Orbis, Spring 2003

Mapping the Past: Historical Atlases

by Jeremy Black

Jeremy Black is professor of history at Exeter University. His books include Maps and History: Constructing Images of the Past (New Haven: Yale University Press, 1997) and Maps and Politics (Chicago: University of Chicago Press, 1997).

A cartographer makes decisions based upon what he has observed or perceived, creating the map’s ostensible story. Whatever the merits of the argument that it is necessary and possible to seek objective cartographic accounts of the modern world, it is clear that selection plays a central role in historical mapping.

Historical atlases, narrative in both form and content, offer the ultimate narrative: the past. The dynamics of this relationship can be appreciated by considering both past atlases and also current work, which is infrequently considered when looking at cartography.

Here we will follow the conventional approach and offer a narrative from past to present, presenting the opportunities and challenges, which include cost and profitability pressures on the production of the atlas, as well as problems of information (what can be mapped?) and space (what among a range of possible topics and treatments should be selected?). The interaction of these problems and opportunities during the course of historical cartography has produced changing narratives. History has been “composed” very differently in thematic subject (the extent to which non-political issues have been treated), scale (the balance between global, international, and regional coverage), and treatment (reflecting changing views of concepts of class, race, [1] ethnography, and gender).

Some atlases make reference to the issues involved in their compilation and production. To the degree that they are self-referential in this way, we are more conscious of the role of choice, and thus, of the atlas and its producer(s) as narrators. This self-referencing can cover a range of matters including choices of what to cover, how best to cover it, organization, and cost or marketing.

Self-referencing is not new, nor is it only a feature of the present preference for postmodernist discourse. For example, Reginald Lane Poole, in his preface to the classic Historical Atlas of Modern Europe from the Decline of the Roman Empire (Oxford, 1902), specifically located his work with reference to the earlier German works that still then defined cartographic quality. He acknowledged his debt to Theodor Menke and outlined his rationale for preferring Menke’s scheme of “presenting Europe as a whole and each country by itself, in a series of maps showing its geography at different periods” to Gustav Droysen’s single chronological series.

Until the twentieth century, historical atlases’ content was predominantly defined in terms of international relations—particularly warfare and shifts in control over territory. The state was assumed to be the crucial unit (and objective) in the historical process, and atlases were accordingly concerned with changing state boundaries, especially the rise and fall of empires. [2] The apparently cyclical character of the rise and fall of empires, especially the Roman empire, gave historical atlases the character of morality tales, mirroring historical works of this period, such as Edward Gibbon’s Decline and Fall of the Roman Empire (London, 1776–88).

A normative pattern for the historical atlas was set in the first half of the nineteenth century, as works appeared that were reprinted in several different countries. The most important atlas initially was the Genealogical, Chronological, Historical and Geographical Atlas (London, 1801; Paris, 1803–4) of Le Sage (Marie-Joseph-Emmanuel-Auguste Dieudonné de Las Cases). An émigré marquis who had become a teacher in London, the pedagogic Las Cases presented his historical maps as part of a compendium of historical knowledge—an approach that became less popular later in the century but has gained popularity over the past few decades. The provision of other forms of information limits the need for the maps to bear a weight of information that is frequently excessive in visual terms and which cannot be readily expressed in a cartographic form. Expanding the definition of the term “map,” Las Cases used what he termed “genealogical and chronological maps” in short genealogical tables and time-charts. He did not view these as different from the geographical maps, but presented them as devices to disseminate knowledge by fixing location. Anticipating modern atlases, Las Cases stressed the complementarity of his devices. The plan of the 1801 edition claimed that pupils would find relative spaces impossible to confuse: they would learn from the historical map the distance of time just as they learned the distance of places on the geographical map; they would determine the situation of Elizabeth with respect to William the Conqueror or George the Third just as easily as they would that of London to Edinburgh or Portsmouth.

Las Cases was explicit about the problems he faced with his sources, especially beyond Europe. He wrote of the difficulty of mapping Africa, which remained largely unexplored. “A burning climate, desolate wastes, moving sands, inhabitants either barbarous or brutish, ferocious animals, venomous reptiles, all conspire to exhibit Africa as the refuse of the world, and the malediction of nature.” Contemporary interest in ending the slave trade and exploring Africa did not lessen this widespread sense of superiority he exhibited. As to Europe, Las Cases wrote that tracing the “Barbarian” invasions of the Roman empire was like “pursuing them through their impenetrable woods and difficult morasses, where as we advance, the traces of their footsteps disappear.” Las Cases’ willingness to draw attention to limitations underlines the pedagogic quality of his work, but he remained an amateur, not academic in his approach. [3] He can be contrasted with Christian Kruse, Professor of History at Leipzig, whose Atlas und Tabellen zur Übersicht der Geschichte aller europáischen Lánder und Staaten (Leipzig, 1802–18) consisted of sequential maps on the same base map of Europe and the Near East, which facilitated comparison. These more academic works did not lack for popularity, either: the fifth edition of Kruse’s atlas appeared in 1834 and a French translation in 1836.

Method and rigor were emphasized in Karl von Spruner’s Historisch-geographischer Hand-Atlas zur Geschichte der Staaten Europa’s von Anfang des Mittelalters bis auf die neueste Zeit (Gotha, 1846), the most influential of nineteenth-century historical atlases. Its focus was resolutely on maps, an approach that would be dominant thereafter, as lithography and common color printing expanded the potential of mapmaking and reduced the cost of an atlas full of maps. [4] In another important departure, Spruner depicted the development of Europe primarily from the perspective of the growth and interaction of its states, rather than the dynasties that had been so important to Las Cases. This theme was especially important in an age of nationalism.

Spruner’s influence was spread by the heavily revised third edition produced by Menke—the Spruner-Menke Hand-Atlas für die Geschichte des Mittelalters und die neueren Zeit (Gotha, 1880)—and by Droysen’s Allgemeiner historischer Handatlas (Bielefeld and Leipzig, 1886). Along with the most popular German school historical atlas, Friedrich Wilhelm Putzger’s Historischer Schul-Atlas (Bielefeld and Leipzig, 1877), [5] these works established the content and style for historical atlases. They became the historical atlas, a form that prevailed until the late twentieth century. Other works were being produced elsewhere, especially in France, but it was Germany that strongly influenced both British and American works. Atlases that claimed in their title to be modeled on Spruner were published in London in 1853 and 1870, and editions of Spruner appeared in London and New York in 1861. And William R. Shepherd’s Historical Atlas (New York, 1911), long the leading American historical atlas of the world, drew heavily on Putzger.

A German historical atlas demonstrated what could be offered in terms of aesthetic appeal, interest, and information, and also what the nineteenth century was prepared to overlook. Fine lines and the plentiful use of color permitted the clear presentation of complex territorial information, the prime desideratum in the period. It was particularly necessary for the mapping of the Holy Roman Empire, with its profusion of states. Mapping offered internally undifferentiated blocs of territory separated by clear linear frontiers.

These maps of the Empire provided nineteenth-century readers a clear depiction of successive stages in Germany’s formation, as well as guidance to a historical landscape that had been swept clean on several occasions. Droysen’s atlas of Germany in the fourteenth century mapped a whole series of states, from the prince bishopric of Trent to the Duchy of Pomerania-Wolgast, that was far more a world of the past than the equivalent maps of France or England: Because coverage in the nineteenth century did not offer any account of internal features, such as demographic distribution or industrial location, Droysen’s map of the British Isles after 1603 was essentially that of the Britain of the readers’ day. Uprisings, both in foreign countries (the English Civil War and the Jacobite risings) and Germany (the Peasants’ War of 1524–25), were ignored. The situation was to be very different when the Communist East Germans came to produce an historical atlas.

A national consciousness was by no means restricted to German historical atlases. The introduction to C. Grant Robertson and J. G. Bartholomew’s Historical and Modern Atlas of the British Empire (London, 1905), entitled “Relation of Geography to History,” declared that this atlas was the first to offer “a correct and intelligent interpretation of the modern physical and political geography of the British empire [based] upon a geographical study of the long historic evolution through which our Empire has passed.” Studying the history and “geographical structure” of the British empire was presented as “a positive duty for every British citizen.” Robertson and Bartholomew traced a longer imperial destiny that proved that “the theory and claims of the modern imperial crown are no pinchbeck creation of the nineteenth century.” Their history of earlier imperial episodes gave British imperialism a legitimating history, one that made it seem natural.

The technique of mapping undifferentiated territorial blocs separated by clear frontiers was expanded as Western imperialism and its associated mapping reduced the rest of the world to a territoriality that was known and comprehensible in Western terms. Indeed, the expansion of European power served to organize and rank the rest of the world. In the introduction to his Historical Atlas in a Series of Maps of the World, as known at Different Periods (London, 1830), London barrister Edward Quin employed color to depict civilization, although this was very much seen in Eurocentric terms. “We have covered alike in all the periods with a flat olive shading … barbarous and uncivilized countries,” he wrote, “such as the interior of Africa at the present moment.”

In general, areas outside Europe were ignored or underrated, or their history presented in terms of Western imperialism. [6] The elision of non-Western history was bluntly presented in the map “The British Settlement of Australasia” in Ramsay Muir’s New School Atlas of Modern History (London, 1911). Muir annotated that map, “Happy is the nation that has no history. Apart from the Maori wars in New Zealand, the only noteworthy features of the history of Australasia are the dates of the successive settlements, and the chief stages of the exploration of the region.” (Eighty-five years later, the New Zealand Historical Atlas edited by Malcolm McKinnon, Auckland, 1997, included ten Papatuanuku plates on pre-contact Maori society.) In addition to imperial themes, Muir also gave considerable attention to warfare, with 22 campaign maps and battle plans in the first black-and-white section of the book. None was of a British defeat, though Crécy, Poitiers, and Agincourt each had a plan.

The notes brought out the significance of the physical features depicted: how for Roman Britain, “Not only the mountains but the forests and marshes exercised a profound influence, breaking up the country into isolated fragments. Observe the skill with which the Roman roads overcame these obstacles.” For the Peninsular War of 1808–13, a British triumph over the French, he tells the reader, “Note how the campaigns were determined by the direction of the river valleys and mountain ranges, and by the greater military roads.” For the colonization of North America, “Note that the barrier of the Alleghanies [sic] shut the English into Lawrence, cutting through the mountain-line, led the French by an easy path to the Great Lakes and the Mississippi Valley.”

Muir reflected current wisdom in emphasizing not only the role of the environment, but also the elements of statehood. The Ottoman Empire, he wrote, was “not a nation state … [A]s the Empire lacked all the elements of unity, it could have no permanence; but for the jealousies of the European powers, it must have been broken up much earlier.”

Muir's themes were echoed in the Cambridge Modern History Atlas (Cambridge, 1912), edited by another distinguished British historian, Sir Adolphus Ward, whose map of the Mysore and Maratha Wars 1792–1804 is a record of British triumphs. Less glorious and successful earlier struggles with the Mysore and Marathas were ignored. The theme of the atlas was Europe’s triumph:

The general idea of the Atlas is to illustrate, in a series of maps of Europe and … other parts of the world associated with the progress of European history, the course of events by which the Europe of the fifteenth century has been transformed into the Europe of the present day.

Muir’s atlas eventually became the Philips’ New Historical Atlas for Students (published in the United States as Hammond’s New Historical Atlas for Students), a work that went through numerous editions since its initial publication in 1911 (the eleventh edition, published in 1969, was reprinted in 1982).

There has been a significant process of change in historical atlases since Philips’ first appeared. Most important is the broadening of range. The scholarly character and overt self-awareness of Charles O. Paullin’s Atlas of the Historical Geography of the United States (Washington and New York, 1932), with its statistical analysis of its own contents, gave way to Clifford and Elizabeth Lords’ Historical Atlas of the United States (New York, 1944), which offered only 300 fairly simple maps, but was more accessible, easier to use, and less expensive than the Paullin. [7] In Paullin, traditional cartographic topics, especially military history, were relegated in importance. The introduction claimed that “most maps illustrating wars and campaigns are confused and difficult to follow, especially when a tangle of lines on them is intended to show troop movement,” [8] but lack of interest, not difficulty, was the crucial issue. The Atlas also differed from its predecessors and most other major inter-war historical atlases in emphasizing themes rather than chronology. The essential organization of the work was by topic, a process that underlines the importance of editorial imagination and direction, whereas with chronology the issue of organization is far less problematic.

The thematic maps included appreciable sections on economic history and cultural development. The first included maps reflecting the available wealth of statistical information; they were not impressionistic. For example, the development of industry was mapped in terms of the increase in the number of wage earners, and also the increase in value added by manufacture, by states, in 1849, 1880, and 1927.

The Atlas’ introduction invited readers to draw relationships between different subjects, comparing the maps on the locations of cities, for example, with “the maps showing relief, physical divisions, climatic factors, and mineral resources, all circumstances upon which the location and grouping of cities are closely dependent.” The Atlas was also sophisticated in its self-awareness: the manner in which issues of inclusion and representation were discussed:

On Plates 61-67A space was economized by marking on the same maps both colonial and state boundaries and the locations of towns and cities … Perhaps somewhat more graphic effect might have been achieved had it been practicable to separate the two elements—political units and cities—and to employ a greater range and variety of symbols to distinguish the towns and cities according to size.

There are of course problems with the Paullin Atlas. It generally ignores Native Americans except for mentioning their “shadowy claims” to land. It encountered difficulties mapping socio-economic history: unemployment and poverty are shown but not employment and wealth. Yet it included topics that were absent from most historical atlases until recent decades.

The broadening of historical atlases’ range has not prevented the continued production of traditional-style historical atlases: modern versions of works such as Spruner’s still appear, especially in Germany. Yet, there has also been a major shift towards a more impressionistic and design-oriented cartography arising from the new desire to produce maps that suggest movement. The most important example is the Times Atlas of World History (1978, 1984, 1989, 1993, and 1999), which has appeared in eighteen languages. Its appearance was an important aspect of the globalization of the historical atlas, which has interacted with an intellectual critique of Eurocentricity [9] and a call for a greater mapping of non-political topics. In combination, these have dominated the agenda in recent decades. So also has a determination to “explain” as much as to show. This clearly differentiates the German school, with its heavy emphasis on location maps, from atlases that use pictures, text, and maps suggestive of causal relationships in order to offer a more explanatory account. For example, the foreword to the Atlas of Irish History (Dublin, 1997), edited by Séan Duffy, states the intention “to spotlight the movements of Ireland’s long and multi-layered history and to explain how its present circumstances have arisen.” The map of the Easter Rising, for example, a street map of central Dublin with the rebel garrisons and the British cordon-marked, is complemented by text that explains its importance. The crisis in Northern Ireland since 1969 is less helpfully mapped, however. It is unclear how best to map the problems and events there, although it would have been useful to include maps showing the degree of sectarianism in, for example, housing as well as other maps of parades and processions comparable to those shown in the Historical Atlas of Canada II. [10] Thus maps could be used to show contested spaces in towns, employing cartography to take note of different accounts of spatiality and spatial links, a particular problem in Northern Ireland. [11]

The desire to explain is possibly linked to a lack of confidence in the ability of maps to do so. The environmentalism that flourished at the turn of the nineteenth and twentieth centuries encouraged the production of a range of work—historical, geographical, and scientific—in which maps played a major role. Some scholars discussed environmental influence in a deterministic fashion, while others preferred a concept of probabilism, but all were united in the view that physical geography counted. [12] Maps were seen as constitutive rather than illustrative of theses. R. F. Treharne and Harold Fullard, the editors of the 1969 edition of Ramsay Muir’s Historical Atlas, emphasized the educational value of physical geography, stating in the introduction that the Muir historical atlases “played a decisive part in establishing historical geography as an integral part of the study and teaching of history in Britain.” In the post-1945 world, however, the emphasis on physical geography declined. For example, the 1969 Muir Atlas dropped much of the physical coloring, arguing that the information was provided in “atlases of physical geography” and that only “where the subject-matter of the map especially requires a physical basis for its understanding” should it be included.

The reduced emphasis on physical geography in historical atlases reflected a number of factors, including globalization, the sharp relative decline of agriculture, and analytical shifts away from materialistic explanations. The net effect has been that as maps have become more innovative in design, there has been less confidence in their ability to explain on their own as opposed to describe. The consequences can readily be seen in the presentation of historical atlases. Those produced by Times Books (now part of HarperCollins), amply demonstrate this. The introduction to the Times Atlas of the Twentieth Century (London, 1996), edited by Richard Overy, a first-rate historian of World War Two, makes no reference to cartography: it could have come straight from an ordinary historical work. The inside cover blurb refers to Overy’s text rather than the maps. This trend was continued in the fifth edition of the Times Atlas of World History. Its title, The Times History of the World (London, 1999), made no reference to maps, a decision apparently taken for marketing reasons (the word “atlas” was thought to put purchasers off), and against the advice of Overy, the academic editor. [13] Cartography appears in the title of John Man’s Atlas of the Year 1000 (London, 1999), but neither the introduction nor the book’s text refer to maps or mapping.

In the Times History of The World, maps are treated as essentially illustrations. The back cover makes no reference to this being an atlas: instead it is presented as “the most important new edition of the most successful history of the world ever published.” The introduction has little to say about cartography, the maps, or their limitations, or how best to relate the maps in the book.

It would be misleading to suggest that all modern historical atlases seeking the popular market are map-unfriendly. The Cassell Atlas of World History (London, 1997), for example, made much of its maps, stating on the inside flap: “They give—as only maps can—a very clear idea of the ways in which the history of humankind has been shaped by geography and environment.” However, even here, there was only one map per spread and much of each spread was covered by text.

Scholarly works continue to place more emphasis on maps than their popular counterparts. Roger Kain’s introduction to the Historical Atlas of South-West England (Exeter, 1999), edited by Kain and William Ravenhill, is a powerful defense of maps, and also explains their methodology:

Maps are the rationale of any historical atlas … [T]he spatial distribution of features plotted out on maps can aid our understanding and help explanation—as demonstrated by the analysis of the distribution of archaeological finds in relation to the components of the physical environment. For some purposes, a map is an indispensable method of communication; for the analysis of town plans, for example, maps are both the source of data and the only practicable way of presenting the results… . The texts … provide a brief discussion of source material for maps where this is especially problematic as in compiling maps of population in the pre-nineteenth-century period. Chapter texts also reveal the mapping methodology and data assumptions where these are not clear from the maps themselves.

This policy is very different from that in more popular works, in large part because mapping methodology and data assumptions are different in character, and neither are seen as of interest to the bulk of the readers. In fact, historical atlases with high scholarly values can sell reasonably well. The New Zealand Historical Atlas edited by Malcolm McKinnon (Auckland, 1997), which was reprinted in 1997, 1998, and 1999 and has sold over 20,000 copies, ably combines scholarship with innovative cartography using a variety of projections and perspectives.

It would be misleading to suggest that commercial considerations do not play a role in scholarly and specialist works. For example, the introduction to the third edition of An Historical Atlas of Suffolk edited by David Dymond and Edward Martin (Ipswich, 1999) explains that the case for a third, extended edition became “irresistible” not only for academic reasons but also because of the “steady commercial demand” that had caused the 1989 second edition to sell out in 1993. The decision to adopt a county format was taken not only for academic, but also for “marketing” reasons, and was not without serious academic problems: many subjects that are mapped preceded county divisions or are in no way affected by them. Indeed, county divisions emulate the problems of showing history in terms of national blocs.

Through my own work in historical atlases, I may be able to add to the narrative of cartographic scholarship. The source base for modern historical atlases is extraordinarily small, largely confined to “material culture”: the atlases themselves. More detail is added by discussion in atlas prefaces and introductions, but, again, what is available is limited. Furthermore, there is an important bias in the evidence. Discussions concentrate on major and self-consciously scholarly and innovative projects, such as the Historical Atlas of Canada, the subject of the three chapters on historical atlases in Joan Winearls’ Editing Early and Historical Atlases (Toronto, 1995). [14] Crucial as these are, they do not cover the whole range of atlases, not least those designed for the wider, more popular market.

In the course of a number of projects, [15] I have been in touch with a large number of cartographers and editors. Based on those discussions, I would question the notion of the cartographer as shaping the protean past and bringing the panoptic eye of scholarship or of his/her own views and suppositions—in short as objective hero or subjective human. It is clear that, in contrast to most other books, the name on the cover—whether of author, cartographer, editor, or all three categories—tells one only so much. This is because the framework of the atlas is set by the publishers. At the most basic, they decide how long the work should be and how many maps it should contain. The last is a crucial point because maps, especially new maps, are expensive, certainly more so than text or pictures.

If the publisher does not initiate an atlas designed for the commercial market, it is most likely to have been produced by a packaging house, some of which also publish under their own imprint, although in general they conceive and create works that are then sold as a “package” to the company which publishes them under its own imprint. Prominent examples include Calmann and King, Andromeda, and Swanston in Britain, and Carta in Israel. Generally, the publisher has relatively little input into the contents of the atlas. For example, Andromeda was responsible for The Cassell Atlas of World History.

The input of major publishing houses comes in terms of the linked issues of size and cost, with the latter establishing how much can be spent on the cartography. This drives the book. It is not a case of “How many maps do you need to show your view of the history of the world?” but “You have 200 spreads each with one big and one small map. Now, how are you going to produce an attractive history?” Attractiveness is important, if the works are to sell in quantities that will permit the heavy production costs to be recouped.

The point at which these considerations come to affect the contents is where market assumptions are involved. The Cambridge Illustrated Atlas of Warfare series, for instance, was designed to be part of a five-volume series, but the nineteenth-century volume proceeded only as far as synopsis stage, and the classical and twentieth-century volumes not even that far. The first volume, The Middle Ages 768–1487 (Cambridge, 1996), edited by Matthew Bennett and Nicolas Hooper, was misleadingly named. It was restricted to Europe with an excursion into the Middle East for the Crusades. The New World, the Orient, and Africa were all neglected, despite the importance of warfare on these continents.

For the early-modern volume, Renaissance to Revolution 1492–1792 (Cambridge, 1996), I, as academic editor, found it difficult to persuade the packagers (Calmann and King) to include a spread on the Orient. For commercial reasons the emphasis was to be on Britain and the United States. Intervention could be more direct than that. In the case of China, I suggested a projection designed to offer a new perspective on Manchu expansion. Instead of the customary north-south map with the Chinese advancing westwards, as if inevitably filling a void, I wanted a west-east map, with west at the top, and the Manchu expanding from the 1680s from their base in the bottom half of the map, north-east into the Amur Valley, north-northeast into Mongolia and then Xianjiang, north-northwest into Tibet and then Nepal, northwest into Burma, and west into Tongking. This was designed to suggest activity on all the landward frontiers. It was, however, dismissed as likely to confuse. The same fate killed the “upside-down” map of the Adriatic in the Times Atlas of World History. Designed to throw new light on the geopolitics of the Venetian maritime empire, it was taken out of the fourth edition on the grounds that it discomforted confused readers. [16] Although maps such as those in the Times Atlas of the Second World War showed how novel alignments could throw light on strategic options and problems, there has been considerable resistance to them. [17]

Second, returning to the Warfare book, I produced a map of India in the eighteenth century designed to show the peripheral nature of the European impact in the first six decades of empire. The standard north-south map places a premium on European penetration, making the relationship between India and the surrounding seas central: India appears primarily as a peninsula. Eye lines focus on Delhi from European coastal positions, such as Bombay, Calcutta, Goa, and Madras. The customary maps also indicate only European victories, such as those of Clive at the Battles of Arcot and Plassey, and organize space and time in terms of British annexations. A totally different narrative of Indian history could be told focusing on European defeats, such as the Convention of Wadgaon in 1779, or Haidar Ali of Mysore’s victory over the British at the Battle of Perumbakam in 1780, or Tipu Sultan’s victory on the Coleroon river in 1782, or the unsuccessful British campaigns against Mysore in 1790 and the summer of 1791.

The major theme of my draft map was the contested succession to the Mughal Empire by a number of expansionist powers: Britain and also the Maratha Confederation, the Nizam of Hyderabad, the Nawabs of Bengal and the Carnatic, and the Sultan of Mysore—to contextualize the Europeans’ role. To do this, one would adopt a perspective in which India opens up from the Khyber Pass, with a central alignment thence via Delhi. Again, however, marketing considerations militatated against this map.

The coverage of military history, too, tends to focus on dramatic episodes rather than less well-known campaigns and on conflict as opposed to crucial aspects of the military background. For example, David Smurthwaite’s Pacific War Atlas 1941–1945 (London, 1995) includes a crowded map of operations in the Dutch East Indies in 1941–42 that contrasts with the more lavish coverage of the Japanese conquest of the Philippine. Smurthwaite and his cartographers also concentrate on the fighting at the expense of the logistical and industrial capabilities of the combatants. This is a fault with most treatments of World War II, which leads to neglect of the crucial relationship between Japan and the United States. There is little point in mapping the sinking of warships but not the disproportionate launching, equipping, and deploying of new vessels. The same is true of airpower.

Lastly, Smurthwaite fails to make sufficient allowance for the multiple assessments of the war. The war looked different from the perspectives of Siam or the Indian opponents of the British or the Soviet Union, and it would be interesting to see this depicted in cartographic form. [18] For example, Map 1.15 of Information China. Volume I (Peking: Social Sciences Publishing House, 1989), depicting the counteroffensive, includes much of China south of Manchuria in the “anti-Japanese base area” and “guerrilla area” on the eve of the offensive, areas that appear in every other map I have seen as unqualifiedly occupied by Japan.

Yet another example of packager intervention over content is the coverage of the Thirty Years’ War (1618–48) in the Cambridge Illustrated Atlas of Warfare: Renaissance to Revolution. For this I roughed out four maps: 1618–29, 1630–2, 1633–9, and 1640–8, attempting to give due weight to the conflict after the Swedish defeat at Nördlingen in 1634 and to cover the wide-ranging campaigns of 1636–39. The packagers dropped the third map, however, late in the process.

Cost is clearly a central concern for packagers, too, and lessens the number of maps in atlases. In his preface to his World Atlas of Military History I. to 1500 (London, 1973), Arthur Banks noted: “Whereas authors obviously want as many maps as possible in their books, publishers try to keep these to a minimum for economic reasons.” But the decline in the number of maps in an atlas is more than a matter of cost. It also reflects a reconceptualization of the atlas away from being a book of maps and towards a work including text, pictures, time-lines, biographies, glossaries, and bibliographies. Again, this is an example of the publisher-driven character of modern historical atlases. In the 1980s and 1990s, publishers showed increasing support for atlases that could be regarded as books in atlas format. For example, Donald Matthew’s Atlas of Medieval Europe (Oxford, 1983), is 240 pages long, but contains only 64 maps. The Atlas of the British Empire, edited by Christopher Bayly (London, 1989), contains only 39 maps in its 244 pages. Mike Corbishley’s The Middle Ages: Cultural Atlas for Young People (Oxford, 1990) contains 26 maps (excluding small insets) for 87 pages. Reviewing Matthew’s book, the distinguished historian of medieval cartography, P. D. A. Harvey observed: “We might reasonably think that an atlas in which maps do not predominate is not an atlas at all.” [19]

The publisher-driven character of modern historical atlases generally takes precedence over scholarly considerations. The Times Atlas of European History (London, 1994) used full-color computer-generated maps to “create a dynamic picture of a continent in flux and reveal the ebb and flow of political fortunes in unprecedented clarity,” it claimed. The Introduction resolutely explained the intention: “This atlas put the politics back in … At each date, frontiers are drawn on identical base maps of the whole continent, allowing the evolution of its nations, states, and empires—their appearance and growth, their contraction and disappearance—to become immediately clear.” But it does not address the precariousness and porosity of frontiers for much of the period covered in the book. This issue troubled not only myself but also Rosamond McKitterick, who covered the Dark Ages. For the 1466 map, I referred to a “trend towards more definite frontiers” and “lands whose status had been ill-defined for centuries,” but the earlier maps had indeed displayed such definition, a problem that can also be seen by contrasting the preface and maps in Dennis P. Hupchick and Harold Cox’s Palgrave Concise Historical Atlas of the Balkans (Basingstoke, 2001).

This discussion of the gap between commercial pressures and scholarly perspectives is not intended as a litany of complaint, but to point to the role of the former in the project’s organization. These accounts of the present situation also provide a necessary background against which to assess earlier historical atlases. The documentary trace of the actual process is very slight for modern atlases. Publishers rarely keep drafts, and, indeed the increased practice of designing maps on screen reduces the number of hard-copy drafts. The situation is of course even worse for the past. Aside from the absence of oral evidence, there is also the attrition of company and scholarly records, not least once atlases lose their shelf life. Even when an institutional record survives, as with the Historical Association in Britain, which was involved with Philips’ Historical Atlas (London, 1927), the surviving archival material is scanty.

Yet there are still the atlases, and they provide a whole series of narratives that it is worth considering. The atlases depict the passage of time and mark in it the location of events through a series of chronologically different maps that the reader is supposed to differentiate by a process of scrutiny. This scrutiny can be aided by cartographic devices, most obviously arrows, and by text, but nevertheless, a significant problem for all static graphic depictions of historical change is that they lack a dynamic dimension and so depict change as a series of stages or “stills,” rather than as a continuum. Historical atlases are especially prone to this, because their overview maps of, for example, the World in 1500 or Europe in 1815, focus on the situation then understood as spatial territory and a locational matrix of power, rather than looking at processes of transformation at that moment. A map of a year will convey more a sense of constancy and less one of alteration than, for example, a map depicting a half-century, and yet overview maps tend to be of the first case.

Even as static representations, maps necessarily depict processes of change, but that is not satisfactorily explained to readers. Leaving aside the more fluid and ambiguous nature of location and line than is generally suggested on maps, the cartographic moment of depiction takes on much of its meaning, and its analytical value, in terms of its context in change. For example, are the frontiers of the state advancing, contracting or static? Is industrial location constant, and if not, why not? This problem led to interest in employing technology, such as slides, to provide a sense of dynamism. [20] Software packages for use on computers followed with, for example, Millennium, a historical atlas of Europe and the Middle East launched in 1993 by Clockwork Software.

Technological developments have been of great value in data collection and analysis, editorial compilation, and scholarly production, [21] but they cannot overcome major problems with data availability, historical analysis, and map compilation. This is an issue with three-dimensional computerized maps. [22] It is still the case that time is seen rather as a setting than a sequence, and that little attention has been devoted to decision-making units and processes, rather than aggregate geographical patterns. [23] More generally, the combination of time and space remains a serious problem.

Alongside the map of a precise moment comes the map designed to cover a broad chronological sweep. This offers the possibility of suggesting a connection or parallel between events or developments. Time, thus, is treated with a degree of flexibility in order to construct a narrative that does not depend on simultaneity. This can create serious problems. There can be questionable combinations, conflations, and juxtapositions.

More widely, this issue relates to the question of how best to fill the map, and whether the very processes of choosing what to depict and how to depict it are misleading. This is arguably a more acute problem with historical cartography, because it is more difficult to challenge the authority of the map than in the case of modern thematic mapping where personal experience can act as a critique. For example, a map of Germany today showing it in terms solely of political violence can be more readily dismissed as slanted than one of the situation in the sixteenth century. The map “Regimes and Rights” in Dan Smith’s The State of War and Peace Atlas (London, 1997), depicted democratic states, including Australia, France, Italy, Germany, Japan, Spain, the UK, and the United States, in terms of “mistreatment by police or prison authorities.” This blanket description depicted them more harshly than Kazakstan, Suriname, Costa Rica, Panama, and Madagascar. [24] Such surprising allocations are much harder to query in historical cartography.

As with non-historical maps, there is the question of what to depict—in this case, which aspects of the past and of change to show—and also the danger that what is shown implies a causality that may be absent, exaggerated, or controverted. Periodization poses similar problems.

If one of the purposes of historical cartography is to pin down generalizations precisely in space, then it is still the case that historical research skills are at a premium and that lacunae evidence can have a serious impact on any mapping process. We know little about past mental geographies: the meanings of location, distance, proximity, space, and territory are part of the past narrative. They explain the spatial context of a given subject and period, and cannot be assumed in modern terms. Illustrations of past maps, showing their awareness and sense of space, can only help so much here. Modern scholarly text is also important: because text lacks the apparent authority of maps, it reminds us that maps have narrators.

Space and distance look the same on the map, established and measured by the scale. But they are not the same. Maps of France, for example, in 1600 and today may employ the same projection, alignment, and scale, but they cover different experiences of space and contrasting meanings of travel. Space and distance are not just a matter of the number of hours it takes to get from, say, Paris to Rouen, but also the attitudes that are created by the nature of travel and the effects of these attitudes on how distance is perceived. The rate of change in perception is not constant. Journeys and concepts of space and time in 1776 were more similar to the situation 223 years earlier than 223 years later. The menace of the dark (as unknown areas were often depicted), when space shrank to the shadowy spots lit by flickering lights, cannot be captured by maps, which offer a panoptic vision. The role of the seasons, not only for travel but also for industry and agriculture, is another limitation on modern historical maps. So also is the impact of a different mental world. The sense of direct providential intervention, of a daily interaction of the human world and wider spheres of good and evil, of heaven and hell, of sacred places, is heavily constricted today by secularism and science. Time, in short, is not an unproblematic third dimension.

Even more than most forms of cartography, historical maps do not exist as obvious images and bodies of knowledge requiring little bar organization for the purpose of reproduction. This process of narration that is required does not lessen the value of historical cartography, but does ensure that it is necessary to understand the relevant web of practices, constraints, and needs. This web is fascinating, worthy of study, and important, not only for our understanding of the past but also for our conceptualization of cartography.

Notes

[1] At one extreme, see the Nazi historical atlases, for example, Bernhard Kumsteller’s Werden und Wachsen, Ein Geschichtsatlas auf völkisher Grundlage (Brunswick, 1938).

[2] See Walter A. Goffart, “Breaking the Ortelian Pattern: Historical Atlases with a New Program, 1747–1830,” in Joan Winearls, ed., Editing Early and Historical Atlases (Toronto: University of Toronto Press, 1995), pp. 49–60.

[3] Walter Goffart, “The Map of the Barbarian Invasions: A Preliminary Report,” Nottingham Medieval Studies, 32 (1988) pp. 49–64, and Goffart, “The Map of the Barbarian Invasion: A Longer Look,” in Marc Anthony Meyer, ed., The Culture of Christendom: Essays in Medieval History in Commemoration of Denis L.T. Bethell (London, Hambledon Press, 1993), pp. 1–27.

[4] Armin Wolf, “What Can the History of Historical Atlases Teach? Some Lessons from a Century of Putzger’s Historischer Schul-Atlas,” Cartographica, 28 (1991), p. 22. See, more generally, Ferjan Ormeling, “Justus Perthes: Europa’s oudste kartografische uitgeverij,” Kartografisch Tijdschrift, 18, 4 (1992), pp. 35–9.

[5] Armin Wolf, “100 Jahre Putzger: 100 Jahre Geschichtsbild in Deutschland (1877–1977),”Geschichte in Wissenschaft und Unterricht, 29 (1978), pp. 702–18.

[6] Brian Hudson, “The New Geography and the New Imperialism, 1870–1918,” Antipode, 9 (1977), pp. 12–19; Felix Driver, “Geography’s Empire: Histories of Geographical Knowledge,” Environment and Planning D: Society and Space, 10 (1992) pp. 23–40; Anne Godlewska and Neil Smith, eds., Geography and Empire: Critical Studies in the History of Empire (Oxford, 1993).

[7] For a less positive interpretation of the Lords’ atlas, Lester J. Cappon, “The Historical Map in American Atlas,” Annals of the Association of American Geographers, 69 (1979), p. 633.

[8] Paullin, Atlas of the Historical Geography of the United States, p. xiii.

[9] J. A. S. Grenville, National Prejudice and International History (Leeds, 1968), p. 9; Keith Hodgkinson, “Standing the World on its Head: A Review of Eurocentrism in Humanities Maps and Atlases,” Teaching History, 62 (1991), pp. 21–3.

[10] Historical Atlas of Canada II: The Land Transformed 1800–1891 (Toronto, 1993), plate 58.

[11] Brian J. Graham, “No Place of the Mind: Contested Protestant Representations of Ulster,” Ecumene 1 (1994), pp. 257–81.

[12] Hereford B. George, The Relations of Geography and History (Oxford, 1901, 5th ed., Oxford, 1924); Albert Perry Brigham, Geographic Influences in American History (New York: Chatauqua, 1903); Ellen Churchill Semple, Influences of Geographic Environment on the Basis of Ratzel’s System of Anthropo-Geography (New York, Holt & Co. 1911); William Gordon East, The Geography Behind History (London, 1932); Lucien Febvre, A Geographical Introduction to History (London, K. Paul, Trench, Trubaer & Co., 1932); John K. Wright, “Miss Semple’s ‘Influences of Geographic Environment’: Notes towards a Bibliography,” Geographical Review, 52 (1962) pp. 346–61; Roy Allen Billington, The Genesis of the Frontier Thesis: A Study in Historic Creativity (San Marino, Calif., Huntington Library, 1971); Allen R. H. Baker, “Reflections on the Relations of Historical Geography and the Annales School of History,” in Allen R. H. Baker and Derek Gregory, eds., Explorations in Historical Geography (Cambridge, 1984); André-Louis Sanguin, Vidal de La Blache 1845–1918: un génie de la géographie (Paris: Belin, 1993).

[13] Ex inf. the editor, Richard Overy, meeting Jan. 6, 1999.

[14] See also Lester J. Cappon, “The Case for a New Historical Atlas of Early America,” William and Mary quarterly 28 (1971); Barbara B. Petchenik, “Cartography and the Making of a Historical Atlas: A Memoir,” American Cartographer 4 (1977); J. McQuilton, “Mapping the Past: Producing an Australian Historical Atlas,” Cartography 16 (1987); K. M. Davies and Malcolm McKinnon, “The Historical Atlas of New Zealand: An Introduction,” unpublished talk; and Anne B. Piternick, “The Historical Atlas of Canada: The Project behind the Product,” Cartographica 30:4 (1993).

[15] Among others, the 1000–1500 section of the Times Atlas of European History (London, 1994); Warfare: Renaissance to Revolution 1492–1792, sold as a Cambridge Illustrated Atlas (Cambridge, 1996); the 1492–1914 section of the Cassell Atlas of World History (London, 1997); the World Atlas of the Past (Oxford University Press, New York, 1999); the DK Atlas of World History (Dorling Kindersley, 1999); The Historical Atlas of Britain: The End of the Middle Ages to the Georgian Era (Sutton Publishing, 2000); and the Penguin Atlas of British and Irish History (2001); as well as consulting for HarperCollins for the fifth and sixth editions of The Times Atlas of World History (1993 and 1999).

[16] Ex. inf. Thomas Cussans, meeting, Feb. 23, 1994.

[17] Two maps I designed to illustrate Jacobite options for a military history of the Jacobite uprising of 1745, Culloden and the ’45 (Gloucester, 1990), one with London directly above Edinburgh and the other with Inverness on a horizontal line with Aberdeen, had to be fought for, as the publishers’ sales representatives did not like this feature when the book was presented at a sales conference. Ex. inf. the publishers, then Alan Sutton, now Sutton Publishing, 1990.

[18] For a limited approach, John Keegan, ed., The Times Atlas of the Second World War (London, 1989), p. 206.

[19] P. D. A. Harvey, “The Medievalist’s Atlas,” Nottingham Medieval Studies, 30 (1986) p. 117.

[20] Clifford L. and Elizabeth H. Lord, Historical Atlas of the United States (2nd edn., New York, 1953), p. iii; A. Breznay, “Projects for Making Comparable the Historical Maps of Different Ages and Matters,” in Making and Use of Historical Maps (Budapest: National Office of Lands & Mapping, 1972), p. 12.

[21] See, for example, Bruce Macdonald, Vancouver: A Visual History (Vancouver: Talonbooks, 1992), p. ix; Kenneth C. Martis and Gregory A. Elmes, The Historical Atlas of State Power in Congress 1790–1990 (Washington, 1993) p. xv; William G. Dean, “Atlas Structures and their Influence on Editorial Decisions: Two Recent Case Histories,” in Winearls, ed., Editing Early and Historical Atlases, pp. 149–52.

[22] Torsten Hágerstrand, Innovation Diffusion as a Spatial Process (Chicago, 1968); Hugh C. Prince, “Time and Historical Geography,” in Tommy Carlstein, Don Parkes and Nigel Thrift, eds., Making Sense of Time (London: E. Arnold, 1978), pp. 17–37; Jonathan Raper, Three Dimensional Applications in Geographic Information Systems (London, 1989); Donna J. Peuquet, “It’s About Time: A Conceptual Framework for the Representation of Temporal Dynamics in Geographic Information Systems,” Annals of the Association of American Geographers, 84 (1994), pp. 441–61.

[23] Arnold R. H. Baker, “Rethinking Historical Geography,” in Baker, ed., Progress in Historical Geography (Newton Abbot, 1972) pp. 15, 26–7.

[24] Dan Smith, The State of War and Peace Atlas (London, 1997), pp. 18–19.