To look forward to a future in which recourse to war has been brought under rational limitation should not lead us into the false view that there have been no limitations on warmaking in the past. The higher political and ethical systems attempted to impose legal or moral restrictions both on the use of war and its usages from early times. The most important limitations on warmaking, however, have always lain beyond the will or power of man to command. They belong within the realm of what the Soviet General Staff used to call ‘permanently operating factors’, and such factors — weather, climate, seasons, terrain, vegetation — always affect, often inhibit and sometimes altogether prohibit the operations of war. Other factors, loosely categorised as ‘contingent’ and including difficulties of supply, provisioning, quartering and equipment, have strictly limited the scope, intensity and duration of warmaking in many periods of human history. As wealth increased and technology developed, some were reduced or largely overcome — the soldier’s rations, for example, may now be preserved in convenient form for almost indefinite periods — but none can be said to have been eliminated altogether. How to feed, how to shelter, how to move an army in the field remain today the first, chief and most persistent problems that a commander has to solve.
Perhaps the effect of both ‘permanent’ and ‘contingent’ factors in limiting the scope and intensity of offensive or defensive operations is best illustrated from naval warfare. Man may fight with his fists on land, but to do even that on the surface of the water he requires a buoyant platform. Purpose-built platforms we must guess, since of their nature they decompose, appeared comparatively late in human history. The earliest to be found has been dated only to 6315 BC and given the effort, probably cooperative effort, needed to construct the simplest raft or dugout, we may presume that the bone and stone tools which provide evidence of man’s earliest industry predate boatbuilding by a very long period indeed.1
Specialised warships, even ships suitable for war, are relatively recent in origin. They have always been expensive to build and they require handling by specialist crews. Their construction and operation therefore demands considerable disposable wealth, probably the surplus of a ruler’s revenue; and if the earliest form of fighting at sea was piratical rather than political in motive, we must remember that even the pirate needs capital to start in business. The first navies may or may not have been anti-piratical in purpose — the advantages conferred by the ability to move forces or supplies along rivers or coasts may have first prompted rulers to maintain warships — but navies are, by definition, more costly than individual ships. Whichever way it is looked at, fighting on water has cost more than fighting on land from the start.
Wealth, or the lack of it, is not the only factor to limit the ability to wage war at will on water; others are weather and deficiencies of propulsive power. Wind comes free and the earliest representation we have of naval warfare — of a fight between warriors of the Pharaoh Rameses III and the Sea Peoples in the Nile delta in 1186 BC — shows the Egyptians in a ship with sails.2 Sailing-ships, however, were not to make suitable fighting platforms before the invention of the gun, since the management of sails precluded engagement at the short ranges where pre-gunpowder weapons could take effect. Oared ships were much more manoeuvrable in encounters where crews sought to close hand-to-hand with swords and spears. The advantages of the oared ship went further: by mounting a ram, and working up to full rowing speed, it could actually sink another if it caught it broadside on, which a wooden sailing-ship stout enough to bear the shock of impact could not do. Light winds would not impart the necessary speed; strong winds raise seas in which no captain with a thought for the survival of his ship would risk such an encounter.
The oared ship had serious deficiencies as a ship of war, however; in confined waters, like those of the Mediterranean, dominated from the second millennium BC onward by a succession of rich states that could afford the manpower costs, it was to set the terms of naval warfare until the coming of the gun. Yet it could not keep the seas in bad weather and so was essentially a summertime weapon. Worse, it could not work away from a port of re-supply for more than a few days at a time, since the hull form which made it fast in smooth waters — long but shallow and narrow — deprived it of the carrying space required to feed and water the large crew needed to row it at ramming speed. True, it was later to be used outside the inland seas as a vehicle of marauding in oceanic waters by nihilists like the Vikings — once they had mastered the technology of deep-keel construction and the technique of star-sight navigation — where it spread terror, devastation and death over coasts and riverine lands hundreds of miles from base. The Vikings, however, flourished in an era when states were weak, particularly at sea, and in any case they depended on the wind to carry their longships to undefended shores, using oars only for auxiliary purposes.
In consequence, as John Guilmartin has demonstrated in his brilliant analysis of Mediterranean naval warfare, galley navies were never autonomous instruments of strategy but extensions, or more accurately partners, of armies on land.3 The inshore wing of a galley fleet normally hinged on the coastward flank of an accompanying army, in operations that were amphibious in the strict sense of the term. The fleet manoeuvred so as to isolate an enemy coastal base from support by its own naval forces, while the army advanced with supplies to positions from which the galleys could be re-provisioned. This symbiosis explains why the great Mediterranean sea battles, from Salamis in 480 BC to Lepanto in AD 1571, were all fought within sight of land. Why, though, once the big-gun sailing-ship came to exercise mastery of the seas — that is, from the sixteenth century onward — were most naval battles still fought within sight of land, or very close to it? Two of the victories won by the greatest of sailing-ship admirals, Nelson, were gained against fleets lying inshore at anchor — the Nile and Copenhagen — while the third, Trafalgar, was the result of an encounter only twenty-five miles off the Spanish coast. The tendency for sailing-fleets to fight inshore had nothing to do with endurance. The wooden man-of-war, unlike the galley, carried stores and water sufficient to keep it at sea for many months, so that as early as 1502 Portuguese ships, which had sailed round the Cape of Good Hope, were able to fight and defeat the fleet of a local ruler off the west coast of India. In the 1650s Cromwell’s admiral, Blake, could campaign in the Mediterranean, where England then had no base, while by the middle of the next century Britain and France were conducting intensive naval campaigns against each other off the east coast of India, six months’ sailing-time from home. Despite their distance from base, all these fleets continued to do battle in coastal waters.
Several reasons combine to explain this circumstance. One is that battle under sail could not be conducted in rough weather (an exception was Quiberon Bay, fought in Atlantic squalls in November 1759) and inshore waters are more often calm than the high seas. Another is that the objects for which naval battles are fought — free access to the high seas from port, protection of coastwise shipping, defence against invasion — have their locus in coastal waters. A third is that sailing-fleets, operating exclusively by visual communication, have extreme difficulty finding each other in great waters. Even with a chain of frigates, the visual link between each was at most twenty miles; many fleets missed each other with the greatest ease, as Nelson found at the Nile in 1798. It is significant that in two real but rare deep-sea encounters — the second battle of Finisterre, 1747, fought 200 miles off Ushant, and the Glorious First of June, 1794, fought again off Ushant but 400 miles into the Atlantic, both between the British and French — the French fleets were in both cases encumbered by convoys, the latter 130 ships strong, covering so large an area of sea that they made a much more prominent target for a pursuer than their escort of warships would have done if sailing alone.
The supersession of sail by steam as a means of propulsion might be thought to have loosened the link between the warship and the land, since a steam warship could manoeuvre to engage even in flat calm, and remained a stable gun-platform at wind speeds that forced sailing-warships to reef and close their gunports. Paradoxically, however, the steam ship actually restored the logistic dependency under which the galley had lain and greatly diminished the operating range of steam fleets relative to that of sailing-fleets. The reason was that until the comparatively late adoption of oil fuel, steam warships burnt coal at an enormous rate — HMS Dreadnought of 1906 emptied its bunkers in five days’ steaming at twenty knots — and so were tied to their coaling-stations.4 A naval power like Great Britain, which acquired its worldwide network of bases in sailing-ship days, could keep fleets in all the oceans because they could coal at hundreds of ports; even so, they were local, not oceanic, in range. A state without such a chain of bases could either not project naval power at all, or was dependent on the goodwill of allies to do so. When Russia sent its Baltic fleet to the Far East in 1904–5, at a time when it was on bad terms with Britain, the ships managed the voyage only by piling their decks so high with coal that, between stops at French colonial ports, they could not have used their guns.
It is an extra paradox that coal-fired fleets, though in theory capable of oceanic encounter (two days’ steaming would carry them 500 miles from land), continued in practice to clash near coasts. In part the same strategic factors affected them, but they also continued, like their sailing predecessors, to be virtually blind until the coming of wireless; indeed, the real extension of their line of sight had to wait upon the arrival of the wireless-equipped, shipborne aircraft. As a result, all the sea battles of the First World War were fought within a hundred miles of land; the pattern repeated itself in the Second World War, despite the advent of radar, the aircraft-carrier and the long-range patrol submarine, and the mastery of the technique of replenishment at sea. The ultimate explanation derives from the vastness of the oceans; fleets could rarely count upon defeating distance in the vasty deep. The American aircraft that sank the Japanese carriers at Midway — one of the few true oceanic encounters in the history of the world — were guided to them by shrewd guesswork; the Bismarck, eventually sunk a thousand miles off Brest in May 1941, had twice shaken off the whole of Britain’s Home Fleet; while the mid-Atlantic battles between Allied escorts and surfaced German U-boats were brought about because large, slow convoys made abnormally conspicuous targets. Given the resistance offered to surveillance systems by the movements of oceanic storms, such as the large weather fronts which the Japanese used to cover their approach to Pearl Harbor in December 1941, and the persisting difficulty of coordinating long- with short-range target-acquisition equipment, the seas may well keep their secrecy for a long time to come.
The facts of the past are more surely and very simply stated. Seventy per cent of the globe’s surface is covered by water, most of it open sea, and most large sea battles have taken place in but a fraction of that area. If we draw up a list of Fifteen Decisive Sea Battles, to match Creasy’s famous Fifteen Decisive Battles of the World, taking ‘decisive’ to mean ‘of durable and more than local importance’, it might read as follows:
Salamis, 480 BC: defeat of Persia’s invasion of Greece
Lepanto, 1571: Muslim advance into western Mediterranean checked
Armada, 1588: Spain’s offensive against Protestant England and Holland frustrated
Quiberon Bay, 1759: Anglo-Saxon success in struggle with France for dominance in North America and India ensured
Virginia Capes, 1781: victory for the American colonists guaranteed
Camperdown, 1797: Dutch naval competition with the British extinguished for good
The Nile, 1798: Napoleon’s ambition to dominate both shores of the Mediterranean and reopen struggle for India thwarted
Copenhagen, 1801: mastery of North European waters conveyed to Britain
Trafalgar, 1805: Napoleon’s naval power finally destroyed
Navarino, 1827: inaugurated dissolution of Ottoman empire in Europe
Tsushima, 1905: established Japan as dominant power over China and in North Pacific
Jutland, 1916: collapsed Germany’s ambition to operate an oceanic navy
Midway, 1942: denied Japan control of the Western Pacific
March convoy battles, 1943: forced withdrawal of Germany’s U-boats from the Battle of the Atlantic
Leyte Gulf, 1944: established incontestable power of the United States over Imperial Japanese Navy
These are shorthand notes of the significance of the battles chosen; what is remarkable about this list — though experts may quarrel with it — is how close and often the sea battles cluster in the same corner of the map. Camperdown, Copenhagen and Jutland, for example, were all fought within 300 miles of each other; Salamis, Lepanto and Navarino, the first and last separated by 2300 years in time, took place near the Peloponnese at points scarcely more than a hundred miles apart. The Armada battle, Quiberon Bay and Trafalgar were fought within a hundred miles of Longitude Five West, between Latitudes Fifty and Thirty North, a comparatively tiny patch of the globe — much of it occupied by dry land. The Virginia Capes were to be the scene of much naval fighting after 1781, as Tsushima had been before 1905, notably during the Mongol offensive against Japan in 1274–81, while the coast where the Nile was fought had been a magnet for naval operations since the Pharaohs. Of the fifteen ‘decisive’ sea battles cited, therefore, it emerges that only two, Midway and the March convoy battles, took place in previously unviolated water far from land.
Equally, most of the globe’s dry land has no military history. Tundra, desert, rain forest and the great mountain ranges are as inhospitable to the soldier as to the traveller; indeed, even more so, for the soldier’s necessities are more cumbersome. Military manuals may contain entries on ‘desert’ or ‘mountain’ or ‘jungle’ warfare, but the truth is that to attempt to fight in terrain that is waterless or roadless defies nature and that such fighting as does take place is usually mere skirmishing between expensively over-equipped specialists. Rommel’s and Montgomery’s desert armies of the Second World War clung to the coast of North Africa; Japan’s conquest of the dense forest land of Malaya in December 1941-January 1942 was achieved along the colony’s excellent roads and by amphibious ‘hooks’ down its coastline; China’s seizure of parts of India’s mountain frontier in 1962, when attacks were mounted at heights above 16,000 feet, was staged by troops who had acclimatised for a year on the Tibetan plateau — many of the Indian defenders who had recently come up from the plains were incapacitated by altitude sickness.
In all, about seventy per cent of the world’s 60,000,000 square miles of dry land is either too high, too cold or too waterless for the conduct of military operations. The poles, North and South, demonstrate the effect of such conditions with starkness. The Antarctic continent’s inaccessibility and the extreme climatic conditions that prevail there secluded it from warmaking for millennia, though several states laid claim to territory; the icecap, moreover, is known to cover valuable mineral deposits. Since the signing of the Antarctic Treaty of 1959, all territorial claims have been put into abeyance and the continent has been declared demilitarised. The North Pole, by contrast, is not demilitarised and, indeed, the icecap is regularly undersailed by nuclear-propelled submarines. But the length of the polar night — three months in winter — the extreme winter cold and the absence of any resource of value makes it improbable that fighting will ever be conducted on its surface. The most northerly military incidents to have taken place on land in polar regions were the skirmishes fought in 1940–3 to capture or defend weather stations, set up by German or Allied parties, on the east coast of Greenland and on Spitzbergen, near Latitude Eighty North; casualties were inflicted by both sides but, under the attack of the elements, they were at times compelled to assist each other to survive.5 Beyond that, intense military activity has been concentrated into a fraction even of that space where conditions do favour the movement and maintenance of armed forces. Battles not only tend to recur on sites close to each other — the ‘cockpit of Europe’ in northern Belgium is one such area, the ‘Quadrilateral’ between Mantua, Verona, Peschiera and Legnano in northern Italy another — but have also frequently been fought on exactly the same spot over a very long period of history.
The most arresting example is Adrianople, now Edirne, in European Turkey, where fifteen battles or sieges have been recorded, the first in AD 323 and the last in July 1913.6*
Edirne is not now and never has been a large city; its population remains under 100,000. Its curious distinction as the most frequently contested spot on the globe has been conferred on it not by its wealth or size but by its peculiar geographical position. It stands at the confluence of three rivers, whose valleys provide avenues of movement through the mountains of Macedonia to the west, Bulgaria to the north-west and the Black Sea coast to the north, and which then flow to the sea through the only extensive plain in the most southeasterly tip of Europe. At the other side of the plain stands the great city of Constantinople (Istanbul), on a site chosen by Constantine for his capital because it was the most easily fortified position on the Bosphorus, which separates Europe from Asia. Adrianople and Constantinople are therefore strategically twin cities, together guarding movement from the Black Sea to the Mediterranean and from southern Europe to Asia Minor, or vice versa. Because Constantinople defied attack from the sea, particularly after the building of the walls of Theodosius in the early fifth century, all invaders of southern Europe from Asia Minor found themselves compelled to land in the plain to its rear; invaders starting north of the Black Sea were driven to hug its western shore by the barrier of the Carpathians on their inland flank, and so also ended up in the plain of Adrianople; while invaders from Europe, drawn by the prize of Constantinople, the richest place in the Western world between the fall of Rome and its sack by the Crusaders in 1204, had no choice but to cross the same plain in their approach march. Adrianople, in short, is the European end of what geographers call a land bridge, by which Asia gives on to Europe along two major routes, and it was fated to be fought over whenever there was a major outflow of military force, east-west or west-east, by way of either; in the circumstances, it is not surprising that the city never grew to any size.
Few other places exemplify the effect of permanent or contingent factors on the course of warfare as well as Adrianople does; nevertheless, in a weaker form their influence can be traced throughout the course of history on most landscapes where military activity has been high. Large rivers, highland barriers, dense forests form ‘natural frontiers’ with which, over time, political boundaries tend to coincide; the gaps between them are avenues along which armies on the march are drawn. Once through such gaps, however, armies rarely find themselves free to manoeuvre at will, even if no apparent obstacles stand in their way. A more subtle geography comes into play, reinforced by climate and the season, and adapted by the roadmaker and bridge-builder, even if not by the fortification engineer. Thus the German Blitzkrieg into France in 1940, apparently an unconfined romp across open country once the tanks that led it had broken the barrier of the Ardennes forests and the River Meuse, turns out to have followed very closely the line of Route nationale 43, which for much of its length is the Roman road laid out soon after Caesar’s conquest of Gaul in the first century BC.7 Neither the Romans nor those who built on their work made a point of quarrelling with geography; we may infer, therefore, that the German tank commanders, whatever their illusion of pursuing a free trajectory, were in fact obeying topographical dictates as old as the last reshaping of the earth’s surface in northern France, laid down at the retreat of the glaciers 10,000 years earlier.
A similar pattern of obedience to the laws of nature presents itself in the study of the German army’s campaign into Russia in the year following the Blitzkrieg in France. Western Russia appears to offer the invader, particularly a mechanised invader, free movement at will. Between its 1941 frontier and the three cities of Leningrad (St Petersburg), Moscow and Kiev, 600 miles distant, the surface of the land nowhere rises above 500 feet, while the rivers that cross that enormous and almost treeless plain tend to flow with the line of advance rather than across it. Nothing solid should impede the onset of the invader. Nothing solid does. In the centre, however, rise two of Russia’s greatest rivers, the Dnieper and the Niemen, flowing respectively into the Black and Baltic seas; their headwaters, with many tributaries, combine to form the Pripet marshes, a swamp 40,000 square miles in extent and so resistant to the conduct of military operations that its position on the situation map became known to German staff officers as the ‘Wehrmacht hole’ (Wehrmachtloch), containing no German military units of worth whatsoever. As a result it became a main base area for Soviet partisan operations against the Wehrmacht’s rear areas and, doubtfully effective though such operations were, a source of persistent unease to the German army as long as its front line in Russia was drawn further to the east.
The Wehrmachtloch, though a permanent feature of the Russian theatre of war, was a minor influence on German operations. A major and repetitive factor was the appearance of the seasonal swamp, created by the spring snowmelt and the autumn rains, across its whole front of engagement. The rasputitsa, as Russians call the twice-yearly liquefaction of the face of the steppe, brings military movement to a halt for a month at a time. As Golikov, the Soviet commander of the Voronezh Front, signalled to a subordinate who had enquired about the prospect of a counter-offensive reaching the line of the Dnieper in March 1943, ‘There are 200–230 miles to the Dnieper and to the spring rasputitsa there are 30–35 days. Draw your own conclusions.’8 The inescapable conclusion was that the onset of the snowmelt would outpace the Soviet advance, leaving the line of the Dnieper in German hands. So, indeed, it turned out. But the rasputitsa more often worked to the Germans’ disadvantage. It was prolonged in the spring of 1941, thus delaying the start of the invasion by several critical weeks, and again in the autumn, forcing a postponement of the advance on Moscow. The late arrival that year of the winter frosts, which restore a weight-bearing crust to the surface of the steppe, left the Wehrmacht’s tanks literally bogged down just too far from Moscow to ensure the capture of the capital by the timetabled date. Tsar Nicholas I called January and February ‘two generals in whom [Russia] can confide’;9 the March and October rasputitsa proved better generals to Russia in 1941 and may, indeed, have rescued it from catastrophe that year.
How may the discussion be summarised so far? What is clear is that the congruence of ‘permanently operating’ and contingent factors — climate, vegetation, topography and the alterations that man has made to the natural landscape — imposes on Mercator’s projection of the world map a sharp division between military and non-military zones, the latter vastly exceeding the former in extent. Organised and intensive warfare has been carried on over extended periods of time along an irregular but continuous band of the world’s surface lying between the tenth and fifty-fifth degrees of latitude in the northern hemisphere, and stretching from the Mississippi valley in North America to the Philippines and their outliers in the western Pacific, or from 90 degrees West of Greenwich to 135 degrees East. The Times Atlas of the World classifies vegetation into sixteen categories, including (before land-clearing for agriculture) Mixed Forest, Broadleaf Forest, Mediterranean Scrub and Dry Tropical Forest.10 If a line is drawn to enclose those four vegetation zones in the northern hemisphere and the land and sea routes between them, one may quickly see that almost all of history’s battles have been fought within the space the line encloses and very few outside. If the battle locations are dated by month, a seasonal concentration will superimpose itself, varying from place to place with highs and lows of temperature and rainfall, and dates of harvesting. By way of illustration, the first three battles of Adrianople were fought in July, August and July respectively and the last three in August, March and July; March is unusually early for campaigning even in the southern Balkans, when the rivers run high with snowmelt, but the other dates, immediately following the Mediterranean harvest, are exactly as to be expected.
Is it true, then, that the zone of organised warfare coincides, inside seasonal variables, with that which geographers call ‘the lands of first choice’, those easiest to clear of forest and yielding the richest crops when brought under cultivation? Does warfare, in short, appear cartographically as nothing more than a quarrel between farmers? In the sense that serious warmaking requires wealth, and intensive agriculture has always yielded the largest and most consistent return on any of man’s activities until very recent times, there is something to that view. On the other hand, while farmers are implacable in disputes about boundaries and water rights, and sturdy fighters when called to arms by superiors, they are also, by common observation, implacable individualists who abandon servitude to their animals and their fields only with the greatest reluctance. Marx regarded the peasants as ‘irredeemable’, by which he meant that he saw no prospect of enlisting them in the revolutionary armies with which he hoped to overthrow the capitalist order.11 Mao thought differently; and Victor Davis Hanson, in his breathtakingly original study of warmaking in classical Greece, is persuasive that it was the small landholders of the Greek city states who invented the idea of the ‘decisive battle’ as Westerners have practised it ever since. Nevertheless, Marx had a point. The farmer is indeed rooted in his plot, his village and his grumbles, and naturally resists the summons to march to some distant border between the lands of first choice and the unploughed region that lies beyond, however good the reason that he should.
We should note that plough people of the same language and religion rarely fight each other on a major scale. On the other hand borders between ploughed and unploughed land, throughout the temperate zone, are very frequently defined by long and expensive works of fortification: the Roman Antonine Wall just short of the Highland Line in Scotland; the limes marking the line between plough and forest in Roman Germany; the fossatum Africae which defended the fertile Maghreb from Saharan raiders; the Roman ‘Syrian’ frontier of forts and military roads separating plough from desert along the line of the Jordan and the Tigris-Euphrates headwaters; the Russian cherta lines running for 2000 miles from the Caspian Sea to the Altai mountains as a defence against steppe raiders; the Habsburg Military Frontier in Croatia separating the plains of the Sava and Drava from the Turkish-controlled mountain zone to the south; above all, the Great Wall of China, built to exclude the steppe nomads from the irrigated lands of the Yangtse and Yellow rivers on so extensive a scale and over so long a period that archaeologists have as yet failed to map all its complexities.12
These fortified boundaries suggest a fundamental tension between the haves of ploughed land and the have-nots of soils too thin, cold or dry to be broken for cultivation. To recognise that tension is not to fall into the false perception that the motive underlying major warmaking is mere expropriation. Man the warrior is a more complex being than that. Cultivators who are ethnic kin do fight each other, sometimes with the deadliest ferocity; have-nots from the wastes beyond the fertile zone may fight apparently for an idea alone — for example, the Arab followers of Muhammad expropriated with a will, but it was the urge to extend the frontiers of the House of Submission, rather than a base material motive, that drove them to their extraordinary exploits. The greatest of conquerors, Alexander of Macedon, was already comfortably established as ruler of the cities of Greece before he set off to the ends of the earth, and seems to have pillaged the Persian empire largely for the pleasure of it. The Mongols, even wider-ranging than Alexander in their assaults on settled states, showed virtually no capacity to consolidate the fruits of their victories: some of the descendants of the Diadochi, Alexander’s generals, were still in power in Bactria 300 years after his death, while none of the regimes founded by Genghis or his immediate successors lasted for more than a century. Tamerlane, a Tartar claiming Mongol ancestry — from Genghis, no less — appears to have valued the rich lands he overran absolutely not at all but, like a slash-and-burn harvester, to have moved on as soon as he had exhausted the soil where he ravaged.
Yet to note that the have-nots often misuse what they expropriate is not to invalidate the general point that the tide of war tends to flow one way — from poor lands to rich, and very rarely in the opposite direction. That is not simply because poor lands offer little worth fighting over; it is also because fighting in poor lands is difficult, sometimes impossible. Poor people from what William McNeill calls ‘food-deficit areas’ — desert, steppe, forest, mountains — will fight among themselves, and their fierce military skills have been valued and purchased by the rich for as long as we have records of organised warmaking. Hence the exotic names — hussar, uhlan, jäger — that some European regiments proudly bear to this day, and the even more exotic scraps of barbaric clothing — bearskin caps, frogged jackets, kilts and lionskin aprons — that continue to be worn for ceremony. The warfare of poor peoples, nevertheless, was limited in scope and intensity by their very poverty. It was only when they broke into the rich lands that they were able to accumulate the stocks of provender which made deeper penetration, and eventual conquest, a possibility. Hence the wealth and labour expended by cultivators in fortifying their borders, to exclude the predators before they could make serious trouble.
The causes underlying the operation of ‘permanent’ and ‘contingent’ factors on warfare may therefore be seen to be exceedingly complex. Man the warmaker is not an agent of unbounded free will, even though in warmaking he may burst the limits that convention and material prudence normally impose on his behaviour. War is always limited, not because man chooses to make it so, but because nature determines that it shall be. King Lear, railing at his enemies, may have threatened to ‘do such things — what they are yet I know not — but they shall be the terrors of the earth’; as other potentates in straitened circumstances have found, however, the terrors of the earth are hard to conjure up. Wealth lacks, the weather worsens, the seasons turn, the will of friends and allies fails, human nature itself may revolt against the hardships that strife demands.
Half of human nature — the female half — is in any case highly ambivalent about warmaking. Women may be both the cause or pretext of warmaking — wife-stealing is a principal source of conflict in primitive societies — and can be the instigators of violence in an extreme form: Lady Macbeth is a type who strikes a universal chord of recognition; they can also be remarkably hard-hearted mothers of warriors, some apparently preferring the pain of bereavement to the shame of accepting the homeward return of a coward.13 Women can, moreover, make positively messianic war leaders, evoking through the interaction of the complex chemistry of femininity with masculine responses a degree of loyalty and self-sacrifice from their male followers which a man might well fail to call forth.14 Warfare is, nevertheless, the one human activity from which women, with the most insignificant exceptions, have always and everywhere stood apart. Women look to men to protect them from danger, and bitterly reproach them when they fail as defenders. Women have followed the drum, nursed the wounded, tended the fields and herded the flocks when the man of the family has followed his leader, have even dug the trenches for men to defend and laboured in the workshops to send them their weapons. Women, however, do not fight. They rarely fight among themselves and they never, in any military sense, fight men. If warfare is as old as history and as universal as mankind, we must now enter the supremely important limitation that it is an entirely masculine activity.
* Adrianople I was fought between the Roman Emperor Constantine and the pretender Licinius, who approached from west and east respectively; at Adrianople II, 378, one of the catastrophes of history, the Emperor Valens and the last great Roman army were overwhelmed by the Goths, who had invaded the empire across the Danube (in flight from the Huns, a horse people, who had broken out of the steppe); at Adrianople III, 718, the recently arrived Bulgars defeated a Muslim army attempting to take Constantinople from the rear — an outcome of crucial importance for Christian Europe; Adrianople IV, V and VI were fought by the Bulgars in their attempts to attack Constantinople, in 813, 914 and 1003; Adrianople VII, 1094, was a battle between a Byzantine emperor and a pretender; at Adrianople VIII, 1205, the Bulgars defeated the Crusader Baldwin, who had made himself Byzantine emperor, and the Doge Dandolo (whose family house in Venice is now that city’s most expensive hotel); Adrianople IX, 1224, ended in a victory by the restored Byzantine imperial house over the Bulgars; Adrianople X, 1255, was an internal Byzantine struggle; Adrianople XI, 1355, ended in victory by the Byzantines over the Serbs, who had recently become a Balkan military power; Adrianople XII, 1365, marked a successful stage in the Ottomans’ advance from Asia Minor into Europe; following the consolidation of Ottoman power, there was no further battle until 1829, when at Adrianople XIII a Russian army took the city from them; in the last two battles, in 1913, Ottoman Turkey first lost and then regained Adrianople from the Serbs and Bulgars.