoiding alignments along modern political boundaries. We have also hedged our view by showing large areas as "possible included or tenuously held," a form of equivoca- tion most other authors did not allow themselves. Regrettably, given the limited space at our disposal, in setting forth our interpretation of the extent of Ku&stod;ā&ntod;a domains in the text for plate III.C.2, we can do no more than outline the principal reasons for our view.

While there are probably few, if any, other ancient states delineated in this atlas that pose problems as formidable as those relating to the Ku&stod;ā&ntod;as, there were sev- eral apparently powerful, but very likely ephemeral, polities about which we know so little that it would be pointless even to try to map them. The Aulikaras of Mālava, who seem to have defeated the mighty Hū&ntod;a king Mihirakula and forced the Hū&ntod;as to withdraw from India, are one such power. For the remaining ancient states within our purview we are dealing with varying degrees of certainty or, more precisely, uncertainty. The evidence we present in the text for each of our views— which is almost never as full as we would wish to make it—should be weighed on its individual merits. We can offer no guidelines or helpful generalizations as to when the reader ought to repose faith in us and when he should remain highly skeptical.

Several caveats, however, are in order. First, the inscriptions on which the limits of states are largely based are frequently panegyric; they laud and exaggerate the achievements of monarchs and rarely, if ever, take note of defeats or failure to attain goals. If, then, we are to learn at all of the failures of king X it may only be through the claims of king Y, with whom he interacted. Second, conquering an area or marching an army through it is not the same as holding it, not to mention integrating it into the dominions of the putative conquering state. A single inscrip- tional claim, unsupported by corroborating evidence, offers a slim foundation for any firm conclusions about the territorial extent of specific polities. Third, a system of stable, clearly delimited state boundaries, like that which characterizes most of the world today, simply did not exist in ancient India. Properly, we should be speaking of "frontiers."² States were in constant flux and the degree of allegiance kings could command from vassals or provincial governors generally wanted with increasing distance from centers of power and often changed radically from one generation to the next, depending on the strength of character and the organiza- tional ability of particular sovereigns. Fourth, the constitutions of states varied with time and place. Monarchies were the rule, but not a few states were republi- can oligarchies. Most states were unitary, but some were confederal. Some, such as the Mauryan Empire, were highly centralized; others, the Guptan Empire, for example, found it expedient to diffuse political authority widely. Hence, showing an area as being "within" a particular state does not connote the same thing from one map to another. Finally, in the portrayal of the maximum extent of a partic- ular state, it should not be supposed that the entire area outlined was under the control of that state at any given date, though in certain instances that might well have been the case. More commonly, states attained their maximum extent in dif- ferent quarters at different times, perhaps gaining in one sector during the very period when they were waning in another.

The political plates in sections III through VI of the atlas relegate the represen- tation of frontiers to inset maps, if they are shown at all (the map of Mughal ad- ministration on plate VI.A.2 being the sole exception to this rule). Also charac- terizing those plates, as previously noted, is the fact that they relate not to a single point in time, but to a period that may extend up to several centuries. The unwary map user may, in surveying those plates, impute to them far more stability than is warranted, although a careful reading of the map notes, dynastic chronology, and the atlas text should disabuse one of any such misconceptions. For an illus- tration of the remarkable shifts one might find in a relatively short time span, we refer the reader to the map plate entitled "Political Flux on the Northwest Thresh- old of South Asia, c. 130 B.C. to A.D. 78" (p. xxxiv). This series of maps, adapted with only minor changes from a set provided by Dr. K. Walton Dobbins, is based primarily on the find-spots, mint locations, and dates (known and supposed) of coins of the individual polities or sovereigns included thereon. The time span of the individual maps ((c)–(s)), it will be noted, ranges from a minimum of only seven years to a maximum of twenty-five. If, as many scholars suggest, the prove- nance of ancient coins (of which thousands have been considered in the making of Dobbins's maps) provides a fair guide to the regions under the political control of the issuing authorities, and if Dobbins's coin identifications and datings and those of other numismatists on whom he partially relies are correct, then we have in the map series a vivid representation of the swift ebb and flow of power in the politically turbulent area and period under review.³